Myosins 1e/f at the podosome base regulate podosome dynamics and promote macrophage migration

Abstract

Cells of the monocyte lineage form specialized membrane-associated, actin-rich structures, called podosomes. Podosomes play important roles in cell adhesion and migration as well as the proteolytic degradation of the extracellular matrix. While podosomes are always closely associated with the plasma membrane, the structural components linking the podosome core, composed of branched actin, to the membrane are not fully understood. In this study we show that class I myosins, Myo1e and Myo1f, localize to a specific region of podosomes, underneath the podosome core and near the ventral plasma membrane, and that this localization is mainly mediated by the Myo1e/f TH2 domains. Respective knockdowns or knockouts of Myo1e/f lead to increased podosome size, altered turnover and lateral mobility, which is likely due to Myo1e/f regulating the attachment of core actin filaments to the plasma membrane. In addition, Myo1e/f double knockout macrophages were characterized by a reduction in 3D and 2D migration, even though these cells exhibited increased ability to degrade the extracellular matrix. Along with the other membrane-associated podosome components, such as the transmembrane protein MT-MMP and the GPI-anchored DNase X, Myo1e and Myo1f mark the membrane-proximal region of podosomes. We propose to label this region as the podosome “base”, an additional substructure joining the current trifecta of the podosome cap, core, and ring.

1. Introduction

Macrophages are cells of the innate immune system that actively migrate along the surfaces of endothelial cells, across the endothelial and basement membrane barriers, and throughout tissues and interstitial spaces to perform their immune surveillance functions. They are highly dynamic cells that can employ a variety of cell migration modalities. Of note, efficient directional migration of macrophages has been proposed to be based on the stabilization of leading edges by podosomes, specialized actin-rich adhesion structures, on their ventral surfaces (Wiesner et al., 2014). In addition to forming cell-matrix adhesion sites, podosomes also serve as the sites of proteolytic enzyme secretion, allowing macrophages to invade and breach barriers (Linder et al., 2023).

Podosomes consist of > 300 different proteins (Cervero et al., 2012) that form a characteristic tripartite organization involving an F-actin-rich core surrounded by a ring of adhesion proteins and topped by a cap structure (Linder et al., 2023). The F-actin core of podosomes includes the Arp2/3 complex (Linder et al., 2000) that promotes assembly of branched actin networks, Arp2/3 activators belonging to the WASp protein family (Linder et al., 1999), actin cross-linker L-plastin (Zhou et al., 2016), as well as the actin-binding protein cortactin (Tehrani et al., 2006) and Tks family scaffold proteins (Seals et al., 2005), among others. The discontinuous ring structure contains proteins such as integrins (Marchisio et al., 1988), paxillin (DeFife et al., 1999), and vinculin (DeFife et al., 1999; Marchisio et al., 1988), while the cap structure contains mostly actin crosslinking or bundling proteins such as α-actinin (van den Dries et al., 2019), fascin (Van Audenhove et al., 2015), or LSP1 (Cervero et al., 2018). Altogether these proteins cooperate to regulate podosome life cycle and function, but the role of many podosome components remains unclear.

Myosin 1e (Myo1e) and the closely related myosin I isoform myosin 1f (Myo1f) were first identified as podosome components by mass spectroscopy of podosome-enriched fractions from primary human macrophages (Cervero et al., 2012). These class I myosins are monomeric motors that can bind actin filaments via a motor domain and plasma membrane via positively charged tail domains (Giron-Perez et al., 2019). Class I myosins include long-tailed (Myo1e and Myo1f, containing proline-rich and SH3 domains in their tails) and short-tailed myosins, whose tails consist only of the membrane-binding regions (McConnell and Tyska, 2010). In mammalian and yeast cells, class I myosins preferentially associate with Arp2/3-nucleated branched actin networks and nearby membranes in structures such as endocytic actin patches and phagocytic cups (Barger et al., 2019; Sirotkin et al., 2005). While Myo1e was previously detected in invadopodia of v-src-transformed fibroblasts and podosomes of RAW macrophages (Ouderkirk and Krendel, 2014; Zhang et al., 2019), the specific functions of Myo1e and Myo1f at podosomes and their contributions to macrophage migration have not been elucidated.

In the present study, we have interrogated the potential functional roles of Myo1e/f in podosome dynamics and macrophage migration using human macrophages differentiated from primary blood monocytes as well as murine bone-marrow derived macrophages (BMDMs) from the wild-type and Myo1e/Myo1f single and double knockout (dKO) mice (Barger et al., 2019; Kim et al., 2006; Krendel et al., 2009). Human cells readily form podosomes in large numbers, which facilitates statistical analysis, while murine knockout macrophages allow podosome analysis in the complete absence of our proteins of interest. Therefore, we combined the two approaches to enable a comprehensive analysis of myosin I functions. We examined the effects of Myo1e/f depletion by siRNA and Myo1e/f knockout on regulation of podosome size, turnover, and lateral mobility and found that the absence of Myo1e/f was associated with reduced macrophage migration in 2D and 3D. Moreover, we found that Myo1e and Myo1f, but not other class I myosins tested, localized to the ventral surface of the podosome core, a region that in macrophages also contains other plasma membrane-associated proteins. In line with its restricted localization underneath the podosome core and its specific components (see Discussion), we propose to name this part of the podosome architecture as the podosome “base”.

2. Results

2.1. Myo1e and Myo1f are enriched at the podosome base in macrophages derived from primary human monocytes

First, we examined the endogenous Myo1e and Myo1f localization at podosomes of human macrophages by immunofluorescence using myosin isoform-specific antibodies (Fig. 1). Both myosins localized to the F-actin rich podosome core labeled with phalloidin (Fig. 1A–F, J–M), but appeared concentrated at the lower part of the podosome core, close to the ventral surface. We hereby refer to this region as the podosome “base”. To test for potential chromatic aberration, we also stained control cells with both Alexa-568 and Alexa-488 labeled phalloidin. These results show that there is indeed a slight chromatic aberration between the green and red channels (Fig. 1G–I, N,O). However, this cannot account for the pronounced difference in location (and shape) between the Myo1e/Myo1f signals and the F-actin core of podosomes (Fig. 1A–F). Analysis of fluorescence intensities using the Poji macro (Herzog et al., 2020) confirmed that the peaks of F-actin and Myo1e/f intensity align along the central axis of podosomes and are equally wide, however, the signals for Myo1e/f show a maximum intensity in the confocal Z-layer between 0.6 – 0.8 μm above the substrate, while F-actin peaks at 1.0 μm (Fig. 1J–M). Like the endogenous proteins, exogenously expressed EGFP-tagged Myo1e and Myo1f also localized to podosomes (Fig. 2C–E, C′–E′). In contrast, none of the fluorescently tagged (EGFP or mEmerald) short-tailed class I myosins (Myo1a,b,c,d,g) were enriched at podosomes (Supp. Fig. 1).

Fig. 1.

Endogenous Myo1e and Myo1f localize to the base of macrophage podosomes. (A–F) Confocal micrographs of primary human macrophages stained for endogenous Myo1e (A–C) or Myo1f (D–F), and for F-actin using Alexa-Fluor-488 phalloidin, to label podosome cores (A–F). White boxes in (A,D) indicate regions of interest shown as single channels in the upper and lower left corner, respectively, with corresponding XY (B,E) or XZ views (C,F). Note localization of Myo1e and Myo1f to the podosome base, a region mostly beneath the F-actin-rich podosome core. (G-I) Test for potential chromatic aberration by staining primary human macrophages with both Alexa-Fluor-568 and −488. Note that the slight shift between the red and green channels cannot account for the difference between Myo1e/f localization and the actin-rich core of podosomes (A-F). Scale bars: 10 μm for (A,D,G), 1 μm for (B,C,E,F,H,I). (J-O) Analysis of Myo1e and Myo1f localization at podosomes. Radial fluorescence intensity profiles of podosomes stained for F-actin (K,M,N,O) and Myo1e (J) or Myo1f (L). Profiles were calculated using PoJi, a Fiji-based tool (Herzog et al., 2020), and represent the mean ± s.e.m. of all podosome profiles imaged at indicated confocal z planes, with 0 μm set as the most ventral F-actin signal of podosomes; n = 6542 podosomes, per z section, collected from 12 different cells (J, K) and n = 6118 podosomes, per z section, collected from 10 different cells, (L,M); n = 4330 podosomes per z section, collected from 8 different cells (N,O). Note that Myo1e and Myo1f at podosomes show the maximum fluorescence intensity at z-plane of 0.6–0.8 μm, whereas F-actin peaks at z-planes of 1.0 μm, and also extends to the higher optical planes (not shown; see A-F for complete visualization).

Fig. 2.

Localization of Myo1e and Myo1f to podosomes requires their TH2 domains. (A,B) Domain structure of Myo1e (A) and Myo1f (B). Both myosins contain an N-terminal motor domain, followed by a neck domain (containing an IQ motif) that binds calmodulin or other light chains, and a C-terminal tail, containing lipid-binding tail homology-1 (TH1), proline-rich tail homology-2 (TH2), and Src homology-3 (SH3) domains. Presence/absence of the respective EGFP-tagged expression constructs at macrophage podosomes is indicated by “±”. (C-T, C’-T’) Confocal micrographs of primary human macrophages overexpressing GFP-fused full length or partial constructs of Myo1e (C,F,I,L,O,R) or Myo1f (C’,F’,I’,L’,O’,R’), as indicated, with co-expression of lifeact-RFP, to label podosome cores. Insets show magnified images of individual podosomes, indicated by white boxes, shown as single GFP or RFP channels, with respective XY merges (D’,G’,J’,M’,P’,S’,D’,G’,J’,M’, P’,S’). Note absence from podosomes of constructs lacking the tail (“motor+IQ”), and the marked decrease in localization to podosomes by constructs lacking the TH2 domain (ΔTH2). Scale bars: 10 μm.

These data show that both endogenous and overexpressed forms of Myo1e and Myo1f localize to macrophage podosomes. Indeed, both motors are specifically enriched at the ventral side of podosome cores, which we name the podosome base. This localization is close to the ventral plasma membrane, which is likely correlated with the ability of both myosins to bind membrane lipids (Feeser et al., 2010; McIntosh and Ostap, 2016).

2.2. The TH2 domains determine localization of Myo1e and Myo1f to podosomes

Myo1e and Myo1f contain a motor/head domain that binds ATP and F-actin, a neck domain (IQ) that binds calmodulin, and a tail, containing a lipid-binding tail homology-1 (TH1), proline-rich tail homology-2 (TH2), and Src homology-3 (SH3) domains (Fig. 2A,B). We have previously demonstrated that the TH2 domain of Myo1e is necessary for its localization to podosomes, and that the TH1 domain promotes its membrane enrichment/ventral localization (Ouderkirk and Krendel, 2014; Zhang et al., 2019). In contrast, localization requirements have not been previously examined for Myo1f. By overexpressing EGFP-tagged domain deletion mutants of Myo1e and Myo1f together with the RFP-labeled F-actin marker Lifeact, we tested the requirements for specific protein domains in localization of both myosins using live-cell imaging to avoid fixation artifacts. We found, for both Myo1e and Myo1f, that constructs containing only the N-terminal motor and IQ domains (“motor + IQ”) (Fig. 2 F–H, F’–H’; Suppl. Fig. 2C,D,C’,D’) were not enriched in podosomes, while constructs containing only the C-terminal half comprising the TH1, TH2 and SH3 domains (“tail”) were enriched in podosome cores (Fig. 2 I–K, I’–K’; Suppl. Fig. 2E,F,E’,F’). Further dissection of the C-terminal tail regions showed that constructs lacking the TH1 domain lost their specific localization to the podosome base (Fig. 2L–N, L’–N’; Suppl. Fig. 2G,H,G’,H’), with the respective Myo1f deletion construct also being more diffusely distributed in the XY plane (Fig. 2L’–N’; Suppl. Fig. 2G’,H’). Importantly, constructs lacking the TH2 domain (Fig. 2O–Q, O’–Q’; Suppl. Fig. I,J,I’,J’) showed marked reduction in their localization to podosomes. Moreover, deletion of the SH3 domain of Myo1e or Myo1f did not interfere with their localization to the podosome base (Fig. 2R–T, R’–T’; Suppl. Fig. 2K,L,K’,L’). Collectively, these results show that the TH2 domains of Myo1e and Myo1f are important for the localization of both myosins to macrophage podosomes. In the case of Myo1f, podosomal localization is also in part determined by the TH1 domain. We also found that actin nucleation-promoting factor WASp colocalized with Myo1e at the podosome base (Supp. Fig. 3).

2.3. Myo1e and Myo1f regulate podosome properties

To test how Myo1e/f contribute to the regulation of podosome organization and dynamics, we examined the effects of the knockdown of these proteins in human macrophages, using two distinct siRNAs for each myosin to control for potential off-target effects (Suppl. Fig.4A,B). Analysis of live-cell imaging videos of human macrophages expressing Lifeact-RFP to label podosome cores showed that Myo1e knockdown increased the size of podosome cores by ~ 50 % (0.38 ± 0.07 μm² for controls, and values ranging from 0.56 ± 0.17 μm² to 0.57 ± 0.2 μm² for respective depletions) (Fig. 3A), while treatment with all siRNAs led to reduced podosome lifetime by ~30 % (7.2 ± 3.1 min for controls, and values ranging from 4.9 ± 1.5 min to 5.4 ± 1.5 min for respective depletions) (Fig. 3B). Further analyses showed that the lateral displacement of podosomes during their lifetime was also reduced in these cells by ~20 % (1.0 ± 0.29 μm for controls, and values ranging from 0.77 ± 0.17 μm to 0.87 ± 0.23 μm for respective depletions) (Fig. 3C).

Fig. 3.

Myo1e/f regulate podosome size, lifetime, and lateral movement, but not reformation. (A-C) Analysis of podosome core area (A), lifetime (B) or lateral mobility (C) in macrophages treated with siRNAs for Myo1e or Myo1f. Ordinary one-way ANOVA, ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05, ns = statistically not significant; at least 17 cells / condition were evaluated in 30 min videos, collected from 3 different donors. (D-G) Analysis of podosome numbers per cell (D), podosome-covered cell area (E), correlation analysis between the podosome-covered area of a specific cell and the respective cell area (F), and podosome core area (G) in macrophages depleted simultaneously (dKD) for both Myo1e and Myo1f. Unpaired t-test, ****P < 0.0001; 18 cells / condition, collected from 3 different donors (D). Unpaired t-test, **P < 0.01; at least 120 cells / condition were evaluated, collected from three different donors (E). Correlation coefficients are significantly different, according to Fisher-Z-test, ****P<0.001 (F). (G) Analysis of podosome core area by ImageJ macro in cells simultaneously depleted (dKD) for both Myo1e and Myo1f. Unpaired t-test, ****P < 0.0001; 18 cells / condition, from 3 different donors. (H-K) Podosome reformation assay to analyze the ability of podosomes to reform over time (15, 30, 90 min) after PP2-mediated dissolution, in macrophages simultaneously depleted for both Myo1e and Myo1f (dKD), with DMSO treated cells as controls. F-actin of podosome cores was stained with Alexa488-coupled phalloidin, and podosomes are visible as μm-sized black dots within cells. Red arrows in DMSO controls indicate large F-actin aggregates forming after dKD. Yellow boxes indicate regions of interest shown enlarged as insets in lower right corner (H); Scale bars = 10 μm, 1 μm for insets. ImageJ macro-based evaluation of the number of podosomes per cells (I), cell area (J) and related podosome density (K). Multiple unpaired t-test; ****P < 0.0001, *P < 0.05. At least 120 cells / condition were evaluated, collected from 3 different donors.

To further probe the potential influence of Myo1e and Myo1f on podosome genesis, we performed podosome reformation assays, which involve the complete disassembly of podosomes by the Src tyrosine kinase inhibitor PP2, followed by subsequent washout and podosome reassembly (Cervero et al., 2013). Because single knockdowns did not show clear effects on podosome reformation (not shown), we tested double knockdowns of Myo1e and Myo1f together. At time point 0 (no addition of PP2), human macrophages depleted for both Myo1e and Myo1f (Suppl. Fig. 4B) showed a reduction of podosome numbers by 35 % (296 ± 66 podosomes/cell for controls, and 191 ± 61 podosomes/cell for double depletions) (Fig. 3D), which was accompanied by a similarly reduced podosome covered area, compared to controls (Fig. 3E). Plotting cell data based on the ratio of the podosome-covered area to cell area further showed that depletion of Myo1e/f led to a reduced correlation, indicating a decreased coupling between podosome-covered area and cell size (r = 0.88 for control cells vs. r = 0.69 for Myo1e/f depleted cells) (Fig. 3F). Similarly to individual knockdowns (Fig. 3A), double knockdown cells showed a similar increase in podosome core size (Fig. 3G). In addition, we observed increased formation of larger actin structures, possibly corresponding to podosome clusters or enlarged podosomes, in double depleted cells (Fig. 3H). At different time points after PP2 washout (15 min, 30 min, 90 min), Myo1e/f depleted cells showed reduced numbers of podosomes, compared to controls (Fig. 3I). However, as double-depleted cells also showed reduced cell size, although not to a significant degree, during the wash out (Fig. 3J), the density of podosomes (number of podosomes / 100 μm²) did not change significantly during the reformation (Fig. 3K).

Collectively, these data show that both Myo1e and Myo1f affect the size, lifetime, and lateral mobility of podosomes, but do not appear to play a direct role in podosome assembly. The single myosin KD studies suggest that Myo1e may play a more dominant role in regulating podosome parameters than Myo1f. However, both myosins influence the number of podosomes formed per cell.

2.4. Myo1e and Myo1f regulate podosomes and 2D/3D cell migration in murine macrophages

Having established the influence of Myo1e and Myo1f knockdown on podosomes in human macrophages, we next wanted to verify and expand on key findings using murine knockout macrophages. Similar to the results in knockdown human macrophages, double knockout of Myo1e/f led to increased size of podosomes (Fig. 4A,B), which was associated with an increase in podosomal F-actin, as determined by fluorescence intensity measurements (Fig. 4C). Analysis of single myosin 1 knockout macrophages showed a similar trend towards larger podosomes and increased F-actin content to a lesser extent than loss of both myosins (Fig. 4B,C). These data suggest that Myo1e and Myo1f may be able to substitute for loss of the other, but both individually contribute to regulating F-actin at podosome cores.

Fig. 4.

KO of Myo1e and Myo1f results in formation of podosome aggregates and increases matrix degradation. A, BMDMs derived from Myo1e/f dKO mice contain podosome clumps and aggregates. Macrophages were fixed and stained with phalloidin and anti-vinculin. (B,C,D) Graphs comparing (B) area and (C) fluorescence intensity of F-actin podosome cores of WT, eKO, fKO, and dKO BMDM podosomes (n = 100–250 podosomes measured in 10–20 cells). Data is derived from one representative experiment conducted in triplicate; each graph displays mean+SD and individual data points. * p < 0.05, **** p < 0.0001, compared to the WT by one-way ANOVA. B, The average area of podosome cores is increased in the Myo1f KO and Myo1e/f dKO BMDMs. C, Podosome mean F-actin intensity is higher in BMDM lacking Myo1e, Myo1f, or both Myo1e and Myo1f than in the WT cells. (D-G), gelatin matrix degradation by BMDMs. D, areas of matrix degradation are associated with F-actin-containing podosomes. E-F, Fraction of BMDMs containing podosome rosettes (E) or associated with the areas of matrix degradation (F) is higher in the dKO BMDM cultures. (N = 138 WT cells and 219 dKO cells pooled from 4 independent experiments). G, the area of matrix degradation associated with individual cells is higher in the dKO cells. * P < 0.05 compared to the WT; **** P < 0.0001 compared to the WT.

To test the degradative ability of the murine podosomes, we utilized a fluorescent gelatin degradation assay. Murine-derived macrophages were plated on fluorescent-gelatin coated coverslips and activated with IL-4 to form podosomes (Cougoule et al., 2012). Podosome-mediated degradation was confirmed by counterstaining with phalloidin (Fig. 4D). Compared to control macrophages, dKO cells displayed podosome rosettes more frequently with a higher incidence of matrix degradation (Fig. 4E,F). Quantification of degraded area per cell area further showed that dKO macrophages exhibited increased matrix degradation (Fig. 4G).

Since podosome-mediated matrix degradation is required for macrophage invasion in a 3D context (Van Goethem et al., 2010), we tested the ability of Myo 1e/f dKO cells to migrate towards the chemoattractants M-CSF/FBS in 3D Matrigel. Surprisingly, in spite of the increased matrix degradation in 2D (Fig. 4), the number of dKO cells that migrated through Matrigel in 3D (Fig. 5A), as well as the maximum migration depth (Fig. 5B), were significantly reduced compared to the WT cells. This finding may reflect the fact that macrophage 3D invasion is a complex process that relies not only on matrix degradation, but also on cell motility and adhesion capacity (Wiesner et al., 2014), which may also be altered by the Myo1e/f dKO.

Fig. 5.

dKO BMDM interferes with BMDM migration. (A, B) The ability of dKO BMDMs to migrate into 3D Matrigel inserts was reduced compared to the WT cells. (A) The percentage of migration (dot) of BMDMs from 5 WT and 6 dKO mice is represented with the mean (bar) and SD. (B) The distribution of the distances migrated in Matrigel by individual cells obtained from 5 WT and 6 dKO mice is shown. (C) Tracks of WT and dKO macrophages randomly migrating on fibronectin-coated glass substrate over 20 hrs. (D) Quantification of tracks in (C) reporting cell velocity. Representative results of 4 independent experiments, n = 42 WT and 26 dKO cells. (E). Representative images (phase contrast) of WT and dKO macrophages migrating into a gap created by scratching. Scale bar = 100 μm. (F) Quantification of gap area closure over 6 h (N = 1 experiment, with data points representing 4 different FOVs). * P < 0.05 compared to the WT; ** P < 0.01 compared to the WT;**** P < 0.0001 compared to the WT.

We thus additionally tested the effects of Myo1e/f on macrophage migration in 2D. To assess random migration in 2D, we tracked cell trajectories over the course of 20 h (Fig. 5C). dKO macrophages were significantly less motile and appeared to have difficulty detaching from the substrate or sufficiently polarizing to form a leading edge (Supp. Video 1). Tracking cell movement showed that the velocity of dKO cell movement was significantly lower than that of the WT cells (Fig. 5D). Finally, a scratch assay was used to measure the rate of cell migration into a scratch in the densely plated layer of cells. The rate of gap closure was significantly lower in the dKO cultures compared to WT cultures (Fig. 5 E–F). Overall, we find that macrophage migration in both 3D and 2D settings is significantly reduced in the absence of Myo1e/f.

Supplementary material related to this article can be found online at doi:10.1016/j.ejcb.2025.151514.

3. Discussion

Class I myosins, which contain both actin-binding motor domains and membrane-binding tail domains, localize to actin-membrane contact sites and have been linked to vesicular trafficking, membrane deformation during clathrin-dependent endocytosis, cell migration, and regulation of cell adhesion (McIntosh and Ostap, 2016). Positively charged tail homology domains of these myosins (TH1 domains) promote association with phospholipids, in some cases exhibiting very specific lipid-binding preferences (Myo1c) (Hokanson and Ostap, 2006) and in others having broader phospholipid binding specificity (Myo1e) (Feeser et al., 2010). A subclass of class I myosins is formed by the so-called long-tailed class I myosins, whose tails include additional domains, such as the positively charged and proline-rich TH2 domain and the SH3 domain that can bind proline-rich domains (PRDs). These myosins are found in a variety of organisms, including yeast, amoeboid protozoans, and vertebrates, with mammalian genomes encoding two long-tailed myosins, Myo1e and Myo1f (Kim and Flavell, 2008).

We have previously shown that Myo1e localizes to the ventral podosome surface in src-transformed fibroblasts, RAW mouse macrophages, and A375 melanoma cells (Ouderkirk and Krendel, 2014; Zhang et al., 2019) and that both Myo1e and Myo1f can be identified in podosome-enriched macrophage fractions by proteomic analysis (Cervero et al., 2012). We have also previously observed both Myo1e and Myo1f in phagosome-associated podosomes (Barger et al., 2019).

In this paper, we have dissected Myo1e and Myo1f localization to macrophage podosomes in more detail and identified defects in podosome dynamics and macrophage migration that are associated with their depletion. We find that both Myo1e and Myo1f, but not other class I myosins, localize to the region of podosomes that is close to the ventral plasma membrane of cells. Previously identified podosome components that share a similar localization are the matrix metalloproteinase MT1-MMP (El Azzouzi et al., 2016), DNase X (Pal et al., 2021), and, possibly, the hyaluronan receptor CD44 (Chabadel et al., 2007). Cdc42 (Linder et al., 1999) is another likely candidate for base localization, considering that Cdc42 in podosomes is activated by membrane phospholipid binding to its GEFs (Qi and Yu, 2025) and can, in general, associate with membranes through prenyl- or palmityl anchors (Wirth and Ponimaskin, 2022). Of note, the podosome component WASp (Linder et al., 1999) co-localizes with Myo1e at the base of podosomes (Suppl. Fig. 3), likely through its association with PIP2, which is required for its activation (Higgs and Pollard, 2000), and Myo1f has been shown to interact with WASP by proximity biotinylation in myeloid cells (Arden et al., 2025). Most of these proteins are attached to at least one leaflet of the ventral plasma membrane, either through a transmembrane domain or through a GPI anchor, in the case of DNaseX, whereas Myo1e/Myo1f interact with the membrane phospholipids via their tail domains.

In line with this restricted localization beneath the podosome core and its specific components, we propose to label this part of the podosome architecture as the podosome “base”. The podosome base would thus join the current trifecta of podosome core, ring and cap as an additional substructure with a defined molecular composition (Fig. 6). Currently the concept of the podosome base, as described here, relies primarily on the common localization of the molecular components. However, we propose that the restricted localization of Myo1e/f directs their unique function at podosomes. These motor proteins, which bind plasma membrane via their tails and move towards the membranefacing barbed ends of actin filaments using their motor domains, may serve as a flexible/sliding linker between the plasma membrane and actin filaments at the ventral podosome surface. Myo1e/f could thus maintain the proximity of actin filament barbed ends to the membrane, as well as the appropriate orientation/angle of actin filaments to support podosome dynamics and protrusive activity (Fig. 6, discussed in detail later).

Fig. 6.

Schematic representation of the podosome base and its components. The podosome base represents the lower region of the podosome core whose components are associated with the plasma membrane (PM). In addition to Myo1e and Myo1f, which bind the PM through their lipid binding domains, known components are transmembrane proteins such as the matrix metalloproteinase MT1-MMP, and the cell adhesion molecule CD44, the GPI-anchored DNase X, and the actin nucleation promoting factor WASp, which requires PIP 2 binding for its activation. A further likely, but formally unproven, component is the small GTPase CDC42, an activator of WASp, that can associate with the PM through PM-bound GTPase exchange factors (GEFs) or membrane anchors such as prenyl- or palmityl anchors. For details, see Discussion. Modified from (Weber et al., 2022).

Similarly to the previous localization studies (Barger et al., 2019; Ouderkirk and Krendel, 2014; Zhang et al., 2019), Myo1e/f localization to macrophage podosomes depended on the TH2 domain. None of the short-tailed class I myosins we tested, which lack TH2 and SH3 domains, localized to podosomes, further highlighting the importance of the myosin C-terminal tail region for podosome localization. Both TH1 and TH2 domains of Myo1e and Myo1f are highly positively charged, which presumably promotes their interactions with membrane phospholipids. Interestingly, the TH1 domain does not appear to be required for recruitment of Myo1e to podosomes, while it contributes to the localization of Myo1f to the podosome core. In addition to carrying positive charges, the Myo1e/f TH2 domains may be able to interact with SH3 domain-containing proteins via the proline-rich motifs in the TH2 domain and may also bind actin filaments in an ATP-independent fashion similarly to the long-tailed amoeboid myosins I (Rosenfeld and Rener, 1994). Whether these activities further contribute to their role in podosome localization or dynamics remains to be determined.

Simultaneous knockout of Myo1e/f in mouse macrophages or knockdown of these myosins in human macrophages resulted in formation of enlarged podosomes or podosome aggregates concomitant with a reduction in the number of podosomes per cell. Unlike the overexpression of the TH1-TH2 domains of Myo1e, which blocks macrophage podosome formation and suppresses gelatin degradation by cancer cells (Zhang et al., 2019), Myo1e/f depletion did not result in the loss of podosomes or their substrate-degrading activity. It is possible that the overexpression of the podosome-localized TH1-TH2 fragment may act in a dominant-negative fashion by disrupting additional protein-protein and protein-lipid interactions within nascent podosomes and further impairing other cellular functions. In contrast, Myo1e/f depletion does not show such a global effect on the overall podosome assembly but instead disrupts specific steps in podosome dynamics and reorganization.

We also find that individual knockdowns of Myo1e or Myo1f in human macrophages lead to a reduction in podosome lifetime and lateral mobility. Podosome reformation experiments showed that the decreased mean lifetime (Fig. 1B) is not associated with the overall reduced de novo formation of podosomes. Myo1e/f activity may contribute to higher podosome turnover rates and organization via several possible mechanisms. We have previously shown that macrophages lacking Myo1e/f exhibit reduced membrane tension (Barger et al., 2019). Artificially increasing membrane tension by osmotic swelling causes macrophage podosomes to disassemble, while artificially decreasing membrane tension causes podosomes to cluster (Rafiq et al., 2019). This could explain our observations in Myo1e/f dKO or dKD cells, where podosomes are enlarged or form clusters. These findings are also reminiscent of our observations on the effects of Myo1e/f KO on phagocytic actin waves/phagocytic podosomes, where the lack of Myo1e/f was associated with excessive branched actin assembly and the replacement of individual phagocytic podosomes with large actin aggregates (Barger et al., 2019). In addition, by mediating actin-membrane interactions, Myo1e/f may also regulate the angle of actin filament-membrane contacts or modulate bending or compressive forces on actin filaments (Jasnin et al., 2022), which in turn would affect the density of branched actin networks (Li et al., 2022; Risca et al., 2012) or even promote actin debranching (Pernier et al., 2020). Indeed, a recent study has identified a role for the motor activity of a class I myosin in promoting formation of sparser/less densely branched actin networks around beads coated with Arp2/3-activating proteins (Xu et al., 2024). Actin-based motility of Myo1e/f, possibly coupled with PIP-dependent anchoring to the plasma membrane, could also contribute to the observed roles of Myo1e/f in lateral mobility of podosomes by moving actin networks relative to the membrane or moving membrane PIPs relative to the underlying actin structures.

We have also found that Myo1e/f dKO macrophages exhibit slower cell migration in both 2D and 3D settings. This finding agrees with the previous studies reporting defects in cell migration with Myo1e/f loss in various cell types. For example, cancer cells lacking Myo1e migrate more slowly in scratch-wound assays and transwell assays (Garone et al., 2023; Tanimura et al., 2016). Defects in immune cell migration have been observed in B cells and neutrophils from Myo1e KO mice (Giron-Perez et al., 2020; Vadillo et al., 2019), intraepithelial T-lymphocytes from Myo1f KO mice (Martinez-Vargas et al., 2023), and neutrophils from Myo1f KO mice (Kim et al., 2006; Salvermoser et al., 2018), as well as macrophages in which interactions between Myo1e and LSP1 were disrupted (Scharinger et al., 2021). The mechanisms underlying these defects vary suggesting cell type-specific myo1e/f functions that may correlate with differences in individual myosin expression levels (Maravillas-Montero, 2012). Here, we provide a new, comprehensive examination of the effects of myo1e/f depletion on both cell migration and matrix degradation in macrophages.

The fact that Myo1e/f dKO macrophages are more degradative could stem from their reduced motility, i.e. because KO macrophages are moving less, they are degrading more of the area beneath them. Given that recruitment of podosomes to the leading edge plays an important role in macrophage symmetry breaking and migration initiation (Cervero et al., 2018), the disruption of podosome organization and dynamics in Myo1e/f knockdown cells could explain the observed reduction in migration of these cells. Whether the observed defects in podosome dynamics represent the sole reason for reduced cell migration upon loss of Myo1e/f requires further study since Myo1e/f may also promote lamellipodial protrusion (Tanimura et al., 2016) or modulate intracellular signaling (Giron-Perez et al., 2020; Heim et al., 2017). Overall, our findings suggest that Myo1e/f are crucial not only for podosome dynamics but also for coordinating the balance between adhesion, degradation, and motility required for efficient macrophage migration.

4. Materials and methods

4.1. Expression constructs

EGFP-tagged and mEmerald-tagged Myo1e/f constructs as well as deletion constructs have been previously described (Barger et al., 2019; Bi et al., 2013). LifeAct-TagRFP was purchased from Ibidi (Martinsried, Germany).

4.2. Antibodies and staining reagents

The following commercial antibodies were used: mouse monoclonal anti-vinculin (Sigma, H-VIN), mouse monoclonal anti-Myo1f (Santa Cruz, sc-376534), rabbit anti-Myo1e (RRID:AB_2909514) (Skowron et al., 1998), rabbit polyclonal anti-GAPDH (Proteintech, 10494–1-AP).

Secondary antibodies labeled with Alexa Fluor-488 and Alexa Fluor-568 and fluorescently-labeled phalloidin were purchased from Invitrogen. HRP-linked F(ab’)2 fragment donkey-anti rabbit (NA9340v) and HRP-linked sheep-anti mouse IgG (H + L) (NA931v) were purchased from GE Healthcare.

4.3. Macrophage preparation

4.3.1. Mouse macrophages

All experiments involving mouse bone marrow isolation were performed according to the animal protocol approved by the Upstate Medical University IACUC. Myo1e and Myo1f KO mice were maintained on the C57Bl/6 background and all experiments used age- and sex-matched bone marrow preparations. BMDM isolation and differentiation has been previously described (Barger et al., 2022).

4.3.2. Human macrophages

Peripheral blood monocytes were isolated from buffy coats (obtained from University Medical Center Hamburg-Eppendorf, Hamburg, Germany) as described previously (Wiesner et al., 2010). The buffy coats were purified using Lymphocyte Separation Medium (LSM), a Ficoll-based medium. Then, the monocytes were positively selected for CD14. Monocytes were differentiated in culture for at least seven days, under addition of 20 % human autologous serum.

4.4. Cell Culture

BMDMs were maintained in DMEM supplemented with 10 % FBS, 1 % antibiotic/antimycotic (Life Technologies), and 20 ng/ml recombinant murine M-CSF (Biolegend; 576404). Human macrophages were cultured in RPMI medium supplemented with 20 % autologous serum, penicillin, and streptomycin at 37 °C, 5 % CO₂.

4.5. Human macrophage transfection

For transfection experiments, differentiated macrophages, at days 10–14 of culture, were transiently transfected using the Neon^® Transfection System (Thermo Fisher Scientific, Waltham, MA), an electroporation-based system, with standard settings (1000 V, 40 ms, 2 pulses) and a concentration of [100 nM] for siRNA and 0.5 μg per 1 × 10⁵ cells for plasmids.

4.6. 3D cell migration

Matrigel was slowly thawed on ice, deposited into the Transwell inserts (8μm-pore size polyester membrane, Corning 353097) in a 24-well plate (100 μl of 10 mg/ml Matrigel per insert), and allowed to polymerize at 37°C. Matrices were rehydrated overnight with RPMI without FBS. Differentiated BMDM were serum-starved for 3 hrs and added to the top of the Matrigel plug in RPMI medium supplemented with 2 % FBS and 20 ng/ml mouse M-CSF (4 × 10⁴–5 × 10⁴ cells per insert). The bottom well was filled with RPMI medium containing 10 % FBS and 20 ng/ml mouse M-CSF, and cells were allowed to migrate for 2 days (approximately 48 hrs). Brightfield images of cells at the surface and within the Matrigel were collected as Z-stacks with 30 μm intervals using a Nikon Eclipse TE2000-E2 multimode TIRF microscope equipped with PRIME-95B cMOS camera (Photometrics). Quantification was performed as in (Van Goethem et al., 2010). Specifically, % migrated cells (Fig. 5A) was determined by counting all cells in a region of interest in a Z-stack (all slices) while counting cells in the topmost slice (0) as non-migrated; the ratio of cells that have migrated to lower slices to the total cell number was then calculated. Two independent sets of experiments were performed, using BMDM from a total of 5 individual WT mice and 6 dKO mice. We used the mean percentage of migrated cells (relative to total cell number) from 2 to 3 inserts per mouse as technical replicates.

4.7. Immunostaining

4.7.1. Mouse macrophages

Cells were fixed using fresh 4 % paraformaldehyde/PBS for 15 min. After washing away fixative, cells were permeabilized in 0.1 % Triton X-100/PBS for 3–5 min. Cells were blocked for 30 min at room temperature with 5 % normal goat serum/3 % BSA dissolved in PBS with 0.05 % Tween-20 and 0.1 M glycine. Cells were exposed to primary antibodies at the appropriate dilutions for 1 h at room temperature. Cells were then washed three times for 5 min with PBS. Secondary antibodies and fluorescent phalloidin were then applied for 30 min at room temperature. Cells were then washed again for three times, 5 min each before mounting with Prolong Diamond Antifade Mountant (Invitrogen). NucBlue Fixed Cell ReadyProbes Reagent (Invitrogen) was used according to the manufacturer’s instructions.

4.7.2. Human macrophages

2-week old macrophages were seeded on 12-mm glass coverslips (~7 × 10⁴ cells per coverslip) and incubated overnight at 37 °C, 5 % CO₂, and 90 % humidity. To fix, the coverslips were dipped for 1–3 s in glacial methanol (−20 °C) then placed in 3.7 % formaldehyde for 10 min. The coverslips were then placed in PBS and stored at 4°C overnight. The cells were permeabilized for 5 min using 0.5 % Triton X-100, followed by three washes with PBS containing 0.05 % of Triton X-100. Then, the samples were incubated for 1 hr in blocking solution (2 % BSA + 5 % goat serum in PBS), followed again by three washes. They were then incubated with primary antibody solution for 2 hrs, followed by three washes. Finally, they were incubated for 1 hr with secondary antibody solution, which included a fluorescently-labelled phalloidin. After three PBS washes and a dip in dH₂O, the coverslips were mounted on glass slides using Mowiol heated to 37°C. Primary antibodies were used at the following concentrations: anti-Myo1e (1:100), anti-Myo1f (1:100).

4.8. Western blotting

Cells were lysed with RIPA buffer (150 mM NaCl, 1 % Triton X-100, 0.5 % sodium deoxycholate, 0.1 % SDS, 50 mM Tris–HCl (pH 8.0)) and vortexed. Equal amounts of protein samples were then mixed with 4×Laemmli sample loading buffer and examined by standard immunoblotting procedure using NuPAGE 4–12 % Bis-Tris gels (Invitrogen, Thermo Fisher Scientific), iBlot2 dry blotting system (Thermo Fisher Scientific) and the above-mentioned primary and HRP-conjugated antibodies, as indicated. When needed, nitrocellulose membranes were mild-stripped by extensive washing with buffer (200 mM glycine, 3.5 mM SDS, 1 % Tween-20, (pH 2.2)) before reblocking and reprobing membranes with primary and secondary HRP-conjugated antibodies. Protein bands were visualized by using Super Signal Pico or Femto kit (Pierce) and X Omat AR films (Kodak). Results were scanned and protein band intensities quantified with Fiji distribution of ImageJ (NIH, Bethesda, MD).

4.9. Gelatin degradation assay

Gelatin was purchased from Sigma while Oregon-Green-488-labeled gelatin was purchased from Thermo Fisher, and gelatin-coated coverslips were prepared as previously described (Diaz, 2013). WT and dKO BMDM were plated on fluorescent gelatin and treated with 20 ng/ml IL-4 to induce podosome formation and ECM degradation (Cougoule et al., 2012). After 24 h, cells were fixed using fresh 4 % paraformaldehyde/PBS for 15 min. After washing away fixative, cells were permeabilized in 0.1 % Triton X-100/PBS for 3 min and stained with Alexa Fluor-568 phalloidin for 30 min. During washes, cells were counterstained with NucBlue and carefully mounted using Prolong Diamond Antifade Mountant. All plating and staining steps were performed under reduced lighting.

4.10. Microscopy

4.10.1. Mouse macrophages

Live cell imaging to examine BMDM motility was performed on a True MultiColor Laser TIRF Leica AM TIRF MC system equipped with an Andor DU-885K-CSO-#VP camera and a 10X objective (Leica HCX PL FLUOTAR 10x/0.30NA PH1). Environmental conditions were maintained by an Okolab temperatureand CO₂-control system. Spinning disk images were collected using a PerkinElmer UltraView VoX Spinning Disc Confocal system mounted on a Nikon Eclipse Ti-E microscope equipped with a Hamamatsu C9100–50 EMCCD camera, a 100X (1.4 N. A.) PlanApo objective, and controlled by the Volocity software.

4.10.2. Human macrophages

Imaging of Myo1e/f deletion constructs and Myosins 1 a/b/c/d/g was performed in a live cell set-up using a Visitron spinning disk microscope equipped as follows: Nikon Eclipse TiE microscope; 60X Apo TIRF Oil objective with NA 1.49; Photometrics Prime 95B camera; Yokogawa CSU W-1 spinning disk unit (50 μm pinhole size); VisiView software. The same microscope was also used to image fixed samples stained for endogenous Myosins I e/f. Images from the PP2 reformation assay were acquired using a confocal laser scanning microscope (DMI 6000 with a TCS SP5 AOBS confocal point scanner; Leica) equipped with an oil-immersion HCX PL APO 63X NA1.4–0.6 objective. Where needed, Z-stacks were acquired at 0.20 μm intervals.

4.11. Image analysis

4.11.1. Mouse macrophages

Fiji (Schindelin et al., 2012) was used to process and analyze all microscopy images. To make Z-projections, Z-stacks that displayed the podosomes were flattened using average intensity projections. The same stacks were used for all channels in an image. The contrast on the images were then enhanced to 0.1 % and normalized. To generate intensity profiles, a podosome was selected, and line scans of fluorescent intensity were generated using a linear selection and the “plot profile” function in Fiji was used to obtain intensity values along the line. For podosome core size measurements in Fiji, regions of podosome clusters were initially cropped and duplicated. On the duplicated image, the Threshold function was used to manually set the intensity threshold by the Default method to isolate podosomes using the original phalloidin-stained image for comparison (Supp Figure 5). Using the thresholded masked image, the area and intensity of the podosome actin cores was then measured using the Fiji Measure function. Masked areas of 1 pixel were not considered. Podosome rosettes were excluded from this analysis. The areas of gelatin degradation were similarly measured, using the Threshold function for demarcation and the Fiji Measure function for quantification. Scratch migration analysis was performed using the MRI Wound Healing plugin, and tracking random motility was conducted using the Manual Tracking and Chemotaxis Tool plugins.

4.11.2. Human macrophages

Detailed analysis of several podosome profiles at different Z-positions was performed using the Fiji-based macro code termed ‘Poji’ (Herzog et al., 2020). PP2 reformation assay and measurements of number of podosome per cell and podosome core areas were performed as previously described (Cervero et al., 2013). Tracking of podosome lifetime and podosome lateral mobility was performed using the Fiji-based Trackmate plugin (Ershov et al., 2022). All podosomes detected in the first frame are recorded as individual ROIs with XY coordinates and then tracked over time frame-by-frame. At the end of the process, tracks connecting individual ROIs are created and lifetime or lateral mobility of tracks can be extrapolated respectively by measuring the difference in time between the last (podosome dissolution) and first (podosome formation) timepoints of a track, or by calculating the path length covered by single ROIs. F-actin aggregates, defined as size > 4 μm² and with a circularity < 0.5, were thus not considered in the analysis.

4.12. Statistical analysis

Comparisons between WT and dKO BMDM were carried out using an unpaired two-tailed t-test for independent samples, with differences between genotypes considered statistically significant at p < 0.05. For multiple comparisons (WT, eKO, fKO, dKO), data were analyzed using a one-way ANOVA with Tukey’s post-hoc test, with statistical significance set at p-value < 0.05. For primary human macrophages, comparisons between controls and individual siRNAs were performed using ordinary one-way ANOVA with Dunnett’s multiple comparisons test, whereas comparisons between controls and Myo1e/1f double knockdown were performed using unpaired two-tailed t-test. For the podosome reformation assay, multiple unpaired t-tests were used. For quantification of knockdown efficiency by western blot, the intensities of specific proteins were first normalized for respective GAPDH intensities, then, after setting the control samples to 100 %, the protein expression of Myo1e/1f (single and double knockdowns) was calculated relative to the control, and the resulting data analyzed by one sample t and Wilcoxon test. In all cases the statistical significance was set at p-value < 0.05. Statistical analyses and graphing were performed using GraphPad Prism software.

Appendix A. Supporting information

Supplementary data associated with this article can be found in the online version at doi:10.1016/j.ejcb.2025.151514.

Data availability

Data has been uploaded to Fig Share at doi 10.6084/m9.figshare.30058336.