Podosome initiation and assembly is highly regulated, both spatially and temporally. Dendritic cells best exemplify the temporal regulation of podosome formation. Following activation by an antigen or inflammatory cytokine, immature dendritic cells have a small window of time, roughly 6 hours, during which podosomes are able to form. This window correlates with the development of immature dendritic cells into mature dendritic cells that are no longer able to form podosomes [1]. During this maturation process, immature dendritic cells migrate from their source, through the bloodstream or lymphatic system, into the spleen or lymph nodes where they are needed to present antigens to T-cells. The function of podosomes during the maturation process is still under investigation, however it has been suggested that their matrix-degrading ability promotes dendritic cell migration by breaking down connective tissue barriers [2].

Spatially the nature of the extracellular matrix (ECM) and the distribution of ligands within it, have been shown to affect the initiation of podosome assembly, as demonstrated by experiments with macrophages on fibronectin. Uniform distribution of fibronectin was shown to promote podosome assembly, however when fibronectin spots were created on a micropattern, podosomes tended to form outside of these spots [3]. The somewhat contradictory nature of these results are not yet fully understood, however they do illustrate that the detection of different geometrical arrangements of ECM ligands by cells affects their ability to initiate podosome formation.

Podosome initiation occurs in response to interactions between ECM ligands, such as fibronectin and fibrinogen [3], with cell surface integrins. Distinct integrins are recruited to the adhesive ring structure of podosomes, namely integrin β2 in dendritic cells and macrophages and integrin β3 in osteoclasts [4]. Integrin β1 is also postulated to localize to podosomes, but to membrane sites underlying the actin core [5]. Integrin activation is therefore key to podosome initiation. The cell surface glycoprotein and receptor, CD44, is also reported to engage the ECM and anchor the actin core of the podosome to the underlying membrane [6].

The Arp2/3 activator, WASP (Wiskott Aldrich Syndrome protein) is suggested to regulate integrin receptor clustering during podosome initiation, by an as yet unknown mechanism [1]. This is inferred from dendritic cells lacking WASP that show a diffuse distribution of integrin β2, rather than the concentrated, ring distribution observed during podosome formation. Although these cells are able to maintain adherence to the underlying substrate via integrin β1-containing focal adhesions, they show decreased adherence to the integrin β2 ECM ligand, ICAM-1 (Inter-Cellular Adhesion Molecule 1)[1].

The regulation of podosome formation has so far been limited to discussing the biophysical events of protein-protein interactions. However, recent experiments have provided insight into upstream regulatory events that control the initiation of podosomes at the genetic level. This refers specifically to micro RNAs (miRNA). miRNAs are short (21-25 nucleotides in length) strands of non-coding RNA that negatively regulate gene expression. They do so through binding complimentary sequences in the 3’ untranslated regions of messenger RNA (mRNA) and thereby prevent translation of the mRNA transcript (reviewed in [7]).

miR-143 and -145 have been shown to prevent podosome formation in vascular smooth muscle cells (VSMCs) and so regulate their transition from a differentiated, contractile state to a de-differentiated, migratory state [8]. This was demonstrated both in vitro and in vivo through the use of miRNA-143/145 knockout mice, whose VSMCs showed an increased propensity to form podosomes. Further analysis confirmed the need for miR-143 and miR-145 down-regulation to permit podosome formation. The involvement of miRNAs in podosome formation in other cell types has yet to be studied, however this example demonstrates the need for post-transcriptional regulation in order to prime the cell to respond appropriately to extracellular stimuli.

Following detection of the required external stimulus, via mechanosensing integrins, the actin network is commandeered to facilitate construction of the podosome. Actin assembly is known to occur both from the extension of pre-existing lamellipodial networks, as well as de novo [9].

It is widely believed that Arp2/3-mediated nucleation is the major means by which the podosome actin cytoskeleton is built. This is evidenced by the colocalization of actin, Arp2/3 and the potent activator of Arp2/3, WASP (Wiskott Aldrich Syndrome protein), to these structures [10], as well reduced podosome formation following sequestration [11] or small molecule inhibition [12] of Arp2/3. Notably, WASP has not been detected in focal adhesions (FAs), highlighting a key point of divergence between the protein constituents of podosomes and FAs [4].

The activity of WASP is central to podosome formation, as illustrated in cells lacking full length WASP that consequently also lack podosomes. Resulting defects include reduced bone resorption in osteoclasts [13] and impaired migration of dendritic cells [14] and macrophages [15]. The migratory defects observed in WASP-deficient immune cells are in large part responsible for the systemic, immunological deficits observed in boys presenting with the X-linked disease, Wiskott Aldrich Syndrome (WAS) [14].

The upstream regulation of WASP-activated, Arp2/3-mediated actin nucleation involves the RhoGTPase, Cdc42. The importance of Cdc42 in podosome formation is evidenced by the mutant protein studies showing constitutively inactive Cdc42 greatly reduces podosome formation [14]. The mechanism underlying this is reliant on the interaction between Cdc42 and WASP, which activates the Arp2/3 complex [16]. It should be noted that Cdc42 is suggested to have a dual role in podosome formation [14], firstly in promoting actin nucleation and secondly in determining the distribution of podosomes in migrating cells. The latter function is evidenced by the expression of constitutively active Cdc42 resulting in defects in podosome clustering and their polarized arrangement along the length of the cell [14]. Cdc42 activity must therefore be tightly regulated in order to promote both of these activities.

In addition to the known activation of Arp2/3 by Cdc42–WASP, a second weaker activator of Arp2/3 also resides in podosomes (and invadopodia), namely cortactin. Although cortactin is key to the formation of invadopodia, with depletion of this protein greatly diminishing the number of invadopodia [17] its importance in podosome formation is less clear. In leukocytes, a primary model for the study of podosomes, the cortactin homologue HS-1 (hematopoietic lineage cell-specific protein 1) is dispensable for podosome formation but is instead required for the polarized distribution of podosomes during migration [18]. Cortactin is also suggested to be required to transport vesicles for matrix metalloproteinase-mediated degradation of the ECM. This is dependent on the cortactin-binding domain of WIP (WASP interacting protein), a protein essential for WASP-mediated actin polymerization during podosome formation [19].

Collectively these studies support Arp2/3-mediated nucleation as the primary means for initiating construction of the podosome architecture. Electron microscopy (EM) studies have detected branched actin filaments comprising both the actin core [9] and the radial actin network [20] – though in the latter instance there is still no clear consensus on whether the filaments are indeed branched or not [9].

There is some speculation over the involvement of formin-mediated nucleation of actin within podosomes although the investigations into this are still in their infancy. Evidence for formin-mediated nucleation in podosomes comes from a single study that has observed the formin, FRL1, localizing to a cap-like structure on top of the actin core of macrophage podosomes [21]. Further investigation is needed to confirm this finding and resolve the issue of its apparent localization to the pointed ends of the actin filaments, as opposed to the barbed ends where formins are conventionally known to act. It should be noted that a model for the polarity of radial actin filaments of podosomes has been posited, whereby the barbed ends are orientated towards to the actin core of the podosome, as inferred from scanning EM in conjunction with myosin ‘S1’ fragment labeling [9].

In this model actin polymerization, at the face of the actin core, generates forces that push against the core and consequently promotes podosome protrusion [9]. In this context, it is plausible that formins, localizing as expected at the barbed ends of actin filaments, could be found encasing the tops of the actin cores. However it is important to note that this model is yet to be comprehensively tested and therefore remains a speculative hypothesis. The prevailing theory for actin assembly during podosome formation therefore still centers around Arp2/3-mediated nucleation.

The specific cascade of events leading to disassembly and turnover of the podosome architecture are not fully understood. Podosome disassembly is suggested to involve myosin IIA-induced contractions, affecting first the adhesive ring and then the actin core, as illustrated in dendritic cells where myosin IIA is the predominant myosin isoform [23].

Rho GTPases, which are known to regulate podosome formation [14], are also implicated in podosome disassembly, upstream of myosin IIA activity. Specifically, the downregulation of Rac1 and Cdc42 activity and the upregulation of RhoA activity are suggested to regulate myosin IIA, with RhoA targeting myosin IIA through its effector Rho kinase [23]. Conversely, the inhibition of myosin II activity has been shown to promote podosome formation, with a concomitant loss of focal adhesions [24]. An inverse relationship between the ability of a cell to form and maintain podosomes and its ability to form and maintain FAs has been demonstrated in several instances [24, 25, 26].

The calcium-dependent cysteine protease, calpain, is also implicated in podosome disassembly [27]. Inhibition of this enzyme in dendritic cells promotes the stabilization of podosomes and the accumulation of podosome components, including F-actin, WASP (Wiskott Aldrich Syndrome protein), β2 integrin and the integrin-associated proteins talin, paxillin and vinculin. Furthermore several podosome resident proteins, namely Pyk2 (protein tyrosine kinase 2 beta), WASP and talin, were shown to be cleaved by calpain, providing a mechanism by which calpain can induce disassembly of the podosome architecture and promote dendritic cell motility [27]. The calpain-dependent cleavage of WASP only occurs following the release of WASP from the WIP–WASP complex (personal communication, Prof Gareth Jones).