Podosomes in migrating microglia: components and matrix degradation

Abstract

Background: To perform their functions during development and after central nervous system injury, the brain's immune cells (microglia) must migrate through dense neuropil and extracellular matrix (ECM), but it is not known how they degrade the ECM. In several cancer cell lines and peripheral cells, small multi-molecular complexes (invadopodia in cancer cells, podosomes in nontumor cells) can both adhere to and dissolve the ECM. Podosomes are tiny multi-molecular structures (0.4 to 1 μm) with a core, rich in F-actin and its regulatory molecules, surrounded by a ring containing adhesion and structural proteins.

Methods: Using rat microglia, we performed several functional assays: live cell imaging for chemokinesis, degradation of the ECM component, fibronectin, and chemotactic invasion through Matrigel™, a basement membrane type of ECM. Fluorescent markers were used with high-resolution microscopy to identify podosomes and their components.

Results: The fan-shaped lamella at the leading edge of migrating microglia contained a large F-actin-rich superstructure composed of many tiny (<1 μm) punctae that were adjacent to the substrate, as expected for cell-matrix contact points. This superstructure (which we call a podonut) was restricted to cells with lamellae, and conversely almost every lamella contained a podonut. Each podonut comprised hundreds of podosomes, which could also be seen individually adjacent to the podonut. Microglial podosomes contained hallmark components of these structures previously seen in several cell types: the plaque protein talin in the ring, and F-actin and actin-related protein (Arp) 2 in the core. In microglia, podosomes were also enriched in phosphotyrosine residues and three tyrosine-kinase-regulated proteins: tyrosine kinase substrate with five Src homology 3 domains (Tks5), phosphorylated caveolin-1, and Nox1 (nicotinamide adenine dinucleotide phosphate oxidase 1). When microglia expressed podonuts, they were able to degrade the ECM components, fibronectin, and Matrigel™.

Conclusion: The discovery of functional podosomes in microglia has broad implications, because migration of these innate immune cells is crucial in the developing brain, after damage, and in disease states involving inflammation and matrix remodeling. Based on the roles of invadosomes in peripheral tissues, we propose that microglia use these complex structures to adhere to and degrade the ECM for efficient migration.

Figure 1

Cultured rat microglia display chemokinesis, chemotaxis, and invasion. (A) To monitor chemokinesis (random migration) in two dimensions, cultured rat microglia were imaged for 45 minutes under phase-contrast microscopy at 37°C and 5% CO₂. A representative video frame capture, in which the numbered cells were tracked and analyzed (results in (c)). Inset: higher-magnification images show a lamella and extension of the leading edge in a migrating cell, but not in an immobile bipolar microglial cell. Scale bar: 80 μm (upper image), 40 μm (lower image). (B) Higher-magnification phase-contrast image showing a large ring-like structure in the lamella (circled), and a uropod at the trailing edge (arrow). Scale bar: 20 μm. (C) Each of the 28 microglial cells marked in (a) had a lamella at some time during the monitoring period. During a 10-minute examination period, these cells were classified as migrating in the direction of the lamella, in the opposite direction, or not moving. (D) Quantification of microglia transmigration through pores of 8 μm diameter in filters in the upper well of Transwell™ chambers. Microglia displayed chemotaxis toward 300 μM ATP added to the lower well. Values on bars are mean ± standard error of the mean (SEM) (n = 5). (E) Microglia invaded through Matrigel™-coated pores in the BioCoat Matrigel™ Invasion chambers, and invasion was increased by adding 300 μM ATP to the lower well. Values on bars are mean ± SEM (n = 6).

Figure 2

Lamellae in microglia have a large F-actin ring, comprised of many small punctae. F-actin was labeled with phalloidin (green) in all images. (A) Low-magnification image of many migrating microglia on a glass coverslip. Most lamellae have a large, donut-shaped F-actin ring (* indicates examples), which we call a podonut. Scale bar: 40 μm. (B) Higher-magnification, deconvolved image shows a podonut in the front half of a migrating microglial cell. Scale bar: 10 μm. (C), (D) F-actin rings were predominantly associated with microglial lamellae. Values expressed as mean ± standard error of the mean from three separate cultures. Most microglia with a lamella had a large F-actin ring (a podonut) (87 ± 5% of such cells (c)). As further demonstrated in Figure 3, the large ring was made up of many podosomes. When podosomes were present, they (and the large F-actin ring) were located in the lamella (95 ± 2% of such cells (d)). (E) The large F-actin rings contain hundreds of tiny punctae of F-actin. A representative deconvolved image of a single podonut shows that the large ring is composed of many tiny (<1 μm diameter) F-actin-rich punctae. Scale bar: 5 μm. Inset: a deconvolved three-dimensional (3D) reconstruction shows that the punctae of F-actin were located near the substrate-contacting cell membrane. (F) Microglia formed similar large, F-actin rings in their lamellae; nuclei stained with 4′,6-diamidino-2-phenylindole, (DAPI, blue) when cultured on Ultra-Web™, which is a 3D mesh of polyamide nanofibers (upper image). Right: higher magnification image of several microglia with large F-actin rings. Scale bar: 100 μm (left), 20 μm (right).

Figure 3

The F-actin-rich punctae are podosomes. (A) Microglia were stained for F-actin (phalloidin, green), the plaque protein talin (red), and the nuclear stain 4′,6-diamidino-2-phenylindole (DAPI; blue). A representative image of the large F-actin ring (podonut) in front of the nucleus shows co-localization with talin. Scale bar: 5 μm. Right: higher-magnification deconvolved images of the two boxed regions, color-separated and merged to show that the dense F-actin ring is comprised of many tiny punctae of F-actin and talin. Individual podosomes are seen as a core of F-actin, surrounded by a ring of talin (arrows show examples). Scale bar: 1 μm. (B) Microglia were immunostained for the plaque protein, talin (green), the actin nucleator actin-related protein 2 (Arp2; red), and labeled with the nuclear stain DAPI (blue). Arp2 is enriched in the podosome core, and the plaque protein talin identifies the ring. Lower image: under phase contrast, podosomes were often seen as dark punctae. (C) Prevalence of podosome expression in primary microglia (that is, percentage of cells with podosomes); >65 cells were analyzed per replicate from each of three cultures (mean ± standard error of the mean).

Figure 4

Microglial podosomes are enriched in phosphotyrosine and the Src substrate Tks5. Deconvolved, color-separated and merged images of entire podonuts are shown in the upper panels, and the boxed areas are shown magnified and color-separated below. Scale bar: 5 μm (upper), 2 μm (lower). (A) Immunostaining for phosphotyrosine residues (pTyr, red) is enriched in the F-actin-rich podonut (labeled with phalloidin, green). At higher magnification, some co-localization is seen. (B) Immunostaining shows tyrosine kinase substrate with five Src homology 3 domains (Tks5; green) in microglia podosomes together with the ring marker talin (red). The small punctae of Tks5 are often adjacent to the talin staining.

Figure 5

Microglial podosomes are enriched in Tyr¹⁴-phosphorylated caveolin-1 and nicotinamide adenine dinucleotide phosphate oxidase 1(Nox1). Deconvolved, color-separated and merged images of the lamellar region are shown for two microglia. The boxed areas are magnified in the images at the bottom. Scale bar: 5 μm (upper), 2 μm (lower). (A) Immunolabeling for tyrosine-phosphorylated caveolin-1 (p-Tyr¹⁴Cav1; green) and the podosome ring marker talin (red). (B) Immunolabeling for Nox1 (green) and talin (red). Note: Nox1 staining required methanol fixation.

Figure 6

Microglia with podosomes degrade the extracellular matrix molecule fibronectin. (A) Development of F-actin superstructures (podonuts) with time after plating microglia on coverslips. Upper panels: microglia were labeled for the podosome core component F-actin (phalloidin, red). Podonuts in the lamellae were prevalent after 20 hours of culturing (examples shown by arrows), but not after 2.5 or 10 hours. Lower panels: higher-magnification grayscale images of the boxed regions. Scale bar: 40 μm (upper), 20 μm (lower). (B) Panels indicate the control condition at 20 hours without microglia (left), or 2.5 and 20 hours after adding microglia. Microglia were labeled for F-actin (phalloidin, red), and the substrate, fibronectin, was conjugated to AlexaFluor 488 (green). Representative images from three separate experiments. Scale bar: 40 μm. (C) Higher-magnification, color-separated and merged images showing microglia-sized areas of fibronectin degradation. Stains for F-actin and fibronectin were as in (b), and microglial nuclei were labeled with 4′,6-diamidino-2-phenylindole (DAPI, blue). Scale bar: 10 μm.

Figure 7

Fibronectin degradation corresponds with podosomes. (A) Microglia labeled with the podosome ring component, talin, showing two large clusters of podosomes (arrows). The cell at the right has a large ring of talin (a podonut). Fibronectin degradation (as in Figure 6) is indicated by loss of fluorescence (note: for better contrast, the AlexaFluor 488 was pseudo-colored as blue). Scale bar: 10 μm. (B) Co-localization of podosomes labeled for F-actin (phalloidin, red) with punctate regions of reduced fibronectin, labeled with AlexaFluor 488 (green). Left: combined phase-contrast and fluorescence image of the front of a microglia (nucleus is smooth circle), stained for F-actin (phalloidin), showing the podosome-rich lamella (boxed area). The colored images of the boxed area show F-actin, and the merged image with AlexaFluor 488-labeled fibronectin. Scale bar: 5 μm. (C) Fibronectin loss is not by phagocytosis. A high-resolution, deconvolved three-dimensional projection shows that punctae of fibronectin staining (green) are adjacent to the glass, rather than inside the microglia (4′,6-diamidino-2-phenylindole-stained nuclei, DAPI, blue). Scale bar: 10 μm.