Abstract
Various heterotrimeric Gi proteins are considered to be involved in cell migration and effector function of immune cells. The underlying mechanisms, how they control the activation of myeloid effector cells, are not well understood. To elucidate isoform-redundant and -specific roles for Gαi proteins in these processes, we analyzed mice genetically deficient in Gαi2 or Gαi3. First, we show an altered distribution of tissue macrophages and blood monocytes in the absence of Gαi2 but not Gαi3. Gαi2-deficient but not wild-type or Gαi3-deficient mice exhibited reduced recruitment of macrophages in experimental models of thioglycollate-induced peritonitis and LPS-triggered lung injury. In contrast, genetic ablation of Gαi2 had no effect on Gαi-dependent peritoneal cytokine production in vitro and the phagocytosis-promoting function of the Gαi-coupled C5a anaphylatoxin receptor by liver macrophages in vivo. Interestingly, actin rearrangement and CCL2- and C5a anaphylatoxin receptor-induced chemotaxis but not macrophage CCR2 and C5a anaphylatoxin receptor expression were reduced in the specific absence of Gαi2. Furthermore, knockdown of Gαi2 caused decreased cell migration and motility of RAW 264.7 cells, which was rescued by transfection of Gαi2 but not Gαi3. These results indicate that Gαi2, albeit redundant to Gαi3 in some macrophage activation processes, clearly exhibits a Gαi isoform-specific role in the regulation of macrophage migration.
Introduction
Signals generated by the engagement of chemoattractants with their cognate receptors orchestrate the recruitment and positioning of leukocytes to lymphoid and peripheral organs. Typically, cellular migration occurs along chemotactic gradients, and changes in leukocyte receptor expressions and their cross-talk are critical mechanisms for regulating the movement of such cells to various tissues and sites of inflammation (1–4). Within tissues, chemoattractants also direct the encounter between immune cells, thereby playing a regulatory role in immune homeostasis. Most chemoattractants, including chemokines, complement C5a, fMLP, leukotriene B4, and others (5–8), signal through G protein-coupled receptors (GPCRs) that mediate their functions mainly via the heterotrimeric Gi-proteins (9). Ligand binding to this class of GPCRs triggers Gαi subunits to exchange GTP for GDP, resulting in the dissociation of the Gαi subunit from Gβγ heterodimers. The release of Gαi-associated Gβγ is necessary for triggering directional migration (10).
The Gαi family consists of three closely related members, Gαi1–3, which are characterized by their sensitivity to pertussis toxin (PTx). The Gαi1–3 isoforms share 85–95% of amino acid sequence identity and overlapping expression patterns. Although Gαi1 is primarily found in the brain, Gαi2 and Gαi3 are abundantly expressed in the immune system. Many biological functions of leukocytes, including macrophage phagocytosis and migration, involve Gαi-mediated signaling (11, 12). Previous studies in mice lacking Gαi2 or Gαi3 suggested key roles of these two Gi isoforms in distinct cellular responses (13–16). For instance, Gαi3 deficiency leads to an impaired response of insulin-regulated autophagy in liver cells (16, 17), whereas Gαi2−/− mice show defects in the migration of T and B cells to lymph nodes (18, 19) and of eosinophils to sites of allergic tissue injury (20). Moreover, GPCR-induced activation of platelets is affected in Gαi2−/− mice (21). Gαi2 has also been reported to regulate the migration of neutrophils in acute inflammation (20, 22). In the lung Arthus reaction, capillary leakage and neutrophil influx are reduced in the absence of the Gαi-coupled C5a receptor (C5aR) as well as Gαi2 but not Gαi3 (23). This result suggests a predominant role of Gαi2 for neutrophil recruitment by pathogenic immune complexes. However, the relative contribution of Gαi proteins in the migration control of other myeloid effector cells remained to be clarified.
In this study, we identify Gαi2 as the essential Gαi protein required for homeostatic and inflammation-induced migration of monocytes and macrophages using mice deficient for Gαi2 or Gαi3 in combination with cell knockdown (KD) and rescue approaches. Importantly, remodeling of the actin cytoskeleton and chemotactic migration of macrophages are selectively regulated by Gαi2. Only the rescue of Gαi2 but not Gαi3 restored chemotaxis of RAW 264.7 cells that were silenced for Gαi2, thus providing, to our knowledge, the first definitive evidence for an isoform-specific role of Gαi2 in the regulation of macrophage migration.
Materials and Methods
Mice
Generation of Gαi2- and Gαi3-deficient SV129 mice was described previously (13, 24). These Gαi knockout (KO) mouse strains were backcrossed to C57BL/6J (B6) mice for >11 generations and were strictly kept under IVC conditions. Under these conditions, no signs of intestinal inflammation were visible in Gαi2-deficient mice during the course of the study (15). C5aR-deficient mice backcrossed to C57BL/6J mice for six generations were provided by C. Gerard (Boston, MA) (25). B6 control mice were from Charles River Laboratories. All mice were used at 8–14 wk of age. Animal experiments were conducted in accordance with current laws in combination with the regulations of the local authorities.
Monocyte/macrophage cell counts and flow cytometry
CD11b-positive liver macrophages as well as lavaged macrophages from lung and peritoneum were isolated and quantitated as described previously (26–28). Blood was obtained from anesthetized mice by puncture of the retro-orbital plexus, transferred to EDTA-coated tubes, and counted for blood monocytes by the automate Animal Blood Counter (Scil Animal Care, Viernheim, Germany). Absolute cell numbers of CD115+ and Ly6C+/CD62L+ monocyte subsets were obtained by using BD Truecount counting beads (BD Pharmingen, San Diego, CA). Expression of CD11b, L-selectin (CD62L) adhesion molecules on F4/80-positive blood cells was analyzed with allophycocyanin-conjugated anti-CD11b/CD62L mAbs by flow cytometry, using a FACSCalibur flow cytometer (BD Biosciences, Heidelberg, Germany).
Thioglycollate-induced peritonitis
Peritoneal recruitment of leukocytes was induced using 4% thioglycollate (Life Technologies). After 4 d, mice were killed, and the peritoneal cavity was rinsed 10 times with 1 ml ice-cold PBS/5 mM EDTA. Total cell count of the collected peritoneal lavage fluid was assessed. For quantification of macrophage influx, differential cell counts were performed on cytospins (10 min at 55 × g) stained with May-Grünwald-Giemsa using 300 μl peritoneal lavage fluid.
LPS-induced lung inflammation
Mice were anesthetized with ketamine and xylazin, the trachea was cannulated, and 20 μg Escherichia coli LPS (Sigma-Aldrich) was applied. Control animals received PBS instead of LPS. Mice were killed at 4 d, and bronchoalveolar lavage was performed five times with 1 ml PBS containing 5 mM EDTA at 4°C. For quantification of alveolar macrophage accumulation, differential cell counts were performed on cytospins (10 min at 55 × g) stained with May-Grünwald-Giemsa using 300 μl bronchoalveolar lavage fluid.
In vivo erythrophagocytosis by liver macrophages
C5aR-regulated phagocytosis was induced by a single i.p. injection of the pathogenic anti-erythrocyte 34-3C mAb (IgG2a, 150 μg/mouse), as described previously (26, 28, 29). Mice were sacrificed 2 d later, and the livers were processed for histological examination. Tissues were fixed in 10% buffered formalin, embedded in paraffin, and stained with H&E, according to conventional procedures.
LPS-induced activation of Gαi-deficient macrophages in vitro
Peritoneal macrophage (PM) cells were flushed out of the peritoneal cavity of Gαi2-deficient, Gαi3-deficient, or wild-type (WT) B6 mice, washed twice with PBS/5 mM EDTA, and suspended in RPMI 1640 medium and 10% FCS. The PM cells were allowed to adhere for 4 h in 12-well plates (Corning) at a density of 8 × 105
Transwell migration assay of Gαi-deficient macrophages in vitro
Gαi2- or Gαi3-deficient 0.5 × 105 PM in 100 μl RPMI 1640 medium/0.1% BSA were placed into the insert of a Transwell chemotaxis chamber (8-μm pore size filter; Greiner) and incubated at 37°C for 4 h with RPMI 1640 medium/0.1% BSA supplemented with C5a (50 ng/ml) or MCP-1 (CCL2) (10 ng/ml). For quantification of chemotaxis, Transwell filters were fixed in methanol, stained with May-Grünwald-Giemsa, and mounted on a glass slide. PM cells that attached to the filter were counted by light microscopy. The expression status of the C5a and CCL2 receptors (C5aR and CCR2) on F4/80-positive PM cells was analyzed with Alexa 647-conjugated anti-C5aR (provided by J. Zwirner, Göttingen, Germany) (30) and anti-CCR2 (provided by M. Mack, Regensburg, Germany) (31) mAbs by flow cytometry, using a FACScalibur flow cytometer (BD Biosciences).
Actin rearrangement assay of Gαi-deficient macrophages in vitro
Freshly isolated Gαi2- or Gαi3-deficient 8 × 105 PM cells in RPMI 1640 medium/1% FCS were plated into 12-well plates onto coverslips coated with fibronectin and incubated at 37°C over night. Next day, cells were washed and then stimulated for 4 h with C5a (50 ng/ml). For visualization of cell morphology, PM cells were then fixed with 4% paraformaldehyde for 10 min, permeabilized with 0.5% Triton X-100 for 7 min, washed with PBS containing 3% BSA and 0.05% Triton X-100, and stained for F-actin by incubation with Alexa 488-conjugated phalloidin (Sigma-Aldrich, Munich, Germany) for 1 h. Then, cells were washed briefly, mounted on slides, and analyzed by fluorescence microscopy. The percentage of cells able to undergo morphological changes toward an asymmetric cell shape was used as a measure of actin rearrangement.
Generation and functional analysis of macrophage Gαi2 KD cells
One microgram of pLKO.1-puro plasmid (Sigma-Aldrich) containing a Gαi2 short hairpin RNA (shRNA) with the target sequence in the 3′-untranslated region of exon 9 of Gαi2 (5′-CCTGCTCATTCTCGTAGCTTT-3′) was used for transfection of RAW 264.7 macrophage cells. Plasmid with no shRNA served as mock control. Transfections were done with Lipofectamine 2000 using the recommended protocol (Invitrogen). After transfection, RAW 264.7 cells were selected with 3 μg/ml puromycin, and the expression of the Gαi2 target gene, as well as Gαi3, was validated with SYBR Green real-time PCR analysis using the following Gαi2- and Gαi3-specific primers (Gαi2, sense, 5′-CAACTCCTCCAGCCTAGACC-3′; antisense, 5′-TCTCTCACGCTTC TTGTGCT-3′) (Gαi3, sense, 5′-TGGCTCTCAGTGATTACGA CCTT-3′; antisense, 5′-GGTTCATTTCCTCATCCTCAGC-3′). The Gαi2 shRNA-transfected cells showed a stable KD of Gαi2, and these Gαi2 KD RAW 264.7 macrophages (along with the mock transfectants) were further characterized by immunoblot analysis and used in functional Transwell migration assays.
Immunoblot analysis
Confluent RAW 264.7 cells (Gαi2 KD and mock) were washed in PBS and lysed with lysis buffer (20 mM Tris [pH 7.4], 300 mM NaCl, 3.5 mM EDTA, 1 mM EGTA, and 1% Triton X-100) containing protease inhibitors (one complete EDTA-free inhibitor tablet; Roche) and homogenized by passing through a 26-gauge needle four to six times. Lysates were centrifuged at 13,000 × g for 10 min at 4°C and quantified using the Micro BCA Protein Assay Kit (Pierce). Thirty micrograms of protein from whole-cell lysates was separated by urea-supplemented SDS-PAGE as previously described (32) and blotted onto nitrocellulose membrane. Rabbit anti-Gαi2, anti-Gαi3, and anti-Gβcommon Abs (33) were diluted 1:8,000. The mouse anti-tubulin Ab was diluted 1:1,000 (Sigma-Aldrich). Appropriate HRP-conjugated secondary Abs (anti-rabbit, Cell Signaling Technology; anti-mouse, DakoCytomation) were diluted 1:2,000. Immunoreactive bands were visualized by using an ECL detection system (GE Healthcare). The levels of Gαi2, Gαi3, Gβ1, and Gβ2 were quantified by using densitometric analysis software (Image Lab; Bio-Rad) and were normalized to the levels of anti-tubulin in the same sample. The pixel intensity of immunoreactive bands in samples prepared from Gαi2-shRNA–transfected cells was normalized to the corresponding immunoreactive band intensity identified in an equal amount of lysate prepared from mock-transfected cells.
Gαi2 and Gαi3 expression vectors and transfections
The constructs were generated by cloning PCR-amplified Gαi2 and Gαi3 cDNA sequences from RAW 264.7 cells into the EcoRI and XhoI restriction sites of the pMSCV-IRES-YFP expression vector. The Gαi2 PCR primers were sense, 5′-CGCAGAATTCACATGGGCTGCACCGTGAGCGC-3′, and antisense, 5′-TTACCTCGAGTCAGAAGAGGCCACAGTCCTTCAG-3′. The Gαi3 PCR primers were sense, 5′-CGCAGAATTCATGGGCTGCACGTTGAGC-3′, and antisense, 5′-TTACCTCGAGTCAATAAAGCCCACATTCCTT-3′. Both constructs were verified by sequence analysis. Gαi2 KD or mock control RAW 264.7 macrophage cells were transfected with 1 μg of the Gαi2 and Gαi3 expression vectors using Lipofectamine 2000 (Invitrogen, St. Louis, MO) as delivery cargo. After 2 d, the transfectants were used in functional Transwell migration and immunoblot assays as described before.
Statistical analysis
To analyze differences between more than two normally distributed groups, a one-way ANOVA was performed. Pairwise comparisons were then performed using Tukey’s test. To analyze differences between two normally distributed groups, an unpaired Student t test was used.
Results
Defective homeostatic and inflammation-induced migration of macrophages in Gαi2-deficient mice
Heterotrimeric G proteins of the Gi family have been implicated in signaling pathways regulating immune homeostasis and cell migration. Previously, it was shown that lack of Gαi2 leads to defects in the trafficking of T and B cells to lymph nodes (18, 19) and in the recruitment of eosinophils and neutrophils to sites of inflammation (20, 22, 23). Moreover, increased blood monocyte numbers were noted in Gαi2-deficient mice, whereas monocytes in the bone marrow (BM) are not affected (Ref. 22; data not shown). To determine the potential role of the individual Gαi proteins in the homing of monocytes and/or macrophages, we first examined whether the presence of Gαi2 and Gαi3 is needed for proper distribution of macrophages at different organ sites. Macrophage cell counts in lung, peritoneum, and liver were similar between WT B6 mice and Gαi3-deficient mice (Gαi3 KO). In contrast, Gαi2 KO mice exhibited significantly reduced numbers of tissue macrophages in the compartments tested (Fig. 1A) and increased numbers of total and CD115+ monocytes in peripheral blood (Fig. 1B, left and middle panels, respectively). The expression pattern of CCR2 and the adhesion molecules CD11b and CD62L on monocytes, however, appeared normal in Gαi2 KO, Gαi3 KO, and WT B6 mice (Fig. 1C).
Altered distribution of tissue macrophages and blood monocytes in the genetic absence of Gαi2 but not Gαi3. (A) Macrophages were obtained from the indicated organ sites of B6 and Gαi2, or Gαi3 KO mice and counted as described in Materials and Methods. (B) Blood was collected and analyzed for circulating monocytes by an automatic ABC blood counter (left panel) or by counting beads in respect to monocyte markers, CD115 and Ly6C (middle and right panels, respectively). (A and B) The mean cell count ± SEM (n = 15 mice for each group) is shown. (C) Peripheral blood monocytes from B6 (upper panels), Gαi2 KO (middle panels), and Gαi3 KO mice (lower panels) were analyzed for the surface expression of CD11b, CD62L, and CCR2 by flow cytometry. One representative experiment out of three is shown. *p < 0.05, **p < 0.001, ***p < 0.0005.
At least two distinct peripheral blood monocyte subpopulations are recognized in mice (34), and the Ly6Chi inflammatory monocyte subset specifically expresses CCR2 and CD62L (35). Increased accumulation of this Ly6Chi (Fig. 1B, right panel) subpopulation was also evident in Gαi2 KO mice, suggesting that Gαi2—but not Gαi3—is required for homeostatic and inflammation-induced migration of monocytes and macrophages.
Therefore, we next examined the relative contribution of Gαi2 and Gαi3 in mediating the recruitment of macrophages in two models of acute inflammation, namely thioglycollate-induced peritonitis and LPS-induced alveolitis. Macrophage accumulation to the peritoneum and lung was studied at day 4 by counting macrophage numbers obtained after peritoneal and bronchoalveolar lavages. As shown in Fig. 2A, Gαi2 KO mice showed significantly decreased macrophage elicitation in response to both stimuli, whereas Gαi3 KO mice displayed either normal (thioglycollate) or increased (LPS) macrophage numbers.
Attenuation of agonist-induced recruitment of Gαi2- but not Gαi3-deficient macrophages in vivo and in vitro. (A) Acute inflammation was induced in peritoneum and lung of B6 WT and Gαi2 or Gαi3 KO mice by i.p. injection of 4% thioglycollate (left panel) or intratracheal application of 20 μg LPS (right panel). Control mice received PBS. Agonist-induced macrophage accumulations were analyzed at day 4. Results were obtained from 6–8 mice in each group (*p < 0.05). (B) C5a- and CCL2-induced in vitro chemotaxis was determined by Transwell migration assays of PM cells from the indicated mice (left panel). Data are expressed as mean ± SEM from five experiments performed in duplicate. Surface expression of CCR2 and C5aR was analyzed by flow cytometry (right panel). Representative results from four separate experiments are shown. (C) C5a-induced morphological changes toward an asymmetric cell shape were used as a measure of cell polarization. Both qualitative (left; examples of polarized cells are indicated by the arrows; original magnification ×40) and quantitative (right) examinations of PM cells are shown. Quantitative data are expressed as mean ± SEM from three independent experiments performed in triplicate. *p < 0.05, **p < 0.001.
We also tested the migratory capacity of Gαi-deficient macrophages in response to CCL2 and C5a in standard Transwell chemotaxis assays in vitro (Fig. 2B). Similar to their defect to migrate into inflamed tissue in vivo, PM from Gαi2 KO mice showed reduced CCL2/C5a-induced chemotaxis by CCR2 and C5aR in vitro, as compared with Gαi3 KO and WT PM cells (Fig. 2B, left panel). Recent analyses of lymphocyte migration suggested Gαi-dependent changes in the expression of CC and CXC chemokine receptors regulating the movement of CD4 and CD8 T cells (18). Examining Gαi-deficient PM cells, we analyzed in this study the expression status of C5aR and CCR2 on macrophages by flow cytometry. We found no altered cell surface appearance of the two receptors on these cells from Gαi2 KO and Gαi3 KO mice, as compared with WT controls (Fig. 2B, right panel). Irrespective of normal C5aR cell surface expression, fluorescence microscopic analysis of Gαi-deficient PM cells revealed an impaired C5a-induced cell polarization response in the selective absence of Gαi2 but not Gαi3 (Fig. 2C). Because C5aR is known to couple to both Gαi2 and Gαi3, this result indicates a specific role of Gαi2 in C5aR signaling that leads to actin rearrangement and cell migration, which cannot be compensated by Gαi3.
Normal LPS-induced cytokine release and C5aR-regulated erythrophagocytosis in Gαi2-deficient mice
Gαi-dependent signaling has been shown to play a crucial role in LPS-induced cell activation (36). Thus, we next examined the capacity of Gαi2- and Gαi3-deficient macrophages to respond to LPS by measuring the release of inflammatory mediators. PM cells were cultured overnight, incubated with LPS (1 μg/ml) for 12 h, and analyzed for the production of TNF-α and CXCL2 by ELISA. High levels of these cytokines were detectable in Gαi2 KO PM cells upon LPS stimulation (Fig. 3A). Gαi3 KO PM cells were equally effective, whereas inactivation of both Gαi isoforms by PTx caused an impaired response (Fig. 3A). Taken together with the observation that macrophages express Gαi2 in a stoichiometric excess over Gαi3 (23), the results indicate quantitative differences but functional redundancy of the two Gαi isoforms in LPS-induced macrophage activation.
LPS-induced cytokine release and C5aR-regulated erythrophagocytosis in mice lacking Gαi2 or Gαi3. (A) PM cells from B6 WT and Gαi2 or Gαi3 KO mice were activated for 12 h with LPS and pretreated or not for 2 h with PTx (LPS + PTx) and analyzed for TNF-α (left panel) and CXCL2 (right panel) production by ELISA. Data are expressed as mean ± SEM from two experiments performed in triplicate. (B) Erythrophagocytosis by Kupffer cells (KC) was induced in WT B6, Gαi2 KO, Gαi3 KO, or C5aR KO mice by i.p. injection of 150 μg of the pathogenic IgG2a 34-3C mAb, and liver phagocytosis was assessed on day 2. Results were obtained from four to six mice in each group. Qualitative (left panel; original magnification ×40) and quantitative (right panel) examinations of liver H&E sections from the indicated anemic mice are shown. **p < 0.001.
To further assess the specific role of Gαi2 or Gαi3 in macrophage-mediated immune reactions in vivo, we used the IgG-mediated model of hemolytic anemia. WT, Gαi2 KO, or Gαi3 KO mice were injected with 150 μg anti-erythrocyte band 3 (34-3C) self-reactive Abs, and ingestion of IgG-bound RBC by liver macrophages was monitored as described previously (29, 37, 38). In this type of autoimmune disease model, cellular RBC destruction is induced by simultaneous activation of IgG FcRs and the Gαi-coupled C5aR on Kupffer cells (26, 28). In agreement with the previous observations, WT B6 mice showed a detectable phagocytotic response upon i.p. injection of the 34-3C autoantibody at day 2, whereas KO mice lacking C5aR (C5aR KO) showed reduced RBC destruction (Fig. 3B). In contrast to C5aR KO mice, liver phagocytosis was not significantly ameliorated in Gαi2- or Gαi3-deficient animals (Fig. 3B). These results suggest that C5a-induced chemotaxis in vitro (see Fig. 2B) but not C5aR-regulated phagocytosis of IgG-bound RBCs in vivo is mediated by a Gαi2-selective mechanism.
KD of Gαi2 in macrophages causes defective chemotaxis that is rescued upon Gαi2 but not Gαi3 transfection
Among various Gαi-dependent macrophage functions tested, we identified chemotactic migration to be selectively regulated by Gαi2. To further strengthen our findings on the significance of Gαi2 specificity for cellular migration, we generated a stable KD of Gαi2 in RAW 264.7 macrophages by Gαi2 shRNA expression. These Gαi2 KD macrophages showed markedly reduced Gαi2 mRNA levels (Fig. 4A, left panel) and a reduction of protein expression by ∼75% (Fig. 4B, left panel). Interestingly, we neither observed a concomitant upregulation of Gαi3-specific transcripts (Fig. 4A, right panel) nor of Gαi3 protein (Fig. 4B, right panel). Corresponding to decreased expression of Gαi2, the expression levels of the predominant Gβ isoforms, Gβ1 and Gβ2, were also decreased in Gαi2 KD macrophages (Fig. 4C).
Generation of a Gαi2 KD in macrophages. Detection of Gαi2 KD efficiency in RAW 264.7 cells. (A) Both mock and Gαi2 KD macrophages were assayed for altered Gαi2 and Gαi3 mRNA by real-time RT-PCR. Results are expressed as relative Gαi mRNA expression ± SEM from three independent experiments. (B and C) Protein expression of Gαi and Gβ subunits following treatment of RAW 264.7 cells with shRNA against Gαi2. Representative immunoblots of Gαi2 (B, left), Gαi3 (B, right), Gβ1, and Gβ2 (C) are shown. The graphs show densitometric analysis of the immunoblots obtained from six independent experiments and data are expressed as mean ± SEM. *p < 0.05, **p < 0.01.
Having established selective inhibition of Gαi2 protein expression, we then analyzed the Gαi2 KD effect on cytokine production and cell migration. Comparable to the situation with primary PM cells (Fig. 3A) and mock-transfected cells (Fig. 5A), LPS-induced production of TNF-α and CXCL2 was still detectable in Gαi2 KD macrophages (Fig. 5A). Importantly, however, both chemokinesis (Fig. 5B, white bars) and C5a-induced chemotaxis (Fig. 5B, black bars) were markedly reduced in Gαi2 KD cells as compared with mock-transfected cells confirming a selectivity of Gαi2 for this cellular response. To rigorously check our finding on Gαi2-selective regulation of chemotactic migration, we finally performed rescue experiments in KD cells transfected with Gαi2 or Gαi3 cDNAs. Importantly, Gαi2 KD macrophages regained the ability for C5a-induced chemotaxis only in the presence of cDNA expression vectors encoding for Gαi2 (pGαi2) but not Gαi3 (pGαi3) (Fig. 5C).
Gαi2 KD in macrophages leads to defective chemotaxis that is rescued in Gαi2 but not Gαi3-transfected cells. (A) Gαi2 KD and mock-transfected RAW 264.7 macrophages were stimulated with LPS or received PBS (control), and CXCL2 (left panel) and TNF-α (right panel) cytokine production was determined by ELISA. *p < 0.05. (B) The Gαi2 function dependency of C5aR-mediated chemotaxis was assayed with C5a using Gαi2 KD and mock-transfected macrophages. Data are expressed as the mean ± SEM from three independent experiments performed in triplicate. (C) Chemotaxis of Gαi2 KD or mock control macrophages 48 h after external delivery of Gαi2 or Gαi3 cDNA. The experiments were performed three times in triplicate with similar results. *p < 0.05, **p < 0.001.
Discussion
We have demonstrated previously that, despite the high homology between Gαi2 and Gαi3, the inflammatory cascade in the lung Arthus reaction critically depends on Gαi2 but not Gαi3 (23). The present work significantly extends these findings by providing novel evidence that Gαi2 plays a critical role in the homeostatic distribution of tissue macrophages (Fig. 1) as well as inflammation-induced migration in vivo (Fig. 2A). The in vitro observation that Gαi2 but not Gαi3 is essential for optimal CCL2- and C5a-induced migration (Fig. 2B) further indicates that the reduced recruitment of macrophages in acute inflammation in Gαi2-deficient mice may be due to a Gαi2-specific defect in directed chemotaxis of these cells. However, not all effector functions of macrophages are mediated by a Gαi2-specific mechanism, and Gαi3 is able to substitute for Gαi2 in C5aR-regulated phagocytosis as well as Gαi-dependent cytokine production (Fig. 3). The latter observation is in agreement with the findings of Fan et al. (39), who suggested that Gαi2 and Gαi3 are both involved in LPS-induced cell activation. However, we also note that Gαi3-deficient mice exhibit an increased accumulation of macrophages in lung tissue in response to LPS (Fig. 2A), indicating that Gαi3 may negatively influence certain Gαi2 signaling events. Further studies will be needed to test the possibility of an antagonistic interplay between Gαi2 and Gαi3 in tissue infiltration of macrophages, as has been suggested before for CXCR3-mediated signaling in T cells (40).
Macrophages are present in all tissues, and many of these cells are replaced continually from BM-derived monocytes. A recent study suggested that tissue macrophages can also develop from a nonhematopoietic origin in a Myb-independent manner (41). We do not discriminate between the different origins of macrophages, but we found a reduction of total macrophage counts in lung, liver, and peritoneum in the absence of Gαi2 (Fig. 1A). Interestingly, Gαi2-deficient mice show increased blood monocyte numbers (Fig. 1B), thus indicating an involvement of Gαi2 in the transit of monocytes from blood into tissues. Previous studies suggested a role of CCL2 in the homeostatic migration of the CCR2-positive Ly6Chi monocyte subpopulation (42, 43). CCR2-deficient mice showed reduced Ly6Chi monocyte numbers in the blood but not in the BM, indicating an involvement of CCR2 in steady-state BM egress (44). Gαi2 deficiency, while causing reduced CCR2-mediated macrophage chemotaxis (Fig. 2B), is associated with increased numbers of peripheral blood Ly6Chi monocytes (Fig. 1B), indicating that homeostatic emigration of this inflammatory monocyte subset from the BM can occur in the absence of Gαi2-mediated CCR2 activation.
Changes in the expression of GPCRs are thought to be critical for the migration of immune cells. Previously, a Gαi protein requirement in the regulation of CC and CXCR expression was reported for T lymphocytes (18). Gαi2-negative CD4 and CD8 T cells each displayed reduced levels of CCR7 and CXCR4 as well as CD62L. Importantly, our results demonstrate comparable expression profiles of CD11b, CD62L, and CCR2 on blood monocytes (Fig. 1C) and of CCR2 and C5aR on peritoneal macrophages (Fig. 2B) in WT, Gαi2, and Gαi3 KO mice. Although Gαi2-deficient macrophages normally express CCR2 and C5aR on their cell surfaces, they display a severe defect of CCL2- and C5a-induced chemotaxis (Fig. 2B). Actin reorganization and cell polarization are required events for directional cell movement (45, 46). To our knowledge, our data now show for the first time that both C5a-induced remodeling of the actin cytoskeleton and chemotaxis are strictly dependent on the presence of Gαi2 in macrophages (Fig. 2C).
The diminished chemotactic response exhibited by mice lacking Gαi2 but not Gαi3 may depend on differences in the total amount of the two Gαi proteins. Macrophages, like most other cell types (16), express Gαi2 in much higher concentrations than Gαi3 (23). Importantly, however, low Gαi3 can compensate for the lack of high Gαi2 in LPS-induced macrophage activation and C5aR-regulated erythrophagocytosis (Fig. 3), suggesting that structural rather than quantitative differences of Gαi2 and Gαi3 determine the unique function of Gαi2 in the regulation of macrophage migration. Our RNA interference data support this conclusion. KD of Gαi2 in RAW 264.7 macrophages, while not affecting Gαi3 expression (Fig. 4) and LPS-induced production of TNF-α and CXCL2 (Fig. 5A), causes decreased Gβ1 and Gβ2 protein levels (Fig. 4) and defective chemotaxis (Fig. 5B). Most significantly, only the overexpression rescues of Gαi2 but not Gαi3 restore migration of Gαi2-silenced RAW 264.7 cells (Fig. 5C).
In summary, we have shown that Gαi2 is a crucial regulator of macrophage migration and the specific KD of Gαi2 reflects the phenotype of Gαi2-deficient PM cells with respect to chemotaxis. Importantly, the rescue of Gαi2 but not of Gαi3 could restore this phenotype providing clear evidence for an isoform-specific role of Gαi2. In contrast, Gαi3 was able to substitute for Gαi2 regarding the LPS-induced cytokine production and the phagocytotic activity of liver macrophages. These findings serve as an important basis for the identification of the underlying mechanisms of redundancy versus isoform specificity of Gαi2 and Gαi3. Such mechanisms could be based on a different subcellular localization of these isoforms (16, 47), structural differences in regions different from the receptor- or effector-binding domains, as well as the total amount of Gαi proteins in the cells, with the latter obviously not playing a decisive role in our model.
Disclosures
The authors have no financial conflicts of interest.
Acknowledgments
We thank all members of the Gessner and Nürnberg laboratories for helpful discussions.
Footnotes
This work was supported by Deutsche Forschungsgemeinschaft SFB 587 and GE 892/11-1 (to J.E.G.), Nu 53/7-1 (to B.N.), SFB 612 (to R.P.P. and B.N.), and BE 2813/1-2 (to S.B.-H.). K.W. is a recipient of a graduate fellowship from Deutsche Forschungsgemeinschaft GRK 705 and trained by the international Ph.D. program of the Medical School Hannover.
Abbreviations used in this article:
- B6
- C57BL/6
- BM
- bone marrow
- C5aR
- C5a anaphylatoxin receptor
- CD62L
- L-selectin
- GPCR
- G protein-coupled receptor
- KD
- knockdown
- KO
- knockout
- PM
- peritoneal macrophage
- PTx
- pertussis toxin
- shRNA
- short hairpin RNA
- WT
- wild-type.
- Received March 21, 2012.
- Accepted May 11, 2012.
- Copyright © 2012 by The American Association of Immunologists, Inc.
References
In this issue
