Neuronal Chemorepellent Slit2 Inhibits Vascular Smooth Muscle Cell Migration by Suppressing Small GTPase Rac1 Activation
The Slits are secreted proteins with roles in axonal guidance and leukocyte migration. On binding to Robo receptors, Slit2 repels developing axons and inhibits leukocyte chemotaxis. Slit2 is cleaved into Slit2-N, a protein tightly binding to cell membranes, and Slit2-C, a diffusible fragment. In the present study, we characterized the functional role of Slit2-N in vascular smooth muscle cells (VSMCs) and the cell association properties of 2 truncated versions of Slit2-N. Here, we document for the first time that Slit2-N is a chemorepellent of VSMCs. Intact blood vessels expressed Slit2 and Robo receptors as demonstrated by immunohistochemistry and quantitative real time PCR. Recombinant Slit2-N prevented the platelet-derived growth factor (PDGF)-stimulated migration of VSMCs. Slit2-N also abrogated PDGF-mediated activation of small guanosine triphosphatase (GTPase) Rac1, a member of the Rho GTPase superfamily of proteins involved in regulating the actin cytoskeleton. Furthermore, Slit2-N inhibited the PDGF-induced formation of lamellipodia, a crucial cytoskeletal reorganization event for cell motility. Slit2-N had no effect on the PDGF-mediated increase in DNA synthesis determined by [3H]thymidine uptake, suggesting that VSMC growth is unaffected by Slit2. Analysis of 2 engineered Slit2-N fragments (Slit2-N/1118 and Slit2-N/1121) indicated that 3 amino acids upstream of the putative cleavage site (Arg1121, Thr1122) are involved in the association of Slit2-N to the cell membrane. Our data assign a novel functional role to Slit2 in vascular function and show that cell guidance mechanisms that operate in the developing central nervous system are conserved in VSMCs.
Cell migration is modulated by similar attraction and repulsion factors in the patterning of the nervous and vascular systems.1 In the development of atherosclerosis, cells such as vascular smooth muscle cells (VSMCs) and inflammatory cells follow various migration pathways guided by locally secreted molecules.2 These environmental cues can be diffusible or bound to either the extracellular matrix (ECM) or the cell membrane. Guidance mechanisms in a dynamic cellular microenvironment include chemoattraction and chemorepulsion, which may in turn function in a contact-dependent or -independent manner. Unlike chemoattraction, the contribution of chemorepellent cues as counterregulatory mechanisms in vasculopathies has received little attention. In addition to attraction signals, repulsion cues may also be involved in regulating the extensive VSMC migration associated with the architectural and phenotypic remodeling of injured vessels.3–5
Platelet-derived growth factor (PDGF) is a potent chemoattractant of VSMCs and plays a key role in the migration of medial SMCs in vascular injury models.6 To assess the involvement of chemorepulsion in arterial cell motility, we investigated the effects of Slit2, a neuronal repellent, in PDGF-induced migration of VSMCs. The chemorepellent actions of Slit2 have been primarily investigated in models of axonal migration and in leukocyte chemotaxis.7–9 Slit has been cloned in a diverse range of species from Drosophila to mammals and consists of a family of 3 genes: Slit1, Slit2, and Slit3. They contain 4 tandem leucine-rich repeats (LRRs), 7 epidermal growth factor (EGF)-like repeats, an Agrin-Laminin-Perlecan-Slit (ALPS) conserved spacer motif, and a cystine knot.9 Slit2 is proteolytically cleaved into Slit2-N (≈140 kDa) and Slit2-C (≈55 kDa). Slit2-N remains tightly bound to the cell membrane, whereas Slit2-C is readily diffusible.10 Little is known about the motifs involved in the attachment of Slit2-N to the cell membrane or whether further enzymatic processing generates additional N-terminal fragments in vivo. Slit2-N binds to Robo, a single-pass transmembrane receptor.11,12 The intracellular domain of Robo interacts with proteins that regulate the Rho family of small guanosine triphosphatases (GTPases), which play well-defined roles in cell migration, cytoskeletal organization, and remodeling.13 To gain insight into the mechanisms of Slit2-N membrane binding, we generated 2 truncated versions of Slit2-N and examined their cell association characteristics and biological activity. Our data demonstrate that Slit2-N is a VSMC chemorepellent with no effects on VSMC growth. The small GTPase Rac1 is involved in the intracellular signaling pathway of Slit2/Robo leading to regulation of actin polymerization in VSMCs. Our findings also demonstrate that residues near the Slit2 cleavage site on the C terminus of Slit2-N are involved in the association of Slit2-N to the cell membrane.
Materials and Methods
Harvesting and Processing of Rat Arteries
Our study was approved by the Institutional Animal Care and Use Committee of the Atlanta University Center. We euthanized male Sprague–Dawley rats (350 to 400 g) with CO2 and performed thoracotomies. Animals were perfused with PBS and vascular tissues were collected for RNA extraction and immunohistochemistry as described in the expanded Materials and Methods section available at http://circres.ahajournals.org.
Commercial sources and cell culture conditions for human aortic smooth muscle cells (HASMCs), human aortic endothelial cells (HAECs), human peripheral blood mononuclear cells (PBMCs), rat aortic smooth muscle cells (RASMCs), and rat aortic ECs (RAECs) are included in the online data supplement.
Real-Time PCR Analysis
Primer pairs used to amplify human and rat Slit2 and Robo receptors are listed in the Table. Real-time PCR protocols are described in the online data supplement.
The immunohistochemical detection of Slit2 in rat carotid arteries and RAECs was performed using a goat anti-human Slit2 IgG antibody (Ab) (G-19, 1/200 dilution, Santa Cruz Biotechnology Inc, Santa Cruz, Calif). The sections were fixed in methanol for 5 minutes. The primary Ab was applied for 1 hour at 37°C, and the slides were incubated with a secondary mouse anti-goat Ab conjugated to Alexa Fluor 488 (Molecular Probes, Eugene, Ore). Cell nuclei were identified with 4′,6-diamidino-2-phenylindole, dihydrochloride (DAPI). Nonimmune IgG was used as a negative control. VSMCs were visualized with an anti-smooth muscle α-actin (α-SMA) Ab (Clone 1A4, Sigma) and the slides stained using ABC-AP (Vector) as described.5 Immunofluorescence images were captured using an Olympus BX60 microscope.
Expression Vectors and Generation of Recombinant Slit2-N/1118
Two pDONR221/Slit2-N constructs from amino acids (aa) 1 to 1121 (MRGVGW to PMVLPR) and 1 to 1118 (MRGVGW to FSPPMV) were generated by cloning the N terminus of human Slit2 from pCMV6-XL5/Slit2 (NM_004778, Origene Technologies, Rockville, Md) into pDONR221 (Invitrogen). The pDONR221/Slit2-N/1118 and pDONR221/Slit2-N/1121 constructs were then recombined separately into pcDNA-DEST40 containing a V5/6xHis tag at the C terminus. Western blot analysis of conditioned media, high-salt extracts, and cell lysates from transiently transfected HEK-293 cells, as well as the generation of a stable cell line overexpressing Slit2-N/1118 are described in the online data supplement.
Migration of PBMCs and HASMCs
Rat PBMCs were isolated by cell-density gradients, and migration was assessed using standard transwell migration inserts. The migration of HASMCs was estimated by a quantitative cell migration assay that combines the use of membrane-based Boyden chambers, and propidium iodide (PI)-staining. The effects of full-length Slit2 were examined by transfecting VSMCs with XL5/Slit2. Detailed protocols are available in the online data supplement.
Growth-arrested HASMCs were exposed to PDGF for 24 hours in the presence and absence of increasing concentrations of Slit2-N/1118. Cells were then labeled with [3H]thymidine for 8 hours, and its incorporation into DNA was measured as described in the online data supplement.
PDGF Receptor Activation Assay
HASMCs were stimulated with Slit2-N/1118 (0.2 μg/mL for 10 minutes) before adding vehicle or PDGF (10 ng/mL) for 2 minutes. The phosphorylation status (tyrosine 751) of the PDGF receptor (PDGFR) was detected by standard immunoblotting using anti-PDGFRβ and anti–phospho-PDGFRβ Abs (1:1000 dilution) (Cell Signaling Technology, Beverly, Mass).
GTPases (Rac1 and Cdc42) were isolated via their specific downstream effectors (Pierce Biotechnology). Briefly, the p21-binding domain (PBD) of the target Pak1, fused to glutathione S-transferase (GST) (GST-PBD), was used to selectively bind GTP-Rac1 and GTP-Cdc42 in HASMC lysates. GTP-Rac1 and GTP-Cdc42 bound to GST-PBD were immobilized on a resin, and the active GTPases eluted with sample buffer. The active GTPases were then detected by Western blotting using Abs specific against Cdc42 and Rac1.
Formation of Lamellipodia in HASMCs
HASMCs were starved for 24 hours and then resuspended in SmBM-2 containing vehicle or 10 ng/mL of PDGF in the presence and absence of Slit2-N/1118 or NSC23766. Cells were plated on fibronectin-coated plates (10 ng/mL). After a 3-hour incubation period at 37°C, HASMCs were fixed with 2% paraformaldehyde, permeabilized with 0.1% Triton X-100, and washed twice with 1% BSA in PBS. The cells were incubated with Alexa Fluor 594 Phalloidin (0.5 μmol/L, Molecular Probes) for 20 minutes to detect F-actin, followed by rinsing with 1% BSA in PBS. Immunofluorescence images were captured using an inverted fluorescence microscope (Olympus IX71).
Results are expressed as mean±SEM. The data were evaluated by t test or 1-way ANOVA. When the overall F test of the ANOVA analysis was significant, a multiple comparison test (Tukey) was applied to determine the sources of differences, which were considered to be statistically significant when P≤0.05.
Slit2/Robo mRNA Expression by Intact Vessels and Cultured Cells
mRNAs encoding Slit2, Robo1, Robo2, and Robo4 were found in intact rat carotid arteries, thoracic aortas, abdominal aortas, and iliac arteries (Figure 1A). Expression levels of Slit2, Robo1, and Robo4 were significantly higher in rat thoracic aortas compared with carotid arteries (Figure 1B). RASMCs and RAECs (Figure 1C), as well as HAECs, HASMCs, and, to a lower extent, human PBMCs (Figure 1D) expressed Slit2 mRNA. Robo1 and Robo4 receptors were expressed by both HASMCs and HAECs. HAECs expressed Robo2, whereas negligible amounts of mRNA encoding for this receptor were detected in HASMCs (Figure 1E). A Robo receptor expression pattern similar to HASMCs and HAECs was found in RASMCs and RAECs, respectively (not shown). Slit1 and Robo3 were robustly and almost exclusively expressed in rat brain (Figure 1A and 1B). Low mRNA expression levels of Slit1 and Robo3 genes were found in carotid arteries and thoracic aortas, whereas abdominal aortas and iliac arteries exhibited undetectable mRNA levels.
Slit2 Protein Localize to Medial VSMCs and the Endothelium of Large Vessels As Well As the Periadventitial Vasa Vasorum
Slit2 protein was expressed in ECs of the main vessel and in scattered adventitial fibroblast–like cells, whereas the medial layer revealed a diffuse Slit2-positive immunostaining (Figure 2A). Adventitial arterioles and veins consistently stained positive for Slit2 (Figure 2B) and α-SMA (Figure 2G). Background staining with a nonimmune IgG isotype Ab was minimal (Figure 2C and 2E). Immunohistochemical staining of cultured RAECs showed strong Slit2 positive staining in the cytoplasm and perinuclear region (Figure 2D).
Recombinant Slit2-N/1121 Remains Tightly Bound to Cell Membranes, Whereas Slit2-N/1118 Is Secreted to the Media
Recombinant Slit2-N/1121 was detected by anti-His Ab and anti-Slit2 Ab exclusively in cell lysates (Figure 3B). No Slit2-N/1121 was detected by either Ab in conditioned media, suggesting that Slit2-N/1121 was not secreted and remained largely associated with cell membranes. Slit2-N/1118 was detected by the anti-His Ab and, to a lower extent, by the anti-Slit2 Ab in cell lysates. However, unlike Slit2-N/1121, Slit2-N/1118 was readily detected by both Abs in conditioned media (Figure 3B). An increase in Slit2-N/1118 yield was observed by high salt extraction; however, Slit2-N/1121 was not detected in either the supernatant or the high-salt extract (not shown). The immunoblotting detection of Slit2-N/1118 with the anti-His Ab consistently showed higher amounts secreted to the media compared with levels detected in cell lysates.
Identification and Biological Activity of Recombinant Slit2N/1118
Slit2-N/1118 secretion to the cell culture media facilitated its isolation and purification. Therefore, we generated a stable cell line to produce larger amounts of recombinant Slit2-N/1118. Anti-His Ab and anti-Slit2 Ab detected Slit2-N/1118 at approximately 140 kDa in purified, dialyzed, and concentrated conditioned media (Figure 4A). Media from the control HEK-293/pcDNA3.1 cell line showed no protein expression (Figure 4B). The purified protein quantitated by silver staining detected a single band of the right molecular size. The biological activity of Slit2-N/1118 was tested in a transwell migration assay using rat PBMCs. SDF-1 induced a 9-fold increase in PBMC migration, which was dose-dependently inhibited by Slit2-N/1118 (Figure 4C). Fifteen nanograms per milliliter Slit2-N/1118 prevented migration of SDF-1–stimulated PBMCs to levels not significantly different from control values (Figure 4C, fifth bar).
Slit2-N Does Not Inhibit PDGF-Mediated Growth of HASMCs
Growth-arrested HASMCs were treated with 10 ng/mL PDGF or vehicle in the presence and absence of increasing concentrations of recombinant Slit2-N and DNA synthesis measured by [3H]thymidine uptake. As shown in Figure 4D, PDGF induced a 9-fold increase in DNA synthesis by HASMCs. Slit2-N/1118 alone did not stimulate DNA synthesis, and it did not inhibit either PDGF-mediated increase in [3H]thymidine incorporation within the range of concentrations tested.
Slit2-N/1118 Inhibits PDGF-Mediated Migration of HASMCs Independent of Interactions With PDGFR
Spontaneous and PDGF-stimulated migration of HASMCs was examined in the presence and absence of purified recombinant Slit2-N/1118. PDGF induced a 16-fold increase in chemotaxis of HASMCs compared with the vehicle alone (PDGF: 383±39; vehicle: 24±4 cells/field) (Figure 5A). Slit2-N/1118 potently inhibited HASMC migration in a concentration-dependent manner at concentrations ranging from 0.1 to 2 μg/mL, with an ED50 of 0.24 μg/mL (Figure 5A). Similar concentrations of Slit2-N/1118 neither stimulated nor inhibited random, spontaneous migration of HASMCs in the absence of PDGF (Figure 5A). To further investigate potential molecular interactions of Slit2-N/1118 with PDGFR on the cell surface, we examined the effects of Slit2-N/1118 on the phosphorylation status of PDGFR by Western blotting using specific Abs against the phosphorylated and nonphosphorylated forms of PDGFRβ. PDGF activated PDGFR in HASMCs at concentrations similar to those having potent chemotactic effects on this cell type (10 ng/mL) (Figure 5B). The incubation of HASMCs in the presence of Slit2-N/1118 (0.20 μg/mL) at concentrations close to the chemorepellent ED50 (0.24 μg/mL) did not prevent PDGFR phosphorylation, indicating that the antichemotactic effects of Slit2-N/1118 in HASMCs are independent of PDGF-mediated activation of PDGFRβ.
Transfection of Full-length Slit2 Abrogates PDGF-Mediated Migration of HASMCs
Unlike Slit2-N/1118, isolation and purification of recombinant full-length Slit2 in amounts required for cell migration and other cell-based assays is restricted by its tightly binding to cell membranes. Thus, to overcome this limitation, we examined the effects of Slit2 on VSMC migration by transfecting full length Slit2 (XL5/Slit2) and a control empty vector (XL5) into HASMCs. Spontaneous HASMC migration in this setting was higher than the levels of random migration observed in untransfected cells (Figure 5A [open circles] and 5C [first bar]). Nonetheless, PDGF induced a significant 2.1-fold increase in directional migration (P<0.002). XL5/Slit2 significantly inhibited random (P<0.02) as well as PDGF-stimulated migration of HASMCs (P<0.001) (Figure 5C).
Involvement of Small GTPase Rac1 in Mediating the Chemorepellent Actions of Slit2-N/1118 in HASMCs
Using GST-PBD fusion proteins expressing the downstream effectors of Rac1 and Cdc42, we detected their active forms (GTP-Rac1 and GTP-Cdc42) by Western blotting. GTP-Rac1 was strongly activated by PDGF at concentrations that induced a potent chemotactic effect (10 ng/mL). The maximum activation of Rac1 was observed at 2 minutes (Figure 6A, top). The PDGF-mediated activation of Rac1 was transient and incubation times longer than 2 minutes (5 and 10 minutes) exhibited significantly lower levels of activated Rac1. Total Rac1 levels were not modified by PDGF at any time point (Figure 6A, second panel). GTP-Cdc42 was constitutively active in HASMCs and PDGF had little or no effect on the levels of activated GTP-Cdc42 or total Cdc42 (Figure 6A, third and fourth panels). We then tested the effects of Slit2-N/1118 in the presence and absence of PDGF. Although the constitutive expression levels of GTP-Rac1 were low, Sit2-N/1118 alone induced a discernible inhibitory effect on GTP-Rac1 at 5, 10, and 20 minutes, whereas total levels of Rac1 remained unchanged (Figure 6B). HASMCs were stimulated for 2 minutes with PDGF (10 ng/mL) in the presence of increasing concentrations of Slit2-N/1118 (25 to 200 ng/mL). A concentration-dependent inhibition of GTP-Rac1 without changes in total Rac1 was induced by Slit2-N/1118 (Figure 6C). These results suggest that Rac1 is a signal transduction molecule involved in the chemorepellent activity of Slit2-N/1118 in HASMCs.
Rac1 Inhibition Prevents PDGF-Mediated Migration of HASMCs
To further test the hypothesis that the inhibition of Rac1 would abolish the PDGF-mediated chemotactic effects on HASMCs, we used NSC23766, a specific Rac1 inhibitor.14 Western blot analysis of activated Rac1 revealed that GTP-Rac1 was concentration-dependently inhibited by NSC23766, whereas total Rac1 was unmodified (Figure 7A, first and second panels). GTP-Cdc42 and total Cdc42 levels were not affected by NSC23766 at the same doses used to inhibit Rac1 (50 to 100 μmol/L) (Figure 7A, third and fourth panels). The addition of NSC23766 inhibited the PDGF-induced activation of Rac1 in a concentration-dependent manner (10 to 100 μmol/L) without affecting total Rac1 (Figure 7B). NSC23766 also elicited inhibition of PDGF-mediated HASMC migration in the transwell migration assay in a concentration-dependent fashion (10 to 200 μmol/L), with an ED50 of 60 μmol/L (Figure 7C).
Slit2-N/1118 and NSC23766 Inhibit PDGF-Induced Formation of Lamellipodia in HASMCs
We seeded HASMCs in fibronectin-coated plates. Unstimulated HASMCs were spread out and exhibited an irregularly oriented stellate morphology with well-defined stress fibers detected by Alexa Fluor 594 Phalloidin staining (Figure 7D, top left). The incubation of HASMCs with PDGF resulted in the formation of lamellipodia (filamentous actin-rich protrusions) and cell body elongation, suggestive of migration (Figure 7D, bottom left). Stress fibers were considerably less discernible, and even disrupted, in HASMCs stimulated by PDGF in areas close to the leading lamellipodial edge. The addition of Slit2-N/1118 or NSC23766 prevented the formation of lamellipodia and returned the morphological state of HASMCs to the original irregularly spread, stellate shape, with abundant stress fibers across the entire length of the cell (Figure 7D, top and bottom right).
As a basic cellular feature that occurs in a wide range of organisms during development and pathological states, migration requires “attraction,” “repulsion,” and “steering” signals for cells to reach their distal targets. Slit2, a chemorepellent molecule originally described in Drosophila, has been shown to inhibit neuronal, leukocyte, and EC migration.8,10,15 These studies prompted us to investigate whether chemorepellent biological signals were involved in modulating VSMC migration. We found transcriptional expression of Slit2 and Robo receptors in intact vascular tissues and in cultured ECs and VSMCs. Immunohistochemical localization of Slit2 in intact vessels indicated that adventitial arterioles and veins, as well as medial VSMCs and luminal ECs, expressed Slit2 protein. In a separate study in balloon-injured rat carotid arteries (H.D.L. et al, unpublished data, 2005), we have demonstrated upregulation of Slit2, Slit3, Robo1, Robo2, and Robo4 mRNA expression, which is associated with the adventitial infiltration of leukocytes that we have reported in this model.16 We hypothesize that Slit2 chemorepellent actions may be involved in modulating chemokine-mediated attraction of leukocytes in the adventitial vasa vasorum of injured vessels, where inflammatory cells are recruited minutes after endothelial denudation. In the current study, leukocytes were greater than 2 orders of magnitude more sensitive than VSMCs to the inhibitory effects of Slit2-N/1118. Leukocytes migrate at a faster rate than VSMCs. The migratory behavior of leukocytes can be monitored in seconds,8 whereas the morphology of migrating VSMCs is evident hours after applying a chemotactic stimulus. Thus, the migratory nature of inflammatory cells, compared with quiescent VSMCs, may account for the higher sensitivity of leukocytes to repulsion signals.
However, a functional association between Slit2/Robo and leukocyte chemorepulsion in naive vessels is unlikely, as these vessels exhibit no leukocyte infiltration. Similarly, Slit2-mediated repulsion of cells displaying no apparent migratory behavior, SMCs in the media layer and in adventitial arterioles of naive arteries, is an improbable event. Thus, the functional role of Slit2 in intact vascular structures is unclear. We speculate that Slit2 may be constitutively expressed by naive vessels to prevent random infiltration of leukocytes and/or to preserve the nonmigratory state of VSMCs. Of note, leukocytes were particularly sensitive to the chemorepellent actions of Slit2 and forced expression of Slit2 into HASMCs inhibited spontaneous, random migration in the absence of PDGF. A role for Slit2 in VSMC growth is unlikely as Slit2-N/1118 had no influence on PDGF-stimulated VSMC DNA synthesis. These findings suggest that Slit2 inhibits VSMC migration without affecting other key functions as previously reported in leukocytes.8 However, our data cannot rule out the effects of full-length Slit2, Slit2-C, or other Slit2-N fragments on VSMC proliferation mediated by PDGF or by additional mitogens.
PDGF is a potent chemoattractant of VSMCs.6 Slit2-N/1118 inhibited PDGF-mediated chemotaxis of HASMCs in a concentration-dependent manner. Although the receptor mediating the actions of Slit2 in HASMCs remains to be identified, our expression data suggest that Robo1 and Robo4 are likely candidates. Slit2-N/1118 did not prevent PDGFRβ activation in VSMCs, indicating that Slit2 actions are not accounted for by molecular interactions between Slit2 and PDGF or PDGFRβ. Transfection of HASMCs with a vector encoding full-length Slit2 prevented PDGF-induced migration, suggesting that HASMCs may synthesize the protease that cleaves Slit2. The activation of small Rho GTPases by PDGF17 and the identification of Cdc42 as the primary Slit/Robo signaling pathway in neuronal migration18 led us to test the effects of Slit2-N/1118 on small Rho GTPases in VSMCs. Unexpectedly, PDGF did not have a significant effect on Cdc42 activation in HASMCs. However, Slit2-N/1118 showed a concentration-dependent inhibition of PDGF-stimulated levels of activated Rac1, and Rac1 inhibition by NSC2376614 was sufficient to prevent PDGF-mediated migration of HASMCs. These results suggest that Rac1 is involved in the Slit2/Robo pathway in HASMCs. Nonetheless, our data cannot exclude the contribution of additional pathways downstream of Robo, including RhoA GTPase. Slit/Robo GTPase activating proteins (srGAPs) bind to the intracellular domain of Robo and promote the hydrolysis of GTP-Cdc42.18 In HEK cells expressing Robo1, srGAP1 binds to Cdc42 and RhoA.18 However, srGAPs neither bind to nor regulate Rac1,18 and therefore they are unlikely to provide a direct link between Robo and Rac1 in VSMCs. An additional family of RhoGAPs, the Vilse proteins, also bind to the intracellular domains of Robo receptors, and they promote hydrolysis of GTP-Rac.19 Therefore, Vilse’s human homolog19 or other yet unidentified VSMC-specific RhoGAPs may be responsible for Rac1 inactivation in HASMCs.
The small GTPase Rac1 has been shown to govern the assembly of PDGF-mediated formation of lamellipodia,20 one of the first steps in the coordinated cellular processes that lead to directional locomotion.21 Slit2-N/1118 and NSC23766 inhibited PDGF-mediated lamellipodia formation, actin cytoskeleton disruption, and cell body elongation in HASMCs. Although GTPases are best known as regulators of the actin cytoskeleton, they also control cell polarity and microtubule dynamics.22 Thus, Slit2-N/1118-mediated inhibition of Rac1 may not be confined to reorganization of the cell cytoskeleton.
Like Hedgehog proteins,23 Slit2 undergoes a proteolytic cleavage yielding 2 fragments that exhibit different cell-association characteristics.7,10 The processed fragments may then have different diffusion rates, and different binding interactions with Robo receptors (eg, cell-to-cell contact versus contact independent) and ECM proteins.24,25 It is not yet clear whether Slit2-N remains fully associated to the membrane or is partially released from expressing cells. Identification of the structural motifs involved in the cell association of Slit2-N will aid in understanding its chemorepulsive actions. We found that Slit2-N/1121, the putative endogenous cleaved product, is tightly bound to the cell membrane, supporting a cell contact–dependent mechanism of action, as previously suggested for dSlit at the Drosophila midline.26 Unlike Slit2-N/1121, Slit2-N/1118, a shorter, yet biologically active, fragment, was secreted to the media. Thus, preservation of Leu1119, Pro1120, and Arg1121 conferred on Slit2-N/1121 the ability to remain anchored to the cell membrane. Our data extend the original observations by Brose et al and Nguyen Ba-Charvet et al10,27 and suggest that Leu1119, Pro1120, and Arg1121 are involved in the physical binding of Slit2-N to the cell membrane. Partial detection of Slit2-N/1118 in cell lysates suggests that additional aa upstream of Leu1119 may also be involved in this association.
It has been suggested that Slit2-C might be responsible for the long-range repulsive effects of Slit2 on spinal motor axons10 and mesodermal cells.7 Under this model, it is difficult to reconcile the nature of the receptor mediating the Slit2-C repulsive actions as this fragment does not bind Robo receptors. Another hypothesis advocates that small amounts of the N-terminal fragment may diffuse away from the source cells to elicit the long-range effects.10 This mechanism of action, which has been documented for the N-terminal fragment of Hedgehog,28 assumes that distant target cells display higher sensitivity to lower concentrations of Slit ligand.10 Alternatively, long-range, contact-independent effects may be explained by cleavage of shorter and more diffusible Slit2-N fragments (eg, Slit2-N/1118) by tissue-specific proteases. Whether these mechanisms operate in the vascular microenvironment is unknown. Therefore, identification of the vascular cleaving protease(s), as well as the proteolytic fragments, is fundamental to our understanding of Slit2-mediated arterial cell chemorepulsion.
In the central nervous system, Slit2 interacts with Netrin-1,29 and other neuronal chemorepellents including Nogo-B have been recently shown to have similar antichemotactic functions in VSMCs.30 The roles of these embryonic axonal path-finding molecules in vascular remodeling and vessel guidance is emerging rapidly.31 In summary, our data indicate that a chemorepellent system originally described in neuronal development has been evolutionary conserved in adult arterial tissues. Further studies will be required to decipher the precise molecular mechanisms that govern Slit/Robo-mediated chemorepulsion of arterial cells in the extracellular environment.
This work was supported in part by funding from the Georgia Tech/Emory Center (GTEC) for the Engineering of Living Tissues, an ERC Program of the National Science Foundation under award no. EEC-9731643.
Original received October 16, 2005; revision received December 20, 2005; accepted January 13, 2006.
Libby P. Inflammation in atherosclerosis. Nature. 2002; 19–26;420:868–874.
Majesky MW, Giachelli CM, Reidy MA, Schwartz SM. Rat carotid neointimal smooth muscle cells reexpress a developmentally regulated mRNA phenotype during repair of arterial injury. Circ Res. 1992; 71: 759–768.
Mulvihill ER, Jaeger J, Sengupta R, Ruzzo WL, Reimer C, Lukito S, Schwartz SM. Atherosclerotic plaque smooth muscle cells have a distinct phenotype. Arterioscler Thromb Vasc Biol. 2004; 24: 1283–1289.
Okamoto EE, Couse T, de Leon H, Vinten-Johansen J, Goodman RB, Scott NA, Wilcox JN. Perivascular inflammation after balloon angioplasty of porcine coronary arteries. Circulation. 2001; 104: 2228–2235.
Wang KH, Brose K, Arnott D, Kidd T, Goodman CS, Henzel W, Tessier-Lavigne M. Biochemical purification of a mammalian slit protein as a positive regulator of sensory axon elongation and branching. Cell. 1999; 19: 96:771–784.
Gao Y, Dickerson JB, Guo F, Zheng J, Zheng Y. Rational design and characterization of a Rac GTPase-specific small molecule inhibitor. Proc Natl Acad Sci U S A. 2004; 101: 7618–7623.
Hay C, Micko C, Prescott MF, Liau G, Robinson K, de Leon H. Differential cell cycle progression patterns of infiltrating leukocytes and resident Cells after balloon injury of the rat carotid artery. Arterioscler Thromb Vasc Biol. 2001; 21: 1948–1954.
Lundstrom A, Gallio M, Englund C, Steneberg P, Hemphala J, Aspenstrom P, Keleman K, Falileeva L, Dickson BJ, Samakovlis C. Vilse, a conserved Rac/Cdc42 GAP mediating Robo repulsion in tracheal cells and axons. Genes Dev. 2004; 18: 2161–2171.
Pichon S, Bryckaert M, Berrou E. Control of actin dynamics by p38 MAP kinase - Hsp27 distribution in the lamellipodium of smooth muscle cells. J Cell Sci. 2004; 117: 2569–2577.
Lee JJ, Ekker SC, von Kessler DP, Porter JA, Sun BI, Beachy PA. Autoproteolysis in hedgehog protein biogenesis. Science. 1994; 266: 1528–1537.
Nguyen Ba-Charvet KT, Brose K, Ma L, Wang KH, Marillat V, Sotelo C, Tessier-Lavigne M, Chedotal A. Diversity and specificity of actions of Slit2 proteolytic fragments in axon guidance. J Neurosci. 2001; 21: 4281–4289.
Eichmann A, Makinen T, Alitalo K. Neural guidance molecules regulate vascular remodeling and vessel navigation. Genes Dev. 2005; 19: 1013–1021.