Increased Acute Inflammation, Leukotriene B₄-Induced Chemotaxis, and Signaling in Mice Deficient for G Protein-Coupled Receptor Kinase 6

Polymorphonuclear neutrophils (PMN) ³ play a central role in defense against infection. PMN migrate from the circulation to inflammatory sites, where they kill invading pathogens. A large number of chemoattractants and chemokines are involved in the regulation of PMN recruitment. Although chemotactic factors can be lipids, proteins, and peptides, most of them exert their action via binding to G protein-coupled receptors (GPCR). Little is known about the cellular mechanisms that are responsible for the directed migration of PMN toward chemoattractants. It has been suggested that receptor desensitization, internalization, and re-expression play an important role in the directed migration of cells, although there are also reports indicating that receptor recycling is not required per se for chemotaxis to occur (1, 2, 3).

Several kinases play a role in the desensitization and internalization of GPCR upon agonist activation. GPCR kinases (GRK) are involved in the rapid, agonist-dependent uncoupling of the receptor from G proteins in a process called homologous desensitization. Phosphorylation of agonist-stimulated receptors by GRK promotes the binding of proteins from the arrestin family, which can prevent further G protein activation and enhance receptor internalization (4, 5, 6). More recently, it has been shown that arrestins can also serve as scaffolding proteins and are involved directly in activation of certain signaling pathways (7). GPCR are also sensitive to phosphorylation by second messenger-activated protein kinases such as protein kinase C and protein kinase A. In general, phosphorylation and desensitization of GPCR via second messenger-dependent protein kinases are slower processes than GRK-dependent desensitization and do not require agonist binding to the receptor (6).

The family of GRK consists of seven members, GRK1–7. GRK 2, 3, 5, and 6 are expressed at high levels in cells of the immune system (8, 9). Interestingly, the level of expression of GRK2, 3, and 6 in leukocytes is known to be tightly regulated (10). For example, we have shown that in patients with rheumatoid arthritis or in rats after development of adjuvant arthritis or experimental autoimmune encephalomyelitis, the level of GRK2, 3, and 6 is reduced by 50–80% in leukocytes from the blood or spleen (11, 12, 13). In addition, there is evidence that cytokines as well as oxygen radicals can regulate the level of expression of GRK in leukocytes and other cell types (11, 14, 15). Moreover, differentiation of the pro-myelocytic cell line HL-60 toward the neutrophil lineage results in increased expression of GRK6, suggesting a functional role for this GRK in PMN (9).

Leukotriene (LT) B₄ is one of the most potent chemoattractants for PMN and is known to stimulate adhesion of leukocytes to endothelium, chemotaxis, and PMN degranulation (16, 17). LTB₄ has been implicated in the pathology of chronic inflammatory diseases like asthma and rheumatoid arthritis as well as in acute inflammation. High levels of LTB₄ have been found in acute and chronic inflammatory lesions. In animal models for arthritis, skin inflammation, and peritonitis, LTB₄ antagonists reduce clinical symptoms (18, 19, 20). Moreover, in mice deficient for the LTB₄ receptor BLT1, the inflammatory response in a peritonitis model is significantly reduced, suggesting an important role for LTB₄ in PMN chemotaxis in vivo (21).

Recently, it has been shown in overexpression systems that the human receptor for LTB₄ (hBLT1) can be phosphorylated by GRK6 (22). In cells overexpressing the hBLT1, overexpression of GRK6 results in reduced LTB₄ receptor signaling, demonstrating that GRK6 can desensitize this receptor. Threonine 308 within the cytoplasmic tail of hBLT1 was shown to be crucial for LTB₄-induced desensitization of the receptor (22). To investigate the possible role of GRK6 in inflammation and in LTB₄ reactivity of PMN, we used mice with a targeted deletion of the GRK6 gene. The cytoplasmic tail of the murine BLT1 also contains a threonine residue at position 308 in a region that is highly homologous to hBLT1(23). We demonstrate that in the absence of GRK6, the in vivo acute inflammatory response to topical application of arachidonic acid (AA) is increased. Moreover, we show that the in vitro chemotactic response of PMN toward LTB₄ is increased in the absence of GRK6.

We used GRK6-deficent mice and littermate controls from a mixed C57BL/6 × SVJ/129 background (24). Offspring of GRK6^+/− × GRK6^+/− mice were used and in all experiments we used pairs of wild-type (WT) and GRK6^−/− littermates to control for genetic background. Animals were genotyped by PCR and housed in the Central Animal Facility of the University of Utrecht under specific pathogen-free conditions. The animal committee of the University Medical Center Utrecht approved all animal experiments.

Heparinized blood was collected via the orbital plexus for composition analysis. For chemotaxis and cell signaling assays, peripheral blood was collected by cardiac puncture. RBC in the peripheral blood were lysed in hypotonic saline. For transendothelial migration assays, PMN were obtained by injecting 1 ml of a 1 mg/ml zymosan solution (Sigma-Aldrich, St. Louis, MO) in PBS i.p. After 4 h, the mice were euthanized and a peritoneal lavage was performed using 9 ml of ice-cold RPMI 1640 containing 2% FBS and 2 mM EDTA. After centrifugation, any remaining RBC were lysed. For some experiments bone marrow leukocytes were collected by flushing femurs with PBS.

MPO activity was measured according to a modified protocol of Bradley et al. (25). Ear biopsies were homogenized in 300 μl of PBS containing 0.5% hexadecyltrimethylammonium bromide. After two freeze/thaw cycles and sonification, homogenates were centrifuged at 10,000 × g for 15 min and MPO activity was assayed by incubating 20 μl of diluted sample with 180 μl of 50 mM sodium phosphate buffer (pH 6.0) containing o-dianisidine (0.2 mg/ml) and H₂O₂ (0.0006%). The OD at 450 nm was determined. MPO activity was normalized to standardized dilutions of homogenate of purified human PMN and is expressed in arbitrary units.

AA (20 μl, 100 mg/ml in acetone; Sigma-Aldrich) was applied to the left ear and 20 μl of vehicle was applied to the right ear. After 1 or 6 h, 5-mm diameter biopsies were taken and weighed. Tissue was fixed in 4% paraformaldehyde and tissue sections were stained with H&E for histological examination. Alternatively, biopsies were processed for determination of MPO activity. In some experiments, PMA (1 μg in 20 μl of DMSO; Sigma-Aldrich) was applied.

PMN chemotaxis was quantified using a Transwell system (polycarbonate filters with pore size 5 μm). Cells were placed in the top well in a total volume of 100 μl, and 600 μl of chemoattractant or medium (RPMI 1640 with 0.5% FBS) was added to the lower chamber. After a 60-min incubation at 37°C, cells in the bottom chamber were collected and the number of migrated PMN was determined.

Transendothelial migration (TEM) assays were performed as previously described (26). Briefly, 10⁵ Ea.hy 926 endothelial cells were plated onto 24-well Transwells and incubated for 3 days. Monolayer integrity was determined by assessing diffusion of [¹⁴C]mannitol (Amersham, Arlington Heights, IL) from the top chamber to the bottom chamber within the Transwell, and those Transwells in which the amount of ¹⁴C cpm detected in the bottom well was <35% of control were used. The medium at the bottom was replaced with medium with or without LTB₄ and neutrophils were added to the top chamber. Following 4 h of incubation at 37°C, cells in the lower chamber were harvested and stained with anti-mouse Gr-1 conjugated to PE (BD PharMingen, San Diego, CA). The number of migrated PMN was determined by flow cytometry.

Cells were loaded with Fluo-3-AM (Molecular Probes, Eugene, OR) for 30 min at room temperature. Subsequently, cells were stimulated in assay buffer (145 mM NaCl, 5 mM KCl, 1 mM Na₂HPO₄, 1 mM CaCl₂, 0.5 mM MgSO₄.7H₂O, 5 mM glucose, and 10 mM HEPES, pH 7.4) and fluorescence data for PMN were collected by flow cytometry. Data represent mean fluorescence intensity for 20–40 cells collected in 1-s periods. In some experiments, cells were stimulated for 5 min with LTB₄ (100 nM) and subsequently placed on ice, washed extensively with ice-cold medium, resuspended in assay buffer, and restimulated with LTB₄ or fMLP (Sigma-Aldrich).

Cells were stimulated with LTB₄ at room temperature and fixed in 3% formaldehyde/0.3% saponin in PBS for 20 min. F-actin was stained by addition of 0.2 μM FITC-phalloidin (Sigma-Aldrich) and fluorescence intensity of PMN was determined by flow cytometry.

Data are expressed as mean and SEM and were collected in three independent experiments using at least two pairs of animals in each experiment. Data were analyzed by Student’s t test or two-way ANOVA as indicated. A value of p < 0.05 was considered statistically significant.

Topical application of AA induces an acute inflammatory response that involves both vascular leakage with edema as well as influx of PMN. To get more insight into the role of GRK6 in the in vivo acute inflammatory response, we analyzed the inflammatory response to topical application of AA in WT and GRK6^−/− mice. Edema was determined by measuring changes in ear biopsy weight at 1 and 6 h after application of AA. Fig. 1,A shows representative examples of the increase in ear thickness of WT, GRK6^+/−, and GRK6^−/− animals. The increase in ear biopsy weight was significantly higher in GRK6^−/− animals than in WT animals at both time points (Fig. 1,B, 1 h: WT, 50.3 ± 9.9%; GRK6^−/−, 90.6 ± 11.3%, p < 0.01; 6 h: WT, 22.9 ± 2.5%; GRK6^−/−, 37.5 ± 5.2%, p < 0.05). The increase in ear weight in GRK6^+/− animals (1 h, 79 ± 15.1%, n = 6; 6 h, 30.9 ± 4.4%) was consistently between that of WT and GRK6^−/− animals, suggesting a gene dosage effect. To quantify cellular infiltration into the ear after application of AA, we determined MPO activity in ear biopsy homogenates (25). MPO activity in ear biopsies from GRK6^−/− mice collected at 6 h after application of AA was significantly higher than in ear biopsies from WT mice (Fig. 1,D, WT: 1.67 ± 0.22 U; GRK6^−/−: 2.96 ± 0.35 U, p < 0.01). The increased cellular infiltrate in GRK6^−/− ears after application of AA is illustrated in Fig. 1 C. MPO activity in untreated ear biopsies was below detection level. MPO activity per PMN as determined in peripheral blood samples was not different between WT and GRK6^−/− mice (data not shown). In addition, the number of PMN per milliliter of blood was similar in WT and GRK6^−/− animals (WT, 0.64 ± 0.11 million PMN/ml, n = 11 and GRK6^−/−, 0.75 ± 0.11 million PMN/ml, n = 11, p = 0.5). Thus, the increased PMN influx into the ears of GRK6^−/− animals after topical application of AA cannot be attributed to increased PMN numbers in the peripheral blood.

The acute inflammatory response to application of PMA did not differ between GRK6^−/− and WT animals. The PMA-induced increase in ear biopsy weight was 50 ± 16% in WT and 62 ± 3% in GRK6^−/− animals (n = 6, p = 0.5). Moreover, MPO activity determined 6 h after application of PMA was similar in WT and GRK6^−/− animals (WT, 1.2 ± 0.4; GRK6^−/−, 1.2 ± 0.3, n = 5, p = 0.99), indicating that the increased response to AA in GRK6^−/− mice was specific for AA.

The response to AA is highly dependent on the formation of the LTs LTB₄ and LTC₄, of which LTB₄ is thought to be responsible for the attraction of PMN to the site of inflammation after AA application (21, 27, 28). We therefore assessed the in vitro chemotactic response of peripheral blood leukocytes from WT and GRK6^−/− animals toward LTB₄ in a Transwell chemotactic assay. Basal motility of PMN (with medium in both upper and lower wells) was not significantly different for WT or GRK6^−/− cells (Fig. 2,A). When we tested the chemotactic response to 100 or 1000 nM LTB₄ (LTB₄ in the lower well), significantly more GRK6^−/− PMN reached the lower well than WT cells, indicating that the chemotactic response to LTB₄ is significantly increased in GRK6^−/− cells. However, the chemokinetic response to 100 and 1000 nM LTB₄ (LTB₄ in upper and lower wells) was also higher in GRK6^−/− PMN than in WT cells. In addition, when we used 100 nM LTB₄, we did observe the expected difference between chemotaxis (LTB₄ in lower well) and chemokinesis (LTB₄ in lower and upper wells) using WT cells but not when we used GRK6^−/− cells. Similar results were obtained when we analyzed the chemotactic response of GRK6^−/− bone marrow leukocytes to LTB₄ (Fig. 2,B). Both the chemotactic and chemokinetic response of GRK6^−/− bone marrow PMN toward LTB₄ were higher than the response of WT bone marrow PMN (Fig. 2 B). Moreover, also using GRK6^−/− bone marrow PMN we did not observe a significant difference between chemotaxis and chemokinesis at 100 nM LTB₄.

It is possible that the increased chemotaxis in fact is primarily due to increased LTB₄-induced random migration (or chemokinesis) rather than directed migration. To further address this issue, we also determined peritoneal exudates PMN chemotaxis toward LTB₄ in a TEM assay (Fig. 2 C). Less than 0.01% of input WT or GRK6^−/− cells were able to cross the endothelial monolayer without LTB₄ driving the response in the TEM assay. In addition, for both WT and GRK6^−/− cells <0.1% of input reached the lower well when LTB₄ was applied to both the lower and upper chambers. Thus, this assay allows a better assessment of chemotactic responses. LTB₄ applied to the bottom chamber induced the migration of 3.4 times as many GRK6^−/− PMN than WT PMN (10.2% vs 3.0% of input, respectively, p = 0.03), confirming that chemotactic responses to LTB₄ were increased in the absence of GRK6. Similarly, when we analyzed the response of bone marrow-derived PMN in a TEM assay using 1–1000 nM LTB₄, the response of GRK6^−/− cells was consistently higher than that of WT PMN (1.2- to 2-fold as many GRK6^−/− PMN than WT PMN reach the lower well depending on the concentration of LTB₄).

Polymerization of actin is one of the early responses of PMN to stimulation with chemoattractants and is thought to be required for the initiation of chemotaxis. Therefore, we investigated whether the increased chemotactic response to LTB₄ stimulation of PMN is also reflected in increased LTB₄-induced actin polymerization. LTB₄ induces rapid polymerization of actin in PMN from WT and GRK6^−/− animals (Fig. 3,A). Maximal LTB₄-induced actin polymerization in GRK6^−/− PMN is not different from the maximum response in WT cells. However, actin polymerization was more sustained after stimulation of GRK6^−/− PMN with LTB₄ (Fig. 3 A).

The rapid and transient LTB₄-induced increase in intracellular calcium is a major player in BLT1 signaling (29). To investigate whether BLT1 signaling is changed in the absence of GRK6, we investigated LTB₄-induced calcium responses in WT and GRK6^−/− PMN. Stimulation of PMN with LTB₄ induces a rapid and robust increase in intracellular calcium concentration, which peaks within 10–15 s after activation. The peak level of intracellular calcium is similar in WT and GRK6^−/− cells (Fig. 3 B). However, similarly to the prolonged actin polymerization, in GRK6^−/− cells a prolongation of the rise in intracellular calcium level was observed. These data suggest that BLT1 desensitization is at least partially impaired in GRK6^−/− PMN.

Next, we investigated the capacity of LTB₄ to induce unresponsiveness to restimulation with LTB₄ in WT and GRK6^−/− cells. PMN were stimulated with 100 nM LTB₄ and the calcium response was recorded for 3 min. Cells were kept in the presence of LTB₄ for another 2 min and then washed extensively with ice-cold buffer. Cells were resuspended in assay buffer and a second baseline calcium level was recorded, followed by restimulation with 100 nM LTB₄. As expected, using this procedure with WT cells, restimulation with LTB₄ does not induce a second calcium response (Fig. 4,A). This is not due to inability of the cells to mount a second calcium response, since fMLP can still induce a robust increase in intracellular calcium in these cells. The data in Fig. 4 B clearly demonstrate that under these conditions, cells from GRK6^−/− animals also become refractory to restimulation with LTB₄. Similar to WT PMN, pre-exposure of GRK6^−/− PMN to LTB₄ does not prevent the response to subsequent stimulation with fMLP.

The results presented in this study demonstrate that the acute inflammatory response to topical application of AA is increased in GRK6^−/− animals, both at the level of edema formation as well as at the level of cellular infiltration. Moreover, the in vitro chemotactic response to LTB₄ is significantly increased in GRK6^−/− cells. Finally, we show that LTB₄-induced calcium signaling and actin polymerization are prolonged in PMN from GRK6^−/− animals.

The inflammatory response after topical application of AA depends on the formation of the unstable intermediate LTA₄ by the enzyme 5-lipoxygenase. LTA₄ is further metabolized into either LTB₄ or cysteinyl LTs including LTC₄ and LTD₄. LTB₄ is a potent chemoattractant for PMN and the cysteinyl LTs increase vascular permeability. 5-Lipoxygenase-deficient animals cannot produce LTB₄ and LTC₄ and therefore have reduced inflammatory responses to topical application of AA (28, 30). Moreover, both AA-induced ear swelling as well as AA-induced PMN influx into the ear are significantly reduced in BLT1^−/− mice, suggesting a major role for this receptor in the increased vascular permeability as well (21). In GRK6^−/− mice, we observed both increased edema formation as well as increased influx of PMN after topical application of AA. The increased influx of PMN may well be the result of increased responsiveness of these cells to LTB₄ in GRK6^−/− cells, as was observed in chemotaxis and signaling experiments in vitro. Edema formation is the result of vascular effects of AA metabolites, especially LTC₄. In view of the reduced edema formation in BLT1^−/− mice in this model, it is possible that the increased ear swelling in GRK6^−/− mice either directly or indirectly results from increased responsiveness of BLT1 to LTB₄. However, it may well be possible that the LTC₄ receptor is also regulated by GRK6 and that increased responsiveness of LTC₄ receptors on endothelial cells is responsible for the increased edema. Our data show that the response of GRK6^−/− mice to topical application of PMA is not different from WT mice. PMA activates protein kinase C, resulting in general stimulation of multiple cell types including T cells, neutrophils, vascular cells, and mast cells. It has been reported that the inflammatory response to PMA is independent of the formation of LTB₄ (28). The normal response to PMA therefore suggests that the increased response to application of AA is not the result of a general increase in inflammatory activity in the GRK6^−/− mice. Binding of LTB₄ to the BLT1 activates phospholipase C via pertussis toxin-sensitive and -insensitive G proteins (31). The subsequent increase in cytosolic calcium concentration is an important signaling pathway for this receptor. We show here that both the calcium response and LTB₄-induced actin polymerization are prolonged in GRK6^−/− PMN. These data strongly suggest that desensitization of the BLT1 naturally expressed in PMN is at least partially dependent on the presence of endogenous GRK6. In line with our observations, it has been shown recently that LTB₄-induced inositol 1,4,5-triphosphate accumulation in hBLT1-transfected COS-7 cells was inhibited by overexpression of GRK6 (22). Interestingly, in those studies not only overexpression of GRK6, but also of GRK2 or GRK5 could enhance agonist-induced desensitization of hBLT1 in COS-7 cells. Leukocytes are known to express a high level of GRK2 and also express GRK5 (8, 9), and it is possible that the endogenous levels of these kinases compensate for lost GRK6 in GRK6^−/− PMN. Our data suggest, however, that other kinases that are naturally expressed in PMN do not fully compensate for GRK6, since calcium signaling and actin polymerization in response to LTB₄ are both prolonged in the absence of GRK6.

We have recently presented evidence for altered signaling after activation of the lymphocyte chemokine receptor CXCR4 in the absence of GRK6 (26). GRK6^−/− lymphocytes respond to stromal cell-derived factor 1 (SDF-1) with increased signaling as determined by SDF-1-induced GTPase activity, demonstrating that in the same GRK6^−/− animals desensitization of the chemokine receptor CXCR4 is also impaired. At the same time, the chemotactic response of T and B lymphocytes to SDF-1 is decreased in the absence of GRK6. In contrast, we describe in the present study that increased signaling to the chemoattractant LTB₄ in GRK6^−/− PMN is associated with increased LTB₄-induced chemotaxis. Moreover, we now have data ⁴ showing that the chemotactic response of GRK6^−/− PMN to SDF-1 is increased as well. These data suggest that GRK6 deficiency enhances both SDF-1- and LTB₄-induced chemotaxis of PMN, but reduces SDF-1-induced chemotaxis of lymphocytes. Apparently, the role of GRK6 in chemotaxis is dependent on the cell type rather than the stimulus. It has been suggested that lymphocytes and PMN use different mechanisms to migrate. Lymphocytes increase locomotion by reducing the duration of migration breaks, while PMN migrate by reducing the frequency of migration breaks (32). Whether these differences are related to differences in regulation of chemotaxis by GRK6 remains to be established.

Although it has been reported that signaling through hBLT1 mutants lacking specific GRK6 phosphorylation sites is increased due to reduced desensitization (22), there are no data describing the role of GRK6 in preventing a response to a second stimulation with LTB₄. Therefore, we investigated the effect of pre-exposure of cells to the agonist LTB₄ on the calcium response to restimulation with LTB₄ in GRK6^−/− and WT PMN. Surprisingly, we did not observe any difference between WT and GRK6^−/− PMN: in both cases, 5-min exposure of cells to LTB₄ prevents a calcium response to restimulation with LTB₄, whereas cells do still respond to another stimulus (fMLP). These data show that GRK6 is not absolutely required to render PMN refractory to restimulation with LTB₄. Apparently, kinase-independent mechanisms or other kinases (including other GRK), can render cells refractory to restimulation of BLT1 but do not compensate for the role of GRK6 in terminating signaling to calcium and actin polymerization or in inhibiting chemotaxis.

Investigating the response of GRK6^+/− animals in vivo indicated that there was a gene dosage effect: edema formation and PMN influx into the ear were increased in GRK6^+/− animals, although not to the same extent as in GRK6^−/− animals. Gaudreau et al. (22) have shown in a GRK6 overexpression system that the extent of inhibition of the LTB₄-induced inositol 1,4,5-triphosphate accumulation depends on the absolute level of GRK6. These data suggest that both in vitro as well as in vivo not only the presence of GRK6, but also its level is important in determining cellular responsiveness. Similarly, we have shown recently that 50% reduction of GRK2 in T cells from GRK2^+/− animals results in a significant increase in CCR5 signaling to phosphatidylinositol 3-kinase and calcium as well as increases in the chemotactic response of these cells to CCR5 agonists. ⁵ The observation that partial reduction in GRK levels in inflammatory cells has functional consequences is of particular importance in view of our earlier studies showing that in humans with adjuvant arthritis and in rats with rheumatoid arthritis or experimental autoimmune encephalomyelitis the expression of GRK6 in leukocytes is reduced by ∼50% (11, 12, 13). Our present data suggest that this reduction in leukocyte GRK6 expression during an ongoing inflammatory response may play an important role in the efficient recruitment of inflammatory cells to sites of inflammation.