Podosomes are tiny, multi-molecular structures with two key properties that can aid in cell migration through tissue. They provide anchorage and traction mediated by attachment to the ECM, and localized ECM degradation. Podosomes are distinguished by having a two-part architecture. The F-actin-rich core is surrounded by a ring containing integrins and adhesion-plaque proteins, including talin, vinculin and paxillin [49–51]. We recently discovered that the lamellae of migrating microglia contain many podosomes, often arranged into a large ring that we called a ‘podonut’ . Individual podosomes were identified as tiny (< 1 μm diameter) punctae with a core with F-actin and its nucleator, Arp2/3 that is surrounded by a ring of talin. Microglia with podosomes degraded the ECM component, fibronectin. This was seen as a loss of fluorescence in cell-sized patches at low magnification, and as podosome-sized punctae (approximately 1 μm) at high magnification. The present study contributes several novel findings concerning podosome structure and regulation.
Podosomes are highly dynamic and continually assemble, mature and disassemble [7, 8]. There is limited information about processes regulating their rapid turnover. Podosomes in normal cells and invadopodia in cancer cells form only after cell adhesion. A key initiating factor is thought to be cell attachment to the substrate through integrins [9, 51] but this is not sufficient. Of note, myeloid-lineage cells (including our observations on microglia) are apparently unique in spontaneously forming podosomes upon cell attachment. While short-lived, podosome stability involves regulation of the actin cytoskeleton [52, 53]. The podosome core contains many actin-regulating molecules. These include actin nucleators (Arp2/3 complex, formins), binding proteins (coronin, tropomyosin), filament crosslinkers (caldesmon, α-actinin) and polymerization activators (cortactin, WASp and its regulators) [8, 9, 50]. In addition, activation of integrins and receptor tyrosine kinases can induce intracellular cascades involving c-Src, protein kinase C and Rho- GTPases . Several components of podosomes (and invadopodia) are regulated by tyrosine kinase signaling [51, 54]. Thus, it is not surprising that phosphotyrosine residues are highly enriched in podosomes, including those in microglia .
Initially, we addressed the role of Ca2+ entry based on evidence that Ca2+ regulates cell migration and cell-substrate adhesions [12, 13]. For instance, in human breast cancer cells, turnover of focal adhesions was disrupted by reducing external Ca2+ . That study also showed that substrate adhesion and migration were impaired by the drug, SKF96365, which blocks several Ca2+-permeable channels, including CRAC. Migrating cells maintain a descending intracellular Ca2+ gradient from the trailing uropod to the leading lamellum [13, 56]. Brief, localized Ca2+ rises (Ca2+ flickers) in lamellipodia can aid cell steering . For microglia, we found that Ca2+ influx was required for formation of podosomes (and podonuts). The pharmacological profile implicates CRAC channels in podosome formation, microglia migration into a scratch wound, transmigration through open pores, and invasion through Matrigel™. Thus, it is notable that the core of individual podosomes contained Orai1, which is the pore-forming subunit of CRAC . CRAC channels open when Orai1 interacts with the ER-molecule, STIM1, which is oligomerized following depletion of intracellular Ca2+ stores (recently reviewed in ). Oligomerization is rapidly reversed (approximately 2 min) when stores are replenished, and therefore the STIM1-Orai1 interaction is transient. Podosomes are also highly dynamic, with lifetimes as short as 2 min [8, 11]. Despite both processes being short-lived, we found a close association of STIM1 with podosomes, and some clear co-localization in the podosome ring. Although it would be interesting to know if podosomes transiently interact with functional CRAC channels in response to localized depletion of Ca2+ stores, it will be difficult to study such transient interactions.
Previous evidence linking podosomes to specific routes of Ca2+ entry is limited. Information derives mainly from over-expression studies and is often conflicting. Podosome formation in a neuroblastoma cell line was induced by over-expressing and activating the Ca2+-permeable channel, TRPM7 , and is consistent with a dependence on intracellular Ca2+. Another study addressed TRPM7 but did not examine podosomes . Having found that TRPM7 produced Ca2+ flicker activity and was stretch-activated, the authors proposed that this channel responds to cell adhesion, traction and migration. We previously demonstrated robust expression of native TRPM7 channels in primary rat microglia . In microglia, TRPM7 was not stretch-activated, and instead produced a large current under a wide range of activation conditions . Here, we show that TRPM7 was not enriched in podosomes, and that blocking it did not affect podosome formation. There are also conflicting data regarding the role of Ca2+ in podosome formation. Two earlier studies found that elevating intracellular Ca2+reduced podosome numbers, but direct comparisons are difficult. One study was in a macrophage cell line transfected with the Ca2+-permeable channel, TRP Vanilloid 2 . The second was in chicken osteoclasts under several conditions that raised intracellular Ca2+, including high extracellular K+ or activators of voltage-gated Ca2+ channels. Because microglia lack voltage-gated Ca2+ currents [3, 61], this finding is not expected to translate.
We examined subcellular localization of the Ca2+-activated K+ channel, SK3, because of our previous findings that SK3 was increased in activated rat microglia in vivo and regulated their classical activation in vitro . We surmised that SK3 likely acts by maintaining a driving force for Ca2+ entry through Orai1/CRAC channels. No previous reports have linked SK3 channels to podosomes. Here, we found that SK3 was highly enriched in the podosome core. Importantly, its accessory molecule, CaM, was present both within and around the podosomes. This is significant because the Ca2+ sensitivity of SK channel gating is conferred by CaM, which is bound to the channel’s carboxy-terminus [21, 22]. SK channels open after Ca2+ binds to CaM . Here, we found that the SK3 channel inhibitor, NS8593, did not affect migration through open holes but dramatically reduced microglia invasion through Matrigel™. Thus, we conclude that SK3 is involved in matrix degradation. SK3 has been implicated in migration of some cancer cells. In one study, an SK3-dependent membrane hyperpolarization increased the motility of melanoma cells . In another, SK3 was expressed in tumor breast biopsies and a highly metastasizing mammary cancer cell line, but not in non-tumor breast tissue . The latter study showed that SK3 blockers inhibited migration, depolarized the cell, and reduced intracellular Ca2+. We found one report on non-cancer cells. SK3 was present in lamellipodia and filopodia of neural progenitor cells . Pharmacological treatments implicated the channel in formation of cellular projections, which are structures used to explore the local environment, interact with other cells and for migration.
Several other Ca2+-regulated molecules have been identified in podosomes but mainly in transformed cells. Caldesmon was found in Src-transformed fibroblasts , calponin in the A7r5 smooth-muscle cell line , gelsolin in Src-transformed fibroblasts and monocyte-derived cells , and Pyk2 in the MB1.8 osteoclast cell line . We found that microglial podosomes are highly enriched in the Ca2+-binding molecule, Iba1, which is a marker used to identify microglia and infiltrating macrophages in the CNS [37, 38]. Iba1 has not previously been reported in podosomes or invadopodia but was of interest because it cross-links actin filaments in a Ca2+ dependent manner . Previously, Iba1 was found in membrane ruffles and phagocytic cups of MG5 cells, a microglia cell line from p53-deficient mice . However, Iba1 is not characteristic of all F-actin-rich structures, and is not in filopodia or stress fibers . We speculate that Iba1, which is in the core of microglial podosomes, might stabilize them by cross-linking F-actin. Other Ca2+-regulated actin cross-linking proteins might play a similar role. For instance, α-actinin is present in the podosome ring and core of macrophages, osteoclasts, monocytes, and Src-transformed fibroblasts [51, 71–73].
Finally, we present a scheme to relate the literature on podosome formation and roles to the present study and our recent paper . A key initiating factor in podosome formation is cell attachment to the substrate through integrin binding. The subsequent signaling is thought to activate Src, and promote phosphorylation and activation of substrates that include caveolin-1 [74, 75] and Tks5 [76, 77]. Phosphorylated Tks5 can act as an organizer, recruiting other proteins, including Nox1 , which is an enzyme that generates reactive oxygen species. We recently showed that Tks5 and Nox1 are constituents of microglial podosomes . Src can activate PLCγ  and release soluble inositol-1,4,5-triphosphate (IP3), which evokes depletion of Ca2+ stores. This is followed in some cell types by interaction of STIM1 with Orai1 and activation of CRAC channels. We previously showed that store depletion activates CRAC currents in rat microglia . Here, we show that Orai1 and STIM1 are both present in and around microglia podosomes. Ca2+ influx, most likely through CRAC channels, regulated microglia podosome formation, migration and invasion through Matrigel™. Invasion also required SK3 channels, which we discovered were present in podosomes, along with their gating molecule, CaM. We propose a working model in which localized Ca2+ elevation caused by CRAC channels (Orai1 + STIM1) activates Ca2+-dependent SK3 channels. The resulting K+ efflux is expected to hyperpolarize the membrane and help maintain a driving force for Ca2+ entry. Ca2+ entry is then expected to regulate multiple downstream effector molecules that contribute to cell migration.