Requirement for Vav Proteins in Post-Recruitment Neutrophil Cytotoxicity in IgG but Not Complement C3-Dependent Injury¹

Neutrophils express two major opsonic receptors, FcγR and complement receptor 3, also known as the β₂ integrin Mac-1 (CD11b/CD18), that engage IgG and complement fragment iC3b-opsonized targets, respectively (1). Recent data suggest that tissue injury associated with IgG and/or complement deposition in tissue may be attributed to the activity of these receptors. A deficiency in activating FcγRs protects from immune complex (IC)⁴-mediated inflammation in the kidney (2, 3), joints (4), skin (5, 6), lung (7, 8), and heart (9). A deficiency in Mac-1 is associated with a reduction in complement C3-mediated inflammation in the kidney (10) and skin (11, 12). Both FcγRs and Mac-1 promote neutrophil recruitment and cytotoxicity (13, 14, 15). FcγR-induced mast cell secretion of chemokines and cytokines promote neutrophil recruitment and ensuing sequelae (5, 16). FcγRs on neutrophils may also directly contribute to the recruitment to deposited ICs (13). Mac-1 binding of its endothelial ligands, ICAM-1 and 2, promotes neutrophil recruitment in a number of inflammatory models (14). The importance of Mac-1-mediated neutrophil cytotoxicity was defined in a model of thrombohemorrhagic vasculitis by demonstrating the absence of hemorrhage in Mac-1-deficient mice despite normal neutrophil recruitment (12). However, the contribution of neutrophil FcγRs to cytotoxicity in vivo has been difficult to ascertain because genetic deficiency in activating FcγRs leads to a decrease in neutrophil recruitment in all IgG mediated-inflammation models studied to date (3, 8, 17, 18, 19). Thus, for FcγRs, it remains unclear whether the ability of these receptors to activate neutrophil cytotoxicity vs their ability to simply bind tissue-deposited ICs to allow cellular recruitment is important in tissue injury in vivo.

The Vav guanine exchange factors link diverse cell surface receptors to activation of Rho GTPases including RhoA, Cdc42, and Rac (20). Studies in primary neutrophils have suggested that Vav proteins 1 and 3, the predominant Vav isoforms in phagocytes, are required for the function of both the β₂ integrins and FcγRs (21, 22). Vav 1,3-deficient (−/−) neutrophils exhibited defects in Mac-1-dependent sustained adhesion to serum-coated plates and the ingestion of serum-opsonized Escherichia coli, which were associated with altered Rac1 and RhoA activation (21). Rac proteins 1 and 2, the predominant Rac isoforms in phagocytes, are activated by β₂ integrin ligation (23), but their role in β₂ integrin neutrophil functions remains poorly understood. Vav 1,3^−/− and Rac 1,2^−/− neutrophils exhibited severe defects in several FcγR-mediated functions including IgG-mediated phagocytosis, adhesion, and NADPH oxidase-dependent reactive oxygen species generation, suggesting that Vav regulation of Rac is required for these functions (22, 24). Despite the evidence that Vav proteins link FcγRs, Mac-1, and more recently the pattern recognition receptor TLRs (25) to downstream effector responses, the in vivo relevance of these findings is not known.

In this study, we exploited knockout mice and neutrophil reconstitution approaches to examine the contribution of Vav and Rac GTPases in the IgG-mediated reverse passive Arthus (RPA) reaction, a type III hypersensitivity response leading to edema and hemorrhage that requires FcγR but did not depend on Mac-1. In addition, the role of Vav and Rac was evaluated in the local Shwartzman reaction (LSR), a model of thrombohemorrhagic vasculitis characterized by neutrophil accumulation, hemorrhage, fibrin deposition, thrombosis, and elastase release that is dependent on neutrophil Mac-1 and complement C3 deposited within the vessel wall (12, 26). This was coupled with an in vitro assay evaluating the role of Vav and Rac in Mac-1/complement iC3b-dependent neutrophil phagocytosis. We provide evidence that Vav in neutrophils is essential for FcγR/IgG-dependent tissue injury in the skin and lung in vivo at a step downstream of neutrophil recruitment. Rac-deficient mice were also resistant to the skin RPA reaction. The uncoupling of neutrophil recruitment and cytotoxicity demonstrates that FcγR signaling activates neutrophil effector functions that lead to tissue injury in IC disease. In contrast, neither Vav nor Rac were required for Mac-1/complement-dependent neutrophil cytotoxic functions either in vitro or in vivo. These data indicate selectivity for the Vav/Rac pathway in IgG-mediated tissue injury.

C57BL/6J wild-type (WT) mice served as controls for Mac-1^−/−/C57BL/6 mice (12, 27), and Fcγ chain^−/−/C57BL/6 mice (Taconic Farms). The sources of C57BL/6J/129Sv Vav1^−/−, Vav3^−/−, and Vav 1,3^−/− mice (28, 29) as well as their WT cohorts are described previously. To minimize the development of strain-dependent genetic differences between lines, we produced WT and single and double Vav isoform-deficient mouse lines from Vav1,3 heterozygous crosses, which were then bred separately for three generations. Previously generated and characterized conditional Rac1^−/−, Rac2^−/−, and Rac1,2^−/− mice were maintained as a mixed strain population (C57BL/6 and 129Sv) as previously described (30). Homozygous lines were derived from heterozygous crosses as described for Vav^−/− mice and were bred separately for five to seven generations. Mice were maintained in a virus- and Ab-free facility at the Eugene Braunwald Research Center animal housing facility at Brigham and Women’s Hospital. Mice used for each experiment were between 8 and 20 wk of age and age and sex matched. All experiments in this study were approved by the Harvard Medical School Animal Care and Use Committee.

For in vitro studies, bone marrow mouse neutrophils (BMN) were isolated as described previously (22). For reconstitution of mice with neutrophils in the RPA reaction, neutrophils were purified from bone marrow as previously described (12), which resulted in a population that was 95% pure as assessed by Wright-Giemsa stains of cytospun samples.

For cutaneous RPA reactions, anesthetized age-matched female mice were injected intradermally with rabbit IgG anti-chicken egg albumin Abs (60 μg/30 μl; Cappel), followed immediately thereafter by an i.v. injection of chicken egg OVA (500 μg/mouse; Sigma-Aldrich). The intradermal injection of PBS served as a negative control. In cases where edema was measured, the solution of chicken egg albumin contained 0.15% Evans blue dye (Sigma-Aldrich). The skin was harvested 4 h later.

For i.v. reconstitution, 7.5 × 10⁶ BMN were given through the tail vein. One hour later, the RPA reaction was initiated as described in the previous section. The skin was harvested 3.5 h later.

Edema was evaluated by measuring the vascular leak of Evans Blue. Harvested skin was incubated with dimethylformamide and Evans blue in the supernatant was quantified by measuring the absorbance at 595 nm. Specific edema formation was measured by subtracting the absorbance in the PBS-injected site from that of the IgG-challenged site in the same mouse. The amount of hemorrhage was assessed 8 h after IC challenge and quantified as described previously (31). The data are reported as the average diameter of the purpuric spot of anti-OVA IgG minus the PBS- injected site of the same mouse.

Skin tissues were harvested 2 or 8 h after IC challenge. Tissues were fixed in 3.5% paraformaldehyde and then paraffin embedded. Six-micrometer sections were deparaffinized and subjected to the chloroacetate esterase reaction to identify neutrophils. Neutrophil infiltration was evaluated by counting extravascular neutrophils within the site of injection as described elsewhere (12).

The RPA reaction in the lung was induced as described previously (32). Briefly, 800 μg of chicken OVA (Sigma-Aldrich) in 200 μl of PBS was injected via the tail vein in anesthetized mice. One hundred micrograms of rabbit anti-chicken OVA (Cappel) in 100 μl of PBS was instilled into the trachea immediately thereafter. At the end of 4 h, mice were euthanized and bronchoalveolar lavage (BAL) was performed with 1 ml of HBSS plus EDTA. Sham-treated mice were injected intratracheally with buffer alone.

RBC and total white blood cells in the BAL fluid were counted using a hemacytometer. A differential count was obtained on Giemsa-stained cytospun samples. Total BAL neutrophils were determined by multiplying the white blood cell counts by the percentage of neutrophils. Albumin in the BAL was used as an indicator of lung edema and was determined using an ELISA-based method (Bethyl) as previously described (33).

Age-matched Mac-1^−/− male mice were i.v. reconstituted with 5 × 10⁶ bone marrow donor neutrophils (Vav WT, Vav1,3^−/−, Rac WT, or Rac1,2^−/−) and were subjected to the LSR as described previously (12). In select experiments, skin air pouches were generated and LSR was induced in the pouches (12). The air pouch was lavaged with 3 ml of ice-cold PBS and assayed for elastase release.

An equal number of neutrophils and volume of lavage fluid were added to 24-well plates. Twenty micromolar fluorogenic elastase substrate N-methoxysuccinyl-Ala-Ala-Pro-Val-7-amido-4-methylcoumarin (Sigma-Aldrich) was used to measure elastase activity as previously described (12). The fluorescence at 2-min intervals over a 60-min time period was monitored using a plate-based fluorometer with 380 nm excitation and 460 nm emission (Molecular Devices) at 37°C. Samples were then lysed in Triton X-100 to calculate total cellular elastase activity. Elastase activity was proportional to the slope of the increase in fluorescence over time. Percent release of elastase in response to neutrophil adhesion was calculated by dividing elastase activity by the elastase activity detected following lysis of the sample.

The procedure was originally described to evaluate IgG-mediated phagocytosis (34). A similar approach was adopted but iC3b opsonization was used (35). Briefly, 20% SRBC (Valeant Pharmaceuticals) in PBS (v/v) were opsonized with subagglutinating rabbit IgM anti-sheep RBC (RDI) at room temperature (RT) for 45 min. IgM-coated SRBC were pelleted, washed, and incubated with 20% serum of C5-deficient human serum (Sigma-Aldrich) in PBS with Ca²⁺ and Mg²⁺ for 30 min. After a wash with PBS, IgM-C3 SRBC (C3-RBC) were resuspended in PBS.

To activate β₂ integrins, 1 × 10⁶ bone marrow mouse neutrophils were treated with 1 μg/ml PMA for 15 min at RT in PBS with 1 mM CaCl₂ and 0.5 mM MgCl₂. RBC were then spun gently, layered on top of the neutrophil pellet, and then respun for an additional 5 min at RT to promote binding. Neutrophils and C3-RBC were then resuspended using a wide bore pipette and incubated at 37C for 1 h. The cell mixture was pelleted and RBC were lysed with H₂O followed by the addition of cold PBS. The cells were repelleted, resuspended in PBS, and visualized under the microscope to assess the total number of internalized RBC in 100 neutrophils.

To measure binding of RBC to neutrophils, RBC were labeled with PKH fluorescence dye as per the manufacturer’s instruction (Sigma-Aldrich). PKH-labeled SRBC were then opsonized with IgM and/or IgM and C3. WT and Mac-1 mutant neutrophils were preincubated with allophycocyanin- conjugated anti-Gr-1 Ab (BD Pharmingen) and pretreated with PMA (1 μg/ml) for 15 min at RT, followed by incubation with opsonized or nonopsonized RBC. The mixture was pelleted, resuspended, and incubated on ice for 30 min. The samples were directly analyzed using a flow cytometer and gated for both Gr-1- and PKH-positive signals. The percentage of Gr-1-positive neutrophils that costained with PKH was quantitated.

Quantitative data were represented as mean ± SEM or SD as indicated. Evaluation of statistical significance was performed using a two-tailed unpaired Student’s t test. Values of p < 0.05 were considered statistically significant.

Vav-deficient mice and their WT cohorts were subjected to the cutaneous RPA reaction to examine the role of Vav in FcγR-dependent functions in vivo. The reaction, induced by an intradermal injection of Ab and i.v. delivery of Ag, is characterized by neutrophil accumulation, edema, and hemorrhage. It is dependent on FcγRs (18) and neutrophil function (36), as confirmed in our studies using Fcγ chain-deficient (Fcγ chain^−/−) mice, which lack all activating FcγRs and neutrophil immunodepletion approaches, respectively (data not shown). WT mice and mice deficient in Vav 1, Vav 3, or Vav 1,3 were subjected to the RPA reaction and evaluated for the development of edema, as assessed by the leakage of Evans blue i.v. injected into the mice at the time of Ag delivery. At 4 h after induction of the RPA, Vav 3^−/− mice exhibited a 50% reduction in Evans blue extravasation in comparison to WT mice, whereas a deficiency in Vav1 alone had no effect. Importantly, Vav 1,3^−/− mice exhibited minimal edema compared with WT mice (Fig. 1, A and B). Furthermore, hemorrhage development was significantly reduced in Vav 1,3^−/− animals compared with WT counterparts (Fig. 1 C).

To determine whether Vav on neutrophils was responsible for edema, Vav 1,3^−/− mice were i.v. injected with freshly isolated WT or Vav 1,3^−/− neutrophils and the reconstituted mice were subjected to the RPA reaction. Adoptive transfer of WT neutrophils into Vav 1,3^−/− mice resulted in significant edema while transfer of Vav 1,3^−/− neutrophils resulted in minimal leakage (Fig. 2,A). These data indicate that Vav 1,3 in neutrophils play a critical role in the development of IC-induced tissue injury. Next, the RPA reaction was evaluated in mice lacking Rac 2 or both Rac 1 and 2 to explore the possibility that the phenotype observed in Vav 1,3^−/− mice could be explained by a defect in a functional interaction between Vav and Rac proteins in neutrophils in vivo. Mice deficient in Rac 2 or Rac 1 alone developed edema while Rac 1,2^−/− mice were completely protected (Fig. 2, B and C). This suggests that Rac1 and Rac2 functionally overlap. We have previously shown that FcγR cross-linking induced Rac activation is significantly reduced in Vav 1,3^−/− neutrophils (22). Thus, our in vivo data infer that Vav regulation of Rac 1,2 activity is required for IC-induced tissue injury. A reduction in edema may be a consequence of diminished neutrophil accumulation in tissue and/or defective cytotoxic functions (37). Histological analysis revealed no significant difference in neutrophil accumulation in Vav 1,3^−/− mice compared with WT mice, and a partial reduction in neutrophil influx was observed in Rac 1,2^−/− mice (Fig. 2 D). Thus, Vav is not required for neutrophil influx while Rac may contribute, but is not essential for neutrophil accumulation in this model.

FcγR engagement leads to Mac-1 activation and cooperation of these two receptors is required to sustain FcγR-mediated murine neutrophil adhesion (10) and other phagocytic functions (38). Given these reports and our previous data that Vav proteins are required for Mac-1-dependent functions in vitro (21), we examined the contribution of Mac-1 to the RPA reaction. Mac-1 deficiency resulted in an increase in Evans blue leakage compared with WT counterparts (Fig. 3). Although the mechanisms for the increase needs further investigation, these results demonstrate that Mac-1 is not required for the development of the RPA reaction and rules out the possibility that Vav proteins promote IgG-induced tissue injury by modulating Mac-1.

To examine Vav in a more clinically relevant model of IC disease, Vav^−/− mice were subjected to the RPA reaction in the lung, which shares several features of acute lung injury in patients (39) and requires neutrophils and FcγRs (40, 41). The phenotype in Vav^−/− mice was similar to that observed in the skin RPA model. Although mice with Vav 1 deficiency alone were similar in susceptibility to WT mice, Vav 1,3^−/− mice were protected from acute lung injury as assessed by a significant reduction in albumin and RBC content in the BAL fluid, which are readouts of edema and hemorrhage, respectively (Fig. 4, A and B). A partial reduction in edema and hemorrhage was observed in Vav 3^−/− mice. The number of infiltrated neutrophils in the BAL fluid of WT-, Vav 1,3-, Vav 1-, and Vav 3-deficient mice was equivalent (Fig. 4, C and D). Thus, similar to the cutaneous RPA reaction, Vav was not required for IgG-induced neutrophil accumulation but was critical for IgG/FcγR-induced tissue injury.

Vav is required for Mac-1/complement-dependent functions in vitro including sustained adhesion to serum-coated plates and phagocytosis of serum-opsonized E. coli (21). To evaluate the physiological relevance of these findings in vivo, we examined the contribution of Vav and Rac on neutrophils to the development of the LSR, a response that is critically dependent on Mac-1 expressed on neutrophils, complement C3 deposition within the vessel wall, and neutrophil elastase release (12). The LSR does not require FcγRs since Fcγ chain-deficient mice display a normal LSR response (D. Mekala and T. N. Mayadas, unpublished data). The LSR is induced by the intradermal injection of LPS followed 24 h later by the delivery of TNF at the same site, which results in hemorrhage that is associated with neutrophil accumulation and neutrophil elastase release. Neutrophils isolated from either Vav 1,3^−/−, Rac 1,2^−/− mice, or their respective WT counterparts were i.v. injected into Mac-1-deficient (Mac-1^−/−) mice that are resistant to developing the LSR. The reconstituted mice were then subjected to LSR. This approach allowed us to evaluate the contribution specifically of the Vav/Rac pathway in neutrophils in Mac-1/complement-dependent neutrophil cytotoxicity in vivo. Mac-1^−/− mice injected with Vav 1,3^−/− or Rac 1,2^−/− developed hemorrhage to a similar extent as mice reconstituted with WT neutrophils (Fig. 5, A and B), whereas Mac-1^−/− mice given Mac-1-deficient neutrophils failed to develop disease (Fig. 5, A and B) as previously described (12). Extracellular elastase activity was quantitated in reconstituted Mac-1^−/− mice by real-time analysis of elastase substrate cleavage in neutrophils and fluid retrieved from air pouches in which the LSR was induced. Elastase activity was comparable in Mac-1^−/− mice reconstituted with Vav 1,3^−/− or WT neutrophils (Fig. 5 C). Thus, Vav and Rac proteins on neutrophils are dispensable for the development of hemorrhage, a major aspect of disease in the LSR as well as elastase release, which contributes to tissue injury in this model (12).

In light of our data that Vav- and Rac-deficient neutrophils were able to promote LSR, we revisited complement-mediated phagocytosis in vitro to reassess the role of these proteins in this process. Vav was required for neutrophil phagocytosis of complement-opsonized E. coli in vitro (data not shown), as published previously (21). However, Vav and Rac proteins were dispensable for PMA-mediated neutrophil uptake of complement C3-opsonized RBC (C3-RBC). C3-RBC were generated by incubating IgM-coated RBC with fresh serum deficient in complement C5 (Fig. 6,A), which results in complement activation and subsequent iC3b deposition but avoids RBC lysis (42, 43). Unlike the phagocytosis of serum-opsonized E. coli, which is only partially Mac-1 dependent, uptake of C3-RBC depended exclusively on Mac-1 and complement opsonization. That is, Mac-1^−/− neutrophils or WT cells incubated with unopsonized RBC exhibited no detectable phagocytosis (Fig. 6,B). Furthermore, phagocytosis was only detected when cells were PMA stimulated (data not shown). Previous studies in human monocytes and neutrophils indicate that C3-RBC bind to Mac-1 on the surface of resting cells but that subsequent phagocytosis requires PMA (44, 45, 46). To examine whether this also applies to murine BMN, we exploited FACS analysis to examine the binding of the fluorophore-conjugated C3-RBC to the surface of neutrophils in the absence and presence of PMA. Resting WT neutrophils exhibited no binding to either RBC alone or C3-RBC. Upon PMA stimulation, binding of C3-RBC significantly increased. This was Mac-1 dependent, because Mac-1^−/− BMN exhibited no significant binding to C3-RBC under these conditions (Fig. 6 C). Thus, PMA stimulation in murine neutrophils promotes Mac-1 binding and subsequent phagocytosis of C3-RBC. Together our studies suggest Vav and Rac are not required for PMA-induced binding and phagocytosis of C3-RBC, which are strictly Mac-1-dependent processes. The results of this in vitro assay correlated well with the lack of effect of Vav and Rac deficiency in the development of the LSR, which is dependent on Mac-1 and complement C3 (12).

A deficiency in activating FcγRs, by virtue of deletion of the common γ-chain subunit, results in protection from several IgG-mediated, neutrophil-dependent diseases, including acute glomerulonephritis, arthritis, acute lung injury, and autoimmune skin disorders (3, 4, 7, 8, 17, 18). In all of these models, a reduction in neutrophil influx was observed in mice deficient for FcγRs, thus precluding conclusions about the role of these receptors in subsequent steps of neutrophil activation that lead to tissue injury. Our data using mice with defects in neutrophil FcγR signaling preserve the ability of FcγRs to bind ICs, thus allowing us to distinguish these different roles for the receptors in inflammatory disease. Thus, our work suggest that FcγR binding of tissue-deposited IC is required for neutrophil migration, but the signaling reactions from the FcγRs is critical in cellular activation leading to tissue injury (Fig. 7). The restoration of disease in Vav 1,3^−/− recipients reconstituted with WT but not Fcγ chain^−/− neutrophils provides compelling evidence that FcγR signaling through Vav proteins in neutrophils is responsible for IC-induced tissue injury (Fig. 7). However, we cannot rule out the possibility that the two mutations, Vav and FcγR, disrupt two entirely separate parallel pathways. Previous characterizations of gene knockout mice including Fcγ chain (18), C5a (47), urokinase-type plasminogen activator receptor (48), ICAM-1 (49), L-selectin (31), Syk (50), CD18 (51), and elastase-cathepsin-G double knockout (52) in the skin and/or lung RPA reaction have shown that reductions in edema and hemorrhage correlate closely with reductions in the numbers of transmigrated neutrophils in tissue. However, this is not the case when a signaling deficiency involves the Vav proteins, because neutrophils lacking these proteins are able to migrate normally into the tissue but fail to become activated. The uncoupling of FcγR-mediated neutrophil transmigration from cytotoxicity at the level of Vav provides the first genetic evidence that these two processes can be dissociated. In contrast, Vav/Rac proteins were not required for complement C3 and complement receptor Mac-1-mediated inflammation, which demonstrates selectivity of the Vav/Rac axis for IgG-mediated cytotoxicity in vivo.

Neutrophil diapedesis to sites of IgG-mediated tissue injury in the skin and lung has been associated with vessel injury (8, 51). Neutrophil accumulation in the BAL is likely dependent on macrophage and partially on mast cell functions since elimination of these cells reduced neutrophil accumulation in the alveolar space (53). In the skin, FcγRIII and C5aR on mast cells and leukotrienes have been primarily implicated in promoting neutrophil recruitment (16, 54, 55). However, a variable role for mast cells in neutrophil recruitment has been reported in the skin RPA. Some studies demonstrate a significant reduction in neutrophil accumulation in mast cell-deficient mice (55), while others including our own data not shown suggest that mast cell-deficient mice exhibit only a partial reduction in neutrophil accumulation and edema compared with their WT counterparts in the skin RPA (16, 54). This indicates that other mast cell-independent mechanisms contribute to neutrophil accumulation. Indeed, we and others have shown that neutrophils may be recruited via their own FcγRs by directly tethering to deposited ICs (3, 13, 56). In summary, FcγRs on both mast cells and neutrophils may contribute to neutrophil recruitment. A possible model is that FcγRs induced release of mast cell mediators such as TNF may prime FcγRs on circulating neutrophils for enhanced binding to ICs (57). The phenotype in Vav-deficient mice suggest that the process of neutrophil diapedesis alone is not sufficient for causing tissue injury during IgG-mediated inflammation as previously suggested. Rather, the interaction of transmigrated neutrophils with perivenular and/or tissue ICs are likely primarily responsible for the ensuing changes in vascular integrity. The requirement for Vav 3 alone in neutrophil cytotoxicity suggests that although certain functions of Vav proteins are redundant; individual Vav proteins in neutrophils also serve specialized functions as shown in lymphocytes and osteoclasts (28, 58). Previous work has not established which specific Rho family GTPases serves as a direct substrate for each Vav family member to mediate a specific function. Our studies suggest that Vav links to Rac proteins in neutrophils to mediate neutrophil cytotoxicity in vivo (Fig. 7). It is noteworthy that unlike Vav-deficient mice, Rac deficiency results in a 50% reduction in neutrophil accumulation at the site of RPA induction. This reduction alone is likely not the primary explanation for the observed decrease in edema in these animals, which is far greater than would be expected from a partial loss of neutrophil recruitment. Indeed, mice lacking the neutrophil recruitment receptors L-selectin (31) or ICAM-1 (49) exhibit a 50% reduction in neutrophil accumulation that is associated with only a 40% decline in edema in the skin RPA. Thus, we deduce that Rac plays an additional role downstream of neutrophil recruitment in promoting IgG-mediated neutrophil cytotoxicity in our studies.

Our previous work has shown that hemorrhage in the LSR relies on Mac-1 interaction with complement deposited in the vessel wall and the coupling of Mac-1 to Syk and Src family kinases to trigger elastase release (12). The normal LSR response in both the Vav- and Rac-deficient mice demonstrates that the Vav/Rac axis is not required for Mac-1/complement C3-mediated tissue injury. The dispensability of Vav and Rac in the LSR was surprising given that these proteins are required for neutrophil phagocytosis of complement-opsonized E. coli (21). However, PMA-induced phagocytosis of complement-opsonized RBC, which is strictly dependent on Mac-1 and complement fixation, did not rely on Vav and Rac. E. coli is a complex target that engages neutrophil LPS receptors such as CD14 and TLR2, which activate Mac-1 (59). The observed defect in engulfment of serum-opsonized E. coli in Vav- and Rac-deficient neutrophils may be the result of ineffective Mac-1 activation. The emerging model of integrin activation is that agonist-induced conformational changes in the integrin increases affinity for ligand. Integrin clustering induced by multivalent ligand binding is coupled with agonist-induced intracellular signals that contribute to integrin redistribution. These steps, modulated by the actin cytoskeleton, are responsible for adhesion strengthening (60). Vav 1,3^−/− neutrophils are able to bind serum-opsonized E. coli to their surface (21). This indicates that Vav is not required for integrin affinity changes, thus leaving a possible role for Vav in integrin lateral mobility and subsequent productive adhesion required for C3-E. coli phagocytosis. The fact that C3-RBC phagocytosis in vitro assay correlated better with the development of tissue injury in the LSR than the complement-opsonized E. coli uptake assay suggests that the C3-RBC phagocytosis is a more reliable predictor of in vivo cytotoxicity. We propose that PMA, used to promote C3-RBC phagocytosis, increases β₂ integrin lateral diffusion (61, 62) likely through protein kinase C-dependent effects on the cytoskeleton (45, 63), thus bypassing any role for Vav in Mac-1 integrin activation. Indeed, in agreement with the results in C3-RBC, PMA treatment rescued the defect in phagocytosis of C3-opsonized E. coli in Vav 1,3^−/− neutrophils (21). Unlike human neutrophils, PMA is essential for target binding in murine neutrophils. This difference may be attributed to species-specific differences in the activation state of integrin. Our observed correlation of results from phagocytosis of C3-RBC in vitro and outcomes in the LSR in vivo provides support for the idea that PMA-stimulated C3-RBC phagocytosis may serve as a reliable predictor for complement-mediated vasculitis in vivo. This hypothesis is also supported by our finding that PMA-induced neutrophil phagocytosis of C3-RBC requires Src family and Syk kinases (A. Utomo and T. N. Mayadas, unpublished data), which are essential for LSR development (12). How can PMA-induced C3-RBC phagocytosis be a predictor of a physiological process in vivo? We suggest that PMA-induced activation of Mac-1 is served by multiple stimuli (e.g., TLR and G protein-coupled receptor agonists) encountered in vivo during the inflammatory response. However, the subsequent phagocytic step in vitro may be akin to Mac-1-mediated adhesion to complement C3 leading to vessel injury. Thus, signaling pathways identified in the in vitro assay may predict their contribution to vessel injury in vivo. Spreading or sustained adhesion on opsonized surfaces has been coined as “frustrated” phagocytosis, because the surface of the coated plates mimics a topology of an infinitely large target particle (64). Although Vav-deficient neutrophils cannot sustain adhesion/spreading on IgG- or iC3b-coated surfaces (21, 22), they can phagocytose iC3b-coated SRBC. Thus, the requirements for neutrophil phagocytosis rather than spreading correlate with those needed for mediating tissue injury in vivo. In contrast to the results in neutrophils, Vav and Rac proteins were recently shown to be essential for phagocytosis of complement-coated RBC by murine macrophages (Ref. 35 and A. Utomo and T. N. Mayadas, unpublished data). These differences in intracellular signaling pathways may be used as explanations for why neutrophil phagocytosis is more proinflammatory (i.e., results in greater reactive oxygen species generation) than macrophages. This dichotomy in use of the Vav/Rac axis by the same receptor family in two different phagocyte populations may also provide an opportunity to differentially target specific functions in one or the other phagocytic cell type.

There are likely differences in the mechanisms of tissue injury induced by the LSR and the RPA reaction. In the LSR, neutrophil interaction with complement within the vessel leading to elastase release is responsible for the observed edema and hemorrhage (12). Conversely, in the Arthus-type lesion, extravascular PMN stimulated by perivascular or tissue ICs are likely responsible for impairment of vascular integrity (65). FcγR engagement by ICs initiates the production of cytokines and lipid mediators. Of these, leukotrienes, platelet-activating factor, and bradykinin likely promote plasma extravasation in the RPA reaction as inhibitors or knockouts of these known vascular permeability factors or their corresponding receptors significantly blunts the RPA reaction (65). Of these factors, we speculate that bradykinin is a plausible candidate for Vav-dependent edema because tissue kallikreins, which generate kinins from low- and high-molecular mass kininogens, are present in neutrophils as are the kininogens themselves (66, 67).

In conclusion, by separating neutrophil recruitment and activation at the level of signaling molecules downstream of FcγRs, we demonstrate that effector functions triggered by FcγR engagement are central to disease pathology in IC-mediated processes. Furthermore, we have shown that Vav selectively links one of the two major neutrophil opsonic receptors to specific cytotoxic functions in vivo. There has been a recent focus on targeting signaling molecules as an approach to reduce inflammation in vivo (68). Targeting Vav in neutrophils has the potential of specifically reducing IgG-mediated tissue injury in immune-mediated diseases while retaining neutrophil recruitment, a strategy that would minimize overall immune suppression.

We are grateful to Ling Xiao (Brigham and Women’s Hospital) for excellent technical assistance and mouse husbandry.

The authors have no financial conflict of interest.