Macrophage mesenchymal migration requires podosome stabilization by filamin A

Abstract

Filamin A (FLNa) is a cross-linker of actin filaments and serves as a scaffold protein mostly involved in the regulation of actin polymerization. It is distributed ubiquitously, and null mutations have strong consequences on embryonic development in humans, with organ defects which suggest deficiencies in cell migration. We have reported previously that macrophages, the archetypal migratory cells, use the protease- and podosome-dependent mesenchymal migration mode in dense three-dimensional environments, whereas they use the protease- and podosome-independent amoeboid mode in more porous matrices. Because FLNa has been shown to localize to podosomes, we hypothesized that the defects seen in patients carrying FLNa mutations could be related to the capacity of certain cell types to form podosomes. Using strategies based on FLNa knock-out, knockdown, and rescue, we show that FLNa (i) is involved in podosome stability and their organization as rosettes and three-dimensional podosomes, (ii) regulates the proteolysis of the matrix mediated by podosomes in macrophages, (iii) is required for podosome rosette formation triggered by Hck, and (iv) is necessary for mesenchymal migration but dispensable for amoeboid migration. These new functions assigned to FLNa, particularly its role in mesenchymal migration, could be directly related to the defects in cell migration described during the embryonic development in FLNa-defective patients.

FIGURE 1.

FLNa^−/− BMDMs have decreased abilities to perform three-dimensional mesenchymal migration, to form podosome rosettes, and to degrade gelatin FITC. BMDMs from WT or FLNa^−/− mice were seeded on thick layers of matrices of Matrigel (A) or fibrillar collagen I (B), and the percentage of cells migrating into the matrices was quantified (mean ± S.D. of four independent experiments). Pictures of WT (a′ and b′) and FLNa^−/− cells (a″ and b″) migrating, respectively, in Matrigel (a′ and a″) and fibrillar collagen I (b′ and b″) (z is the depth value of cells in the focal plan marked by an arrow). C, FLNa^−/− BMDMs embedded into Matrigel are defective in forming cell protrusions (see supplemental Movies 1 and 2). Scale bar, 10 μm. D, BMDMs from WT or FLNa^−/− mice were seeded on fibronectin-coated coverslips for 16 h and stained for F-actin. The arrow points to a typical podosome rosette. E, quantification of cells making podosome rosettes in conditions of D (arrow) (mean ± S.D. of four independent experiments). F, macrophages seeded on coverslips coated with gelatin FITC for 16 h were then stained for F-actin. The arrow points to a non-degrading cell, and the dashed lines show the degraded areas. G, quantification of cells degrading the matrix, in conditions of F (mean ± S.D. of four independent experiments). H and I, quantification of the surface of degraded gelatin FITC per cell surface, expressed as the mean area degraded per cell (μm²) in H, and the percentage of degraded area per cell surface in I, in condition of F (mean ± S.D. of four independent experiments, at least 25 cells were quantified per condition). ***, p < 0.001.

FIGURE 2.

FLNa is localized at podosomes and accumulates with integrin and Hck at podosome rosette of human MDMs. MDMs plated on fibrinogen were stained for microscopy observation for FLNa and F-actin (A), or CD11b and F-actin (B), or Hck and F-actin (C); insets are magnification of areas depicted by the white squares (scale bar, 10 μm). D, normalized fluorescence intensity profiles along the white dotted line (a′, b′, c′) in A, B, and C, respectively. hFLNa was observed at individual podosomes forming a ring around the F-actin core (A, arrowhead) and at podosome rosettes (A, arrow), where CD11b (B, arrow) and Hck (C, arrow) also accumulated.

FIGURE 3.

Macrophages with inhibited FLNa expression have a defect in podosome and podosome rosettes formation. A, RAW264.7 macrophages (Control) or RAW264.7 cells stably expressing shRNA against mFLNa, and those rescued with stable expression of hFLNa were all seeded on vitronectin-coated coverslips and treated with IFN-γ to enhance the formation of podosomes, fixed and stained for F-actin and FLNa. Cells counted as forming podosomes and podosome rosettes shown by an arrow presented numerous actin dots (≥5) and rosettes (≥1), respectively (scale bar, 10 μm). B, quantification of cells with podosomes or podosome rosettes (mean ± S.D. of three independent experiments). C, Western blot against hFLNa, mFLNa, and actin and quantification of mFLNa in three experiments. AU, arbitrary unit. D, RAW264.7 macrophages control or transfected with mFLNA shRNA alone or together with a vector expressing hFLNA were transiently transfected with mCherry-Lifeact to reveal F-actin. The lifespans of podosomes were then evaluated using time-lapse microscopy and are plotted for each cell type (mean ± S.D. of three independent experiments, five to 10 podosomes analyzed per cell in at least three cells per experiment). ***, p < 0.001.

FIGURE 4.

The formation of podosome rosettes triggered by expression of Hck^ca in fibroblasts requires FLNa. MEF-3T3 fibroblasts and MEF-3T3 fibroblasts expressing human p59Hck^ca/p61Hck^ca were stained for mFLNa and F-actin and observed by confocal microscopy. A, control cells did not form podosome rosettes, whereas the cells expressing p59 Hck^ca/p61Hck^ca showed FLNa accumulation at podosome rosettes (arrowheads). B, MEF-3T3 fibroblasts expressing p59Hck^ca/p61Hck^ca-GFP and stained for FLNa and F-actin showed accumulation of Hck and FLNa and F-actin at podosome rosettes (arrowheads). Fluorescence intensity profiles along the white dashed line are shown. C, expression of p59Hck^ca/p61Hck^ca-GFP in NIH3T3 stably expressing shRNA against luciferase (control) or against mFLNa were stained for F-actin (scale bar, 10 μm). D, quantification of p59Hck^ca/p61Hck^ca-GFP expressing cells forming podosome rosettes (mean ± S.D. of three independent experiments).

FIGURE 5.

Src kinase activity is required for podosome formation and for filamin A localization to podosomes. MDMs plated on fibrinogen (A) and treated with SU6656 (B) were stained for hFLNa and F-actin, before acquisition of confocal micrograph series (a and b, respectively) (z-step = 0.1 μm) (scale bar, 10 μm). a′ and b′ show the average of the F-actin and FLNa fluorescence staining of at least 100 podosomes from control and SU6656-treated cells (scale bar = 1 μm). a″ and b″ show fluorescence intensity profiles of the averaged podosomes along the white dashed line in (a′) and (b′), respectively. C, quantification of human MDMs with podosomes or podosome rosettes when seeded on coverslips that were either uncoated or coated with fibrinogen (Fg) and treated with the Src inhibitor SU6656 (mean ± S.D. of three independent experiments). D, RAW264.7 macrophages or stably expressing a shRNA against mouse Hck were (or not) transiently transfected with a human Hck-GFP coding vector (to rescue Hck) and were transfected transiently with LifeAct-mCherry coding vector to stain F-actin. Podosome lifespans, measured by time-lapse microscopy, are plotted for each cells type (mean ± S.D. of three independent experiments, five to 10 podosomes analyzed per cell in at least three cells per experiment).