Engulfment of Apoptotic Cells Is Negatively Regulated by Rho-mediated Signaling*

The rapid and efficient phagocytosis of apoptotic cells plays a critical role in preventing secondary necrosis, inflammation as well as in tissue remodeling and regulating immune responses. However, the molecular details of engulfment are just beginning to be elucidated. Among the Rho family GTPases, previous studies have implicated a role for Rac and Cdc42 in the uptake of apoptotic cells by phagocytes, yet the role of Rho has remained unclear. Here, we present evidence that Rho-GTP levels decrease during engulfment. RhoA seems to negatively affect basal engulfment, such that inhibition of Rho-mediated signaling in phagocytes enhanced the uptake of apoptotic targets. Activation of endogenous Rho or overexpression of constitutively active forms of Rho also inhibited engulfment. By testing mutants of RhoA that selectively activate downstream effectors, the Rho-kinase seemed to be primarily responsible for this inhibitory effect. Taken together, these data suggest that inhibition of Rho- and Rho-kinase-mediated signaling might be important during engulfment, which could have important implications for several clinical trials involving inhibition of the Rho kinase.

Apoptosis or programmed cell death occurs during development and throughout life as part of normal tissue homeostasis. The prompt removal of apoptotic cells by phagocytes prevents their secondary necrosis and inflammation and also plays a critical role in tissue remodeling and immune response regulation (1). The general features of apoptosis and engulfment of cell corpses are highly conserved through evolution, which further underscores their importance. It is noteworthy that, compared with other types of phagocytosis such as that mediated via the Fc receptor (FcR), 1 the phagocytosis of apoptotic cells does not normally lead to production of pro-inflammatory mediators. The failure to remove apoptotic bodies has been implicated as a cause for different types of chronic inflammation and a predisposition to cancer, as well as autoimmune diseases (2)(3)(4). There is now increasing evidence that a distinct set of evolutionarily conserved cellular receptors, ligands, and signaling molecules facilitates the uptake of apoptotic cells.
In the past few years, a number of studies have shown a central role for the Rho-family GTPases and their downstream effectors during different types of phagocytosis (5)(6)(7)(8)(9)(10)(11)(12)(13). An important role for Rac and its upstream activators has been demonstrated from worm to mammals in the engulfment of apoptotic cells (10, 12, 14 -17). Previously, we have observed that the CrkII/Dock180/ELMO proteins lead to activation of Rac and promote Rac-dependent phagocytosis (11). Studies using overexpression of a dominant-negative form of Cdc42 and mice lacking the Cdc42 effector WASP (Wiskott-Aldrich syndrome protein) have suggested a positive role for Cdc42 in the uptake of apoptotic cells (6,9,14). However, the role of RhoA in phagocytosis of apoptotic cells is less clear. RhoA and the activation of Rho-kinase have been shown to be critical for complement receptor (CR)-mediated phagocytosis but not FcR-mediated uptake (6,18). Because FcR-and CR-mediated phagocytosis utilize different Rho-family members (6,18), and because the outcome of engulfment of apoptotic cells is distinct from the above two types of phagocytosis (19), we examined the contribution of RhoA during engulfment. Here we present evidence that RhoA activity is down-regulated during engulfment. Moreover, RhoA seems to negatively affect basal engulfment, such that inhibition of RhoA-mediated signaling enhances the uptake of apoptotic cells. Among the Rho effectors, the Rho-kinase seems to be primarily responsible for this inhibitory effect on engulfment.

MATERIALS AND METHODS
Cell Culture-The phagocytic LR73 Chinese hamster ovary cell line was cultured as described previously (12). The J774 macrophage cell line was cultured in RPMI 1640 medium supplemented with 10% fetal calf serum, 10 mM HEPES, pH 7.4, 0.05 M ␤-mercaptoethanol, 4.5 g/liter glucose, and antibiotics.
Rac-GTP and Rho-GTP Pull-down Assay-Bacterially produced glutathione S-transferase-Cdc42/CRIB domain of p21-activated kinase or glutathione S-transferase-Rho-binding domain of rhotekin proteins bound to glutathione-Sepharose beads was incubated with lysates of LR73 for 1 h at 4°C. The beads were washed, and the binding of Rac-GTP and Rho-GTP in the lysates to the beads was analyzed by immunoblotting for Rac and Rho, respectively.
Transfection and Immunoblotting-For transient transfections, Li-pofectAMINE 2000 reagent (Invitrogen) was used as per the manufacturer's recommendations. Briefly, 100,000 LR73 cells were plated in 24-well plates. The cells were transfected the next day by incubation with plain medium containing 1.5 l of LipofectAMINE 2000 and the indicated plasmids for 7 h before washing and were incubated with fresh medium for 16 -20 h before performing the phagocytosis assays. When cotransfecting multiple plasmids, carrier DNA was added to match plasmid concentration in the different samples. In some experiments, the transfection was done in triplicates (using the same master mix of transfection reagent plus plasmids), with two wells used for the engulfment assays and the third well being used to analyze the expression of proteins by immunoblots. The lysis and immunoblotting were performed as described previously (12).
Phagocytosis Assays-16 -20 h after transfection, the cells were incubated with 2 m carboxylate-modified red fluorescent beads (Sigma and Molecular Probes, Eugene, OR) or with apoptotic thymocytes labeled with CM-Orange (Molecular Probes), as described previously (12). 40 -50% of thymocytes were Annexin V positive (i.e. apoptotic), and less than 5% of Annexin V positive cells were propidium iodide positive (i.e. necrotic). The thymocytes were washed and plated (at 0.7-1 ϫ 10 6 cells/well in 300 l of growth medium) on top of LR73 cells or J774 cells. Routinely, the assays were carried out in duplicates or triplicates for each condition with a 50-min incubation for thymocytes. The wells were then aspirated and washed twice with cold phosphate-buffered saline. The cells on the plate were trypsinized, resuspended in cold medium (with 0.1% sodium azide), and analyzed by flow cytometry. Unengulfed thymocytes were gated out by their forward and side scatter. Routinely, 10,000 -20,000 events were collected, and the data were analyzed using CellQuest software. The controls included the use of live thymocytes (not treated with dexamethasone) and apoptotic thymocytes incubated at 4°C. Although we trypsinized the cells prior to analysis, the fluorescence-activated cell sorter assay could not distinguish between fluorescence derived from bound and that from engulfed thymocytes. However, as determined by confocal microscopy, the majority of fluorescent phagocytes scored in the fluorescence-activated cell sorter assay represent cells that have engulfed the thymocytes or are in the process of engulfment (data not shown). For engulfment assays with 2 m carboxylate-modified beads, the cells were incubated with beads in serumfree medium for 2 h. After washing, the cells on the plate were trypsinized, resuspended in cold medium with 0.5% sodium azide, and 10,000 -20,000 cells were analyzed for each point by two-color flow cytometry (12). Forward and side-scatter parameters were used to distinguish the unbound beads from cells. The transfected cells were recognized by their GFP fluorescence (12).
The ADP-ribosyltransferase C3 (intact toxin, Sigma) was pre-incubated with the phagocytes at 5 g/ml for 16 h, and the beads were added with fresh serum-free medium for 2 h. The intact Clostridium difficile toxin B (provided by Dr. Chang Hahn, University of Virginia; Dr. Charalabos Pothoulakis, Harvard Medical School; and Dr. Popoff, Institut Pasteur) was pre-incubated at 10 ng/ml for 15 min and was present throughout the engulfment assay. Y-27632 (Calbiochem, San Diego, CA) was used at 10 M or 30 M (as indicated) and added along with the beads onto the cells for 2 h during uptake. ML-7 (Calbiochem) was used at 5, 10, and 30 M (as indicated) and added along with the beads onto the cells for 2 h during uptake.
Microscopy-The indicated plasmids were transfected into LR73 cells using LipofectAMINE 2000 reagent. 20 h after transfection, the cells were incubated for 2 h in the presence or absence of the Rho inhibitor Y-27632 in serum-free medium. LR73 cells were fixed in 3.7% paraformaldehyde and permeabilized with phosphate-buffered saline/ 0.1% bovine serum albumin/0.1% Triton X-100. The permeabilized cells were blocked with phosphate-buffered saline/1% bovine serum albumin prior to staining with Alexa Fluor 568-phalloidin (Molecular Probes) and analyzed by using an Olympus confocal microscope. The images shown are representative of multiple cells with similar phenotype on the same slide and are representative of two independent experiments.

RESULTS
Down-regulation of Rho Activity during Engulfment-To assess the role of Rho in the phagocytosis of apoptotic cells, we examined the activity of endogenous Rho during engulfment by performing a Rho-GTP pull-down assay. Phagocytic LR73 cell lines were incubated for different times with 2 m carboxylatemodified latex beads, which we have shown previously can function as a simplified target that mimics the negative charge on apoptotic cells (11,12). This simplified target was chosen to avoid contamination by Rho proteins derived from the target cells. The level of Rho-GTP in the phagocyte began to decrease after 30 min and was dramatically decreased after 2 h of incubation with the targets (Fig. 1A). Interestingly, we had observed previously that maximal uptake of these surrogate apoptotic targets occurs at the 2-h time point in our assays (11,12). In contrast, the level of Rac-GTP was increased in the phagocytes at the 30-min and 2-h time points under similar conditions (Fig. 1B). This result suggested a reciprocal regulation of Rho-GTP and Rac-GTP levels and raised the possibility that inhibition of Rho activity might be beneficial during phagocytosis.
Inhibition of Endogenous Rho Promotes Phagocytosis of Apoptotic Cells-To further examine the benefit of inhibition of Rho activity during phagocytosis, a phagocytosis assay was performed to test the effect of C3 exoenzyme (from Clostridium botulinum) (20 -22), which specifically inactivates Rho proteins through ADP-ribosylation but does not affect Rac or Cdc42. Pre-treatment of J774 murine macrophage cells with C3 exoenzyme for 16 h increased their ability to engulf apoptotic thymocytes ( Fig. 2A). Pre-treatment of LR73 cells with the C3 exoenzyme likewise significantly enhanced the engulfment (fold increase of 1.23 ϩ 0.09, p Ͻ 0.01, n ϭ 3). In comparison, C. difficile toxin B, which inhibits all Rho family GTPasesinhibited engulfment, similar to what we have seen previously ( Fig. 2B; Ref. 12). Transient transfection of LR73 cells with a plasmid encoding the C3 exoenzyme also led to a similar enhancement of uptake (Fig. 2B). In the same experiment, transient transfection of an activated form of Rho (Rho Q63L ) partially inhibited the uptake of apoptotic cells ( Fig. 2B and see below). In addition to the effect on uptake of apoptotic cells, C3 pre-treatment of LR73 phagocytes also enhanced the uptake of 2 m carboxylate-modified beads (Fig. 2C). Because the uptake of beads and apoptotic cells were both regulated similarly by Rho-mediated signaling, we have used the carboxylate-modified 2-m beads in many of the subsequent experiments.
Activation of Endogenous Rho Inhibits the Engulfment-The above data implied that inactivation of endogenous Rho promotes engulfment, possibly because of the removal of a basal inhibitory effect due to Rho. We then asked whether forced activation of endogenous Rho, for example through Rho-specific GEFs, would lead to inhibition of phagocytosis. To achieve this, we expressed an oncogenic form of the Rho-specific GEF Lbc, denoted ONC4A (composed of only the DH-PH regions of Lbc; see supplemental Fig. 1), which has been shown to lead to Rho-GTP loading within cells and subsequent oncogenic transformation (23). Overexpression of ONC4A is expected to allow the cycling of RhoA between GTP and GDP bound states and also to permit regulation by Rho-GAPs during phagocytosis (and thereby overcome the limitations of constitutively active mutants of RhoA). Expression of the ONC4A form of Lbc in LR73 phagocytes, but not a DH-PH-deleted mutant of Lbc (PS1; see supplemental  (Fig. 3B), whereas a mutant of Vav2 lacking the catalytic activity (denoted C-ter; see supplemental Fig. 1) had no effect.
We then examined whether the inhibition of phagocytosis observed above was caused by a defect in the binding of targets to the phagocytes or their internalization. Overexpression of the ONC4A or PS1 did not affect particle binding to the phagocytes compared with control vector-transfected cells (assessed by performing the phagocytosis assay in the presence of 1% azide or at 4°C to prevent internalization but not binding). This finding suggested that forced activation of endogenous Rho in the phagocytes causes inhibition of uptake rather than inhibition of the initial binding of the targets (Fig. 3C).
Overexpression of Constitutively Active RhoA but Not Rac and Cdc42 Inhibits Engulfment-Transient transfection of LR73 cells with an activated form of Rho (Rho Q63L ) partially inhibited the uptake of apoptotic cells (Fig. 2B) as well as 2-m carboxylate-modified beads (Fig. 4A). In contrast, a weak dominant-negative form of Rho (Rho T19N ) did not inhibit engulfment. Although in some experiments Rho T19N increased the uptake, this was not consistently observed. The better and more consistent enhancement with C3 likely reflects a higher degree of endogenous RhoA inhibition, as the Rho T19N was not very potent in our hands.
We have previously shown that expression of the adapter protein CrkII (homologue of the worm CED-2) significantly enhances engulfment by a Rac-dependent mechanism (12). Coexpression of Rho Q63L with CrkII potently inhibited the CrkIImediated enhanced uptake (Fig. 4A). In contrast, the Rho T19N did not inhibit the CrkII-mediated uptake. As a control, the dominant-negative Rac T17N inhibited the CrkII-mediated uptake, as has been shown previously ( Fig. 4A; Refs. 11,12). Comparable inhibition of engulfment was also seen when the Rho Q63L plasmid was transfected into LR73 cells stably expressing CrkII (data not shown). As another approach, coexpression of Rho Q63L also inhibited uptake mediated by ELMO1/ Dock180/CrkII proteins (which have been shown to promote maximal Rac activation and maximal phagocytosis) (Ref. 11 and data not shown). These data suggested that the inhibition of phagocytosis caused by overexpression of activated Rho are dominant under these conditions. It is noteworthy that overexpression of activated Rho did not affect particle binding to the phagocytes (see Fig. 3C).
When we compared the effect due to constitutively active Rho with other small GTPases, the inhibitory effect due to active Rho appeared specific, because expression of a constitutively active Rac1 (Rac Q61L ) or constitutively active Cdc42 (Cdc42 Q61L ) did not cause an inhibition (Fig. 4B). Moreover, overexpression of Intersectin, a Cdc42-specific GEF (25,26), promoted engulfment (Fig. 4B). This is in distinct contrast to the effect seen with activation of endogenous Rho through ONC4A. The difference in the ability to engulf between activation of endogenous Cdc42 through Intersectin and activated Cdc42 Q61L likely reflects the lower potency of the "constitutively active" Cdc42 mutant (as has been observed by others with this mutant). Thus, forced activation of endogenous Rho or overexpression of constitutively active forms of Rho, but not Rac1 or Cdc42, inhibited engulfment. Taken together with the enhanced uptake seen when Rho was inactivated through the C3 toxin, these data suggested that Rho proteins negatively regulate basal uptake, which may need to be overcome for optimal engulfment.
Rho-kinase as a Key Rho Effector in Negative Regulation of Engulfment-GTP-bound Rho has been shown to bind a number of different downstream effectors such as p140mDia, Rhokinases, citron, rhotekin, rhophilin, and protein kinase N to mediate various functional outcomes ( Fig. 5A; Refs. [27][28][29]. Expression of a different constitutively active form of RhoA (RhoA G14V ) in J774 macrophage cells again led to a potent inhibition of uptake (similar to that seen with RhoA Q63L ). However, a RhoA G14V mutant carrying a concurrent T37Y mutation, which fails to interact with all of its known effectors, abolished the inhibitory effect due to RhoA G14V (Fig. 5B). This result suggested that inhibition by the active forms of RhoA involves interaction of RhoA with its effectors.
Other investigators have successfully used effector domain mutants of RhoA, which can interact with some effectors but not others, to test the involvement of particular downstream effectors ( Fig. 5A; Refs. 30,31). We observed that the RhoA G14V/ F39V mutant, which can interact with mDia and Rho-kinase but not the other known effectors (32)(33)(34), was still able to potently inhibit engulfment of beads when overexpressed in J774 macrophages (Fig. 5B). However, the RhoA G14V/F39A mutant that retains the ability to activate mDia, but not Rho-kinase, completely lost its inhibitory effect on engulfment. This pointed to Rho-kinase (also referred to as ROK and ROCK) as a potential mediator of the inhibitory effect due to Rho. Essentially similar results were obtained with the same set of Rho mutants when LR73 cells were used as phagocytes (Fig. 5C).
Although the above data suggested that Rho-kinase was likely the primary mediator of the Rho-dependent inhibition of engulfment, the F39A mutant of Rho can also bind mDia (31,35). Because in one case, antagonism between mDia and Rhokinase was observed (36), and mDia can also affect microtubules (31), we examined whether engulfment is affected by activation of endogenous mDia. To address this question, we overexpressed the Dia-autoregulatory domain (DAD) of mDia (see supplemental Fig. 1). It has been shown previously, that overexpression of the isolated DAD region alone can relieve the intramolecular auto-inhibition within mDia between its DAD domain and the Rho-GTP binding domain, thereby leading to Where the error bars are not visible, they were too small to be apparent. endogenous mDia activation (31). Overexpression of DAD did not inhibit or enhance engulfment of carboxylate modified beads used as targets (Fig. 5C). This is consistent with the lack of an inhibition of engulfment seen with the Rho G14V/F39A mutant, which has been shown to activate mDia, but not Rhokinase. Moreover, overexpression of other constitutively active forms of mDia (such as ⌬N3 fragment lacking the autoinhibitory domain) (36) also did not affect engulfment (data not shown). On the other hand, overexpression of a dominantnegative mutant of Rho-kinase-␣ (dnRok␣) increased the basal uptake (see supplemental Fig. 2), suggesting a role for Rhokinase in the basal inhibitory signaling via Rho.
Rho-kinase Inhibitor Y27632 Promotes Phagocytosis-To further address the role of Rho-kinase in engulfment, we tested the effect of the drug Y-27632, a widely used inhibitor of Rhokinase (37,38), during phagocytosis. Treatment of phagocytes with Y-27632 enhanced engulfment of beads as compared with untreated cells (Fig. 6A). This was reminiscent of the increased engulfment when Rho was inactivated with the C3 exoenzyme, which also promoted engulfment, suggesting that the inhibition of basal uptake was likely mediated through endogenous Rho-kinase. The more robust enhancement of uptake due to Y-27632 compared with C3 exoenzyme likely reflects the greater permeability of Y-27632 into the cells. The increased uptake due to Y-27632 was not due to increased binding of the particles to the Y-27632 treated cells (data not shown). Although Y27632 can also inhibit PRK2 (close relative of protein kinase N) under some conditions (39), our data using effectordomain mutants of RhoA (see Fig. 5 above), suggest that pro-tein kinase N does not play a role in Rho-mediated effects on engulfment.
We then tested whether inhibition of engulfment due to the activation of endogenous Rho by ONC4A mutant of Lbc can be reversed by inactivation of the Rho-kinase. Treatment of cells with Y-27632 fully reversed the inhibitory effect due to activation of endogenous Rho by ONC4A (Fig. 6B). Stress fiber formation has long been associated with activation of RhoA (40). When we examined LR73 cells transfected with ONC4A, with or without Y-27632 treatment, the strong stress fibers induced by ONC4A were completely abolished by treatment with Y-27632 (Fig. 6C). The stress fibers formed by activation of endogenous Rho and the disruption of these stress fibers with Y-27632 treatment correlated with the inhibition or enhancement of uptake, respectively. Taken together, these data suggest Rho-kinase as a primary Rho effector in mediating the inhibition of engulfment.

Effect of Rho/Rho-kinase during Engulfment of Apoptotic Cells Differs from Phagocytosis Mediated via the Complement Receptor and Fc
Receptor-A recent report showed that during CR-mediated phagocytosis, activation of Rho and Rho-kinase were critical for recruitment of the Arp2/3 complex, leading to uptake of complement coated targets (18). Interestingly, activation of Rac or Cdc42 was not required for CR-mediated phagocytosis. The authors also showed that activation or inhibition of the Rho/Rho-kinase signaling had no effect on FcR-mediated phagocytosis (18). Our data presented above suggest that during engulfment of apoptotic cells, Rho appears to have a negative effect, whereas Rac and Cdc42 both play a positive role. We tested the involvement of the Arp2/3 complex in apoptotic cell engulfment by expression of Scar-WA. This Scar-WA is a mutant form of the Scar/WAVE protein, involved in delocalization and activation of the Arp2/3 complex preventing its recruitment by endogenous proteins of WASP family (see supplemental Fig. 1; Refs. 18, 41). Expression of Scar-WA resulted in a strong inhibition of basal uptake as well as inhibition of the enhanced engulfment seen with overexpression of constitutively active Rac Q61L or CrkII (Fig. 7A). This suggests that engulfment of apoptotic cells also requires Arp2/3 recruitment, which likely occurs through Rac and/or Cdc42.
Rho-kinase has been shown to affect the phosphorylation status of myosin light chain (MLC) by inhibiting the MLC phosphatase, and through direct phosphorylation of MLC itself, thereby promoting actomyosin assembly and cell contraction (42). Our data suggested that the increased contractility of the cells mediated by Rho/Rho-kinase, possibly through phosphorylation of the MLC, might negatively influence engulf-ment. However, MLC can also be regulated by MLCK which can also regulate cellular contractility (43). Interestingly, previous studies in fibroblasts have suggested that phosphorylation of MLC may be regulated spatially within the cells, such that MLC phosphorylation in the periphery of the cells was more dependent on MLCK, whereas Rho-kinase activity was required for MLC phosphorylation and cell contraction at the center of the cell (44 -46). When we tested the effect of ML-7 (10 M), a selective inhibitor of MLCK (47), it partially inhibited the uptake of the carboxylate-modified beads (Fig. 7B). This suggested that whereas Rho-kinase mediated activity negatively affects engulfment; the MLCK activity at the cell periphery (and by inference, myosin-dependent contraction), likely promotes engulfment. Interestingly, ML-7 was also found to partially inhibit/delay FcR-mediated phagocytosis, whereas Rho or Rho kinase pathways had no effect on FcRmediated uptake (18). This role for MLCK activity during FcRmediated phagocytosis was thought to be due to a role for FIG. 5. Rho-kinase as an effector for Rho-mediated inhibition of engulfment. A, different effector domain mutants of Rho used in this study (in the constitutively active G14V Rho background) and the selective Rho effectors that interact with these mutants. B and C, plasmids encoding GFP alone, GFP-tagged constitutively active Rho G14V , or the indicated effector domain mutants (all of them used at 0.5 g) were transiently transfected into J774 cells (B) or LR73 cells (C). The engulfment was analyzed by flow cytometry as described above. Where the error bars are not visible, they were too small to be apparent. Results are representative of at least four independent experiments. actomyosin contractility in closure of phagocytic cups/phagosome pinching (18,48). Taken together, certain downstream features of uptake, such as Arp2/3 recruitment and MLCK activity are commonly used between all three types of phagocytosis. In contrast, Rho/Rho-kinase mediated signaling is required for CR-mediated phagocytosis, and has no effect on FcR-mediated phagocytosis. However, the inhibitory activity of Rho/Rho-kinase mediated signaling appears to be unique to apoptotic cell engulfment.

DISCUSSION
Most of the earlier studies have primarily focused only on the positive regulatory roles of Rho-family GTPases in phagocytosis. In this report, our data suggest a basal inhibitory effect of RhoA and its downstream effector Rho-kinase during engulfment. The increased engulfment following C3 exoenzyme-induced Rho inactivation suggests that Rho (contrary to Rac) is not essential for mediating engulfment of apoptotic cells. Machesky and colleagues have reported that Rho-kinase is not required for engulfment via FcR-mediated phagocytosis, but is critical for CR-mediated uptake (18). Yet, our observations suggest a basal inhibitory role for Rho kinase-mediated signaling during engulfment of apoptotic cells. This again underscores a fundamental difference in the signaling pathways regulating FcR-and CR-mediated phagocytosis versus engulfment of apoptotic cells, and might also relate to the inflammatory versus anti-inflammatory outcome following the different modalities of uptake.
Although the precise signals that lead to basal activation of Rho during apoptotic cell engulfment are currently unknown, our data suggest that such signals likely need to be overcome or inactivated during engulfment and implies modulation of Rhokinase activity. Rho-kinase has been shown to regulate the phosphorylation status of myosin light chain (MLC) by inhibiting the MLC phosphatase and/or by direct phosphorylation of MLC itself. In turn, the alteration of MLC phosphorylation regulates the actomyosin assembly leading to cell contraction. It is possible that inhibition of Rho-kinase activity and a decrease in cell contractility could account for the increased engulfment of apoptotic cells. Interestingly, a recent report suggested that "don't eat-me" signals delivered through CD31/ PECAM-1 on live cells are critical for the phagocyte to discriminate between live and apoptotic targets in vivo (49). Whether such "don't eat-me signals" on live cells might lead to activation of Rho in the phagocyte, and thereby inhibit engulfment, remains to be seen. Conversely, inhibition of Rho activity could occur through membrane proteins such as Cadherins, which have been shown to inhibit Rho-activity through regulation of p190-Rho-GAP, a Rho-specific guanine nucleotide activating protein (GAP) (50,51). We considered the possibility that Rho-GAPs might be activated during engulfment, which in turn could lead to inhibition of Rho activity. Because tyrosine phosphorylation has been linked to engulfment of apoptotic cells (6,9,14), and p190Rho-GAP functions downstream of tyrosine kinases (27) Note that the forced activation of endogenous Rho by ONC4A induces the formation of stress fibers, which is reversed by Y-27632. mutant of p190Rho-GAP (52) had no detectable effect on basal or enhanced engulfment of apoptotic targets (data not shown). This observation could indicate that the p190Rho-GAP is not involved during apoptotic engulfment, the conditions tested do not reveal a true role for p190Rho-GAP or that other Rho-GAPs might play a role during engulfment.
Our data are consistent with a model where Rac (or Cdc42) and Rho pathways have mutual antagonist effects (53,54). In a number of systems, simultaneous activation/deactivation of multiple Rho-family GTPases appears to regulate the functional outcome (28). For instance, a balance of spatiotemporal activation and downstream signaling of Rho (which regulates cell contractility) and Rac and Cdc42 (which regulate cell membrane protrusions), regulate directional cell motility (55). This allows repeated cycles of cell protrusion at the leading edge with simultaneous cell contraction at the rear end of the cell to allow directional migration (56). Engulfment of apoptotic cells, a complex process involving interaction between two cells, could also involve such crosstalk between the Rho-family GTPases within the phagocytes. When we coexpressed constitutively active forms of both Rac and Rho, overexpression Rac Q61L was able to overcome the effect of Rho partially and vice versa, in different concentrations (see supplemental Fig.  3). One potential mechanism of Rac-mediated inhibition of Rho activation in fibroblasts was recently reported to occur through regulation of reactive oxygen species (ROS) (57). This effect of Rac was also dependent on p190Rho-GAP. As stated above, when we examined the potential role of p190Rho-GAP through overexpression of wild type or dominant negative mutant of p190Rho-GAP, there was no detectable effect on engulfment. Thus, other mechanisms of crosstalk between Rac-and Rhomediated signaling pathways may be involved.
Several different experiments indicated that promoting Rho activity inhibits engulfment through activation of a Rho-kinase dependent signaling pathway. Although the most obvious explanation for this inhibition is the Rho-kinase mediated effects on cell contractility, surprisingly, two drugs that antagonize actomyosin contractility (ML-7 and Y27632, via inhibition of MLCK or Rho-kinase, respectively), had contrasting effects on engulfment of apoptotic targets. This could be reconciled based on several recent studies where differences between the effects of inhibiting MLCK versus Rho-kinase were observed (44 -46). Two different groups, using fibroblasts models, have observed a greater or more prominent role for Rho-kinase/Rho-kinase in regulating myosin contractility and stress fiber formation at the center of cells (45,46). In comparison, the MLCK activity seems to affect myosin phosphorylation in the periphery of the cells, leading to increased stress fiber formations and cell protrusions in the periphery. Consistent with this notion, in our experiments the inhibition of Rho-kinase with Y27632 destabilized the stress fibers formation, yet it did not affect cell protrusions in the periphery (Fig. 6). ML-7, an inhibitor of MLCK, was also shown to inhibit FcR-mediated phagocytosis, although Rho or Rho-kinase mediated signaling had no effect on this process (18). This effect of ML-7 on FcR-dependent phagocytosis has been linked to a role for actomyosin contractility during closure of phagocytic cups/phagosome pinching (48,58). The latter steps would be in common with apoptotic cell phagocytosis. We hypothesize that the specific recognition of apoptotic targets may lead to destabilization of Rho-dependent contractility through the inhibition of basal RhoA/Rhokinase signaling at the center of the cells, whereas MLCK-dependent contractility in the periphery is likely required for the proper uptake of targets.
The linkage between inhibition of Rho-and Rho-kinase-mediated signals and promotion of phagocytosis of apoptotic cells could have clinical relevance in certain cases of chronic inflammation. Interestingly, in animal models, long-term inhibition of Rho-kinase prevents development of atherosclerotic lesions and even causes regression of the vascular lesions in vivo (59 -62). Recent human clinical trials also showed that inhibition of Rho-kinase has a beneficial effect in myocardial ischemia in patients (63). These beneficial effects of the drugs Y-27632 and fasudil were thought to result from effects on the vascular smooth muscle cells, i.e. from either inhibition of proliferation or induction of apoptosis (61,62). Whether the promotion of apoptotic cell clearance (due to Y-27632) may have accompanied anti-inflammatory signaling at the local site also needs to be examined. Overall, our studies suggest an unexpected regulatory role for RhoA and the Rho-kinase during apoptotic cell uptake and that removal of the inhibitory signal through RhoA might be an integral part of the signaling that occurs during engulfment. FIG. 7. Phagocytosis of apoptotic cells requires Apr2/3 complex recruitment and is also inhibited by ML-7 treatment. A, LR73 cells were transfected with GFP alone or GFP together with plasmids encoding either CrkII or constitutively active Rac1 (as positive controls), with or without Scar-WA, a mutant of Scar/WAVE proteins that activates and delocalizes Arp2/3 complex. The uptake of 2-m carboxylatemodified beads was analyzed. The data are representative of three independent experiments. B, untreated, Y27632treated (10 M), or ML-7-treated (5, 10, 30 M) LR73 cells were analyzed for engulfment of 2-m carboxylate-modified beads.