Sos1 Regulates Macrophage Podosome Assembly and Macrophage Invasive Capacity This information is current as of October 15, 2015.

Anna Baruzzi, Sabrina Remelli, Erika Lorenzetto, Michela Sega, Roberto Chignola and Giorgio Berton

Supplementary Material

http://www.jimmunol.org/content/suppl/2015/10/06/jimmunol.150057 9.DCSupplemental.html

Subscriptions

Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscriptions

Permissions Email Alerts

Submit copyright permission requests at: http://www.aai.org/ji/copyright.html Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/cgi/alerts/etoc

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 9650 Rockville Pike, Bethesda, MD 20814-3994. Copyright © 2015 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606.

Downloaded from http://www.jimmunol.org/ at Univ of Otago Med Lib on October 15, 2015

J Immunol published online 7 October 2015 http://www.jimmunol.org/content/early/2015/10/06/jimmun ol.1500579

Published October 7, 2015, doi:10.4049/jimmunol.1500579 The Journal of Immunology

Sos1 Regulates Macrophage Podosome Assembly and Macrophage Invasive Capacity Anna Baruzzi,*,1 Sabrina Remelli,*,1 Erika Lorenzetto,† Michela Sega,‡ Roberto Chignola,‡ and Giorgio Berton*

M

acrophages are a heterogeneous population of cells widely distributed in the body and present in the lymphoid tissue, in serous cavities, in the interstitium of different organs, and in the perivascular space (1, 2). More than 40 years ago the proposal was advanced that they form, together with monocytes circulating into the blood and their bone marrow precursors, a mononuclear phagocyte system (1). Although the name given to this system emphasized their being phagocytic as the trait distinguishing macrophages from other cells, such as endothelial cells, fibroblasts, and other perivascular cells, previously included in wider systems (the reticuloendothelial or the reticulohistiocyte systems) (1), subsequent studies revealed that macrophages perform additional functions underscoring their capability to regulate innate and adaptive immunity and tissue repair. Notably, studies performed in the last two decades implicated *Department of Pathology and Diagnostics, Section of General Pathology, University of Verona, 37134 Verona, Italy; †Department of Neurological and Movement Sciences, University of Verona, 37134 Verona, Italy; and ‡Department of Biotechnology, University of Verona, 37134 Verona, Italy 1

A.B. and S.R. contributed equally to this work.

Received for publication March 13, 2015. Accepted for publication September 3, 2015. This work was supported by Italian Association for Cancer Research Grant IG-10267 (to G.B.) and a fellowship from the Ministero dell’Istruzione, dell’Universita` e della Ricerca (to M.S.). Address correspondence and reprint requests to Prof. Giorgio Berton, Department of Pathology and Diagnostics, Section of General Pathology, University of Verona, Strada Le Grazie 8, 37134 Verona, Italy. E-mail address: [email protected] The online version of this article contains supplemental material. Abbreviations used in this article: BMDM, bone marrow–derived macrophage; ECM, extracellular matrix; GEF, guanine nucleotide exchange factor; hMDM, human monocyte-derived macrophage; LPA, lysophosphatidic acid; RIPA, radioimmunoprecipitation assay buffer; siRNA, small interfering RNA. Copyright Ó 2015 by The American Association of Immunologists, Inc. 0022-1767/15/$25.00 www.jimmunol.org/cgi/doi/10.4049/jimmunol.1500579

macrophages in the pathogenesis of a few of the most common pathologies affecting the human population, including atherosclerosis and obesity-induced type 2 diabetes, and highlighted their role as cells promoting tumor growth and dissemination (reviewed in Refs. 3–5). One central aspect of mononuclear phagocytes is to be endowed with a strong migratory ability responsible for migration of monocytes from the blood to extravascular compartments and of macrophages within the interstitium, two forms of cell movement that are believed to represent an example of migration in two or three dimensions, respectively (reviewed in Refs. 6, 7). Accumulating evidence points to the existence of different modalities of three-dimensional macrophage migration within the tissue, which are dictated by the density of the extracellular matrix (ECM) and are referred to as the ameboid, which occurs in loose matrices, and the mesenchymal, which occurs in dense matrices, migration mode (6–8). Features of the mesenchymal migration mode are the acquirement of an elongated cell morphology, integrin-dependent adhesion to the ECM, and proteolytic degradation of ECM components. Notably, these last two events have been shown to depend on the organization of podosomes, cell projections consisting of an F-actin core enriched with actinbinding and polymerizing proteins, as well as integrins and integrin-associated signaling proteins, which form a privileged site of traffic of protease-containing vesicles (7, 9–11). Podosomes that were described more than 25 years ago in osteoclasts (12) and have been characterized in both mononuclear phagocytes (reviewed in Refs. 6, 7) and lymphocytes (13) are thought to represent a sort of normal counterpart of invadopodia, podosomelike structures implicated in adhesion to and degradation of the ECM by neoplastic cells (reviewed in Refs. 9, 11). Signal transduction implicated in podosome assembly and function in macrophages has been addressed in some detail. The

Downloaded from http://www.jimmunol.org/ at Univ of Otago Med Lib on October 15, 2015

Podosomes are protrusive structures implicated in macrophage extracellular matrix degradation and three-dimensional migration through cell barriers and the interstitium. Podosome formation and assembly are regulated by cytoskeleton remodeling requiring cytoplasmic tyrosine kinases of the Src and the Abl families. Considering that Abl has been reported to phosphorylate the guanine nucleotide exchange factor Sos1, eliciting its Rac-guanine nucleotide exchange factor activity, and Rac regulates podosome formation in myeloid cells and invadopodia formation in cancer cells, we addressed whether Sos1 is implicated in podosome formation and function in macrophages. We found that ectopically expressed Abl or the Src kinase Fgr phosphorylate Sos1, and the Src kinases Hck and Fgr are required for Abl and Sos1 phosphorylation and Abl/Sos1 interaction in macrophages. Sos1 localizes to podosomes in both murine and human macrophages, and its silencing by small interfering RNA results in disassembly of murine macrophage podosomes and a marked reduction of GTP loading on Rac. Matrix degradative capacity, three-dimensional migration through Matrigel, and transmigration through an endothelial cell monolayer of Sos1-silenced macrophages were inhibited. In addition, Sos1- or Abl-silenced macrophages, or macrophages treated with the selective Abl inhibitor imatinib mesylate had a reduced capability to migrate into breast tumor spheroids, the majority of cells remaining at the margin and the outer layers of the spheroid itself. Because of the established role of Src and Abl kinases to regulate also invadopodia formation in cancer cells, our findings suggest that targeting the Src/Abl/Sos1/Rac pathway may represent a double-edged sword to control both cancer-invasive capacities and cancer-related inflammation. The Journal of Immunology, 2015, 195: 000–000.

2

Sos1 REGULATES MACROPHAGE CYTOSKELETON DYNAMICS

Materials and Methods Reagents and Abs Abs used for immunoblotting experiments and their sources were: antiAbl and anti-Sos1 (Santa Cruz Biotechnology, Santa Cruz, CA), antiphosphotyrosine (Millipore, Billerica, MA), and anti-Fgr (rabbit antiserum) and anti-Rac1 (clone 23A8) (Upstate, Lake Placid, NY), anti-actin (SigmaAldrich, St. Louis, MO). Abs used for immunofluorescence experiments and their sources were: anti-vinculin (clone hVin-1; Sigma Aldrich), anti-Sos1 and anti-Rac1 (Abcam, Cambridge, U.K.), and anti-active Rac1 and antiactive Ras (NewEast, Malvern, PA). Secondary goat anti-mouse IgG1 Alexa 488–conjugated, goat anti-mouse IgM Alexa 488–conjugated, and goat anti-rabbit Alexa 647–conjugated Abs were from Invitrogen (Carlsbad, CA). Rhodamine phalloidin from Cytoskeleton (Denver, CO) was used to label F-actin. Imatinib/Gleevec/STI-571 was from Santa Cruz Biotechnology. Lysophosphatidic acid (LPA) and GNF-2 were from SigmaAldrich.

Cell isolation and culture Bone marrow–derived macrophages (BMDMs) were isolated from femurs and tibias of 8 wk-old wild-type C57BL/6J mice and cultured in DMEM supplemented with 15% FBS and 10% L929 cell-conditioned medium as previously described (17, 31). After 7–8 d of culture, macrophages were

detached by scraping and then plated for 24 h on fibronectin- or gelatinFITC–coated coverslips in DMEM supplemented with 1% FBS. Blood was collected from healthy volunteers, who provided their written, informed consent according to a protocol approved by the local Ethical Committee, at the Blood Bank of the University of Verona. Monocytes were isolated from buffy coats by Ficoll-Hypaque and Percoll (GE Healthcare Life Science, Uppsala, Sweden) density gradients. To generate human monocytederived macrophages (hMDMs), we cultured monocytes at 37˚C in 5% CO2 for 2 h in RPMI 1640 medium (Lonza, Basel, Switzerland) without FBS; then the medium was replaced by RPMI 1640 supplemented with heatinactivated, low-endotoxin FBS (10% final concentration), 2 mM L-glutamine, and 20 ng/ml M-CSF (Peprotech, Rocky Hill, NJ). Cells were maintained in culture for 6–8 d before their use. The culture medium was usually renewed on the third or fourth day. For fluorescence microscopy (confocal analysis of podosome formation), hMDMs were left untreated (M0) or stimulated with 20 ng/ml IL-4 (Peprotech) for 18 h to obtain M2 polarized macrophages. During the polarization process, 10 ng/ml M-CSF was maintained in the culture medium. Human macrophages were then detached with Accutase (Millipore, Billerica, MA) for 30 min at 37˚C, suspended in RPMI 1640 with 1% FBS, 2 mM L-glutamine, 10 ng/ml M-CSF, and 20 ng/ml IL-4, and plated on fibrinogen (10 mg/ml)-coated coverslips for 24 h. COS-7 cells were from ATCC (Rockville, MA) and cultured in DMEM supplemented with Glutamax (Orange, Cambrex, Switzerland) and 10% FBS. BaF3 cells expressing Bcr/Abl were cultured in RPMI 1640 medium with 10% FBS.

Transfection and RNA interference For COS-7 cell transfection experiments, 60–80% confluent cells in 100-mm dishes were transfected using Metafectene PRO reagent from Biontex (M€unchen/Laim) according to the manufacturer’s instructions. A construct containing a constitutively active Abl mutant lacking the autoregulatory SH3 domain (28) was kindly donated by Prof. G. Scita (IFOM and Department of Health Sciences, University of Milan). Constructs containing full-length Fgr, myristoylation-defective Fgr, or Fgr with a point mutation causing a lysine-to-arginine [Fgr(K279R)] substitution inactivating the c-Fgr kinase activity have been described (32, 33). Transfection of BMDMs was carried out by electroporation using the Nucleofector technology of Amaxa (Koel, Germany) according to proposed protocols. Cells were transfected with control nonsilencing siRNA pool or mouse-specific ON-TARGET plus siRNA reagents targeting Abl or Sos1 (Dharmacon, Lafayette, CO). For fluorescence microscopy (confocal analysis of podosome formation) and assays of gelatin degradation and transendothelial migration, cells were detached by scraping after 48 h from transfection and plated on fibronectin-coated or gelatin-coated coverslips, or endothelial cell–coated FluoroBlok Inserts, for a further 24 h (see later). For immunoblot analysis, cells were assayed after 72 h of culture. An aliquot of BMDMs used for the different assays was lysed to control for the efficacy of Abl or Sos1 silencing by the siRNA-specific reagent. Mean percent of Abl or Sos1 expression in BMDMs treated with siRNA targeting Abl or Sos1 was 37.8 6 11 and 37.3 6 2% compared with control siRNAtreated BMDMs, respectively.

Immunostaining and confocal analysis Macrophages were detached as described earlier and plated in 18-mm glass coverslips precoated with fibronectin or fibrinogen. Cells were fixed with 4% (w/v) paraformaldehyde for 30 min and, after washing with PBS and quenching with 50 mM NH4Cl, permeabilized with PBS-0.1% Triton X-100. After quenching with 1% BSA for 30 min, cells were stained with primary Abs for 1 h, washed with PBS, and then stained with secondary Abs and rhodamine-phalloidin for 30 min, followed by DAPI (Sigma Aldrich) for 10 min. Images of the ventral, substrate-adherent face of the cell were collected using the SP5 confocal microscope from Leica Microsystems (Mannheim, Germany) with a 633 objective (HCX PL APO l blue 1.4NA OIL). Images were processed for brightness and contrast with Adobe Photoshop.

Gelatin degradation assay For gelatin degradation assays, coverslips were incubated with poly-Llysine for 20 min, washed with PBS, and then incubated with 0.5% glutaraldehyde for 15 min. After washing with PBS, coverslips were put on a drop of 0.2 mg/ml Gelatin-FITC in PBS/2% sucrose, left for 10 min, and washed again with PBS. BMDMs were plated for 24 h on gelatin-FITC– coated coverslips. Images were collected using a DM6000B fluorescence microscope (Leica Microsystems) with a 203 (0.5 NA) objective. Mean area of gelatin degradation was quantified by counting a degraded area in 15–20 different fields containing approximately the same number of cell nuclei.

Downloaded from http://www.jimmunol.org/ at Univ of Otago Med Lib on October 15, 2015

myeloid leukocyte-specific Src-family kinase member Hck has been demonstrated to regulate podosome formation and podosomedependent ECM degradation (8, 14, 15), an action that may in part depend on its capability to phosphorylate the Wiskott–Aldrich syndrome protein (16). In addition, we reported that the cytoplasmic tyrosine kinase Abl, which forms a tripartite complex with the Src kinases Hck or Fgr bound to integrins (17), redistributes to podosomes, and its silencing or pharmacologic inhibition results in podosome disassembly and reduction of macrophage threedimensional migration and ECM degradative capacity (18). Notably, Abl has been also implicated in invadopodia formation in neoplastic cells (19, 20). Downstream targets of Src kinases and Abl regulating macrophage podosome assembly are poorly understood. However, both Src and Abl impinge on Rac activation (17) and Rac1 and Rac2 regulate macrophage podosome formation (21). In addition, Rac proteins have been reported to regulate assembly of podosomes and invadopodia in different cell types (see Refs. 10 and 22–26 for a review of this topic). Hence we reasoned that the Src/Abldependent signaling pathway leading to podosome formation must involve activation of a guanine nucleotide exchange factor (GEF) for Rac. Among Rac-GEFs, Sos1, a dual GEF for Ras and Rac (27), is a likely candidate to act within this signaling pathway. In fact, tyrosine phosphorylation of Sos1 by Abl is sufficient to elicit its Rac-GEF activity (28). In addition, palladin, a phosphoprotein regulating cytoskeleton dynamics and able to interact with Eps8, an activator of the Rac-GEF activity of Sos1, was shown to localize to podosomes (29). Notably, Sos1 binds to Tks5, a podosome and invadopodia component (9), in epithelial A549 cells and Src-transformed NIH 3T3 cells (30). We report that Sos1 is localized in both murine and human macrophage podosomes, and its silencing by small interfering RNA (siRNA) results in an almost complete disassembly of murine macrophage podosomal rosettes. In addition, we found that Sos1 is indispensable for Rac1 activation and its redistribution to podosomes. Silencing of Sos1 resulted in impairment of macrophage ECM degradative capacity, three-dimensional migration through Matrigel, transmigration through an endothelial layer, and infiltration into tumor spheroids. Collectively, these findings highlight a new, essential component of macrophage podosomes playing an indispensable role in regulation of three-dimensional migration.

The Journal of Immunology In vitro migration assay

In vitro adhesion assay For in vitro adhesion assays, after 72 h from transfection, BMDMs were labeled with the fluorescent dye Celltracker CMFDA (Invitrogen) according to the producer’s instructions. Cells detached as earlier described were plated on fibronectin-coated wells, and fluorescence was read immediately with a Victor Multilabel Plate Reader (PerkinElmer) to obtain the total cell fluorescence signal. After 30 min, macrophage monolayers were washed twice with PBS and the fluorescence of adherent cells was read.

Immunoprecipitation and immunoblotting BMDMs or COS-7 cells were lysed in radioimmunoprecipitation assay buffer [50 mM Tris (pH 7.5), 150 mM NaCl, 1% (v/v) Triton X-100, 1% Na-deoxycholate, 0.1% NaDodSO4 (SDS), 5 mM EDTA, 1 mM DTT, 10 mM Na2VO4, 10 mM phenylarsine oxide, and one tablet of Complete Mini EDTA-free protease inhibitor mixture from Roche Molecular Biochemicals for 10 ml lysis buffer). After clarification in a microfuge, lysates were incubated with 2–5 mg Abs bound to 20 ml protein G immobilized to Sepharose (GE Healthcare) for each sample, and precipitated proteins were washed three times with RIPA buffer. Samples were separated on SDSPAGE gels and transferred to nitrocellulose Hybond C (Amersham). After quenching with Odyssey blocking buffer (LI-COR Biosciences, Lincoln, NE), blots were incubated overnight at 4˚C with primary Abs, followed by donkey anti-rabbit or goat anti-mouse Abs, and then analyzed with Odyssey Infrared Imaging System (LI-COR Biosciences). For analysis of protein expression in whole lysates, cells were lysed with sample buffer (25 mM Tris, pH 6.8, 50 mM 2-ME, 1% SDS, and 5% glycerol), and samples were separated on SDS-PAGE gels before Western blotting analysis.

Rac1 activity assay GTP loading on Rac1 was assayed by the Rac1 Activation Assay Biochem Kit according to the manufacturer’s instructions (Cytoskeleton, Denver, CO). Proteins were separated by SDS-PAGE and, after blotting, membranes were probed with anti-Rac1 Ab.

Spheroids invasion assay Cells from the human breast cancer cell line T47D (ECACC number 85102201) were used to generate multicellular tumor spheroids using the liquid overlay technique. In brief, exponentially growing cells were seeded (500/well) in 96-multiwell plates precoated with an agar gel layer (50 ml of a 0.7% w/v solution of agar in complete medium) and grown in RMPI 1640 medium supplemented with 2 mM glutamine, 35 mg/l gentamicin, and 10% heat-inactivated FBS (34). Once the spheroids reached ∼400 mm in diameter [as checked by using an EVOS (Life Technologies) digital microscope], 104 BMDMs, previously marked with the cell tracker green CMFDA (Invitrogen, Carlsbad, CA), were added to each well. In experiments in which we examined the migration mode of BMDMs, after labeling with CMFDA, cells were incubated for 1 h before addition to spheroids with a mixture of inhibitors of mesenchymal (100 mM E64C, 5 mM GM6001, 2 mM pepstatin, 6 mM leupeptin) or amoeboid migration (20 mM Y27632) (35). Inhibitors were kept in the incubation medium for

the whole length of the assay. After 24 h, single spheroids were collected, washed repeatedly with PBS, and fixed with 4% (w/v) paraformaldehyde for analysis under a Leica TCS-SP5 (Leica Microsystems, Mannheim, Germany) confocal microscope equipped with a Chameleon Ultra (Coherent, CA) multiphoton laser. Images were collected by using a 203 objective (HCX APO, 1.00 NA; Leica Microsystems, Mannheim, Germany). Z-stacks of the entire spheroids were acquired by exciting the CMFDA signal at 800 nm and collecting the fluorescence signal in the wavelength range of 510 to 535 nm. To visualize the edge and surface of the spheroids, we collected the confocal reflection signal in parallel by using the 514 nm laser line and operating with the AOBS in reflection mode. Stacks of digital images taken at 5-mm distance along the vertical Z-axis were analyzed using the Image Processing and Analysis package built-in software Mathematica (v. 10; Wolfram Research, Champaign, IL). Color images were first converted to grayscale and then to binary using a global threshold of 0.6. Spots corresponding to fluorescent cells were automatically individuated using the morphological components function with the ConvexHull method that finds nonoverlapping convex hulls of connected foreground image components, and the spots’ centroids determined in px coordinates using the component measurement function and Centroid method. The procedure was validated by plotting the centroids on test images. Coordinate units were then converted knowing that the spatial resolution of digital images was 1 px = 1.44 mm. The vertical coordinate of each spot was carefully calculated from images taken in parallel with the same microscope using the reflection mode to identify the first optical section crossing the surface of the spheroid. The same images were used to measure the mean spheroid diameter (average of the four longest chords) and the three-dimensional coordinates of the spheroid centroid. Because the vertical distance between images in the stack was comparable with cell length, care was taken to distinguish between signals from different cells. The Euclidean distance between the centroids of fluorescent spots was therefore calculated. The signals were then considered as belonging to two different cells if the distance between centroids resulted greater than a macrophage diameter, which was assumed to be, on average, 15 mm. The normalized migration distance of each cell in the tumor spheroid was finally calculated as follows: (R 2 d)/R, where R is the mean spheroid radius and d the Euclidean distance of each cell from the spheroid centroid.

Statistics Statistical significance of differences between the data were analyzed either by the Student t test or ANOVA with Bonferroni’s posttests (see figure legends), performed with GraphPad Prism software (GraphPad Software, La Jolla, CA). Differences between normalized migration distances of macrophages infiltrating tumor spheroids were analyzed with the Kolmogorov– Smirnov nonparametric test.

Results Tyrosine phosphorylation of Sos1 by Abl and Fgr In previous studies we showed that the Src-family tyrosine kinases Hck and Fgr cross-talk with Abl to regulate integrin-dependent macrophage migration, and podosome assembly and function (17, 18). In addition, we found that either the genetic deficiency of Hck and Fgr or the pharmacologic inhibition of Abl resulted in inhibition of chemoattractant-induced Rac activation (17). Considering that Sos1, a dual GEF for Ras and Rac, has been implicated in integrin-dependent Rac activation (36) and its tyrosine phosphorylation by Abl is sufficient to elicit its Rac-GEF activity (28), we addressed the role of Sos1 in the regulation of macrophage cytoskeleton dynamics downstream Src family kinases and Abl. In initial studies (Fig. 1A), we confirmed that in COS-7 cells expressing a constitutive active mutant of Abl, Sos1 is phosphorylated in tyrosine, and imatinib, a pharmacologic inhibitor of Abl, inhibits its phosphorylation but not formation of constitutive Abl/Sos1 complexes (see Ref. 28). An increase in tyrosine phosphorylation of Sos1 was also detected in the murine hematopoietic cell line Baf3 expressing a constitutively active Bcr-Abl fusion protein; notably, both imatinib and GNF-2, a specific allosteric inhibitor of Abl kinases (37, 38), inhibited Sos1 phosphorylation, but not Abl/Sos1 association in these cells (Fig. 1B). Finally, Sos1 tyrosine phosphorylation was increased in COS-7

Downloaded from http://www.jimmunol.org/ at Univ of Otago Med Lib on October 15, 2015

For transendothelial migration assays, H5V cells, an epithelial cell line kindly provided by E. Dejana (FIRC Institute of Molecular Oncology, Milan, Italy), were plated on FluoroBlok Inserts (Falcon) in DMEM supplemented with 10% FBS for 3 d until they formed a confluent monolayer, then activated with 5 ng/ml TNF for 2 h. A total of 105 BMDMs labeled with Celltracker Blue CMAC (Invitrogen) according to the producer instruction, and resuspended in DMEM without serum, were then plated on the FluoroBlok inserts. After 22 h, percentage of cells migrated in the lower chamber was calculated by reading the fluorescence with a Victor Multilabel Plate Reader (PerkinElmer, Waltham, MA). For Matrigel migration assays, BMDMs were detached and starved in DMEM without serum for a total of 3 h. After 2 h of starving, the cells were labeled with the fluorescent dye Celltracker CMFDA (Invitrogen) according to the producer instruction. A total of 105 cells in DMEM without serum were then plated on Bio-Coat Matrigel Invasion Chambers (BD Biosciences) for 24 h. Nonmigrated cells were removed, and migrated cells were counted by reading the fluorescence on the bottom side of the inserts with a Victor Multilabel Plate Reader (PerkinElmer). Cell migration in two dimensions was assessed by scraping a confluent monolayer of BMDMs with a pipette tip as described by Baruzzi et al. (17). Quantification of migrated cells was performed counting cells migrated into the wound in 10 different fields.

3

4

Sos1 REGULATES MACROPHAGE CYTOSKELETON DYNAMICS

cells overexpressing full-length Fgr, but not a myristoylationdefective N-terminal deletion or a kinase defective Fgr mutant (Fig. 1C). Notably, we previously found that N-terminal myristoylation and kinase activity are indispensable for induction of COS-7 cell cytoskeleton dynamics and cell migration by exogenously expressed Fgr (32). A possible link between Sos1 and the Src/Abl signaling pathway we implicated in macrophage migration and cytoskeleton dynamics (17, 18) was addressed by examining Sos1 tyrosine phosphorylation and Sos1/Abl association in wild type and hck2/2fgr2/2 BMDMs (Fig. 1D). Im-

munoprecipitation of Sos1 from wild type BMDM lysates showed that Sos1 and Abl form a complex and are phosphorylated in tyrosine. In contrast, deficiency of Hck and Fgr, which are indispensable for Abl tyrosine phosphorylation in BMDMs (17), resulted in a lack of Abl and Sos1 tyrosine phosphorylation and their association. That also Abl is implicated in tyrosine phosphorylation of Sos1 in BMDMs was suggested by the evidence that treatment with imatinib inhibited, as shown in COS-7 cells expressing a constitutive active mutant of Abl, Sos1 tyrosine phosphorylation (Fig. 1D). We conclude that an

Downloaded from http://www.jimmunol.org/ at Univ of Otago Med Lib on October 15, 2015

FIGURE 1. Sos1 is a downstream target of Abl and Src-family kinases. (A) COS-7 cells expressing a constitutively active form of Abl (AblDSH3) were lysed in RIPA buffer and lysates were immunoprecipitated with an anti-Sos1 Ab as described in Materials and Methods. Western blot analysis was performed with Abs against phospho-Tyrosine (pTyr), Sos1, and Abl. The ratio between the fluorescence signal of the phosphorylated Sos1 versus the total protein in anti-Sos1 immunoprecipitates detected in control cells was set to 100, and values detected in AblDSH3-expressing cells were calculated accordingly. Mean results 6 SD of three independent experiments are shown in the graph. **p , 0.01. Lysates of COS-7 cells expressing a constitutively active form of Abl (AblDSH3) were also immunoprecipitated with anti-pTyr Abs and blotted with anti-Sos1 (A, bottom left). One representative of two experiments performed with equal results is shown. (B) BaF3 cells constitutively transfected with a vector containing Bcr/Abl were treated with 10 mM imatinib mesylate (STI) or 20 mM GNF-2 for 2 h and then lysed in RIPA buffer. Lysates were immunoprecipitated with anti-Sos1 Ab, and Western blot analysis was performed using anti-pTyr, anti-Sos1, or anti-Abl Abs. One representative of two experiments performed with equal results is shown. (C) COS7 cells expressing GFP, full-length Fgr-GFP, a myristoylation-defective N-terminal deletion [Fgr(8-517)-GFP], or a kinase defective [Fgr(K279R)-GFP] Fgr mutant were lysed in RIPA buffer, and lysates were immunoprecipitated with Sos1 Ab or pTyr Ab. Western blot analysis was performed with anti-pTyr and anti-Sos1 Abs. (C, right panel) Mean results of the fluorescence signal of the phosphorylated Sos1 versus the total protein in anti-Sos1 immunoprecipitates, which was calculated as described earlier for (A). Mean results 6 SD of three independent experiments are shown in the graph. ***p , 0.001. The antiSos1 blot of anti-pTyr Ab immunoprecipitates shown in (C, bottom left) represents one representative of two experiments performed with equal results. (D) BMDMs from wild type (WT) or hck2/2fgr2/2 mice (HF) were lysed in RIPA buffer, and lysates were immunoprecipitated with anti-Sos1 Ab. Western blot analysis was performed using anti-pTyr, anti-Sos1, and anti-Abl Abs. One representative of two experiments performed with equal results is shown. (E) BMDMs were treated with 10 mM imatinib mesylate (STI) for 2 h and then lysed in RIPA buffer. Lysates were immunoprecipitated with anti-pTyr Ab, and Western blot analysis was performed using anti-Sos1 Ab. One representative of two experiments performed with equal results is shown. ns, not significant.

The Journal of Immunology integrin-dependent signaling pathway implicating Src kinases and Abl and regulating macrophage migration, matrix degradative capacity, and podosome formation (17, 18) may impinge on Sos1. Sos1 regulates cytoskeleton dynamics in macrophages

FIGURE 2. Sos1 localizes to podosome rosettes in BMDMs and is required for their assembly. (A) Control nonsilencing siRNA pool (Ctr) and siRNA reagents targeting Sos1 were transfected by electroporation as described in Materials and Methods. Cells were lysed after 48, 72, and 96 h from transfection to check silencing efficiency by Western blot analysis. Sos1 expression was controlled in control and Sos1-silenced BMDMs used to perform the whole set of experiments described in this report. In the course of several experiments, Sos1 expression in Sos1-silenced BMDMs was ,30% of that detected in control cells. (B) Control nonsilencing siRNA pool (Ctr) and siRNA reagents targeting Sos1 were transfected by electroporation as described in Materials and Methods and plated on plain tissue culture plastic in macrophage differentiation medium containing 15% FBS (see Materials and Methods). After 48 h, cell morphology was analyzed at original magnification 3400 (taken with an Olympus IX50 inverted microscope and acquired with an Olympus c-7070 wide-zoom Digital compact camera). (C) Control or Sos1silenced BMDMs were plated for 24 h on fibronectin-coated coverslips. F-actin (red) organization was examined by immunofluorescence microscopy. Original magnification 3200. Nuclei (blue) were stained with DAPI. Red arrows point to podosome rosettes in cells transfected with control siRNA. Podosome rosette number was counted in at least 10 fields. Mean results 6 SD obtained in four different experiments is shown in the bottom panel. *p , 0.05. (D) Control or Sos1-silenced BMDMs plated on fibronectin-coated coverslips for 24 h were stained with Abs against vinculin and Sos1, and rhodamine-phalloidin was used to stain F-actin (see Materials and Methods). Images were collected using an SP5 confocal microscope from Leica Microsystems with a 633 objective. Images are representative of those observed in experiments performed with three different cell pools. In each experiment performed, cells present in at least 10–15 fields of duplicate slides were examined.

To address whether Sos1 localizes to podosomes also in human macrophages, we exploited the evidence that treatment of hMDMs plated on fibrinogen with IL-4 induces podosome rosette formation (39). As shown in Fig. 3, in hMDMs plated on fibrinogen in the presence of IL-4, an anti-cortactin Ab stained large circular structures whose rim was enriched with dotlike podosomes. In some cells these structures looked like the outer rim of a plasma membrane sheet delimited by two rings of podosome-enriched regions. Punctate spots stained by anti-cortactin Abs were present in the whole space delimited by the two circular rims. In contrast with cortactin, which was present only in podosomes, anti-Sos1 Abs revealed a diffuse pattern of intracellular staining. However, Sos1 also colocalized with cortactin-rich podosomes in hMDMs. Rac1 is activated in an Sos1-dependent manner and localizes to podosome rosettes The data reported in Fig. 2 show that Sos1 silencing results in alteration of macrophage cytoskeleton dynamics similar to those

Downloaded from http://www.jimmunol.org/ at Univ of Otago Med Lib on October 15, 2015

To directly address the role of Sos1 in regulating macrophage cytoskeleton dynamics, we silenced Sos1 expression by siRNA. As shown in Fig. 2A, siRNA interference caused a marked and prolonged reduction of Sos1 expression in BMDMs. Silenced cells displayed the typical rounded, nonpolarized morphology we previously observed in BMDMs deficient in the expression of Srcfamily kinases or Abl, or treated with inhibitors of these kinases (17) (Fig. 2B). Notably, as found with Src- or Abl-deficient BMDMs, Sos1 silencing resulted in disassembly of actin-rich podosome rosettes (Fig. 2C). Consistent with its role in podosome rosette assembly, Sos1 localized to podosome rosettes, structures that were no more detectable in Sos1-silenced BMDMs (Fig. 2D).

5

6

Sos1 REGULATES MACROPHAGE CYTOSKELETON DYNAMICS

Sos1 regulates macrophage ECM degradation, three-dimensional migration, transendothelial migration, and infiltration into tumor spheroids To establish a relationship between Sos1-dependent podosome formation and macrophage invasive ability, we plated BMDMs on

gelatin-FITC–coated coverslips for 24 h and examined gelatin degradation (Fig. 7A, 7B). siRNA-induced silencing of Sos1 expression resulted in an almost total suppression of matrix degradation. In addition, BMDMs silenced for Sos1 expression displayed a reduced migratory ability through Matrigel matrices (Fig. 8A), a mixture of ECM proteins dictating a mesenchymal modality of migration (7, 8). Considering that transendothelial migration of lymphoid leukocytes from blood to the interstitium (13) requires podosome formation, we also addressed whether Sos1 silencing resulted in a reduced migration of macrophages through an endothelial cell monolayer. As shown in Fig. 8B, silencing of Sos1 resulted in a significant inhibition of BMDM transendothelial migration. Based on the evidence that Abl regulates also two-dimensional BMDM migration (17), we asked whether Sos1 is implicated in regulation of this migration mode (Fig. 8C). BMDMs silenced for Sos1 expression displayed a retard in filling a wound made by scratching a BMDM monolayer. Notably, albeit significant, the reduced two-dimensional migration ability of Sos1 defective BMDMs was less pronounced than that detected in three-dimensional migration assays (Fig. 8) or in two-dimensional migration assays performed with Abl-silenced BMDMs (17 and data not shown). These findings suggest that among GEFs regulating two-dimensional migration, Sos1 plays only a partial role. Considering that the migration assays performed could be affected by adhesion of macrophages to the substratum used, we also addressed whether Sos1-silenced BMDMs adhered as wild type cells to fibronectin. As suggested by the strongly spread phenotype of Sos1-silenced BMDMs (Fig. 2B), quantitative adhesion assays (see Materials and Methods) did not reveal any regulatory role of Sos1 on BMDM adhesion (data not shown). Finally, we asked whether Sos1 regulates macrophage infiltration into tumor spheroids, a well-established in vitro model of small primary or micrometastatic tumors. T47D human breast tumor spheroids were maintained in culture until they grew to ∼400–500 mm in diameter and then incubated with BMDMs prelabeled with CMFDA. After 24 h of incubation, tumor spheroids were analyzed by two-photon confocal microscopy (Fig. 9). As shown in Fig. 9A, flat images of spheroids incubated with control BMDMs showed the presence of several green macrophages inside the tumor. Quantification of cells migrated at different distances from the edge of the spheroid (normalized migration distance, see Materials and Methods) in sections taken every 5 mm showed that, on the whole, ∼20% of macrophages remained at the margin of the spheroid, whereas the rest moved toward its center. Notably, although BMDMs covered different

FIGURE 3. Sos1 localizes to podosomes in hMDMs. hMDMs plated for 24 h on fibrinogen-coated coverslips in the presence of IL-4 were stained with anti-cortactin and anti-Sos1 Abs. Images were collected using an SP5 confocal microscope from Leica Microsystems with a 633 objective. Colocalization is shown in the merged image. Arrows point to a higher magnification image. Images are representative of those observed in experiments performed with cells from two different donors. In each experiment performed, cells present in at least 10–15 fields of duplicate slides were examined.

Downloaded from http://www.jimmunol.org/ at Univ of Otago Med Lib on October 15, 2015

reported in cells where Abl, or Src-family kinases, were deficient or inhibited (14, 15, 18). The established relation among Abl, Src, and Rac in different cell context (17, 28, 36, 40–43) prompted experiments to address whether Sos1 silencing affects Rac activation. Sos1 being a dual GEF acting on both Ras and Rac, we initially examined Ras and Rac activation in BMDMs silenced for Sos1 expression. Confocal microscopy analysis with an Ab specifically recognizing the GTP-loaded form of Ras showed that the macrophage chemoattractant LPA induced a rapid (within 5 min) increase in the detected signal (Fig. 4). However, after 30 min from stimulation, the RasGTP fluorescence was not above the background signal detected in nonstimulated BMDMs. Notably, Sos1 silencing resulted in a total suppression of the signal ensuing from GTP-loaded Ras. As shown in Supplemental Fig. 1, RasGTP fluorescence was already detectable after 1 min from LPA stimulation, but it was transient and waned after 30–60 min. In addition, Sos1 silencing suppressed RasGTP fluorescence at all time points after LPA stimulation tested. Differently from Ras, the fluorescence signal detected with an Ab against Rac1GTP increased with time up to 30 min (Fig. 4, Supplemental Fig. 2). Consistent with an indispensable role of Sos1 in Rac1 activation, its silencing resulted in suppression of the Rac1GTP fluorescence. An essential role of Sos1 in Rac1 activation also emerged from pulldown assays of GTP-loaded Rac1 from BMDM lysates. As shown in Fig. 5, a 30-min treatment with LPA markedly increased the amount of Rac1GTP in BMDMs, and Sos1 silencing reduced this to undetectable levels in both resting and LPA-stimulated conditions. To address the role of the Sos1/Rac1 axis in podosome assembly, we examined distribution of Rac1 and Rac1GTP by confocal microscopy. As shown in Fig. 6, an anti-Rac1 Ab stained podosome rosettes, perinuclear structures, and the nucleus in BMDMs plated on fibronectin. These same areas, but not the nucleus, were also stained by an anti-Rac1GTP Ab. Notably, disassembly of podosome rosettes in BMDMs silenced for Sos1 expression resulted in a diffuse perinuclear and strong nuclear Rac1 signal and in a very faint and nonlocalized Rac1GTP staining. These findings suggest that Sos1 and Rac1 colocalize in podosome rosettes and in these structures Sos1 triggers GTP loading on Rac1.

The Journal of Immunology distances from the margin toward the center, a fraction of them migrated into the deepest part of the spheroid. In contrast, a much higher fraction of BMDMs silenced for Sos1 expression remained at the margin and the outer layers of the spheroid (Fig. 9B). Considering the role played by Abl in phosphorylation and activation of Sos1 (28) and in regulating macrophage cytoskeleton dynamics and migration in two and three dimensions (17, 18), we

also addressed whether pharmacologic inhibition or silencing of Abl also affected macrophage tumor spheroid infiltration. Tumor spheroid infiltration assays run in the presence of the Abl inhibitor imatinib or with cells silenced for Abl expression showed that both maneuvers markedly inhibited the BMDM capability to infiltrate the spheroid (Fig. 9C, 9D). A three-dimensional view of spheroids incubated with control BMDMs (Supplemental Video 1),

Downloaded from http://www.jimmunol.org/ at Univ of Otago Med Lib on October 15, 2015

FIGURE 4. Rac1 and Ras are activated in an Sos1-dependent manner in BMDMs. Control nonsilencing siRNA pool and siRNA reagents targeting Sos1 were transfected by electroporation as described in Materials and Methods. After 72 h, cells were plated on glass coverslips, left to adhere, and stimulated with 3 mM LPA for the indicated times. BMDMs were then fixed and stained with Abs against Rac1GTP or RasGTP (green). F-actin (red) was labeled with rhodamine-phalloidin and nuclei (blue) with DAPI. Images were collected using an SP5 confocal microscope from Leica Microsystems with a 633 objective. Scale bars, 10 mm. Images are representative of those observed in experiments performed with three different cell pools. In each experiment performed, cells present in at least 10–15 fields of duplicate slides were examined.

7

8

Sos1 REGULATES MACROPHAGE CYTOSKELETON DYNAMICS

Sos1-silenced BMDMs (Supplemental Video 2), Abl-silenced BMDMs (Supplemental Video 3), and imatinib-treated BMDMs (Supplemental Video 4) showed that inhibition of the Abl/Sos1 pathway resulted in a clear inhibition of the macrophageinfiltrating capacity. Two-photon confocal analysis of a limited number of spheroids showed that differences in normalized migration distances of control macrophages or Sos1-silenced, Ablsilenced, and imatinib-treated macrophages was highly significant (see legend for Fig. 9). Clear differences in the capability of both Sos1-silenced and Abl-silenced BMDMs to infiltrate into tumor spheroids was also detected examining by confocal microscopy optical slices taken at the middle of the Z-stack of several sphe-

roids (Supplemental Fig. 3) and by quantifying the number of BMDMs infiltrating the spheroid (Supplemental Fig. 4). Considering that macrophage infiltration into tumor spheroids done with the breast cancer cell line SUM159PT was shown to depend on both a mesenchymal and an amoeboid migration mode (35), we asked whether this was the case also in our experimental system. Indeed, the use of mesenchymal migration mode protease mixture inhibitors or the amoeboid migration mode RHO kinase inhibitor Y27632 (see Ref. 35) were both able to partially affect BMDM migration into T47D spheroids (Supplemental Fig. 4), suggesting that heterogeneity in the porosity of the ECM dictates one or the other of the two migration modalities (see Ref. 35 for

FIGURE 6. Rac1 is activated and localizes to podosome rosettes in murine macrophages. Control nonsilencing siRNA pool and siRNA reagents targeting Sos1 were transfected by electroporation as described in Materials and Methods. After 48 h, cells were then plated on fibronectin-coated coverslips, fixed, and stained with specific Abs against Rac1GTP (green) and Rac1 (azure). Rhodamine-phalloidin was used to label F-actin (red). Images were collected using an SP5 confocal microscope from Leica Microsystems with a 633 objective. Images are representative of those observed in experiments performed with two different cell pools. In each experiment performed, cells present in at least 10–15 fields of duplicate slides were examined.

Downloaded from http://www.jimmunol.org/ at Univ of Otago Med Lib on October 15, 2015

FIGURE 5. Sos1 silencing results in reduced Rac1 activation. (A) Control nonsilencing siRNA pool and siRNA reagents targeting Sos1 were transfected in BMDMs by electroporation as described in Materials and Methods. After 72 h, cells were treated with LPA for 30 min, and the GTP-loaded amount of Rac1 assayed by the Rac1 Activation Assay Biochem Kit from Cytoskeleton, according to the manufacturer’s instructions. Western blot analysis was performed using an anti-Rac1 Ab. Mean results 6 SD of three independent experiments are shown in (B). In each distinct experiment, the ratio of the intensity of the anti-Rac1 signal after pulldown of GTP-loaded Rac1 with PAK-PBD affinity beads with that detected in the total cell lysate of control nonsilencing siRNA pool transfected, nonstimulated (ctr) BMDMs was set to 100. The other ratios were then expressed according to this value. ***p , 0.001 by ANOVA with Bonferroni’s posttest.

The Journal of Immunology

9

a discussion of this topic). To address whether deficiency of Sos1 could also affect the amoeboid migration mode was beyond the scope of this article. However, considering that silencing of Sos1 affects protease secretion and three-dimensional migration (Figs. 7, 8), together with the evidence that Sos1 silencing results in a noncomplete inhibition of BMDM infiltration into T47D spheroids (Supplemental Fig. 4C), suggests that the amoeboid migration mode is unlikely to be regulated by Sos1.

Discussion Recruitment of cells of the monocyte/macrophage lineage into the interstitium and their migration therein represents a key step to fulfill their role in host defenses, wound repair, and tissue homeostasis. Macrophage crawling along the endothelium or basement membranes is viewed as a classical two-dimensional migration mode requiring a coordinated cycle of F-actin–based protrusive forces at the cell leading edge, actin-myosin contractility at the cell body, and retraction of the cell rear; in contrast, in tight interstitial matrices, macrophages may acquire an elongated morphology and form long protrusions, moving in a mode that is referred to as mesenchymal migration (reviewed in Ref. 7). Both at the ventral side of the plasma membrane adhering to a cell or an ECM protein layer in two dimensions and along or at the tip of the plasma membrane elongations characterizing cells migrating in three dimensions, macrophages organize dotlike protrusions characterized by a core of F-actin surrounded by a doughnut-like

ring of vinculin (44). Notably, these macrophage structures were recognized as analogs of ventral membrane protrusions identified in fibroblasts transformed with Rous sarcoma virus (45, 46) and initially named podosomes. Subsequent progress in identifying these protrusive structures, enriched in the cytoplasmic tyrosine kinase Src and F-actin, as the site of degradation of the ECM led to naming them invadopodia to emphasize their role in dictating the invasive capacity of neoplastic cells (reviewed in Ref. 9). Although the terms podosome and invadopodium are currently used to refer to these structures in nontransformed or neoplastic cells, respectively, their strict structural and functional similarity suggests they serve the unique role to favor the capacity to degrade the ECM and promote cell migration into the interstitium. In addition, although this has not been directly addressed in the context of mononuclear phagocytes transendothelial migration, it has been proposed that podosomes are required for the migration of lymphocytes through the endothelium via a transcellular route wherein cell recruitment into the subendothelial space occurs through pores opened in the endothelial cells (13). The main feature of podosomes and invadopodia is to be constituted by integrin-containing plasma membrane evagination filled with an actin-reach core surrounded by a ring of actin-associated and signaling proteins (see Refs. 6, 7, 9, 10, 47 for reviews). These last include cytoplasmic tyrosine kinases that are implicated in regulation of cytoskeleton dynamics triggered by integrin, chemoattractant, and growth factor receptors, that is, Src-family

Downloaded from http://www.jimmunol.org/ at Univ of Otago Med Lib on October 15, 2015

FIGURE 7. Sos1 is indispensable for gelatin degradation. Control nonsilencing siRNA pool and siRNA reagents targeting Sos1 were transfected by electroporation as described in Materials and Methods. (A) Cells were plated on gelatin-FITC–coated coverslips for 24 h and then fixed and stained to identify actin and nuclei. Images were obtained by immunofluorescence microscopy. Original magnification 3200. The mean area of gelatin degradation (black areas) was quantified as described in Materials and Methods. Data are shown as mean 6 SD of area degraded around cells visible in 15–20 fields in at least three independent experiments. *p , 0.05 by the Student t test. A zoom on degradative podosome rosettes is shown in (B); images were collected using an SP5 confocal microscope from Leica Microsystems with a 633 objective. Scale bar, 10 mm.

10

Sos1 REGULATES MACROPHAGE CYTOSKELETON DYNAMICS

kinases, Abl kinases, Fak, and Pyk2. Genetic inactivation or inhibitory studies showed that these kinases play a nonredundant, indispensable role in regulation of macrophage migration (reviewed in Ref. 48). In addition, in the context of podosome formation, Src-family kinases and Abl have been reported to be implicated in podosome assembly and functions in macrophages (14, 15, 18). Together with the evidence that the myeloid-specific Src kinases Hck and Fgr bind to both b1 and b2 integrins and Abl (17), these findings point to a key role of an integrin/Src kinase/ Abl pathway in macrophage podosome assembly. Importantly, recent studies demonstrated that Src and Abl/Arg cross talk in invadopodia formation in response to chemokine and growth factor receptors in carcinoma cells (19, 20), thus highlighting the existence of a common signaling module regulating macrophage podosome and tumor cell invadopodia assembly. The downstream components of the Src/Abl kinases regulating formation of macrophage podosomes are not known. Different members of the Rho GTPase family have been implicated in both podosome and invadopodia formation (see Refs. 9, 10, 25, 26 and references contained therein). However, Rac subfamily members seem to play an indispensable role in podosome formation and

organization in myeloid cells (10, 21, 25). Activation of Rac via its loading with GTP is triggered by different GEFs, whose mechanisms of activation are known only in part. Interestingly, some progress has been made in the identification of GEFs acting on Rac and implicated in podosome assembly in myeloid leukocytes. For example, Vav3 plays an indispensable role in osteoclast actin cytoskeleton dynamics (49), and the FERM-domain–containing GEF, FARP2, regulates osteoclast podosome rearrangements and resorbing activity dictating localized activation of Rac1 into podosomes (50). In addition, the Rho GTPase GEF p21-activated kinase 1–interacting exchange factor (aPIX) was reported to localize to podosomes and to regulate their stabilization in human macrophages (51). Considering the role of Src kinases and Abl in regulating both macrophage podosomes (18) and cancer cell invadopodia assembly (19, 20), we reasoned that Sos1, whose tyrosine phosphorylation by Abl dictates the acquisition of its Rac GEF activity (28), was a potentially critical candidate in regulating Rac-dependent podosome assembly downstream this Src/Abl signaling pathway. Notably, Eps8, an activator of the Rac-GEF activity of Sos1, was shown to interact with the phosphoprotein paladin, a podosome-associated protein, in mesenchymal cells

Downloaded from http://www.jimmunol.org/ at Univ of Otago Med Lib on October 15, 2015

FIGURE 8. Sos1 regulates three-dimensional and two-dimensional migratory ability. Control nonsilencing siRNA pool and siRNA reagents targeting Sos1 were transfected by electroporation as described in Materials and Methods. (A) Cells were plated on Bio-coat Matrigel Invasion Chambers (BD Biosciences) for 24 h. Migrated cells were quantified as described in Materials and Methods. Mean results 6 SD of three independent experiments are reported. (B) BMDMs were plated in the upper well of a transwell insert previously coated with an endothelial cell monolayer as described in Materials and Methods. Cells migrated were quantified after 22 h. Mean results 6 SD of three independent experiments are reported. (C) BMDMs monolayers were scratched with a tip of a pipette. Monolayers were photographed immediately after the wound (t0) or after 5 h from wounding (5 h) (original magnification 3200). Quantification of migrated cells was performed by counting cells migrated into the wound in 10 fields in 3 different experiments. *p , 0.05, **p , 0.01 by the Student t test.

The Journal of Immunology

11

Downloaded from http://www.jimmunol.org/ at Univ of Otago Med Lib on October 15, 2015

FIGURE 9. Sos1 and Abl regulate macrophage infiltration into tumor spheroids. Images show the maximum intensity projection of labeled BMDMs (green) and the reflection signal of the surface of the spheroid (red). (A and B) Control nonsilencing siRNA pool (A) and siRNA reagents targeting Sos1 (B) were transfected by electroporation as described in Materials and Methods. After 72 h, cells were labeled with cell tracker CMFDA and coincubated with tumor spheroids for 24 h (see Materials and Methods). (C) BMDMs were pretreated for 1 h with 10 mM imatinib mesylate (STI), and the inhibitor was maintained in the medium during the coincubation with tumor spheroids. (D) BMDMs were transfected with a siRNA targeting Abl and incubated with tumor spheroid as described earlier and in Materials and Methods. After 24 h of coincubation with CFMDA-labeled BMDMs, single spheroids were acquired with Leica TCS.SP5 confocal microscope equipped with a Chameleon Ultra multiphoton laser (original magnification 3200). Right panels report the fraction of cells (y-axis) migrated at a normalized distance from the spheroid border (x-axis). The normalized distance is 0 when the cell adheres to the spheroid surface and 1 when it is in the spheroid’s center (see Materials and Methods for details). Representative image obtained with one spheroid is reported. Analysis of macrophage migration in four to six different spheroids with the Kolmogorov– Smirnov test showed a highly significant difference in distribution of normalized distances between control macrophages and Sos1-silenced, STI-treated, or Abl-silenced macrophages (p , 0.001). In contrast, no difference was found between STI-treated and either Sos1- or Abl-silenced macrophages (p , 0.21 and p , 0.43, respectively).

(29). In addition, Sos1 binds to Tks5, a podosome and invadopodia component (9), in epithelial A549 cells and Src-transformed NIH 3T3 cells (30).

Placement of Sos1 downstream the Src kinase/Abl signaling pathway regulating cytoskeleton dynamics was substantiated by the finding that Src kinases and Abl regulate its tyrosine phosphorylation

12

Sos1 REGULATES MACROPHAGE CYTOSKELETON DYNAMICS Disassembly of podosome rosettes in murine macrophages via silencing of Sos1 results in a reduction of their capability to exert podosome-dependent functions. Degradation of ECM, revealed by gelatin-degradation assays, was virtually suppressed in the absence of Sos1 (Fig. 7). Notably, the same results were obtained by reducing the expression or by inhibiting pharmacologically the Sos1 upstream component Abl (18). In addition, macrophage migration through Matrigel, a matrix through which macrophages migrate using a mesenchymal migration mode, or through an endothelial cell monolayer, which in lymphocytes has been reported to occur at least in part in a podosome-dependent manner (13, 57), was inhibited by ∼50% after silencing of Sos1, that is, to an extent similar to that detected in Abl-silenced cells (18). A reduced macrophage invasive capacity after silencing of Sos1 was found examining infiltration of macrophages into tumor spheroids (Fig. 9). In fact, the capability of Sos1-silenced macrophages to migrate within breast tumor spheroids after 24 h of cocultivation was markedly reduced, with the majority of cells remaining at the margin and the outer layers of the spheroid. Inhibition of macrophage infiltration into tumor spheroids was also obtained by silencing or by pharmacologic inhibition of Abl, an Sos1 upstream component of this signaling pathway (Fig. 9). In studies in which we demonstrated a role of Abl in macrophage podosome assembly and function, we highlighted the fact that, based on the evidence that the Abl/Arg family of cytoplasmic tyrosine kinases also regulate invadopodia formation and function (19, 20), pharmacologic targeting of Abl may result in inhibition of macrophage infiltration into tumors and tumor cell invasive capacity (18). Sos1 can play the same role in macrophages and tumor cells. Indeed, in cancer cells, Sos1 was reported to localize to invadopodia (29, 30) and regulate their migratory and metastatic potential (30, 58, 59). After the demonstration that macrophages migrated into a tumor can stimulate angiogenesis and cancer cell growth, survival, and motility, cancer-related inflammation has emerged as one of the hallmarks of cancer (5, 60, 61). Together with our previous studies (17, 18), this report and the evidence that the Src kinase Hck and its substrate Wiskott–Aldrich syndrome protein are required for infiltration of macrophages into tumor spheroids (16, 62) point to the conclusion that targeting the Src/Abl/Sos1 pathway can indeed represent a double-edged sword to control cancer-related inflammation and cancer cell invasive capacity.

Disclosures The authors have no financial conflicts of interest.

References 1. van Furth, R., Z. A. Cohn, J. G. Hirsch, J. H. Humphrey, W. G. Spector, and H. L. Langevoort. 1972. The mononuclear phagocyte system: a new classification of macrophages, monocytes, and their precursor cells. Bull. World Health Organ. 46: 845–852. 2. Gordon, S., and P. R. Taylor. 2005. Monocyte and macrophage heterogeneity. Nat. Rev. Immunol. 5: 953–964. 3. Moore, K. J., F. J. Sheedy, and E. A. Fisher. 2013. Macrophages in atherosclerosis: a dynamic balance. Nat. Rev. Immunol. 13: 709–721. 4. McNelis, J. C., and J. M. Olefsky. 2014. Macrophages, immunity, and metabolic disease. Immunity 41: 36–48. 5. Noy, R., and J. W. Pollard. 2014. Tumor-associated macrophages: from mechanisms to therapy. [Published erratum appears in 2014 Immunity. 41: 866.] Immunity 41: 49–61. 6. Maridonneau-Parini, I. 2014. Control of macrophage 3D migration: a therapeutic challenge to limit tissue infiltration. Immunol. Rev. 262: 216–231. 7. Wiesner, C., V. Le-Cabec, K. El Azzouzi, I. Maridonneau-Parini, and S. Linder. 2014. Podosomes in space: macrophage migration and matrix degradation in 2D and 3D settings. Cell Adhes. Migr. 8: 179–191. 8. Van Goethem, E., R. Poincloux, F. Gauffre, I. Maridonneau-Parini, and V. Le Cabec. 2010. Matrix architecture dictates three-dimensional migration modes of human macrophages: differential involvement of proteases and podosome-like structures. J. Immunol. 184: 1049–1061.

Downloaded from http://www.jimmunol.org/ at Univ of Otago Med Lib on October 15, 2015

(Fig. 1). First, and in agreement with previous reports (28), we found that a constitutively active mutant of Abl enhances Sos1 tyrosine phosphorylation in COS-7 cells in a manner that is sensitive to inhibition by the Abl selective inhibitor imatinib. Second, tyrosine phosphorylation of Sos1 in hematopoietic cells expressing the fusion protein BCR/Abl is inhibited by two Abl kinase inhibitors. Third, overexpression in COS-7 cells of full-length Fgr, but not a fatty acylation defective or a kinase inactive Fgr mutant (32), results in enhanced tyrosine phosphorylation of Sos1. Finally, in hck2/2fgr2/2 BMDMs, both Sos1 and Abl tyrosine phosphorylation are undetectable and they do not form a stable complex, and imatinib inhibits Sos1 tyrosine phosphorylation in wild type BMDMs. Collectively, these findings suggest the bona-fide conclusion that the integrin-associated Hck/Fgr/Abl complex (17) impinges on Sos1. Intriguingly, the phenotype of macrophages deficient of Src family kinases or Abl is similar (17, 18), suggesting they play an indispensable, nonredundant role in regulation of cytoskeleton dynamics and podosome assembly. The evidence that Src family kinase deficiency results in reduced tyrosine phosphorylation of both Abl and Sos1 (17) (Fig. 1), coupled with the demonstration that Sos1 is an Abl substrate (28), would suggest a Src/Abl/Sos1 hierarchy in this signaling module. However, we cannot exclude that also Src kinases impinge directly on Sos1 and, via their combined action on Abl, either induce Abl/Sos1 interaction or an extent of Sos1 tyrosine phosphorylation critical for the full disclosure of its Rac GEF activity. Notably, also cortactin, that is, another podosome component, is a substrate of both Src (52) and Abl (53). It is worth noting that macrophage cytoskeleton dynamics have been shown to be regulated by different cytoplasmic tyrosine kinases including Src, Abl, Syk, FAK, and Pyk2 (reviewed in Refs. 48, 54). Genetic inactivation of these kinases results in a similar phenotype characterized by reduced polarization, elongation, and migratory ability, thus suggesting that they play a coordinate more than a redundant role in regulation of cytoskeleton dynamics (48). Notably, silencing of Sos1 also leads to a loss of an elongated morphology and the acquirement of a round, nonpolarized one (Fig. 2). In murine macrophages plated on fibronectin and in IL-4–treated hMDMs plated on fibrinogen, Sos1 localizes to podosome rosettes and more discrete actin puncta also enriched with the podosomal proteins vinculin and cortactin, respectively (Figs. 2, 3). Silencing of Sos1 resulted in disassembly of podosome rosettes and suppression of the loading of GTP on Rac1, whereas Rac1-GTP redistributed to podosome rosettes in control macrophages (Figs. 4–6). Considering that Rac1 localizes to the nuclei in both control and Sos1-silenced cells, but the GTP-loaded form of Rac1 was undetectable in the nuclear compartment also in Sos1expressing cells, we conclude that Sos plays a selective role in Rac1 activation in podosomes and a cell compartment including the perinuclear area and the nuclear membrane, but not the nucleus where a substantial portion of Rac1 and Sos1 is localized (see Fig. 6). The strict relation we found among Sos1 expression, Rac1 activation, and podosome rosette formation in macrophages does not exclude a more complex scenario implicating additional GEFs and Rho GTPases in regulation of podosome assembly. In fact, both CdC42 and Rac, and several exchange factors regulating their activation, have been implicated in podosome and invadopodia formation (see Refs. 10, 25, 26). In particular, Vav proteins, GEFs for CdC42 and Rac, have been reported to regulate both myeloid cell podosomes (49, 55) and invadopodia (56). Thus, it is tempting to speculate that Rho GTPases, and some of their GEFs, may actually act in an integrated more than redundant fashion, although only studies addressing this issue in the same cell type might elucidate this issue.

The Journal of Immunology

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47. 48. 49.

50.

51.

52.

53.

54. 55.

56. 57. 58.

59. 60. 61. 62.

macrophage migration promotes matrix metalloproteinase-independent invasion by tumor cells. J. Immunol. 187: 3806–3814. Mettouchi, A., S. Klein, W. Guo, M. Lopez-Lago, E. Lemichez, J. K. Westwick, and F. G. Giancotti. 2001. Integrin-specific activation of Rac controls progression through the G(1) phase of the cell cycle. Mol. Cell 8: 115–127. Choi, Y., M. A. Seeliger, S. B. Panjarian, H. Kim, X. Deng, T. Sim, B. Couch, A. J. Koleske, T. E. Smithgall, and N. S. Gray. 2009. N-myristoylated c-Abl tyrosine kinase localizes to the endoplasmic reticulum upon binding to an allosteric inhibitor. J. Biol. Chem. 284: 29005–29014. Zhang, J., F. J. Adria´n, W. Jahnke, S. W. Cowan-Jacob, A. G. Li, R. E. Iacob, T. Sim, J. Powers, C. Dierks, F. Sun, et al. 2010. Targeting Bcr-Abl by combining allosteric with ATP-binding-site inhibitors. Nature 463: 501–506. Cougoule, C., E. Van Goethem, V. Le Cabec, F. Lafouresse, L. Dupre´, V. Mehraj, J. L. Me`ge, C. Lastrucci, and I. Maridonneau-Parini. 2012. Blood leukocytes and macrophages of various phenotypes have distinct abilities to form podosomes and to migrate in 3D environments. Eur. J. Cell Biol. 91: 938–949. Lewis, J. M., R. Baskaran, S. Taagepera, M. A. Schwartz, and J. Y. Wang. 1996. Integrin regulation of c-Abl tyrosine kinase activity and cytoplasmic-nuclear transport. Proc. Natl. Acad. Sci. USA 93: 15174–15179. Sirvent, A., A. Boureux, V. Simon, C. Leroy, and S. Roche. 2007. The tyrosine kinase Abl is required for Src-transforming activity in mouse fibroblasts and human breast cancer cells. Oncogene 26: 7313–7323. Thomas, E. K., J. A. Cancelas, Y. Zheng, and D. A. Williams. 2008. Rac GTPases as key regulators of p210-BCR-ABL-dependent leukemogenesis. Leukemia 22: 898–904. Zandy, N. L., M. Playford, and A. M. Pendergast. 2007. Abl tyrosine kinases regulate cell-cell adhesion through Rho GTPases. Proc. Natl. Acad. Sci. USA 104: 17686–17691. Marchisio, P. C., D. Cirillo, A. Teti, A. Zambonin-Zallone, and G. Tarone. 1987. Rous sarcoma virus-transformed fibroblasts and cells of monocytic origin display a peculiar dot-like organization of cytoskeletal proteins involved in microfilament-membrane interactions. Exp. Cell Res. 169: 202–214. David-Pfeuty, T., and S. J. Singer. 1980. Altered distributions of the cytoskeletal proteins vinculin and alpha-actinin in cultured fibroblasts transformed by Rous sarcoma virus. Proc. Natl. Acad. Sci. USA 77: 6687–6691. Tarone, G., D. Cirillo, F. G. Giancotti, P. M. Comoglio, and P. C. Marchisio. 1985. Rous sarcoma virus-transformed fibroblasts adhere primarily at discrete protrusions of the ventral membrane called podosomes. Exp. Cell Res. 159: 141–157. Linder, S., and P. Kopp. 2005. Podosomes at a glance. J. Cell Sci. 118: 2079–2082. Baruzzi, A., E. Caveggion, and G. Berton. 2008. Regulation of phagocyte migration and recruitment by Src-family kinases. Cell. Mol. Life Sci. 65: 2175–2190. Faccio, R., S. L. Teitelbaum, K. Fujikawa, J. Chappel, A. Zallone, V. L. Tybulewicz, F. P. Ross, and W. Swat. 2005. Vav3 regulates osteoclast function and bone mass. Nat. Med. 11: 284–290. Takegahara, N., S. Kang, S. Nojima, H. Takamatsu, T. Okuno, H. Kikutani, T. Toyofuku, and A. Kumanogoh. 2010. Integral roles of a guanine nucleotide exchange factor, FARP2, in osteoclast podosome rearrangements. FASEB J. 24: 4782–4792. Gringel, A., D. Walz, G. Rosenberger, A. Minden, K. Kutsche, P. Kopp, and S. Linder. 2006. PAK4 and alphaPIX determine podosome size and number in macrophages through localized actin regulation. J. Cell. Physiol. 209: 568–579. Wu, H., and J. T. Parsons. 1993. Cortactin, an 80/85-kilodalton pp60src substrate, is a filamentous actin-binding protein enriched in the cell cortex. J. Cell Biol. 120: 1417–1426. Boyle, S. N., G. A. Michaud, B. Schweitzer, P. F. Predki, and A. J. Koleske. 2007. A critical role for cortactin phosphorylation by Abl-family kinases in PDGF-induced dorsal-wave formation. Curr. Biol. 17: 445–451. Berton, G., A. Mo´csai, and C. A. Lowell. 2005. Src and Syk kinases: key regulators of phagocytic cell activation. Trends Immunol. 26: 208–214. Spurrell, D. R., N. A. Luckashenak, D. C. Minney, A. Chaplin, J. M. Penninger, R. S. Liwski, J. L. Clements, and K. A. West. 2009. Vav1 regulates the migration and adhesion of dendritic cells. J. Immunol. 183: 310–318. Razidlo, G. L., B. Schroeder, J. Chen, D. D. Billadeau, and M. A. McNiven. 2014. Vav1 as a central regulator of invadopodia assembly. Curr. Biol. 24: 86–93. Carman, C. V., and T. A. Springer. 2008. Trans-cellular migration: cell-cell contacts get intimate. Curr. Opin. Cell Biol. 20: 533–540. Timofeeva, O. A., X. Zhang, H. W. Ressom, R. S. Varghese, B. V. Kallakury, K. Wang, Y. Ji, A. Cheema, M. Jung, M. L. Brown, et al. 2009. Enhanced expression of SOS1 is detected in prostate cancer epithelial cells from AfricanAmerican men. Int. J. Oncol. 35: 751–760. Chen, H., X. Wu, Z. K. Pan, and S. Huang. 2010. Integrity of SOS1/EPS8/ABI1 tri-complex determines ovarian cancer metastasis. Cancer Res. 70: 9979–9990. Mantovani, A., P. Allavena, A. Sica, and F. Balkwill. 2008. Cancer-related inflammation. Nature 454: 436–444. Hanahan, D., and R. A. Weinberg. 2011. Hallmarks of cancer: the next generation. Cell 144: 646–674. Carre´no, S., E. Caron, C. Cougoule, L. J. Emorine, and I. Maridonneau-Parini. 2002. p59Hck isoform induces F-actin reorganization to form protrusions of the plasma membrane in a Cdc42- and Rac-dependent manner. J. Biol. Chem. 277: 21007–21016.

Downloaded from http://www.jimmunol.org/ at Univ of Otago Med Lib on October 15, 2015

9. Murphy, D. A., and S. A. Courtneidge. 2011. The ‘ins’ and ‘outs’ of podosomes and invadopodia: characteristics, formation and function. Nat. Rev. Mol. Cell Biol. 12: 413–426. 10. Dovas, A., and D. Cox. 2011. Signaling networks regulating leukocyte podosome dynamics and function. Cell. Signal. 23: 1225–1234. 11. Linder, S., C. Wiesner, and M. Himmel. 2011. Degrading devices: invadosomes in proteolytic cell invasion. Annu. Rev. Cell Dev. Biol. 27: 185–211. 12. Zambonin-Zallone, A., A. Teti, A. Carano, and P. C. Marchisio. 1988. The distribution of podosomes in osteoclasts cultured on bone laminae: effect of retinol. J. Bone Miner. Res. 3: 517–523. 13. Carman, C. V., P. T. Sage, T. E. Sciuto, M. A. de la Fuente, R. S. Geha, H. D. Ochs, H. F. Dvorak, A. M. Dvorak, and T. A. Springer. 2007. Transcellular diapedesis is initiated by invasive podosomes. Immunity 26: 784–797. 14. Cougoule, C., V. Le Cabec, R. Poincloux, T. Al Saati, J. L. Me`ge, G. Tabouret, C. A. Lowell, N. Laviolette-Malirat, and I. Maridonneau-Parini. 2010. Threedimensional migration of macrophages requires Hck for podosome organization and extracellular matrix proteolysis. Blood 115: 1444–1452. 15. Poincloux, R., C. Vincent, A. Labrousse, J. Castandet, M. Rigo, C. Cougoule, C. Bordier, V. Le Cabec, and I. Maridonneau-Parini. 2006. Re-arrangements of podosome structures are observed when Hck is activated in myeloid cells. Eur. J. Cell Biol. 85: 327–332. 16. Park, H., A. Dovas, S. Hanna, C. Lastrucci, C. Cougoule, R. Guiet, I. Maridonneau-Parini, and D. Cox. 2014. Tyrosine phosphorylation of WiskottAldrich syndrome protein (WASP) by Hck regulates macrophage function. J. Biol. Chem. 289: 7897–7906. 17. Baruzzi, A., I. Iacobucci, S. Soverini, C. A. Lowell, G. Martinelli, and G. Berton. 2010. c-Abl and Src-family kinases cross-talk in regulation of myeloid cell migration. FEBS Lett. 584: 15–21. 18. Baruzzi, A., and G. Berton. 2012. The tyrosine kinase Abl is a component of macrophage podosomes and is required for podosome formation and function. Eur. J. Immunol. 42: 2720–2726. 19. Smith-Pearson, P. S., E. K. Greuber, G. Yogalingam, and A. M. Pendergast. 2010. Abl kinases are required for invadopodia formation and chemokineinduced invasion. J. Biol. Chem. 285: 40201–40211. 20. Mader, C. C., M. Oser, M. A. Magalhaes, J. J. Bravo-Cordero, J. Condeelis, A. J. Koleske, and H. Gil-Henn. 2011. An EGFR-Src-Arg-cortactin pathway mediates functional maturation of invadopodia and breast cancer cell invasion. Cancer Res. 71: 1730–1741. 21. Wheeler, A. P., C. M. Wells, S. D. Smith, F. M. Vega, R. B. Henderson, V. L. Tybulewicz, and A. J. Ridley. 2006. Rac1 and Rac2 regulate macrophage morphology but are not essential for migration. J. Cell Sci. 119: 2749–2757. 22. Razzouk, S., M. Lieberherr, and G. Cournot. 1999. Rac-GTPase, osteoclast cytoskeleton and bone resorption. Eur. J. Cell Biol. 78: 249–255. 23. Burns, S., A. J. Thrasher, M. P. Blundell, L. Machesky, and G. E. Jones. 2001. Configuration of human dendritic cell cytoskeleton by Rho GTPases, the WAS protein, and differentiation. Blood 98: 1142–1149. 24. Furmaniak-Kazmierczak, E., S. W. Crawley, R. L. Carter, D. H. Maurice, and G. P. Coˆte´. 2007. Formation of extracellular matrix-digesting invadopodia by primary aortic smooth muscle cells. Circ. Res. 100: 1328–1336. 25. van Helden, S. F., and P. L. Hordijk. 2011. Podosome regulation by Rho GTPases in myeloid cells. Eur. J. Cell Biol. 90: 189–197. 26. Spuul, P., P. Ciufici, V. Veillat, A. Leclercq, T. Daubon, IJ. Kramer, and E. Ge´not. 2014. Importance of RhoGTPases in formation, characteristics, and functions of invadosomes. Small GTPases 5: e28195. 27. Pierre, S., A. S. Bats, and X. Coumoul. 2011. Understanding SOS (Son of Sevenless). Biochem. Pharmacol. 82: 1049–1056. 28. Sini, P., A. Cannas, A. J. Koleske, P. P. Di Fiore, and G. Scita. 2004. Abldependent tyrosine phosphorylation of Sos-1 mediates growth-factor-induced Rac activation. Nat. Cell Biol. 6: 268–274. 29. Goicoechea, S., D. Arneman, A. Disanza, R. Garcia-Mata, G. Scita, and C. A. Otey. 2006. Palladin binds to Eps8 and enhances the formation of dorsal ruffles and podosomes in vascular smooth muscle cells. J. Cell Sci. 119: 3316–3324. 30. Rufer, A. C., J. Rumpf, M. von Holleben, S. Beer, K. Rittinger, and Y. Groemping. 2009. Isoform-selective interaction of the adaptor protein Tks5/ FISH with Sos1 and dynamins. J. Mol. Biol. 390: 939–950. 31. Suen, P. W., D. Ilic, E. Caveggion, G. Berton, C. H. Damsky, and C. A. Lowell. 1999. Impaired integrin-mediated signal transduction, altered cytoskeletal structure and reduced motility in Hck/Fgr deficient macrophages. J. Cell Sci. 112: 4067–4078. 32. Continolo, S., A. Baruzzi, M. Majeed, E. Caveggion, L. Fumagalli, C. A. Lowell, and G. Berton. 2005. The proto-oncogene Fgr regulates cell migration and this requires its plasma membrane localization. Exp. Cell Res. 302: 253–269. 33. Vines, C. M., J. W. Potter, Y. Xu, R. L. Geahlen, P. S. Costello, V. L. Tybulewicz, C. A. Lowell, P. W. Chang, H. D. Gresham, and C. L. Willman. 2001. Inhibition of beta 2 integrin receptor and Syk kinase signaling in monocytes by the Src family kinase Fgr. Immunity 15: 507–519. 34. Carlsson, J., and J. M. Yuhas. 1984. Liquid-overlay culture of cellular spheroids. Recent Results Cancer Res. 95: 1–23. 35. Guiet, R., E. Van Goethem, C. Cougoule, S. Balor, A. Valette, T. Al Saati, C. A. Lowell, V. Le Cabec, and I. Maridonneau-Parini. 2011. The process of

13