|Home | About | Journals | Submit | Contact Us | Français|
Fascin is an evolutionarily conserved actin bundling protein that localizes to microspikes, filopodia and actin-based protrusions underneath the plasma membrane. fascin has received a lot of attention among cytoskeletal proteins because multiple clinical studies have implicated its expression in cancer progression and metastasis. this may be because fascin is not normally expressed in epithelial tissues and when it is upregulated as a part of a program of cancer cell epithelial to mesenchymal progression it confers special motility and invasion properties on cancer cells. in normal adult tissues, fascin expression is high in neurons and dendritic cells; both cell types have striking large filopodia and are highly motile. it is not clear how fascin promotes invasive motility in cancer cells, but many studies have implicated filopodia formation in motility and we have recently provided new evidence that fascin stabilizes actin bundles in invasive foot structures termed invadopodia in cancer cells Figure 1.1 Here we review some of the evidence implicating fascin in motility, invasion and cancer aggressiveness, and we speculate that by stabilizing actin, fascin provides cells with powerful invasive properties that may confer increased metastatic potential.
Fascin, a uniquely folded actin bundling protein with four beta-trefoil domains, is conserved from amoebas to man.2 Fascin was first described as an actin bundling protein in sea urchin coelomocytes3 and later also found in Drosophila as the singed gene product.4 In Drosophila, fascin is expressed in motile hemocytes (blood cells) and in stable bristles.4,5 Fascin depletion leads to sterility and a singed-looking twisted bristle phenotype in flies.4 Recent data suggests that fascin in stable actin bundles of bristles behaves differently than the more dynamic spikes and filopodia in migrating hemocytes.5 There are three isoforms of fascin in mammals, fascin-1 (referred to from here on as fascin) is widely expressed during embryogenesis in neural and mesenchymal tissues,6,7 but in adults it is largely restricted to specific tissues, including brain, endothelium and testis. Expression of fascin-2 is retina-specific and fascin-3 is expressed only in testis.2
Fascin has two actin filament binding sites and one of these, near the N-terminus, can be switched off by PKC (protein kinase C) phosphorylation.2 While usually phosphorylation activates proteins and signalling pathways, many actin binding proteins are negatively regulated by phosphorylation, including the actin severing protein cofilin and the membrane associated actin linker MARCKS (Myristoylated alanine rich protein kinase C substrate). Fascin forms flexible and deformable actin gels in vitro and has slow dissociation kinetics from actin filaments.8 Actin microspikes formed by fascin bundling are permissive for motility of myosin motors, such as myosin-II and myosin-V, allowing transport of vesicles or other cargo along actin bundles.9
Fascin localizes to leading edges of crawling cells and is important for the assembly of filopodia and spiky protrusions that direct cells along a rigid surface or through a porous filter.2,10–12 Depletion or inhibition of fascin results in reduced migration in these settings and diminished protrusion of filopodia.1,13 Fascin doesn’t seem to be a simple constitutive actin bundler, but rather its regulation by PKC is tightly coupled with integrins and the extracellular matrix. Phosphorylation of human fascin S39 results in loss of actin bundling activity, but retention of PKC-α association.10,14 When cells are plated on the extracellular matrix components thrombospondin-1 (tsp-1, a glycoprotein) or laminin, they produce numerous spiky protrusions, decreased phosphorylation of S39 and inactivation of PKC-α towards fascin.15,16 However, when cells are plated onto fibronectin coated surfaces, they initially protrude long dynamic filopodia that are soon replaced by smooth sheet-like lamelliopdial protrusions and this conversion is thought to depend at least in part on the inactivation of fascin by increased PKC activation due to engagement of the β1 integrins by fibronectin.15,16 Thus fascin’s on-off status could be intimately controlled by the extracellular environment and the engagement of integrins at the cell surface. Interestingly, high levels of stromal tsp-1 have been associated with poor prognosis in both pancreatic ductal carcinoma and cutaneous melanoma.17,18 Perhaps a systematic analysis of how different integrins interface with fascin would be useful to unravel how fascin promotes invasion.
Not only is fascin important for cellular motility and protrusion in 2D, but when cancer cells invade into extracellular matrix in 3D, they use fascin to make long filopodia-like protrusions and to assemble degradative actin-based protrusions called invadopodia1 (and Fig. 1). Invadopodia have been compared with other matrix remodelling structures called podosomes, which are found in motile cells such as dendritic cells, vascular endothelial cells, macrophages and osteoclasts.19 Podosomes are dot-like clusters of actin and signalling proteins on the ventral surface of a cell that are important for adhesion and degradative activity.19 Invadopodia and podosomes contain many of the same proteins, but at least in tissue culture, invadopodia are almost an order of magnitude longer-lived and more degradative than podosomes.1,20 Invadopodia, unlike podosomes, also often appear as actin comets that are tethered near the head whilst the tail spins and appears much like an actin tail on a pathogenic bacterium such as Listeria monocytogenes.1,21,22 Invadopodia also contain specialised integrin clusters believed to be important for adhesion.23,24 The fascin in invadopodia appears to stabilize the actin, as knockdown of fascin increases the mobile fraction of fascin in invadopodia and decreases the lifetime and size of invadopodia.1 Fascin localizes in the head and tail regions of invadopodia and it is unclear where the trapped actin that does not turn over during photobleaching experiments resides, but it could form a shell around or a core within the actin tail.
In 3D gels of collagen and Matrigel, cancer cells expressing fascin form long spiky protrusions and to more efficiently remodel the matrix to invade towards a chemoattractant1 and Figure 2. Depletion of fascin reduced penetration into reconstituted matrix and greatly reduced the spikiness of invading cells.1 Thus, fascin appears to provide cancer cells with an efficient means to assemble stable long-lived invasive protrusions that allow them to invade into matrix.
Multiple studies have reported that fascin is upregulated in more aggressive and metastatic epithelial cancers and that high fascin expression is a significant independent prognostic indicator of poor outcome (Table 1). Studies showing fascin as a significant independent negative prognostic indicator for survival after multivariate analysis include cancers of the liver, ovary, lung, pancreas, colorectal, head and neck squamous cell carcinoma and brain (Table 1). In many studies, fascin staining was higher in poorly differentiated and more advanced stage or grade tumors. Often positive fascin staining was correlated with increased probability of metastasis to lymph nodes. Fascin was sometimes noted to be high in tumor stroma, perhaps due to the presence of motile fibroblasts or dendritic cells. Tumor cells that expressed high fascin were sometimes noted to have low expression of Ki-67, a marker for proliferation. 25,26 Hashimoto et al. pointed out that this may indicate fascin expression selectively in highly metastatic but non-proliferative cells, an idea that warrants further investigation.
In some cases, high fascin expression has been correlated with low E-cadherin (epithelial cadherin) expression, indicating that as cells progress through the epithelial to mesenchymal transition (EMT), they gain fascin whilst losing E-cadherin.27,28 There is evidence that fascin expression is regulated by the wnt activated LEFTCF transcriptional signalling pathway that promotes EMT.11 However, fascin expression is also regulated by cyclic-AMP response element binding protein (CREB) and the aryl hydrocarbon receptor (AhR).29 How these two systems work together to regulate fascin expression in tissues and in cancer remains to be studied further.
It is becoming clear that upregulation of fascin expression confers both normal and cancerous cells with increased motile properties and the ability to assemble spiky actin-based protrusions that aid in invasion and cell motility. Many questions remain as to why fascin appears to be relatively unique among the actin bundling proteins as having an important role in cancer invasiveness. It is likely that other actin modulating proteins work together with fascin to allow cells to increase their motility and change their morphology when they undergo epithelial to mesenchymal transition (for cancer) or when dendritic cells mature. Some studies have suggested that cortactin may cooperate with fascin to promote cancer aggressiveness,1,22,30–33 but this may just be a coincident increase in expression and co-localization in similar cytoskeletal structures, rather than a direct cooperation in a common pathway. Further characterization of cytoskeletal changes accompanying upregulation of fascin at tumor invasion fronts will be informative. It is tempting to speculate that fascin may have other key cellular roles besides organising actin bundles. While fascin is becoming established as a marker for tumor progression, its role in promoting cancer aggressiveness is much less well established or understood. Further testing of fascin’s role in invasion and metastasis in vivo in physiologically relevant models for cancer metastasis are needed to determine how active a role fascin plays in cancer aggressiveness and whether it is a potential clinical target.
We thank Cancer Research UK and Medical Research Council UK G117/569 for core support.
Previously published online: www.landesbioscience.com/journals/cib/article/11556