Podosome Nucleation Is Facilitated by Multivalent Interactions between Syk and ITAM-containing Membrane Complexes

Key Points

• The kinase and tandem SH2 domains of Syk are essential for podosomes in macrophages.

• ITAM-associated membrane complexes are abridged by Syk to nucleate podosomes.

• Gain-of-function variants of Syk that cause autoimmunity form irregular podosomes.

Visual Abstract

Abstract

Immune cells survey their microenvironment by forming dynamic cellular protrusions that enable chemotaxis, contacts with other cells, and phagocytosis. Podosomes are a unique type of protrusion structured by an adhesive ring of active integrins that surround an F-actin–rich core harboring degradative proteases. Although the features of podosomes, once-established, have been well defined, the steps that lead to podosome formation remain poorly understood by comparison. In this study, we report that spleen tyrosine kinase (Syk) is a critical regulator of podosome formation. Deletion of Syk or targeting its kinase activity eliminated the ability for murine macrophages to form podosomes. We found that the kinase activity of Syk was important for the phosphorylation of its substrates, HS1 and Pyk2, both of which regulate podosome formation. Additionally, before podosomes form, we report that the tandem Src homology 2 domains of Syk afforded multivalent clustering of ITAM-containing adaptors that associated with integrins to structure platforms that initiate podosomes. We therefore propose that Syk has a dual role in regulating podosomes: first, by facilitating the assembly of multivalent signaling hubs that nucleate their formation and second, by sustaining tyrosine kinase activity of the podosomes once they form against their substrates. In cells expressing recently identified gain-of-function variants of SYK, podosomes were dysregulated. These results implicate SYK in the (patho)physiological functions of podosomes in macrophages.

Introduction

The dynamic probing and migration of immune cells involves remodeling of the extracellular matrix (ECM) that can otherwise form barriers to tissue surveillance (1). The ECM is also regularly turned over by myeloid cells to remove its damaged components, control its density, and prevent fibrosis (2). These processes generally require extracellular degradation of polymers embedded in the ECM network by transmembrane and secreted matrix metalloproteinases. The activities of matrix metalloproteinases are coordinated in time and space, concentrated in subcellular structures called podosomes (3). Podosomes are adhesive, force-generating protrusions that are distinctive in their composition and found in select cell types including macrophages (4), dendritic cells (5), osteoclasts (6), endothelial cells (7), and neural crest cells (8). The architecture of podosomes is coarsely conserved between cell types, although there are wide ranges in their half-life reported, depending on the context in which they are formed (6, 9). Podosomes are composed of a densely packed highly branched F-actin core that protrudes from the ventral membrane. These cores are surrounded by a ring of adhesion receptors, notably integrins, and their associated adaptors including talin, vinculin, and paxillin (3). Podosomes are also connected via linear, myosin bundled actin filaments that coordinate the arrangement of rosettes or belts (3, 10). These arrangements in turn facilitate bulk ECM degradation, for example, forming the tightly sealed contacts between osteoclasts and bone that surround a bone-resorbing lacunae (6).

Expectedly, podosomes are tightly regulated structures. They are sites of concentrated kinase activity and are particularly rich in tyrosine phosphorylation. Tyrosine kinases including Src family kinases (i.e., Src and Hck), Abl1, and Pyk2 have all been found to contribute to podosome formation (11, 12); Src-transformed cells, or those ectopically expressing constitutively active forms of Src family kinases, can spontaneously form podosomes (13). Conversely, tyrosine phosphatases, e.g., SHP2, can suppress podosome assembly (14). Signaling lipids and their regulatory kinases are also critical for podosomes to form. For example, the production of phosphatidylinositol (3,4,5)-trisphosphate by PI3K containing the PIK3CB catalytic subunit is essential for podosome formation in macrophages (11). The nucleation promoting factors that branch F-actin networks at the cores of podosomes are also established. Most importantly, nucleation by the Arp2/3 complex, stimulated by Wiskott–Aldrich syndrome protein (WASP), which itself is made active by phosphoinositides and Cdc42, is essential to generate the high density of branching to structure the F-actin cores of podosomes (15). Accordingly, patient cells with mutations in the WAS gene that codes for WASP (4) or in essential subunits of the Arp2/3 complex (16) fail to form podosomes.

Clearly, many regulators of podosomes have been firmly established, yet the complete order of events leading to their production in cells has yet to be fully elucidated. Integrin clustering and activation is thought to be an initial step in podosome formation (17). When extended and immobilized by their ligands, these transmembrane adhesion proteins are capable of excluding the glycocalyx including bulky phosphatases such as CD45 and CD145 (18), which could support tyrosine kinase activity. Integrins also interact directly with signaling molecules such as the Fc receptor common γ chain (FcRγ) and DAP12 via transmembrane associations (19). Both FcRγ and DAP12 contain ITAMs that, when phosphorylated, become docking motifs for kinases with SH2 domains such as Syk. The importance for Syk activity in integrin-dependent processes has been reported in several contexts, such as in the degranulation of neutrophils (20) or during complement-mediated phagocytosis in macrophages (21). Although Syk is predominantly associated with the immune response, it is also capable of phosphorylating and activating actin regulators such as the guanine nucleotide exchange factor Vav1 (22), HS1 (23) and other kinases including PI3K and Pyk2 (24). We therefore questioned whether Syk was also involved in initiating podosome formation.

In this study, we identify Syk as essential for podosomes to form in macrophages. We found that Syk had two roles: its tandem SH2 domains facilitated the establishment of ITAM-containing microclusters, while the kinase domain phosphorylated downstream effectors that are essential for actin polymerization and branching stabilization, namely Pyk2 and HS1, respectively. Gain-of-function variants of SYK, found in patients with systemic inflammation, showed hyperphosphorylation that preceded adhesion formation and ultimately caused dysregulation of podosomes when formed on adhesive surfaces. Given the bone erosion that is also observed in these patients and the protective effect of bone integrity upon Syk inhibition in preclinical models of osteoarthritis, our data implicate SYK in the (patho)physiological functions of podosomes in macrophages and possibly also osteoclasts.

Materials and Methods

Cell culture

RAW264.7 cells were obtained from the American Type Culture Collection. The cells were cultured in DMEM (Wisent, catalog no. 319-007) supplemented with 5% FBS (Corning, catalog no. 35-077-CV) and incubated at 37°C under 5% CO₂. Bone marrow–derived macrophages (BMDMs) were derived from the long bones (femurs and tibia) of wild-type (WT) C57BL/6 mice. The mice were sacrificed by cervical dislocation. Bone marrow was isolated by centrifugation into PBS supplemented with penicillin G, streptomycin, and amphotericin B. Bone marrow pellets were resuspended and plated onto 10-cm Petri dishes and incubated in DMEM supplemented with 10% FBS, 10% L929 medium, antibiotic, and antimycotic (Wisent, catalog no. 450-115-EL). After 7 d, adherent cells were lifted with PBS supplemented with EDTA and plated onto 18-mm coverslips. All animal study protocols were approved by the animal care committee of the Hospital for Sick Children.

To generate Syk knockout (KO) clones, RAW264.7 cells were seeded onto 6-well plates and transfected with Guaranteed Predesigned CRISPR guide RNA plasmid (Sigma-Aldrich) against Syk (catalog no. MMPD0000038537). Single GFP-positive live cells were sorted into the wells of a 96-well plate. The clones were expanded and validated for successful knockout via Western blot.

For transient transfections, the plasmids were mixed with FuGENE HD (Promega) in serum-free DMEM and added to cells seeded in 6-well plates or on a 18-mm coverslip glass. The medium was changed after 4 h. The cells were incubated for either 7 h for short transfection or overnight.

Reagents

Mammalian expression vectors were obtained from the following sources: actin–red fluorescent protein (RFP) and actin-GFP (25), SYK-GFP (26), and vinculin-RFP (AddGene no. 50527, a gift from Kenneth Yamada), and FcγRI-YFP (a gift from Alan Schreiber) were described previously. DAP12-GFP was from Sino Biological (catalog no. MG53479-ACG). The SH2-deleted SYK plasmids were made using the Q5 site-directed mutagenesis kit (New England Biolabs, catalog no. E0554S). Patient variants of SYK were sourced from a previous publication (27).

Primary Abs were obtained from the following sources: anti-vinculin (EMD Millipore, catalog no. MAB3574), anti-phosphotyrosine (EMD Millipore, catalog no. 05-1050), anti-SYK (Cell Signaling, catalog no. 13198), anti-GAPDH (AbClonal, catalog no. AC033), anti–phospho-SYK 525/526 (Cell Signaling, catalog no. 2710), anti-HS1 (Cell Signaling, catalog no. 3892), anti–phospho-HS1 (Cell Signaling, catalog no. 4507), anti-PYK2 (BD Biosciences, catalog no. 610549), anti–phospho-PYK2 (Cell Signaling, catalog no. 3291), anti-FLAG (Cell Signaling, catalog no. 2368), anti–phosphor-Src family (Cell Signaling, catalog no. 2101), anti-Src (Cell Signaling, catalog no. 2108), and anti–phospho-ERK (Cell Signaling, catalog no. 9101). Secondary Abs conjugated with AlexaFluor-405, -488, -555, -647, or -HRP were obtained from Jackson ImmunoResearch.

Other reagents include BAY 61-3606 (Cayman Chemical, catalog no. 11423), R406 (Adooq, catalog no. A10769), Acti-Stain phalloidin (Cytoskeleton Inc.; catalog no. PHDG1-A), PF-431396 (Cayman Chemical, catalog no. 17665), calmidazolium chloride (Tocris, catalog no. 2561), and PP2 (Cayman Chemicals, catalog no. 13198).

Western blotting

The cells seeded in 6-well plates were lysed with RIPA buffer (Sigma-Aldrich, catalog no. R0278) supplemented with protease inhibitors (Thermo Scientific, catalog no. A32955) and phosphatase inhibitors (Thermo Scientific, catalog no. A32957). Samples were centrifuged at 14,000 rpm for 10 min at 4°C, and the supernatants were collected. The protein concentrations were determined via Bradford assay (Bio-Rad, catalog no. 5000006). The samples were mixed with Laemmli buffer (Bio-Rad, catalog no. 1610747) and boiled for 5 min at 95°C. A total of 20 μg of protein was loaded into 4–20% gradient acrylamide gels (Bio-Rad, catalog no. 4561095) and run via electrophoresis. The gels were transferred to nitrocellulose membranes via the Bio-Rad Turbo blot system. The membranes were blocked with 5% BSA in TBST for 30 min. The samples were next incubated in primary Ab for 1.5 h to overnight. The blots were washed three times with TBST before incubating in secondary Ab for 30 min. The samples were washed three times with TBST before being imaged on Bio-Rad gel doc system using ECL (Thermo Scientific, catalog no. 32106) or Immobilon Forte HRP substrate (Millipore, catalog no. WBLUF0500).

Aggregated IgG assay

Human IgG (Sigma-Aldrich, catalog no. I4506) diluted in PBS was incubated at 62°C for 15 min. After cooling, 50 μg of aggregated IgG was added to cells seeded in 6-well plates for 15 min before the cell lysates were collected and analyzed by Western blotting as described above.

Immunoprecipitation

The cell lysates were obtained as described above, and 20 μl of lysate was saved to run as a control. Protein A/G–coated agarose beads (Santa Cruz Biotechnologies, catalog no. sc-2003) were added to the lysates and incubated on a rotating mixer at 4°C for 20 min. The beads were pelleted by centrifugation at 5,000 rpm for 1 min at 4°C. The supernatant was transferred to a new tube. GFP-TRAP nanobody–coated agarose beads (BioLegend, catalog no. 689304) were added and incubated on a rotating mixer at 4°C for 90 min. The samples were centrifuged at 5,000 rpm for 1 min at 4°C, and the supernatant was aspirated. The pellet of beads were resuspended in lysis buffer. This was repeated for a total of five washes. After the last wash, the beads were resuspended with 2× Laemmli buffer and boiled at 95°C for 5 min. The samples were run by Western blotting as described above.

Spinning disk confocal microscopy

Images were acquired using two different spinning disk confocal systems. The first from Quorum Technologies Inc. consisted of a microscope (Zeiss Axiovert 200 M), complementary metal-oxide semiconductor camera (Hamamatsu, ORCA-Fusion BT), five-line laser module (Spectral Applied Research) with 405-, 491-, 561-, and 655-nm lines and a filter wheel (Ludl MAC5000). The system was equipped with a custom 2.4× magnification lens. The microscope was operated using Volocity V6.3 software (Perkin Elmer). The images were acquired with a 63×/1.4 NA oil objective (Zeiss). The other system, used to image cells for quantifying podosomes, consisted of a microscope (Zeiss Axio Observer) and a cooled charge-coupled device camera (Photometrics Evolve 512). The microscope was operated using ZEN2 blue edition software (Zeiss). Images were also taken with a 63×/1.4 NA oil objective (Zeiss).

Immunofluorescence

The cells seeded onto 18-mm coverslips were fixed with 4% paraformaldehyde for 15 min. The samples were then permeabilized with 0.1% Triton X-100 in PBS for 5 min and blocked with 2.5% BSA in PBS for 30 min up to overnight. The samples were incubated in primary Ab for 1.5 h, washed three times with PBS, and then incubated in secondary Ab for 30 min. After washing three times with PBS, the samples were imaged with confocal spinning disk microscopes.

Live cell imaging

Transfected cells seeded on 18-mm coverslips were transferred to a Leiden coverslip chamber. The cells were transferred to HBSS (Wisent, catalog no. 311-515-CL) warmed to 37°C. The microscope stage and objective were warmed to 37°C.

Frustrated phagocytosis

For 1 h, 18-mm coverslips were incubated with 0.5 mg/ml human IgG (Sigma-Aldrich, catalog no. I4506) diluted in PBS. The coverslips were washed three times with PBS and then left in serum-free DMEM containing either vehicle or BAY 61-3606 inhibitor. RAW264.7 macrophages were scraped from 6-well plates and pipetted onto the coated glass. The cells were incubated for 5 min before being fixed.

Quantification of podosomes

The basolateral plane of cells stained for F-actin was acquired via confocal microscopy, and the images were analyzed using ImageJ. The images were converted to 8-bit, and a threshold of signal was set to highlight only the brightest puncta. The “analyze particle” function was used to count podosome number and determine podosome area. To quantify the podosome number in frustrated phagocytosis, podosome analysis using the Fiji (Poji) macro was performed (28). Briefly, the images of the basolateral plane of cells stained for F-actin were opened by the software. The entire cell area was set as the region of interest. Prominence was set to 1,000, and smoothing steps were set to 3.

Quantification of ITAM clusters

The cells expressing FcγRI-YFP were imaged lived. A region of interest of 3.5 × 3.5 μm was made in the leading edge of the polarized macrophages using ImageJ to select flat regions of contact with the coverglass. The images were converted to 8-bit and a threshold was set to identify the brightest puncta. The “analyze particle” function was used to count the number of clusters.

Poly-HEMA assay

Poly(2-hydroxyethyl methacrylate) (poly-HEMA) (Sigma-Aldrich, catalog no. P3932) was dissolved in ethanol to 10 mg/ml. A total of 500 μl of Poly-HEMA solution was added to the wells of 6-well plates, and the plates were incubated at 37°C to allow for evaporation. The wells were washed three times with PBS. Transfected RAW264.7 cells were lifted by scraping and transferred to the coated wells and incubated at 37°C with 5% CO₂ for 2 h before the cell lysates were collected for Western blotting. The control cells were also lifted by scraping and plated on untreated wells for the same duration of time.

Statistics

All experiments were conducted with at least three biological replicates. For all quantified experiments, an unpaired t test was applied using Microsoft Excel. When multiple t tests were used, the Bonferroni correction was applied to reduce type I errors.

Results

Syk is required for podosome formation in macrophages

To establish whether there is a role for Syk in podosome formation, we first determined whether the kinase is targeted to these structures. To that end, RAW264.7 macrophages expressing Syk-GFP were fixed and stained for F-actin, which demarcates the cores of podosomes. Syk-GFP was indeed found to localize to the F-actin cores of podosomes that formed on the basal membrane of the macrophages (Fig. 1A). To determine whether Syk kinase activity influenced podosome formation, RAW264.7 cells expressing actin-GFP together with vinculin-RFP to visualize podosome cores and rings, respectively, were imaged live over 10 min. The cells treated with a vehicle control formed and disassembled podosomes dynamically. Once cells were treated with BAY 61-3606, a somewhat specific inhibitor of Syk kinase function, they failed to maintain or form new podosomes (Fig. 1B, Supplemental Fig. 1, Supplemental Movie 1). To further validate that the actin-GFP puncta observed in these live experiments represented bona fide podosomes, we stained for endogenous podosome markers including vinculin and phosphotyrosine together with F-actin. Compellingly, the vinculin rings and phosphotyrosine of podosome cores that formed readily in our cells were absent after 10 min of Syk inhibition by BAY 61-3606 or R406, a different, clinically relevant Syk inhibitor (Fig. 1C–F).

FIGURE 1.

Syk is required for podosome formation in macrophages. (A) RAW264.7 macrophages expressing Syk-GFP. The cells were fixed and stained for F-actin with phalloidin. (B) RAW264.7 macrophages expressing actin-GFP and vinculin-RFP imaged over 10 min. The cells were treated with either vehicle control or 0.5 µM BAY 61-3606 for the indicated times. (C and D) RAW264.7 cells with the indicated treatment stained with phalloidin and with anti-vinculin (C) or anti-phosphotyrosine (D) Abs. The cells were treated with 0.5 μM BAY 61-3606 for 10 min. (E and F) Individual fields of RAW264.7 cells (gray dots) were quantified for the percentage of cells forming podosomes (E) and the number of podosomes per cell when treated with vehicle or Syk inhibitors, 1 μM R406 and 0.5 μM BAY 61-3606 (F). Here and elsewhere, the colored dots represent means of individual experiments (n = 3). (G) Validation of Syk deletion in RAW264.7 clones via Western blot. (H) Phalloidin staining in Syk KO RAW264.7 clones. (I) Percentage of cells forming podosomes in individual fields (gray dots) for WT versus Syk KO RAW264.7 cells (n = 3). All images are XY sections. Scale bars, 5 µm. IB, immunoblot.

To augment findings derived from these pharmacological interventions, we edited RAW264.7 cells using CRISPR-Cas9 to delete the expression of Syk. We confirmed the loss of Syk expression in multiple clones (Fig. 1G). The loss of functional Syk was supported by experiments in which we challenged these cells with aggregated IgG to evaluate tyrosine phosphorylation downstream of Fcγ receptors, a Syk-dependent process (29). Consistently, we found the Syk KO clones were impaired in tyrosine phosphorylation downstream of FcR signaling (Supplemental Fig. 2). The Syk KO clones also failed to form podosomes, which supported the findings made with Syk inhibitors (Fig. 1H). Taken together, these results demonstrated a role for Syk in regulating the formation of podosomes in macrophages.

Syk is required for phagosome-associated podosome-like structures

During the course of Fc-mediated phagocytosis, podosome-like structures form (9, 30). These structures enable the phagocyte to deform targets during internalization (31) and, conceivably, would support degradation and the extraction of smaller components of large structures (e.g., biofilms, adipocytes, etc.) encountered during “frustrated” phagocytosis. The podosomes that form during Fc-mediated phagocytosis are similar to those forming in migrating macrophages but have a shorter half-life (9). Given the established role for Syk in Fc-mediated phagocytosis and our finding that Syk is required for podosomes to form in unstimulated macrophages, we suspected that Syk may also be important for the podosome-like structures that assemble during phagocytosis. To test this, we opsonized coverglass with IgG to observe the plane of phagocytic contact using BMDMs. Upon making contact with the IgG-coated surfaces, vehicle-treated BMDMs spread and formed an expanded ring of podosomes. BMDMs treated with BAY 61-3606 were also capable of spreading but no longer formed the actin puncta or their associated vinculin rings (Fig. 2A–C). RAW264.7 cells also produced phagosome-associated podosomes on IgG-coated surfaces that were absent in Syk KO clones (Fig. 2D, 2E). Thus, the podosomes produced in unstimulated macrophages, as well as those that form during phagocytosis, are Syk-dependent.

FIGURE 2.

Syk is required for podosome-like structures that form during Fc-mediated phagocytosis. (A and B) BMDMs undergoing frustrated phagocytosis on IgG-opsonized coverglass. The cells were treated with vehicle or BAY 61-3606 for the duration of adhesion/spreading and then fixed and stained with phalloidin, anti-vinculin (A), and anti-phosphotyrosine (B). (C) Podosomes per cell during frustrated phagocytosis when treated with vehicle or BAY 61-3606 (n = 3). (D) Frustrated phagocytosis of WT and Syk KO RAW264.7 cells on IgG-opsonized coverglass. The cells were fixed and stained with phalloidin and anti-vinculin Abs. (E) Podosomes per cell during frustrated phagocytosis for WT and Syk KO RAW264.7 cells. All images are XY sections. Scale bars, 5 µm.

The tandem SH2 domains of Syk are required for ITAM clustering in unstimulated macrophages

The need for Syk in the construction of podosomes would seemingly arise from its tyrosine kinase activity. However, Src family kinases are canonical regulators of podosomes, phosphorylate and bind to ITAMs, and share many of the downstream effectors of Syk. Therefore, to determine the unexpected and nonredundant function(s) of Syk related to podosomes, we reintroduced forms of Syk lacking individual domains or kinase activity in Syk KO RAW264.7 cells. Re-expression of a WT form of Syk clearly restored podosomes, as evinced by the reappearance of F-actin–rich cores of the structures (Fig. 3A, 3B). By contrast, and expectedly, the re-expression of a kinase dead form of Syk did not restore podosomes in the KO cells (Fig. 3A, 3B). Although the SH2 domains of Syk are both capable of binding to phosphorylated ITAMs, we found that deletion of either of the two SH2 domains independently did not restore podosomes in the cells (Fig. 3A, 3B). Structural modeling has suggested the ability of individual Syk molecules to abridge the ITAMs associated with Fcε receptors in cis, particularly during partial ITAM phosphorylation (32). We therefore reasoned that multivalent interactions between Syk and ITAM-containing adaptors, implicit in their structural features, may underlie this effect.

FIGURE 3.

The SH2 domains of Syk facilitate ITAM clustering that precedes podosome formation. (A) Syk KO RAW264.7 cells expressing forms of SYK lacking kinase activity or domains fixed and stained with phalloidin. Anti-FLAG staining was used to identify transfected cells. (B) Individual fields of Syk KO RAW264.7 cells expressing forms of SYK (gray dots) were quantified for number of podosomes per transfected cell. n = 3. (C) RAW264.7 macrophages expressing actin-RFP and FcγRI-YFP imaged live. (D) WT and Syk KO RAW264.7 macrophages expressing FcγRI-YFP. (E) FcγRI-YFP clusters in individual WT and Syk KO RAW264.7 cells were quantified (gray dots; n = 3). (F) WT RAW264.7 cells expressing FcγRI-YFP treated with vehicle or PP2 for 20 min in serum-free media. (G) Syk KO RAW264.7 macrophages expressing FcγRI-YFP and forms of SYK lacking kinase activity or domains imaged live. (H) FcγRI-YFP clusters in individual Syk KO RAW264.7 cells expressing forms of Syk (gray dots; n = 3). All images are XY sections. Scale bars, 5 µm.

There are ITAM-containing adaptors known to associate with integrins such as DAP12 and the common γ-chain that associates with Fc receptors (FcRγ), including when it associates with FcεRI (19). We hypothesized that such adaptors may also be present in podosomes. To test this, we expressed FcγRI-YFP, which constitutively associates with the ITAM-containing common γ-chain and requires this association for its normal traffic, together with actin-RFP. We readily observed that clusters of FcγRI-YFP formed in the basal membrane of these cells, and many of these colocalized with actin puncta (Fig. 3C). Notably, small FcγRI-YFP clusters also formed at the leading edge of the macrophages, suggesting that they may precede the formation of podosomes. Staining for phosphotyrosine revealed that these clusters are phosphorylated even when F-actin is not yet present (Supplemental Fig. 3), suggesting that they may be nascent podosomes.

Importantly, FcγRI-YFP clusters were largely absent in Syk KO cells (Fig. 3D, 3E). This suggests that Syk plays a critical role in the clustering of ITAMs before and during their association with podosomes. The source of the signals that support ITAM clustering at rest are, however, unknown. We targeted the activity of Src family kinases (SFKs), which are thought to be responsible for the initial phosphorylation of ITAMs, because they were found to be active in resting macrophages as judged by phosphorylation of their activating tyrosines (Supplemental Fig. 4). Pharmacological inhibition with PP2 reduced SFK activity (Supplemental Fig. 4). Interestingly, FcγRI-YFP clusters were ablated upon treatment for PP2 (Fig. 3F). This demonstrates that tonic phosphorylation of ITAMs by SFK serves to recruit Syk to ultimately promote ITAM clustering.

As we had observed with podosomes, upon rescuing the expression of WT Syk in the knockout cells, the formation of FcγRI-YFP clusters was restored (Fig. 3G, 3H). We also noted that the FcγRI-YFP microclusters were at least partially restored upon expression of the kinase dead form of Syk, suggesting that these microclusters were not ostensibly dependent on Syk activity (Fig. 3G, 3H). Instead, both SH2 domains were required for FcγRI-YFP microclusters to form in the cells (Fig. 3G, 3H). This suggests that the tandem SH2 domains in Syk indeed play a pivotal role in facilitating the formation of nascent ITAM-containing microclusters that form spontaneously in resting macrophages and are likely responsible for facilitating the generation of podosomes. To then establish bona fide podosomes, the kinase activity of Syk, which would be concentrated at these sites, is presumably required to activate and recruit downstream effectors that stimulate actin polymerization and expand integrin adhesion.

Syk phosphorylates HS1 and Pyk2 in unstimulated macrophages

Once docked to ITAMs, Syk can phosphorylate a plethora of downstream effectors including those known to be localized to podosomes, e.g., Pyk2 and HS1. We evaluated the phosphorylation of these proteins in WT and Syk KO cells, as well as cells treated with the Syk inhibitor BAY 61-3606. In unstimulated cells, both HS1 and Pyk2 were found to be phosphorylated (Fig. 4A). Maintaining the steady-state phosphorylation of these substrates required Syk expression and activity (Fig. 4A).

FIGURE 4.

Downstream targets of Syk important for podosome formation. (A) Phosphorylation of downstream targets of Syk in RAW264.7 macrophages at rest when treated with BAY 61-3606 and in Syk KO RAW264.7 cells, determined by Western blotting. The blots were cropped, and the control blots were derived from the same samples. (B) RAW264.7 cells with the indicated treatments stained with phalloidin. The cells were treated with 1 μM PF-431396 for 10 min and 5 μM calmidazolium chloride for 30 min. (C and D) Individual fields of RAW264.7 cells (gray dots) quantified for the percentage of cells forming podosomes (C) and the number of podosomes per cell when treated with vehicle, 1 μM PF-431396, or 5 μM calmidazolium (n = 3) (D). All images are XY sections. Scale bars, 20 µm.

Pyk2 activation is required for podosomes

Pyk2 is made active by phosphorylation and Ca²⁺/calmodulin (33). When active, Pyk2 serves to recruit Src family kinases (34) and activate gelsolin (35), an F-actin–severing protein also implicated in podosome formation. To determine whether Pyk2 activity is also necessary for podosomes, we treated cells either with PF-431396, an inhibitor of both Pyk2 and focal adhesion kinase (FAK) activity, or with the calmodulin inhibitor calmidizolium to prevent Pyk2 activation. Importantly, Pyk2 has a nonredundant role when compared with FAK in macrophage chemotaxis (36). Both the acute inhibition of FAK/Pyk2 kinase activity with PF-431396 and that of calmodulin with calmidizolium abolished podosomes (Fig. 4B–D). These data support a role for a bona fide Syk substrate, Pyk2, in the regulation of podosomes.

Gain-of-function variants in Syk act upstream of cell adhesion and cause dyresgulation of podosomes

Recently, we identified patients exhibiting systemic inflammation and autoimmune disorders with mutations in SYK (27). Five independent, disease-causing variants of SYK harbored mutations in the kinase domain or in its adjacent interdomain region (Fig. 5A). We initially found that the ectopic expression of these variants in HEK293T cells resulted in increases in the phosphorylation of SYK on its activating tyrosines, Y525 and Y526, and hypersensitized responses to pathogen-associated molecules (27). This led us to describe the variants as “gain-of-function.”

FIGURE 5.

Gain-of-function SYK propagates inflammatory signaling from podosome platforms. (A) PyMOL structure of SYK. SH2 domains (green) and kinase domain (blue) labeled. Identified variants of SYK that result in gain-of-function hyperactivity are highlighted (red). (B) Syk KO RAW264.7 macrophages expressing WT or patient variants of SYK. Phosphorylation and activity of SYK variants were determined by Western blot. The blots were cropped, and the control blots were derived from the same samples. (C) Syk KO RAW264.7 macrophages expressing WT or the S550Y variant of SYK cultured on untreated tissue culture (TC) plastic or Poly-HEMA-coated plates for 2 h. Phosphorylation and activity of SYK were evaluated by Western blotting. The blots were cropped, and the control blots were derived from the same samples. (D) Syk KO RAW264.7 macrophages expressing WT or S550Y variants of SYK were probed for the phosphorylation of downstream effectors at rest by Western blotting. The blots were cropped, and the control blots were derived from the same samples. (E) Syk KO RAW264.7 macrophages expressing WT or S550Y variant of SYK, fixed and stained for phalloidin and anti-FLAG to validate transfected cells. (F) Individual fields of Syk KO RAW264.7 cells expressing WT or the S550Y variant of SYK (gray dots) were quantified for the number of podosomes per transfected cell (n = 3). (G) Podosome size was determined in Syk KO RAW264.7 cells expressing either WT or S550Y SYK. (H) Schematic of the proposed model for the dual role of SYK in podosomes. The figure was created using BioRender.com. All images are XY sections. Scale bars, 5 µm. IB, immunoblot.

To determine whether these patient variants of SYK behaved similarly in macrophages, we transfected the different variants into Syk KO RAW264.7 macrophages. We noted that, similar to their expression in HEK293T cells, the increased phosphorylation of the SYK variants was also found when expressed in macrophages, to varying degrees (Fig. 5B). Notably, the pY525/526 SYK signal in the S550Y variant muted the increased signal in other variants when run on the same gel; all variants are in fact increased when compared with WT SYK (27). In addition to SYK, total levels of phosphotyrosine were increased in the cells (Fig. 5B), indicating that downstream targets of SYK were targeted by the variant forms of the kinase.

We had not previously determined whether the gain-of-function variants of SYK act only downstream of receptor engagement or could potentially also act upstream. Syk has been reported to directly phosphorylate ITAMs (37), which would serve to recruit the kinase to the membrane. Supporting this, we found that the gain-of-function versions of SYK were capable of phosphorylating ITAM-containing adaptors including DAP12 (Supplemental Fig. 5). It therefore remained conceivable that gain-of-function versions of SYK act without any receptor engagement. To test whether the elevated phosphorylation observed upon expression of the active forms of SYK was independent of receptor engagement, we coated plates with poly-HEMA to prevent attachment of cells to the plastic. Even for cells on poly-HEMA, held in suspension for 2 h, the phosphorylation of SYK persisted together with downstream phosphotyrosine, suggesting that the activity of SYK was ostensibly ligand-independent (Fig. 5C).

When cells expressing the S550Y gain-of-function version of SYK were allowed to adhere and engage their integrins, we found that the downstream effectors HS1 and Pyk2 were hyperphosphorylated (Fig. 5D). We hypothesized that the increased activity of SYK at podosomes may therefore promote excessive actin polymerization. Convincingly, we found that the expression of the S550Y mutant version of SYK in the cells rescued podosome formation to a similar extent as the WT form; however, these podosomes appeared abnormal: grossly enlarged when quantified and irregular in their shape (Fig. 5E–G). Taken together, these data support a role for SYK in the formation of podosomes and the signaling from podosomes once they are established. The data also implicate podosomes in the pathophysiology associated with SYK gain-of-function.

Discussion

The functions and downstream targets of Syk in response to the engagement of immunoreceptors, i.e., Fc receptors and the BCR, are well studied. Once recruited to the ITAMs of ligated and clustered immunoreceptors, Syk initiates critical pathways that lead to phagocytosis, pathogen killing, and cell proliferation (38–43). Accordingly, the deletion of Syk in the hematopoietic compartment of mice impairs phagocytic uptake (29), the respiratory burst (20), and Ag presentation in myeloid cells (44) and the development and activation of lymphoid cells (45). In this study, we describe that Syk is also active in “resting” macrophages and that this activity is indispensable for the formation of podosomes. Although tonic signaling (i.e., signaling in the absence of ligand) of the BCR (46) and Fc receptors (47) via Syk had been shown to regulate the actin cortex of cells and receptor diffusion, a role for such activity in the formation of podosomes had not been established previously.

How podosomes are nucleated in the first place remains poorly understood. Pre-existing aggregates of transmembrane matrix metalloproteases appear to be sites where podosomes preferably assemble (48). The polarity of signaling lipids in cells is also a critical determinant because the lipid kinases and phosphatases that orchestrate the maintenance of phosphoinositides at the leading edge of migrating cells, i.e., where podosomes form, are critical regulators of their initiation (11). We found that clustering of ITAMs, phosphorylated by SFKs and clustered by the tandem SH2 domains of Syk, is also important for the instigation of podosomes (see model, Fig. 5H). Supporting this model, the tandem SH2 domains of Syk have been reported to extend across and abridge hemi-ITAM containing C-type lectin receptors (49) and Fcε receptors (32). Clustered ITAMs may exclude the highly glycosylated, bulky transmembrane phosphatases like CD45 and CD148 that would otherwise arrest signal transduction from podosomes. Conceivably, this could be either the result of phase separation of the highly glycosylated ectodomains of the phosphatases (50) or caused by the “kinetic segregation” of the bulky phosphatase away from the receptors when force is applied to the contact surface and the membrane has very close apposition from the substratum (51). CD45 is indeed excluded from mature podosomes (18), but the requirement for this exclusion remains to be tested.

The maturation steps that facilitate the transition of signaling microclusters to become full-fledged podosomes are also not established. We noted that the clusters of ITAM-containing receptors formed by Syk outnumbered podosomes in a given macrophage. The microenvironment of the cluster, including the features described above that control lipids and phosphatases, likely dictate the outcome. Integrins would also need to become extended, clustered, and made active, which requires the local activation of Rap GTPases and the association of integrins with actin filaments via adaptors like talin (52). Although the diffusion of integrins in immune cells is heterogeneous, like other transmembrane proteins, their movement is restricted in the plane of the membrane by the submembrane-associated actin cytoskeleton that furnishes the inner leaflet of the plasma membrane (53). Given the important role for Syk in remodeling of these cortical actin polymers (47), it seems likely that the local diffusion of integrins may be affected without the activity of Syk in macrophages. Surprisingly, then, we still observed polarization and spreading of macrophages that did not express Syk, suggesting that there may be less reliance on Syk for various other adhesion receptors.

Once formed, podosomes are sites of dense yet dynamic actin polymerization and play essential functions in remodeling of the extracellular matrix. We found that podosomes continue to concentrate Syk where it phosphorylates other tyrosine kinases like Pyk2 and actin regulators including HS1. Pyk2 had been implicated in podosome assembly before: the kinase was shown to localize to these structures (12), which form less readily in Pyk2^−/− myeloid cells (54). Interestingly, Pyk2 knockout mice also have increased bone density, likely owing to the loss of bone resorption activity of osteoclasts (36, 54), and impaired macrophage motility, further implicating a role for Pyk2 in the degradative and adhesive functions of podosomes. In certain types of tumors, PYK2 is overexpressed where it plays a role in the formation of invadopodia (55), suggesting a broader role for the kinase in podosome-like structures. An investigation into the regulation of invadopodia by Syk should also be made because its expression and activity are also increased in a wide segment of hematopoietic cancers (56, 57), which invade organs and tissues through remodeling of the basement membrane.

Although HS1 localizes to the actin-rich cores of podosomes, its role in regulating these structures is less clear. HS1 would seemingly serve to be important for podosomes to form because it acts as a scaffold to localize factors that promote F-actin branching such as Vav1, WASP, and Arp2/3 (58) and fortifies branching points after they are formed by binding Arp2/3 and antagonizing their debranching coronin proteins (59). Indeed, HS1^−/− mice have cytoskeletal defects emphasized in their lymphocytes that have an impaired ability to form an immune synapse (60). However, podosomes do appear to form in HS1^−/− macrophages but have been reported to show an altered organization, forming loosely packed arrays of these structures that are improperly localized (61). There may well also be overlapping functions between HS1 and other F-actin branching regulators. Still, the phosphorylation of both HS1 and Pyk2 by Syk in the steady state and the control of podosome formation/dynamics of these substrates point toward a broadly important role of Syk at these sites once they form. We suspect that Syk may also serve to recruit and activate PI3K, a lipid kinase reported to be downstream of Syk in other contexts, which is also implicated in podosome formation.

Recently, we described variants in SYK that lead to human disease in which several infantile patients were reported to have systemic inflammation caused by gain-of-function alleles of the kinase (27). In that context, hematopoietic cells were largely implicated as causing disease. A mouse model in which the gain-of-function allele was introduced established that 1) chimeras receiving affected bone marrow were prone to disease and 2) the irradiation and transplant of WT bone marrow rescued phenotypes in the SYK gain-of-function mice (27). So far, one patient with a severe gain-of-function variant of SYK responded positively to transplant (A. M. Muise, unpublished observations), further supporting the notion that a major component of the disease is driven by hematopoietic cell types. The autoimmunity observed in SYK gain-of-function patients is also associated with arthritis, which is accompanied by bone erosion (27). This would implicate potentially dysregulated osteoclasts. We found that gain-of-function forms of SYK rescued podosome formation in Syk null cell lines and that these podosomes were larger. The pathophysiological functions of these podosomes should be further characterized.

Taken together, this study’s results establish a role for Syk in the regulation of podosomes that form in macrophages. Podosomes and podosome-like structures in myeloid cells and transformed cell types have a range of functions that depend, in part, on the matrices they are poised to degrade. If a direct role for SYK is established in patients experiencing bone density loss or metastatic progression of tumors, this may provide the impetus for targeting the SYK pathway in these cases. Such an approach should be made with cation given the role of SYK in the podosomes that form in macrophages as established in this report.

This article is featured in Top Reads, p. 921

Disclosures

The authors have no financial conflicts of interest.