Cytosolic group IVa phospholipase A2 mediates IL-8/CXCL8-induced transmigration of human polymorphonuclear leukocytes in vitro

Background Cytosolic gIVaPLA2 is a critical enzyme in the generation of arachidonate metabolites and in induction of β2-integrin adhesion in granulocytes. We hypothesized that gIVaPLA2 activation also is an essential downstream step for post adhesive migration of PMN in vitro. Methods Migration of PMNs caused by IL-8/CXCL8 was assessed using a transwell migration chamber. PMNs were pretreated with two structurally unrelated inhibitors of gIVaPLA2, arachidonyl trifluoromethylketone (TFMK) or pyrrophenone, prior to IL-8/CXCL8 exposure. The fraction of migrated PMNs present in the lower chamber was measured as total myeloperoxidase content. GIVaPLA2 enzyme activity was analyzed using [14C-PAPC] as specific substrate F-actin polymerization and cell structure were examined after rhodamine-phalloidin staining. Results IL-8/CXCL8-induced migration of PMNs was elicited in concentration- and time-dependent manner. Time-related phosphorylation and translocation of cytosolic gIVaPLA2 to the nucleus was observed for PMNs stimulated with IL-8/CXCL8 in concentration sufficient to cause upstream phosphorylation of MAPKs (ERK-1/2 and p38) and Akt/PKB. Inhibition of gIVaPLA2 corresponded to the magnitude of blockade of PMN migration. Neither AA nor LTB4 secretion was elicited following IL-8/CXCL8 activation. In unstimulated PMNs, F-actin was located diffusely in the cytosol; however, a clear polarized morphology with F-actin-rich ruffles around the edges of the cell was observed after activation with IL-8/CXCL8. Inhibition of gIVaPLA2 blocked change in cell shape and migration caused by IL-8/CXCL8 but did not cause F-actin polymerization or translocation of cytosolic F-actin to inner leaflet of the PMN membrane. Conclusion We demonstrate that IL-8/CXCL8 causes a) phosphorylation and translocation of cytosolic gIVaPLA2 to the nucleus, b) change in cell shape, c) polymerization of F-actin, and d) chemoattractant/migration of PMN in vitro. Inhibition of gIVaPLA2 blocks the deformability and subsequent migration of PMNs caused by IL-8/CXCL8. Our data suggest that activation of gIVaPLA2 is an essential step in PMN migration in vitro.

Because MAP kinase and PI3K also regulate gIVaPLA 2 phosphorylation, we postulated that activation of gIVaPLA 2 might regulate neutrophil migration. The objective of this study was to examine specifically the functional role of gIVaPLA 2 in PMN migration caused by IL-8/CXCL8. IL-8/CXCL8 was applied in concentration causing upstream phosphorylation of ERK-1/2, p38 MAPK and Akt/PKB. We found that inhibition of gIVaPLA 2 activity blocked substantially the transmigration toward IL-8/CXCL8 in a transwell chamber. This study is the first demonstration that activation of gIVaPLA 2 is a critical regulatory step subsequent to upstream activation of signaling kinases in eliciting PMN migration.

Isolation of human PMNs
Venous blood from normal human subjects (20-45 years old) was collected in heparin-containing tubes, and PMNs were isolated by Ficoll-Paque sedimentation as described previously [26,27]. Purity of PMN on H and Estained cytoslides was ~90-95%. Informed written consent was obtained from all volunteers in this study.

Transwell migration assay
PMN migration in transwell microplates was assessed using the standard methods as described previously [28]. Preliminary experiments have established that the num-ber of cells (4 × 10 4 cells) used allow the optimal % cell migration without clogging the pores of transwell filter of the upper chamber. Cells then were preincubated with HBSS, 3 μM -30 μM arachidonyl trifluoromethylketone [TFMK; inhibitor of gIVaPLA 2 [29], or 10 -9 M -10 -6 M pyrrophenone [inhibitor of gIVaPLA 2 [30] for 30 min at 37°C. Treated cells in 50 μl HBSS were transferred onto 5 μm-pore transwell filters positioned on top of the migration chamber. HBSS or 10 ng/ml to 1000 ng/ml IL-8/ CXCL8 was loaded in the bottom chamber (final volume = 310 μl), and the transwell microplates were incubated for 60 min and 90 min at 37°C. The migrated PMNs were treated with 100 μl of HBSS + 10% FBS buffer and 100 μl developing solution [8 ml 100 nM NaH 2 PO 4 (pH = 5.5), 1000 μl 10% hexadecyltrimethylammonium bromide (HTAB), 3 μl 30% hydrogen peroxide, 1000 μl 10% o-dianisidine dihydrochloride]. The reaction was terminated by addition of 50 μl sulfuric acid and myeloperoxidase (MPO) activity was measured at 405 nm in a Thermomax microplate reader (Molecular Devices, Menlo Park, CA). The fraction of migrated PMNs present in the lower chamber was measured as total MPO content. Data were expressed as % cell migration. Maximal, no-toxic inhibitory concentration of TFMK and pyrrophenone were established in initial studies demonstrating blockade of gIVaPLA 2 activity [see also Results].
In separate studies, morphological changes of the nonmigrated cells (top chamber) and migrated cells (bottom chamber) toward IL-8/CXCL8 were examined. The effect of 30 μM TFMK or 10 -6 M pyrrophenone on cell deformability caused by IL-8/CXCL8 also was examined using confocal microscopy.

Measurement of LTB 4 secretion
Aliquots of 250,000 were activated with saline, 1-1000 ng/ ml IL-8/CXCL8, or 1 μM FMLP (+ 5 μg/ml cytochalasin B) for 15 min at 37°C in a final volume of 250 μl HBSS. The reaction was terminated by centrifugation at 12,000 × g for 1 min. Aliquots of supernatants were assayed with a commercial EIA kit as previously described [24,31].

GIVaPLA 2 activity assay
GIVaPLA 2 activity assay was determined in aliquots of 2 × 10 6 cells incubated for 15 min at 37°C with 3 μM -30 μM TFMK or 10 -10 M -10 -6 M pyrrophenone. Activity was measured at optimal time (30 min) after 100 ng/ml IL-8/CXCL8 [see Results]. This time and concentration were shown in initial studies to cause phosphorylation of gIVaPLA 2 in PMNs. The cell pellets were resuspended in disruption buffer (see above) and immediately sonicated followed by addition of specific substrate ([ 14 C]-PAPC) for gIVaPLA 2 [24,31]. To measure precisely the total gIVaPLA 2 activity, 5 mM dithiotrietol was added to cell lysate to inactivate, if any, the remaining 10-14 kDa secretory PLA 2 enzymes that could interfere with the assay. Thirty min later, the reaction was terminated by adding 560 μl of Dole's reagent (heptane:isopropyl alcohol:1 N H 2 SO 4 ; 400:390:10 by vol), and the radioactivity was measured in a liquid scintillation counter and expressed as picomoles/30 min/10 6 cells [31].

Subfractionation
Freshly isolated PMNs were preincubated with either saline, 100 ng/ml IL-8/CXCL8, or 10 -6 M FMLP for 15 min at 37°C. After washing with PBS, treated cells were centrifuged for 1 min at 400 × g. The pellet were lysed in 50 μl disruption buffer (see above) and put on ice for 10 min. The disrupted pellets, which are mainly nuclear component of the cells, were centrifuged at 500 × g for 1 min. A total of 50 μl of boiling buffer was added to the pellets and boiled for 5 min. The supernatants were centrifuged again at 100,000 × g for 1 h. Eight μl of loading buffer was added to the collected supernatant, which is the cytoplasm fraction, and was boiled for 5 min. Samples were loaded onto SDS-PAGE and membrane was probed with pAb against cPLA 2 . The translocation of cytosolic gIVaPLA 2 to the nuclear component of the cells was detected by an enhanced chemiluminescence (Amersham, Arlington Heights, IL).

Change in cell shape and F-actin polymerization
Change in cell shape and F-actin polymerization were examined in migrated cells. HBSS or rhodamine-phalloidin [32] was added to the paraformadehyde-fixed cytoslides containing samples and changes in cell shape and F-actin polymerization were analyzed by confocal microscopy.

Statistical analysis
Experimental data are expressed as mean ± SEM in each group. Student's t-test was used for comparison between two-paired groups. Where multiple comparisons were made, differences on concentration-response curves for the same agonist or inhibitor were compared after Bonferonni correction. Variation between more than two groups was tested using one-way ANOVA followed by Fisher' least protected difference test. Statistical significance was claimed when P < 0.05.

IL-8/CXCL8-induced upstream kinases phosphorylation
We first demonstrated that the 100 ng/ml concentration of IL-8/CXCL8 used in these studies caused phosphorylation of critical upstream kinases, ERK-1/2, p38 MAPK and Akt/PKB (a target protein for PI3K), in the same PMN isolates (Figure 2). Phosphorylated ERK-1/2 was greatest at 0.5-1 min and gradually decreased thereafter. The p38 MAPK and Akt/PKB were constitutively expressed in unstimulated PMNs. Treatment with IL-8/ CXCL8 upregulated the phosphorylation of p38 MAPK and Akt/PKB, which unlike ERK-1/2, was sustained for ≥ 15 min. Total protein for ERK-1/2, p38 MAPK, and Akt/ PKB was stained with respective Ab to demonstrate equal loading of samples.

Effect of IL-8/CXCL8 on [ 3 H]AA release and LTB 4 secretion
Non-stimulated PMNs released minimal amounts of AA and undetectable amounts of LTB 4 . Activation with IL-8/ CXCL8 did not elicit secretion of either AA or LTB 4 in PMNs. All treated PMNs remained viable as assessed by trypan blue exclusion dye analysis. Measurements were performed as described in Methods section.
Immunoblotting analysis of cell-fractional components demonstrated that gIVaPLA 2 is located mainly in cytoplasm of unstimulated PMNs. Application of IL-8/ CXCL8 translocated the cytosolic gIVaPLA 2 to some extent, to the nuclear component of PMNs (Figure 3b). We used FMLP as a positive control since we previously have shown that FMLP causes the translocation of cytosolic gIVaPLA 2 to nuclear membrane in eosinophils [33].

Cell morphology and F-actin polymerization
We next examined the change in cell shape of migrated PMNs co-incubated with HBSS, TFMK, or pyrrophenone prior to buffer control or IL-8/CXCL8 exposure at 90 min. Representative photomicrographs of cell morphology are shown in Figure 6. PMNs in the buffer control chamber retained their globular appearance after 90 min ( Figure. 6a). By contrast, PMNs activated with IL-8/ CXCL8 developed an elongated cell shape with a contracted tail (Figure 6b). Blockade of PMNs with TFMK ( Figure 6c) or pyrrophenone (Figure 6d) prevented the deformability of cell shape caused by IL-8/CXCL8. IL-8/CXCL8 caused F-actin polymerization and translocation of cytosolic F-actin to the inner leaflet of PMN membrane ( Figure 6). However, inhibition of gIVaPLA 2 , which prevented the elongation of PMN (see Figure 6cd), did not block IL-8/CXCL8 -induced F-actin polymer-  ization (Figure 6g-6h). Thus, while activation of gIVaPLA 2 is essential for the change in shape of PMNs, Factin polymerization, another essential step for cell migration, is not regulated by gIVaPLA 2 .

Discussion
The objective of this study was to examine the functional role of gIVaPLA 2 in the regulation of PMNs migration caused by IL-8/CXCL8. Prior studies have reported the signaling role of upstream kinases, ERK-1/2, p38 MAPK, and PI3K [1,5,6], in the initiation of cell migration; however, the downstream regulation of PMN migration elicited by IL-8/CXCL8 has not been elucidated previously.
We used a transwell-migration chamber [28] and determined whether inhibition of activated gIVaPLA 2 by TFMK or pyrrophenone blocked PMN migration caused by IL-8/CXCL8. We also examined the functional role of gIVaPLA 2 in causing in PMN elongation and F-actin polymerization, which both are necessary for PMN migration [34,35]. While the specific mechanism causing the PMN change in cell shape was not elucidated fully in these studies, we found that inhibition of gIVaPLA 2 is sufficient to block change in cell shape caused by IL-8/ CXCL8 even in the presence of F-actin polymerization ( Figure 6).
Studies were designed using IL-8/CXCL8, a potent chemoattractant of PMNs. Prior studies have suggested that IL-8/CXCL8 is rather weak stimulator of human PMNs in comparison to rodent models [36]. Activation of PMNs with IL-8/CXCL8 did not elicit arachidonic acid or LTB 4 secretion in human PMNs. Thus, our findings suggest that IL-8/CXCL8 caused transmigration of PMNs by a process that does not involve activation of arachidonate synthesis.
Cytosolic gIVaPLA 2 is a critical messenger protein for cellular adhesion [21,22,27]. We have shown recently that neutrophil or eosinophil binding to ICAM-1 is mediated through activation of ERK-1/2 and subsequent phosphorylation of gIVaPLA 2 [22][23][24]27]. In all prior cases, we have found that stimuli that upregulate cell adhesion CD11b expression also induce the activation of gIVaPLA 2 [22,24,25,27]. However, the role of gIVaPLA 2 to mediate PMN migration has not been previously reported.
Initial experiments were performed to confirm that upstream kinases, ERK-1/2, p38 MAPK, and PI3K, were activated by the concentration of IL-8/CXCL8 used in these studies ( Figure 2). Immunoblotting analysis demonstrates that IL-8/CXCL8 elicited rapid phosphorylation of ERK-1/2, p38 MAPK, and Akt/PKB (Figure 2), confirming that these kinases were activated in these experiments; however, the downstream signaling pathway for cell migration has not been characterized. In this study, we used two-unrelated pharmacological inhibitors of gIVaPLA 2 , TFMK and pyrrophenone, to elucidate the role of gIVaPLA 2 in cell migration. Transmigration of PMNs was blocked substantially in the presence of upstream phosphorylation of ERK-1/2, p38 MAPK and Akt/PKB (a target protein of PI3K) using inhibitors of activated gIVaPLA 2 . Accordingly, these data indicate a downstream regulatory role for gIVaPLA 2 in in vitro PMN migration subsequent to activation of upstream kinases by IL-8/CXCL8.
Immunoblotting analysis demonstrated that IL-8/ CXCL8 caused phosphorylation and translocation of cytosolic gIVaPLA 2 to the nuclear component of PMNs [27]. It has been shown that gIVaPLA 2 inhibition effectively blocked cell adhesion and secreted mediators after cell activation [21,24,31]. We have demonstrated that inhibition of gIVaPLA 2 blocked both gIVaPLA 2 enzymatic activity ( Figure 4) and cell migration ( Figure 5) elicited by IL-8/CXCL8 in concentration dependent manner. These data thus imply that activated gIVaPLA 2 is an essential intermediate step in PMN migration in vitro.
Prior studies have demonstrated that interference with F-actin rearrangement could contribute to decrease cell migration [37]. We observed that inhibition of gIVaPLA 2 with TFMK or pyrrophenone did not prevent the rearrangement of F-actin assembly elicited by IL-8/CXCL8. F-actin polymerization still was evident around the edges of inner cell membrane (Figure 6g-h). These findings suggest that while IL-8/CXCL8 caused change in cell shape, gIVaPLA 2 does not directly regulate the rearrangement of the actin cytoskeleton in PMNs.
It is important to note some limitations to our in vitro models of PMN migration. We used transwell chamber in vitro to study transmigration of human PMNs. Migration in vitro occurred in the absence of β 2 -integrin ligation, which is the first step (adhesion) in cell migration in vivo [38]. In vivo conditions are a more complex environment, and it is not possible to extrapolate these data directly to the human situation. Studies in vivo, however, do not allow for stimulus isolation to specify mechanisms and sequence of cell migration. In these studies, maximal migration of PMNs from the upper chamber to the lower chamber containing IL-8/CXCL8 was ~50%. This is comparable to other chemoattractants, i.e., FMLP, C5a and LTB 4 [39]. The initial number of cells (5 × 10 4 cells) was constant in all studies, and was sufficient to cover the area of a chamber (96-well chamber) for optimal PMN migration.

Conclusion
Our data demonstrate that gIVaPLA 2 activation caused by IL-8/CXCL8 (subsequent to activation of upstream kinases) may be an essential step in human PMN migration. Change in PMN cell shape and migration correspond to increased activity of cytosolic gIVaPLA 2 and translocation of gIVaPLA 2 to the nuclear membrane.