Autophagy in adhesion and migration

Cell migration is integral to multiple developmental and physiological processes. During gastrulation, one of the earliest developmental phases for which migration is most crucial, the collective movement of sheets of cells leads to the formation of a multi-layered embryo (Keller, 2005). Migration is also important for immune surveillance, allowing immune cells to infiltrate and contribute to the homeostasis of tissues (Imhof and Dunan, 1997). Additionally, during pathological processes such as cancer metastasis, motility supports the dissemination of tumor cells to distant sites (Bravo-Cordero et al., 2012). Given this broad importance of migration throughout biology, considerable attention has been given to the cell biology mechanisms that govern motility. Notably, four principle steps enable the mobility of cells, including: (1) establishment of front–rear polarity; (2) generation of actin-rich protrusions (i.e. lamellipodia and filopodia); (3) formation of cell–extracellular-matrix (ECM) adhesions; and (4) release of adhesions to allow forward movement of the cell (Case and Waterman, 2015; Ridley et al., 2003; Vicente-Manzanares and Horwitz, 2011) (Fig. 1A). Although these steps have often been discretely studied, they are highly interdependent and occur simultaneously, and proper coordination among them is essential for productive migration. For example, the initial polarization of the cell dictates in what direction an actin-rich protrusion will occur, and formation of cell–ECM adhesions stabilizes these front protrusions to ultimately promote migration in a particular direction. Thus, a mechanistic understanding of processes that contribute to cell migration is crucial for a comprehensive knowledge of motility-dependent processes.

Integrin-based cell–ECM adhesions, termed focal adhesions, are widely acknowledged for their crucial role in regulating cell migration. Electron microscopy analysis of motile fibroblasts in the 1970s revealed that these structures are found at the leading edge and connect to filamentous systems within the cell, hinting at an important function for focal adhesions in integrating mechanisms necessary for cell movement (Abercrombie et al., 1971). In addition to transmembrane integrins, focal adhesions comprise scaffolding and signaling proteins that collectively bridge the ECM to the intracellular actin cytoskeleton (Geiger and Yamada, 2011) (Fig. 1B). Super-resolution microscopy shows that these focal adhesion proteins are highly organized into functional layers that mediate the connection of actin with focal adhesions, transmit forces from the actin cytoskeleton, and link integrin cytoplasmic tails to signaling and adapter molecules (Kanchanawong et al., 2010). Moreover, focal adhesions are highly dynamic at the leading edge of migrating cells, undergoing continuous cycles of assembly and disassembly. Owing to these structural and dynamic features, focal adhesions direct migration by mechanically transmitting forces generated by the actin cytoskeleton for movement such that the formation of focal adhesion sites at the front of the cell establishes traction for forward propulsion (Gardel et al., 2010; Wolfenson et al., 2013). Focal adhesion formation begins with the birth of a nascent adhesion at the leading edge of the cell. Nascent adhesions then rapidly disassemble through poorly understood mechanisms or are linked to the actin cytoskeleton through recruitment of scaffolding proteins. Such recruitment occurs in a hierarchical manner (Zaidel-Bar et al., 2004), and it involves phosphorylation events (Zaidel-Bar et al., 2007) and exposure of cryptic binding sites on focal adhesion proteins that undergo tension-induced conformational changes in response to actomyosin-mediated pulling forces exerted on focal adhesions (del Rio et al., 2009). Thus, in response to tension, focal adhesions grow and stabilize.

Focal adhesions must also disassemble in order for efficient displacement of the cell body to occur, and accordingly, disassembly functions as a key contributor to migratory speed. Disassembly of integrin adhesions is regulated by numerous pathways that promote focal adhesion destabilization (Fig. 1A,B). Notably, loss of tension at focal adhesions appears to be a key mechanism that leads to turnover (Stehbens and Wittmann, 2012), and these pathways of disassembly might, in the end, serve as different routes through which this is achieved. For instance, adhesion site turnover can be activated by multiple mechanisms that directly culminate in the physical disruption of the actin–focal-adhesion–ECM linkage. Calpain-mediated proteolysis of the focal adhesion protein talin (of which there are two isoforms, TLN1 and TLN2) causes focal adhesion disassembly (Franco et al., 2004). Because talin directly interacts with both integrins and actin (Ziegler et al., 2008), this mechanism is expected to compromise the connection between focal adhesions and the cytoskeleton, thereby resulting in loss of tension and focal adhesion destabilization. Microtubules are also well-established regulators of focal adhesion turnover. Microtubules target leading-edge focal adhesions (Kaverina et al., 1998), and this association correlates with focal adhesion disassembly (Kaverina et al., 1999), indicating that microtubule-dependent pathways could serve a crucial role in triggering focal adhesion turnover in a highly spatiotemporal manner (Stehbens and Wittmann, 2012). Mechanistically, microtubule-dependent focal adhesion turnover involves trafficking of multiple cargos along microtubules that contribute to focal adhesion dissolution. For example, localized endocytosis of integrins relies on microtubule-mediated transport (Ezratty et al., 2009, 2005). Furthermore, microtubules direct exocytic vesicles to adhesion sites, and matrix metalloproteinases (MMPs) carried by these vesicles contribute to focal adhesion turnover through localized ECM proteolysis (Stehbens et al., 2014). Such spatially restricted cleavage of ECM molecules destabilizes integrin–ECM binding and promotes focal adhesion turnover by interfering with the mechanical connection and force transmission that strengthen focal adhesions, and possibly by enabling integrin endocytosis. Activation of extracellular-signal-regulated kinases [ERKs; also known as mitogen-activated protein kinases (MAPKs)] and myosin light chain kinase (MLCK) downstream of focal adhesion kinase (FAK)–Src signaling is also required for focal adhesion disassembly (Webb et al., 2004). Although it remains uncertain precisely how this pathway promotes focal adhesion turnover, it has been proposed that increased tension could elicit biochemical or structural changes that activate disassembly-inducing signals.

In addition to these mechanisms mediating focal adhesion disassembly, autophagy, a process of lysosomal-mediated cellular degradation, has emerged as an important regulator of focal adhesion turnover. Our recent work demonstrates a direct role for selective autophagy in promoting focal adhesion disassembly through the sequestration of focal adhesion components (Kenific et al., 2016). Similarly, proteins that are implicated in focal adhesion assembly and maturation have been identified as autophagy substrates, suggesting an unexplored role for their autophagy-dependent turnover in focal adhesion regulation (Belaid et al., 2013; Ulbricht et al., 2013). Moreover, autophagy could also support integrin endocytosis (Tuloup-Minguez et al., 2014). In this Commentary, we discuss these pathways of autophagy-dependent focal adhesion turnover, with an emphasis on how selective autophagy at focal adhesions is regulated and how it might cooperate with other mechanisms of focal adhesion turnover. We also speculate on the broader biological implications of autophagic regulation of focal adhesions to underscore the potential importance of this pathway beyond cell migration.

Autophagy is a cellular stress response pathway that is induced by diverse insults, such as nutrient deprivation, hypoxia, and remarkably, loss of ECM contact (Murrow and Debnath, 2013). It involves the formation of a double-membraned vesicle, the autophagosome, which engulfs and sequesters cellular material, and then fuses with the lysosome to result in degradation of its contents (Fig. 2A). The resulting lysosomal-derived metabolites and degradation products are then recycled back into the cell to support survival and sustain biosynthesis during stress (Kaur and Debnath, 2015). An extensive characterization of the pathway in both yeast and mammals has led to the discovery of over 30 evolutionarily conserved autophagy-related genes (ATGs) that collectively function to activate autophagy, to promote autophagosome biogenesis and to enable fusion of autophagosomes to lysosomes. The precise roles of specific ATGs have been reviewed in detail elsewhere (Feng et al., 2014).

Although typically recognized for its role in alleviating cellular stress, autophagy is now also widely acknowledged to be an important quality-control mechanism that occurs at basal levels during normal cellular homeostasis (Kaur and Debnath, 2015; Murrow and Debnath, 2013). In particular, autophagy is the primary pathway through which large protein aggregates and damaged organelles are eliminated (Fig. 2B). The importance of this housekeeping function is underscored by studies in which neuron-specific knockout of essential ATGs leads to neurodegeneration that results from the accumulation of protein aggregates in autophagy-deficient neurons (Hara et al., 2006; Komatsu et al., 2006).

In contrast to starvation-induced autophagy, which involves non-specific bulk uptake of cellular constituents, basal autophagy, as well as many types of stress-induced autophagy [e.g. hypoxia, endoplasmic reticulum (ER) stress], is highly selective in order to prevent the unwanted removal of cellular material. This selectivity is principally determined by autophagy cargo receptor proteins, of which the best-described include p62 (also known as sequestome-1, SQSTM1), neighbor of BRCA1 (NBR1), optineurin (OPTN), and nuclear dot protein 52 (NDP52; also known as CALCOCO2) (Stolz et al., 2014) (Fig. 2B). These molecules mediate incorporation of substrates into autophagosomes by interacting directly with both cargos and microtubule-associated protein 1A/1B light chain 3 (LC3; of which there are three isoforms, MAP1LC3A, MAP1LC3B and MAP1LC3C) on the autophagosomal membrane. Current evidence supports that cargo recognition by autophagy receptors occurs mainly through binding of ubiquitin-tagged substrates to ubiquitin-binding domains (UBDs) within the receptors. The receptors also possess an LC3-interacting region (LIR) motif which targets them to autophagosomes. This means of cargo recognition has been most notably illustrated by recent studies demonstrating that ubiquitin-decorated mitochondria and pathogens are targeted by the UBDs of p62, OPTN and NDP52, and are subsequently incorporated into autophagosomes through the LIR-dependent interactions of these autophagy receptors (Lazarou et al., 2015; Manzanillo et al., 2013; Watson et al., 2012). In addition, ubiquitin-independent mechanisms of cargo recognition also exist. For example, targeting of ER for autophagy is mediated by the receptor FAM134B, which is an ER-resident protein (Khaminets et al., 2015). Autophagy-dependent clearance of mitochondria, termed mitophagy, can also occur through mitochondria-localized receptors, such as Nix (also known as BNIP3L), BNIP3 and FUNDC1, which do not bind to ubiquitin (Liu et al., 2014). Finally, although not well understood, some autophagy cargo receptors, such as the ferritin receptor nuclear receptor coactivator 4 (NCOA4), bind to autophagosomes in a LIR-independent manner (Mancias et al., 2014).

Overall, the diversity of mechanisms governing autophagy cargo recognition and the continually expanding list of autophagy receptors and substrates suggests key roles for selective autophagy throughout cell biology. Ongoing work shows that selective autophagy impacts numerous molecular processes through the targeted regulation of various cargos, which range widely from pathogens, large macromolecular assemblies and organelles to individual protein substrates. Accordingly, the diverse functional roles for autophagy can most frequently be attributed to its ability to sequester and degrade specific substrates. Although our understanding of the contribution of selective autophagy to an array of cellular pathways has grown immensely in recent years, much remains to be learned with regards to how selective autophagy is regulated. To that end, studies probing the biochemical basis of substrate-receptor interactions have revealed roles for certain E3 ubiquitin ligases in marking cargo with specific types of ubiquitin chains that are recognized by particular receptor UBDs (Rahighi and Dikic, 2012). Unlike stress-induced autophagy for which the activating signals have been extensively studied and characterized, the pathways leading to induction of selective autophagy remain largely unknown. Given the roles for individual autophagy cargo receptors in multiple molecular processes, an important unanswered question is how different types of selective autophagy are so exquisitely controlled in time and space to regulate degradation of specific substrates. Recent work showing that huntingtin activates p62-dependent selective autophagy of protein aggregates by functioning as a scaffold to assemble autophagy machinery, p62 and ubiquitin-tagged substrates is among the first to reveal such a regulatory mechanism (Rui et al., 2015), and further studies could uncover a diverse array or a conserved repertoire of selective-autophagy-inducing signals.

Recent work demonstrates a previously unrecognized role for selective autophagy that is mediated by the cargo receptor NBR1 in promoting cell migration by directly regulating focal adhesion turnover (Kenific et al., 2016). Although previous studies have shown that autophagy facilitates cell motility, the underlying mechanisms have remained unknown (Galavotti et al., 2013; Lock et al., 2014). Notably, autophagy-inhibited cells exhibit diminished migratory rates and possess significantly enlarged leading-edge focal adhesions; these findings raise the hypothesis that autophagy inhibition stabilizes focal adhesions, thereby leading to decreased levels of migration (Kenific et al., 2016). In agreement with this notion, autophagy-incompetent cells exhibit decreased rates of focal adhesion assembly and disassembly, as well as increased focal adhesion lifetime during migration. In addition, we have also found that autophagosomes associate with disassembling focal adhesions in migrating cells, and focal adhesion components localize to autophagosomes. Because autophagosomes are known to contribute to the turnover of various cellular components and multi-protein complexes through sequestration of substrates, these findings suggest that autophagy contributes to the destabilization and turnover of cell–matrix contacts by locally capturing focal adhesion proteins.

Additional investigations into the targeting of autophagosomes to focal adhesions have uncovered a key role for NBR1 in mediating this pathway (Kenific et al., 2016) (Fig. 3). NBR1 localizes to leading-edge focal adhesions, and similar to inhibition of autophagy, knockdown of NBR1 impairs focal adhesion turnover. Consistent with its function as an autophagy cargo receptor, NBR1 biochemically interacts with focal adhesion proteins that had also localized to autophagosomes and promotes the targeting of autophagosomes to dynamic focal adhesions at the leading edge of migrating cells. Finally, our gain-of-function experiments have shown that ectopic expression of autophagy-competent, but not autophagy-defective, NBR1 specifically enhances focal adhesion disassembly, leading to reduced focal adhesion lifetime during cell migration (Kenific et al., 2016). Collectively, these findings indicate that NBR1, by interacting with focal adhesions, targets autophagosomes to them, which then results in focal adhesion disassembly to drive turnover of cell–matrix adhesions and promote cell migration.

Defining the detailed mechanisms that regulate recognition of focal adhesions by NBR1 for autophagic targeting is a necessary goal for future study. Similar to other selective autophagy pathways, NBR1-dependent focal adhesion turnover is likely to be highly regulated in order to ensure selective targeting of focal adhesions. Moreover, the turnover of focal adhesions themselves must be spatially and temporally controlled to properly balance tension forces that enable forward movement (Gardel et al., 2010); hence, as with other mechanisms that regulate focal adhesion remodeling, selective autophagy should be modulated in time and space for proper coordination of focal adhesion assembly and disassembly. This hypothesis is supported by findings showing that autophagosomes are specifically associated with disassembling adhesions and that NBR1 specifically enhances focal adhesion disassembly (Kenific et al., 2016).

Similar to other NBR1-dependent selective autophagy pathways, the observed interactions between NBR1 and focal adhesion proteins might be mediated by ubiquitylation (Fig. 3). Indeed, NBR1 requires its UBD to enhance focal adhesion disassembly, supporting a role for this conserved mechanism of autophagy cargo recognition in focal adhesion turnover (Kenific et al., 2016). Although certain focal adhesion proteins with which NBR1 interacts, such as paxillin and FAK, can be ubiquitylated (Huang, 2010; Kenific et al., 2016), whether or not ubiquitin modifications on these particular factors mediate binding of NBR1 to focal adhesions remains unexplored. Nevertheless, E3 ligases have been identified in proteomics-based studies that have defined the integrin adhesome (Winograd-Katz et al., 2014; Wolfenson et al., 2013), further suggesting that autophagy plays a crucial role in dictating the fate of focal adhesions that harbor ubiquitylated proteins. Intriguingly, spatiotemporal regulation of E3 ligases could impart a crucial level of control, by specifying when and where NBR1 and autophagy specifically target ubiquitylated focal adhesion components during focal adhesion turnover.

Resident focal adhesion kinases, such as FAK and Src, have key roles in regulating focal adhesion dynamics, and phosphorylation is emerging as an integral regulator of selective autophagy. In particular, phosphorylation of ubiquitin by the kinase PINK1 enhances the precision and efficiency of mitophagy by activating the E3 ligase Parkin (PARK2) to ubiquitylate proteins on the outer mitochondrial membrane, and PINK1 also recruits the cargo receptors OPTN and NDP52 to these mitochondria (Kane et al., 2014; Koyano et al., 2014; Lazarou et al., 2015). Although the broader importance of phosphorylated ubiquitin beyond mitophagy is unknown, it is tempting to speculate that focal adhesion kinases mediate a similar pathway during focal adhesion turnover; such a modification might coordinate recruitment of both E3 ligases and NBR1 to leading-edge focal adhesions during focal adhesion turnover (Fig. 3).

NBR1 might also interact with focal adhesions through ubiquitin-independent mechanisms. NBR1 is a relatively large multidomain protein and, therefore, has the potential to engage in multiple interactions. Accordingly, an open question is how such interactions are regulated to favor targeting of mature focal adhesions for disassembly. One such mechanism of interaction could involve binding of NBR1 to mechanically unfolded focal adhesion proteins. Because focal adhesions progressively grow owing to tension-induced exposure of protein-binding sites, NBR1, like other focal adhesion proteins recruited in this manner, could accumulate at adhesions, eventually promoting autophagic targeting of stable, mature focal adhesions. Intriguingly, the focal adhesion proteins vinculin and zyxin, which bind to NBR1 (Kenific et al., 2016), are both known to undergo tension-induced conformational changes that could permit this type of interaction. In further support of this notion, NBR1 can interact with the mechanically opened kinase domain of titin in sarcomeres (Lange et al., 2005), which also harbor proteins that undergo tension-induced conformational changes, lending precedent to this potential mode of interaction.

In addition to biochemical and mechanical modulation of the binding of NBR1 to focal adhesion constituents, microtubule-dependent delivery of NBR1 to focal adhesions could serve as another means of temporal regulation of NBR1 and autophagosome targeting to focal adhesions (Fig. 3). As described above, targeting of microtubules to focal adhesions promotes their disassembly. NBR1 can bind to microtubules, indicating that this targeting of microtubules to focal adhesions could facilitate the positioning of NBR1 for interaction with and capture of focal adhesions by autophagosomes (Marchbank et al., 2012). Similarly, the binding of LC3 proteins to microtubules might also contribute to the delivery of autophagy proteins to focal adhesions for disassembly (Mackeh et al., 2013).

Finally, spatial regulation of NBR1 and autophagy specifically at leading-edge focal adhesions might also be crucial for autophagy to ultimately support productive migration. Similar to temporal regulation, the asymmetric targeting of microtubules to leading-edge focal adhesions in migrating cells might favor trafficking of NBR1 and LC3 proteins to this particular dynamic pool of adhesions. Additionally, glycogen synthase kinase 3β (GSK3β) could also have a role in spatially restricting selective autophagy to leading-edge focal adhesions. GSK3β is locally inactivated at leading-edge lamella where focal adhesions turn over (Barth et al., 2008), and phosphorylation of NBR1 by GSK3β inhibits its ability to turnover ubiquitylated proteins (Nicot et al., 2014). This regulation suggests that NBR1 is preferentially activated at leading-edge focal adhesions where GSK3β is not active. Although it has yet to be determined whether NBR1-dependent selective autophagy is limited to dynamic leading-edge focal adhesions, the relationship between NBR1, microtubules and GSK3β provides intriguing evidence to support this possibility.

Overall, NBR1-dependent selective autophagy is likely to involve multiple layers of complex regulation to ensure precise spatiotemporal fidelity of focal adhesion cargo uptake by autophagosomes. Importantly, studies interrogating these modes of regulation might provide crucial insight into and establish general paradigms for regulation of cargo–receptor-mediated autophagy. Indeed, coordinated focal adhesion turnover during cell migration provides a unique, experimentally tractable and physiologically relevant model to address these key issues. In addition to defining these detailed mechanisms, an important goal for future study is to determine how this pathway is integrated with other mechanisms of focal adhesion disassembly. For example, although it is unclear how activation of MLCK by FAK–Src signaling induces disassembly, by increasing tension on focal adhesions, MLCK might enhance the binding of NBR1 to mechanically unfolded focal adhesion proteins. Autophagy could also play a role in calpain-mediated focal adhesion disassembly, as ubiquitylation of cleaved talin fragments is ultimately required for focal adhesion turnover (Huang et al., 2009). Also, as detailed above, NBR1-dependent selective autophagy might serve as an additional pathway through which microtubules induce focal adhesion disassembly. Recent work also indicates that autophagy regulates focal adhesion turnover through NBR1-independent pathways. Phosphorylation of paxillin by Src promotes its interaction with LC3 proteins for autophagic degradation of paxillin, leading to disassembly of focal adhesions (Sharifi et al., 2016). In addition, activated Src can be targeted for autophagic degradation, and this is mediated by Cbl, which acts as an autophagy cargo receptor by directly interacting with LC3 proteins (Sandilands et al., 2012a). Additional characterization of this pathway has also revealed that p70S6K (also known as RPS6KB1) promotes targeting of Src to autophagosomes (Sandilands et al., 2015). Furthermore, autophagic degradation of active Src induces degradation of Ret (Sandilands et al., 2012b), a focal-adhesion-associated kinase with a known role in activating FAK (Plaza-Menacho et al., 2011). Given the important roles of FAK and Src during focal adhesion disassembly (Webb et al., 2004), these pathways could influence focal adhesion turnover during migration. Altogether, NBR1-mediated selective autophagy is likely to cooperate with these other autophagy-dependent and -independent pathways to fine-tune focal adhesion remodeling for optimal cell migration.

Tensional stress that is imposed on focal adhesions by actomyosin contractility is a key driver of their growth. Interestingly, work demonstrating that known regulators of cellular contractility are autophagy substrates points to additional mechanisms of autophagy-dependent adhesion regulation through modulation of tension (Fig. 4A and B). The GTPase RhoA modulates contractility and tension by activating myosin II through Rho-associated protein kinase (ROCK, of which there are two isoforms ROCK1 and ROCK2) (Nobes and Hall, 1999), and it can also be targeted by autophagy. Specifically, activated, GTP-bound RhoA is long-lived, and ubiquitylation marks it for p62-dependent autophagy (Belaid et al., 2013) (Fig. 4A). This pathway appears to have a role in the regulation of cytokinesis, where localized contractile activity is necessary for faithful division of daughter cells. Here, by contributing to degradation of active RhoA, it has been proposed that autophagy limits its activity to the midbody, where the final abscission event occurs. Because RhoA promotes adhesion maturation, degradation of RhoA by autophagy is consistent with work demonstrating that autophagy supports focal adhesion turnover (Fig. 4A). Remarkably, in cells deprived of ECM attachment, RhoA-mediated focal adhesion maturation against unanchored integrins induces tensional-stress-dependent cell death (Ma et al., 2007). Because autophagy protects cells from detachment-induced cell death (termed anoikis) (Fung et al., 2008), these findings support the hypothesis that autophagy-mediated degradation of RhoA promotes resistance to anoikis by limiting focal adhesion maturation and tension at unanchored integrin sites in detached cells.

Filamin A is an actin cross-linker that stabilizes the actin cytoskeleton (Razinia et al., 2012). It also functions to link integrins to actin and is recruited to focal adhesions in a tension-dependent manner to reinforce tensional stresses on focal adhesions. Moreover, in response to tension, filamin A mechanically unfolds, and such conformational changes can promote its ubiquitylation and subsequent targeting by another mechanism of selective autophagy, chaperone-assisted selective autophagy (CASA) (Ulbricht et al., 2013). This particular pathway is mediated by the CASA co-chaperone BAG3, which recruits a protein complex that contains the E3 ligase CHIP (also known as STUB1) to ubiquitylate filamin A. p62 then binds to ubiquitylated filamin A to target it for autophagy-mediated degradation (Fig. 4B). That work also proposes that tension induces autophagy, as shown by an increase in autophagosome-incorporated LC3 proteins coinciding with increased tension, suggesting that localized tension at focal adhesions could be one mechanism through which autophagy is spatially and temporally regulated. However, that study did not definitively establish whether the increased autophagosomal LC3 protein accumulation is due to enhanced autophagosome formation or impaired lysosomal turnover of autophagosomes (Ulbricht et al., 2013). Nevertheless, the targeting of filamin A by CASA supports a functional role of autophagy in promoting adhesion turnover and could serve as a tightly regulated pathway for spatiotemporal control of focal adhesion turnover (Fig. 4B).

In addition to modulation of specific tension regulators, there is also evidence for a role for autophagy in regulating contractile structures, such as midbodies and sarcomeres. As already discussed above, p62-mediated RhoA degradation impacts midbodies (Belaid et al., 2013). Moreover, NBR1 has been shown to have a role in ubiquitin-dependent midbody degradation following cytokinesis (Isakson et al., 2014; Kuo et al., 2011). Although the relevance of this pathway to contractility at midbodies is unexplored, it shows that NBR1 is associated with the eventual turnover of structures that are derived from a contractile apparatus. NBR1-mediated selective autophagy also appears to regulate tensional homeostasis of myofibrils and sarcomeres. Multiple studies have demonstrated that p62 and NBR1 associate with these contractile units found in muscle and have implicated them in the turnover of proteins that localize to these structures (Arndt et al., 2010; Lange et al., 2005; Musa et al., 2006; Perera et al., 2011). The precise mechanisms underlying these observations are unknown, but autophagy-mediated degradation of sarcomeric filamin A in Drosophila appears to be involved (Arndt et al., 2010), and muscle-specific RING finger protein (MURF) ubiquitin ligases have also been implicated (Lange et al., 2005; Musa et al., 2006; Perera et al., 2011).

Collectively, work showing that autophagy regulates diverse contractile structures, such as midbodies, sarcomeres, and, more recently, focal adhesions highlights an unappreciated role for autophagy in serving as a general mediator of cellular tension. Importantly, regulation of common factors and the involvement of NBR1 hint at common biochemical and molecular mechanisms underlying these functions. Thus, future work investigating the role of selective autophagy in focal adhesion turnover could provide detailed insight into these other pathways.

Endosomal trafficking of integrins has long been described to be a crucial regulator of cell migration (Paul et al., 2015). Endosome-mediated integrin transport occurs through both short and long recycling pathways, through which internalized integrins are delivered back to the plasma membrane (De Franceschi et al., 2015; Dozynkiewicz et al., 2012) (Fig. 4C). Furthermore, trafficking of ubiquitylated integrins to the lysosome also impacts cell migration (Lobert et al., 2010) (Fig. 4C). Additional findings suggest that endocytosis of integrins also functionally impacts on cell motility by promoting the disassembly of focal adhesions (Ezratty et al., 2009, 2005). Although the role of this particular pathway in supporting rapid turnover of leading-edge focal adhesions has been debated, the internalization and the re-expression of surface integrins might nonetheless influence migration by dictating the global make-up of surface integrins that are available for ECM binding and focal adhesion formation.

Multiple connections between the endosomal pathway and autophagy have been reported. For example, recycling endosomes appear to act as a membrane source for the growth and maturation of autophagosomes (Longatti et al., 2012). Moreover, regulators of vesicular trafficking and fusion, such as certain Rab proteins, are shared among the two pathways (Ao et al., 2014), and more recent work shows that ATGs have direct roles in the functions of late endosomes that are distinct from those in autophagy (Murrow et al., 2015). Finally, endosomes and autophagosomes are thought to ultimately fuse with each other, forming a hybrid vesicle known as an amphisome, which is eventually turned over by lysosomal degradation (Berg et al., 1998).

In agreement with these findings that point to intersections between endosomal trafficking and autophagy, and the important role of endosome-dependent integrin transport in motility, autophagy has been observed to modulate integrin recycling (Tuloup-Minguez et al., 2014) (Fig. 4C). In particular, induction of autophagy by starvation increases colocalization of β1-integrin-positive vesicles with autophagosomes; this results in enhanced delivery of β1 integrin to lysosomes and correlates with reduced migration. Because the targeting of β1 integrin to autophagosomes has only been observed during extreme nutrient deprivation during culture of cells in Hank's Buffered saline solution (HBSS), the relevance of these findings to migration in broad physiological contexts remains uncertain (Tuloup-Minguez et al., 2014). Moreover, targeting of β1 integrin to the lysosome has been shown to support cell migration, rather than to inhibit it (Lobert et al., 2010). Because extreme starvation is known to disrupt cellular metabolism and impact signaling nodes that regulate diverse cellular pathways (Jewell and Guan, 2013), it is possible that HBSS starvation affects migration (Tuloup-Minguez et al., 2014) through autophagy-independent pathways. Taken together, it is clear that the role of autophagy in integrin trafficking during migration is still an open question and warrants further investigation.

Mounting evidence indicates that autophagy plays an important role in cell migration by specifically regulating the turnover of cell–matrix adhesion sites. Although this new function for autophagy appears to be distinct from its classic role in supporting cell growth and survival, these findings demonstrate that the fundamental ability of autophagy to sequester and degrade cellular constituents also underlies the mechanisms through which autophagy directs focal adhesion turnover, tensional homeostasis and cell migration. Thus, further investigation of the role of autophagy in the control of cell adhesion and migration will not only enhance our understanding of adhesion regulation but also contribute to unveiling the detailed molecular and biochemical mechanisms that govern selective autophagy. Moreover, the close relationships between the autophagy and endosomal pathways hint at additional mechanisms through which autophagy could influence motility.

Focal adhesions undoubtedly serve crucial functions throughout biology, particularly during various migration-dependent processes. Going forward, an important goal is determining whether autophagy regulates these processes in vivo, for instance during developmental tissue morphogenesis. Interestingly, one study that has closely examined the role of autophagy during embryonic cardiac development found that loss of function of ATGs leads to defects in cardiac morphogenesis in both zebrafish and mice (Lee et al., 2014), providing evidence that autophagy-dependent migration might indeed be of physiological importance. Pathologically, elucidating in detail how autophagy-dependent migration and adhesion impacts metastatic progression in cancer remains an important topic for further investigation. Future work addressing these issues is thus anticipated to offer fresh insight into autophagy-dependent biological processes during both development and disease.