|Home | About | Journals | Submit | Contact Us | Français|
Apoptosis is an important biological process required for the removal of unwanted or damaged cells. Mounting evidence implicates the actin cytoskeleton as both a sensor and mediator of apoptosis. Studies also suggest that actin binding proteins (ABPs) significantly contribute to apoptosis and that actin dynamics play a key role in regulating apoptosis signaling. Changes in the organization of the actin cytoskeleton has been attributed to the process of malignant transformation and it is hypothesized that remodeling of the actin cytoskeleton may enable tumor cells to evade normal apoptotic signaling. This review aims to illuminate the role of the actin cytoskeleton in apoptosis by systematically analyzing how actin and ABPs regulate different apoptosis pathways and to also highlight the potential for developing novel compounds that target tumor-specific actin filaments.
Apoptosis or programmed cell death is an essential biological function required during embryogenesis, tissue homeostasis, organ development and immune system regulation.1-4 The importance of apoptosis in organism development is now recognized by the myriad of pathologies associated with the de-regulation of apoptotic signaling pathways leading to cancer, autoimmune diseases and neurodegenerative diseases.5-7 Apoptosis can be triggered by the reception of a death signal or by the removal of an anti-apoptotic signal resulting in a cascade of distinct morphological changes. The “apoptotic” cell is isolated from surrounding tissue (cell rounding), followed by chromatin condensation, organelle compaction, membrane blebbing and the formation of intact apoptotic bodies.8-10 Finally, the externalization of phosphatidyl serine to the outer membrane surface signals to the immune system that the cell is destined for death by phagocytosis.11,12
Successful apoptosis is coordinated by the family of cysteine proteases termed caspases. Caspases cleave numerous cellular substrates by specifically targeting aspartate residues.13 Caspases are synthesized as inactive zymogens or pro-caspases that are activated in response to specific apoptotic stimuli.14,15 Activation of the initiator caspases-8 and -9 occurs via the extrinsic and intrinsic apoptosis pathways respectively.16 The extrinsic pathway is triggered by the ligation of extracellular “death” ligands from the tumor necrosis factor (TNF) family such as CD95/FasL and TNFα with their cognate membrane receptor.17 Ligand binding to the cell surface triggers the intra-cellular association of Fas associated death domain (FADD) with pro-caspase-8 forming the death-inducing signaling complex (DISC).18,19 The accumulation of pro-caspase-8 molecules results in their dimerization and auto-processing to produce active caspase-8.16,20 Caspase-8 can then activate the executioner caspases-3, -6 and -7 which are responsible for widespread proteolytic activity leading to the removal of the apoptotic cell by the immune system.21 (Fig. 1)
The intrinsic apoptosis pathway is chiefly mediated by the Bcl-2 family of pro- and anti-apoptotic proteins (Fig. 1). The Bcl-2 super family act as sentinels of cell well-being that detect stress signals such as DNA damage, cytokine/growth factor withdrawal and anoikis (detachment induced cell death).22 They also ensure the completion of apoptosis by irreversible mitochondrial membrane damage.23 All members of the mammalian Bcl-2 super family contain conserved BH domains and their classification into three functionally distinct groups is governed by the number of BH domains present.24 The Bcl-2-like anti-apoptotic proteins (Bcl-2, Bcl-xL, Bcl-w, Mcl-1 and A1) contain BH 3 and 4 domains and protect cells from apoptosis by guarding the outer mitochondrial membrane (OMM).25,26 The BH3-only proteins (Bim, Bad, Bid/tBid, Bmf, Bik, Hrk and Noxa Puma) which contain a single BH domain, detect cell stress and when activated engage with specific pro-survival Bcl-2 partners neutralizing their pro-survival activity.27,28Table 1 outlines the specific Bcl-2:BH3 interacting partnerships currently established in the apoptosis field. Lastly, the BH1–3 group (Bax, Bak and Bok) regulate mitochondrial membrane permeability and have a specialized capacity to homo-oligomerize upon activation.29,30 In healthy cells Bak is located at the surface of the OMM in complex with Mcl-1 and Bax resides within the cytosol.31 Inactivation of Bcl-2 activity at the OMM via BH3 ligation releases Bak and activates Bax translocation to the mitochondria.32 Bax and Bak oligomerization at the OMM induces the loss of the mitochondrial membrane potential (mtΔΨ) leading to the formation of the mitochondrial permeability transition pore (mPTP).29,33 Formation of the mPTP induces the release of apoptogenic factors such as cytochrome c, smac/Diablo and apoptosis inducing factor, AIF-1.34,35 Cytochrome c release is particularly important as it activates the conformational change in Apaf-1 (apoptotic protease activating factor) required for the activation of the second initiator caspase-9.36,37 Caspase- 9 can also activate caspases-3, -6 and -7 to further provoke the apoptosis response. The actin cytoskeleton has been implicated in regulating apoptosis at multiple stages both upstream and downstream of caspase activation. Knowledge of the nature of actin filament dynamics reveals how actin can both initiate and mediate an apoptotic signal.
The actin cytoskeleton is a structural network of proteins that are essential for multiple biological functions including cell contraction, cell motility, vesicle trafficking, intracellular organization, cytokinesis, endocytosis and apoptosis.38-41 Actin, the major component of the cytoskeleton, is a 42 kDa globular protein (G-actin) that reversibly polymerizes to form filaments (F-actin). In muscle cells actin is a core component of the sarcomere and interacts with myosin filaments to enable force generation required during muscle contraction.42 In non-muscle cells, actin isoforms (β and γ) perform a diverse range of functions that enable cell survival and adaptation to a changing environment.43-46 For a comprehensive review of actin structure and function see reference 47
To cope with the dynamic cellular environment F-actin assembly is in a constant state of flux with G-actin association occurring at the barbed end (+) and dissociation at the pointed end (-).48,49 Actin filament dynamics are regulated by the action of a large group of proteins termed the actin binding proteins (ABPs). ABPs undertake a range of functions including actin severing, depolymerizing, capping, stabilizing and de novo actin polymerization which enables the actin cytoskeleton to constantly adapt to a changing environment.39 The Rho GTPases are an important signaling protein family that regulate ABP function to achieve the formation of higher order structures such as stress fibers (actin/myosin bundles), lammelipodia (membrane ruffles at the leading edge) and filopodia (membrane protrusions).50 Biochemical reactions such as phosphorylation and calcium ion (Ca2+) binding are also essential to the regulation of ABP function as most ABPs exist in an active and inactive form.51 Furthermore phosphoinositides such as PtdIns(4,5)P2 play a pivotal role in regulating actin functions at the plasma membrane by accumulating within lipid rafts and facilitating F-actin polymerization.52 It is clear that actin filament dynamics are tightly regulated at numerous stages which is warranted considering the vast array of functions mediated by the actin cytoskeleton. Studying the role of actin and ABPs in apoptosis demonstrates the importance of regulated actin filament organization.
The actin cytoskeleton has been demonstrated as essential during multiple hallmarks of apoptosis with dramatic changes in actin filament organization accompanying different stages of apoptosis.10,53 Cell rounding, which involves the loss of focal contacts with the extra-cellular environment, requires the formation of a contractile cortex of myosin II decorated actin filaments.54 Retraction of the actin-myosin II cortex significantly alters membrane dynamics resulting in the formation of membrane blebs.54,55 Actin-dependent membrane blebbing is reliant upon Rho GTPase signaling56 with Rho inhibition preventing bleb formation in PC12 cells.57 However, in Jurkat cells, caspases cleave and activate the Rho effector ROCK1, which can regulate actin-mediated membrane blebbing in a Rho-independent manner.58 At the final stages of apoptosis the actin cytoskeleton is degraded and phagocytosis of the apoptotic bodies ensues.55 In vitro microfilament disruption assays utilizing U-937 and HL-60 cells highlighted the importance of actin filament dynamics at the final stages of apoptosis with actin targeting drugs inhibiting apoptotic body formation.59 The important role of actin in the morphological hallmarks of apoptosis is coupled with mounting evidence demonstrating actin as a mediator and initiator of apoptosis signaling.
Actin has been demonstrated as a substrate for cleavage by caspases in mammalian cells, resulting in the formation of actin fragments that are 31 kDa (Fractin) and 14 kDa (tActin).60,61 Transient transfection of 293 T cells with the expression vector of tActin, but not Fractin, resulted in the appearance of morphological hallmarks of apoptosis such as cell rounding and chromatin condensation.61 Furthermore, ectopic expression of tActin induced these morphological changes without activating caspases, indicating that actin fragment-mediated cellular shrinkage is an event downstream of the caspase signaling cascade.61
Manipulating the actin cytoskeleton via drug intervention has further revealed that changes in actin dynamics can also mediate apoptosis. Jasplakinolide is a potent F-actin stabilizing drug, that is derived from the marine sponge Jaspis johnstoni.62 Treatment of Jurkat cells with jasplakinolide resulted in the appearance of distinct morphological and biochemical hallmarks of apoptosis including DNA fragmentation, chromatin condensation and caspase activation, suggesting that actin stabilization elicits an apoptotic response.63 Treatment of leukemic HL-60 cells with jasplakinolide similarly induced distinct nuclear and membrane changes that resembled apoptosis.64 Jasplakinolide also increased the activity of DNase I, which is responsible for the degradation of nuclear DNA strands during apoptosis.65 DNase I binds with high affinity to G-actin monomers simultaneously promoting actin depolymerization and DNase I inhibition.66 The actin polymerizing activity of jasplakinolide may be triggering the release of DNase 1 from G-actin once it has been added to the barbed end resulting in activation of DNase I activity. This hypothesis remains to be investigated. Actin depolymerization has been shown to induce an apoptotic response in numerous cell types. Cytochalasin D belongs to a family of fungal metabolites that binds to the barbed end of F-actin preventing further polymerization.67 Cytochalasin D was shown to induce caspase-3 mediated apoptosis in T lymphocytes68 and enhanced the commitment of Jurkat T cells to apoptosis after cytokine withdrawal.69 Given that cytochalasin D binds to the barbed end of the actin filament, the effects of cytochlasin D are through its disruption of actin filament dynamics and not through a shift in G:F-actin levels. The importance of actin dynamics thus explains why both actin polymerizing and depolymerizing can affect cell survival. Given the profound response of the actin cytoskeleton to changes in dynamics, actin may also play a role in initiating an apoptotic signaling cascade.
A number of studies have demonstrated a role for the actin cytoskeleton in triggering apoptosis upstream of caspases both in the extrinsic and intrinsic pathways. CD95 or FasL is a major ligand of the extrinsic apoptosis pathway. In CD4+ T lymphocytes activation of CD95-mediated apoptosis resulted in the polarization of CD95 at the cell surface. CD95-mediated apoptosis was found to be dependent upon the interaction of actin with CD95 via the actin-associated protein ezrin.70 Ezrin interacts with membrane associated proteins via the FERM domain located within the N-terminus and is tethered to the actin cytoskeleton at the C-terminus.71,72 Thus ezrin, which is anchored to CD95 at the cell membrane of T lymphocytes is thought to transduce an extracellular signal to the actin cytoskeleton initiating an apoptosis cascade (Fig. 2).73,74 A reduction in ezrin expression has been correlated with the stimulation of CD95-mediated apoptosis in H9 stem cells and normal T lymphocytes which contradicts with aforementioned studies.75 The role of ezrin in CD95 mediated apoptosis thus remains in-conclusive. What is known is that ezrin activity is phospo-regulated and its phophorylated status may govern its' function in apoptosis.75
Actin has also been implicated in the initiation of intrinsic mitochondrial-dependent apoptosis spanning yeast, mammals and plants.40,76,77 Studying the role of actin in mitochondrial-dependent apoptosis in yeast has been very advantageous due to the presence of a single actin isoform, ACT1.78 Yeast strains bearing point mutations in ACT1 clearly demonstrated altered susceptibility to mitochondrial damage. Expression of a mutant allele with decreased actin dynamics (act1-159) resulted in F-actin aggregation and increased susceptibility to apoptosis due to the accumulation of reactive oxygen species (ROS) and mitochondrial membrane depolarization. Expression of a yeast mutant with increased actin filament dynamics (act1-157) did not accumulate ROS and survived in long-term culture.76,79 This suggests that a dynamic actin cytoskeleton is essential to the maintenance of the mtΔΨ in yeast and that modulation of actin dynamics contributes to oxidative stress. The capacity for actin to regulate the opening of voltage dependent anion channels (VDACs) has been postulated as a mechanism by which actin regulates mitochondrial integrity. In Neurospora crassa actin stabilization via phalloidin treatment led to the prolonged opening of VDAC pores resulting in the accumulation of ROS and a loss of mtΔΨ.80
In mammalian HeLa cells, cytochalasin D treatment induced caspase-mediated cytochrome c release suggesting that actin has a role in regulating mitochondrial membrane permeability in both yeast79 and mammalian cells.81 Overexpression of the pro-survival protein Bcl-xL inhibited apoptosis in jasplakinolide treated CTLL-20 cells82 and partially attenuated apoptosis in Jurkat cells treated with cytochalasin D69 implicating the Bcl-2 family in the apoptotic action of cytochalasin D and jasplakinolide. The latrunculins are a group of actin depolymerizing agents that are derived from the Red Sea sponge Latrunculia magnificans and sequester G-actin monomers and prevent polymerization via a mechanism that differs from cytochalasin D.83 Latrunculin A was shown to induce caspase-mediated apoptosis in MCF10A epithelial cells84 whereas latrunculin B induced apoptosis in re-perfused rat kidney tissue.85 Latrunculin A treatment of MCF-7 cells was also shown to induce the translocation of Bax to the mitochondrial membrane where it was postulated to undergo oligomerization and mitochondrial membrane pore formation.84 Therefore the mechanism of action of distinct actin targeting drugs may involve the Bcl-2 family of apoptosis proteins. These reports however do not define the specific signaling pathways linking actin filament changes to Bcl-2 activation. Puthalakath and colleagues have demonstrated a direct link between the Bcl-2 family and actin-mediated apoptosis.86 Pro-apoptotic Bmf was found to be sequestered with actin-associated myosin V motors and upon cell detachment or cytochalasin D treatment is released from the cytoskeleton resulting in a mitochondrial-dependent apoptotic cascade (Fig. 2).86 The translocation of G-actin to the nucleus has been demonstrated in Latrunculin B treated mast cells87 and hepatocytes treated with protein synthesis inhibitors.88 Given the apoptotic effects of Latrunculin B in rat kidney tissue85 nuclear actin translocation could induce an apoptotic response. However a significant increase in apoptotic cells could not be detected in both studies suggesting that nuclear actin translocation was unable to induce apoptosis. As noted by Utsumi et al.,89 a role for tActin upstream of caspases has also been identified whereby tActin was subjected to post-translational N-myristoylation, targeting it to the mitochondria during apoptosis (Fig. 2).89 A possible explanation for the presence of actin in both the initial and final stages of apoptosis may involve a positive feedback loop. Initial disruption of the actin cytoskeleton may lead to downstream caspase activation which causes permanent actin filament fragmentation (t-Actin). N-myristolyated tActin may then be responsible for the amplification of apoptosis by inducing irreversible mitochondrial membrane damage. Slee and colleagues have demonstrated that feedback amplification can occur after apoptosis induction via the downstream cleavage of the pro-apoptotic protein Bid.90 Thus downstream events may play an important role in sustaining an apoptotic cascade and tActin may be a crucial player in this process. Actin filament dynamics are greatly dependent upon ABP regulation suggesting that ABPs may also play a role in actin-mediated apoptosis.
Several ABPs have been postulated as biomarkers of apoptosis due to alterations in their expression leading to cell death signaling pathways. The ABPs that have been studied in relation to apoptosis are ADF/Cofilin (actin dynamizing), thymosin β (actin sequestering), coronin-1 (actin branching), filamin (actin branching), gelsolin (actin severing and capping), tropomyosin (actin stabilizing) and myosin II (actin filament contraction or bundling).
The ADF/cofilin family regulate actin filament turnover by severing and depolymerizing existing actin filaments thus may increase the G:F-actin ratio.91 LIM and testicular kinases (LIMK and TESK I and II) phosphorylate ADF/cofilin at the Ser 3 residue inhibiting G- and F-actin binding92 whereas slingshot homolog (SSH) and chronophin (CIN) de-phosporylate ADF/cofilin activating cofilin.93 In relation to apoptosis, cofilin has been demonstrated to be translocated to the mitochondrial membrane in response to the kinase inhibitor staurosporin resulting in the release of cytochrome c and morphological hallmarks of apoptosis. Expression of a phophorylated (inactive) cofilin mutant abolished this mitochondrial targeting of cofilin emphasizing the requirement for active de-phosphorylated cofilin in apoptosis.94 Oxidation of cofilin by taurine chloramine similarly induced mitochondrial translocation of cofilin resulting in the opening of the mPTP and cytochrome c release.95 Mutations that removed any of the cysteine residues within cofilin inhibited mitochondrial targeting of cofilin and oxidant-induced apoptosis.95 Since oxidation of cysteine residues in cofilin resulted in the formation of intermolecular disulphide bonds96 intermolecular cysteine oxidation may be essential for the mitochondrial targeting of cofilin. Recent studies have identified novel cofilin residues that drive F-actin stabilization induced by nutritional depletion resulting in the accumulation of ROS, mitochondrial fragmentation and Ras hyperactivation.97 This supports the hypothesis that the actin cytoskeleton is an important biosensor of environmental stresses such as oxidative stress. Second, conserved positively charged residues on cofilin that are not actin binding were shown to be essential for respiratory function further highlighting the potential role of cofilin in sensing oxidative stress.97 The formation of actin-cofilin rods is a second apoptosis-related role for cofilin whereby ATP depletion resulted in the formation of short actin/cofilin rods.98,99 Actin-cofilin rods were able to prevent apoptosis by slowing mtΔΨ depletion in hippocampal neurons over a short period of time.100 However over an extended period of time, this protective mechanism was abolished resulting in rapid loss of mtΔΨ and subsequent apoptosis. Thus persistent actin-cofilin rods contribute to the loss of synapse activity in the neurons of patients suffering from neurodegenerative conditions. The short-term pro-survival role for cofilin in neurons specifically may be a biological conditioning mechanism to reduce the mitochondrial damage experienced by neurons affected by oxidation, micro-ischemia or glutamate excitotoxicity.98 Whether this short-term pro-survival role for cofilin exists in other cell types remains elusive. Cofilin has also been demonstrated to mediate the apoptosis of hippocampal neurons due to its activation by the scaffold protein RanBP9. Elevated levels of RanBP9 have been implicated in the production of amyloid β peptide which is known to cause neurodegeneration with cofilin expression being essential to RanBP9-mediated apoptosis.101
Thymosin β prevents polymerization by attaching to and sequestering G-actin.102 Elevated expression of thymosin β10 in ovarian tumor cells has been correlated with an increase in sensitivity to apoptosis. The presence of a second ABP, E-tropomodulin, inhibited apoptosis by competing with thymosin β10 for actin binding highlighting the inter-related dependency of ABPs in regulating actin-mediated apoptosis.103 Thymosin β10 was also shown to accelerate the apoptosis of fibroblasts by disrupting stress fiber formation which further supports the pro-apoptotic role of thymosin β10.104
Filamin promotes orthogonal actin branching which strengthens the cell membrane during cellular movement.105 Filamin cleavage by the T lymphocyte enzyme granzyme B induced an apoptotic response in Jurkat cells that was caspase-independent.106 In a separate study, filamin-mediated apoptosis of platelet cells was shown to be dependent upon caspase-3 activation in vivo.107 This contradictory result in regards to caspase dependency may reflect alternative effects based upon the type of apoptotic stress induced. The former study specifically looked at the physiological process of granzyme B activity whereas in the latter study exogenously expressed caspase-3 was utilized (non-physiological process). The utilization of physiologically relevant conditions is therefore important when studying apoptosis pathways.
Coronin-1 regulates the function of the actin nucleating and branching ABP Arp2/3 and is involved with lammelipodial formation required for cell motility.108 Knockout mouse studies demonstrated that coronin-1−/− cells show an impairment of T lymphocyte migration to the thymus due to an elevated level of apoptosis detected by annexin V staining.109 Elevated cytochrome c levels were also detected in coronin-1−/− T cells, suggesting that coronin-1 may regulate the survival of migrating cells such as T lymphocytes. A proteomics approach employed by Moriceau and colleagues further identified the presence of a cleavage product of coronin-1 after apoptosis induction suggesting that coronin-1 cleavage may be a downstream response to apoptosis signaling similar to actin.110 Expression of full length coronin-1 inhibited mitochondrial-mediated apoptosis of mature neutrophils further supporting the pro-survival role of coronin-1 in hematopoietic cells.110
Myosin II is an ATP-dependent non-muscle motor that interacts with actin filaments producing a contractile force that is essential during cell rounding and migration.111 Maintenance of myosin II tension is also crucial to the formation of the contractile ring during cytokinesis.112 Myosin II activity is regulated by the phosphorylation proteins such as myosin light chain kinase (MLCK) and Rho kinase.111 Studies have demonstrated a non-redundant role for myosin II phosphorylation in regulating apoptosis in endothelial and epithelial cells. TNFα, a regulator of extrinsic apoptosis, is also responsible for vascular endothelial barrier dysfunction. TNFα triggered the apoptosis of endothelial cells accompanied by the phosphorylation of myosin II leading to an increase in stress fiber formation and the appearance of para-cellular gaps indicative of endothelial barrier dysfunction.113 Inhibition of myosin II phosphorylation reduced TNFα-induced stress fiber formation and attenuated caspase-8 levels in vitro.113 As noted by Jin et al.114 myosin II may regulate TNFα mediated endothelial apoptosis by translocating TNF-receptor to the membrane surface. Further analysis of 3D microvessels revealed that vascular endothelial permeability occurred independently of Rho kinase activity implicating other regulatory elements (e.g., phosphoinositides and Ca2+) in actin/myosin II-dependent vascular permeability in vivo.115 Myosin II phosphorylation is also essential for the extrusion of apoptotic epithelial cells from the epithelial barrier during embryonic tissue development. UV irradiation of monolayer MDCK epithelial cells induced the formation of an actin-myosin ring around the edge of apoptotic cells indicative of cell rounding. As this ring of actin and myosin contracted, neighboring live cells moved into the space surrounding the dying cell thus closing the epithelial gap and extruding the apoptotic cell simultaneously. Rho kinase inhibition prevented the extrusion of the apoptotic cell highlighting the importance of myosin II phosphorylation to epithelial cell apoptosis and implicating cross-talk signaling between the actin cytoskeleton of the dying cell and the live neighboring cells.116
Gelsolin is a potent actin severing protein that caps the barbed-end of F-actin in the presence of Ca2+ preventing further barbed-end polymerization.117 Gelsolin has been implicated in apoptosis with caspase-3 activation producing an N-terminal gelsolin fragment (N-Gelsolin) with un-regulated actin filament severing capacity.118 As noted by Chhabra et al., N-Gelsolin specifically induced apoptosis by severing the G-actin:DNase I complex resulting in the nuclear localization and activation of DNase I.119 The mechanism by which N-Gelsolin releases G-actin bound DNase I remains unknown. N-Gelsolin has also been demonstrated as a pro-survival protein upstream of the mitochondria with its N-myristoylation preventing etoposide induced apoptosis.120 Elevated expression of gelsolin protected Jurkat cells from apoptosis induced by a variety of mitochondrial targeting agents121,122 and also prevented apoptosis in neuronal cells with enhanced actin stabilization abolishing this pro-survival effect.123 Silencing of gelsolin expression in Ras-mutated HCT116 colon cancer cells induced butyric-mediated apoptosis, via caspase activation further supporting the pro-survival role of gelsolin.124 Resistance to apoptosis was found to be driven by the capacity of gelsolin to inhibit the opening of VDACs, thus preventing mtΔΨ loss and downstream cytochrome c release.125 It is therefore hypothesized that gelsolin may protect against apoptosis in certain cell types (i.e., neurons, cancer cells), however this hypothesis has not been further explored. What remains certain is that caspase-3 activation releases a pro-apoptotic fragment of gelsolin which completely abolishes its pro-survival role at the mitochondrion and results in the release of DNase-1 from G-actin, but not in the presence of cofilin.119 Given that gelsolin regulation may involve other ABPs such as tropomyosin, the role of gelsolin in apoptosis may also depend on other proteins within the actin cytoskeleton.126
Tropomyosin is a dimerized helical polymer that winds around actin filaments providing structural stability and diverse functioning of actin filaments.127,128 Tropomyosin isoforms can be classified as high molecular weight (HMW) or low molecular weight (LMW) depending on the gene promoter utilized.129 Muscle tropomyosins specifically regulate myofibril contraction whereas non-muscle or cytoskeletal tropomyosins are known to regulate numerous cytoskeletal functions due to their spatial and temporal regulation.127,130 Cytoskeletal tropomyosins have been demonstrated to modulate the activity of other ABPs that are previously mentioned to be involved in apoptosis. Tm5NM1 expression in neuroepithelial cells was found to induce the recruitment of myosin IIA to stress fibers131 and simultaneously displacing ADF interaction with the actin filament.132 Conversely elevated levels of the HMW isoforms TmBr3 and Tm3 in neuroepithelial cells promoted ADF interaction with actin filaments resulting in the formation of filopodia which promote cell migration.131,132 This suggests that certain tropomyosin containing filaments are marked by specific ABP interactions which may be important in apoptosis. Anoikis is a specialized form of apoptosis that is activated when cells dependent on anchorage for survival (e.g., epithelial and endothelial cells) are placed in an anchorage-independent environment.133 Anoikis represents an important homeostatic function that prevents the migration of detached cells to a foreign location. Studies in mammary epithelial carcinomas have demonstrated that a significant downregulation in the HMW isoform Tm1 correlated with an increased resistance to anoikis perpetuating the survival of mammary carcinoma tissue in vitro.134 Restoring Tm1 expression in cultured mammary carcinoma cell lines (MCF-7 and MBA-MB231) led to the generation of distinct actin stress fibers and re-sensitized cells to anoikis.135 The reversion of Tm1 expression was Rho-kinase dependent and resulted in the appearance of more distinct cadheren/catenin containing cell-cell junctions thus enabling the cell to communicate with the extra-cellular environment.135,136 Tm1 can therefore act as an important sensor of the extra-cellular environment with unfavorable conditions leading to Tm1-mediated anoikis.
In summary ABPs are essential in regulating numerous key apoptotic processes such as cell rounding, membrane blebbing, caspase activation and mitochondrial membrane permeabilization. ABPs are also important in regulating specialized death pathways such as anoikis and epithelial cell extrusion. This further highlights the importance of actin filament dynamics in regulating apoptosis signaling via modulation of ABP function. As tumor cells have developed mechanisms to evade apoptosis, the transformed phenotype has been used extensively to further characterize the role of actin and ABPs in apoptosis signaling pathways.
The actin cytoskeleton is dramatically re-modeled upon cellular transformation permitting metastatic properties such as anchorage-independent growth and enhanced cell migration.137 A stabilized actin cytoskeleton has been demonstrated as an activator of Ras signaling resulting in apoptosis driven by the production of ROS and loss of the mtΔΨ.138 This discovery highlights the actin cytoskeleton as a trigger of Ras signaling and given the importance of Ras in tumorogenesis, actin may mediate tumor-associated processes such as cell migration via Ras signaling.138 ABPs such as tropomyosin, gelsolin and cofilin show varied expression profiles in both malignant cell types and in virally transformed cells implicating these proteins as potential biomarkers of malignancy. The exact phenotypic changes in ABPs associated with transformation remains complex. The expression of gelsolin in breast, urothelial and oral carcinomas has been described as biphasic with an early downregulation in gelsolin followed by a substantial elevation in gelsolin with prolonged metastasis.126,139,140 Cofilin overexpression has been confirmed in numerous tumor cell types with elevated cofilin levels potentially correlating with tumor cell migration.141,142 However a downregulation in cofilin was found in the highly metastatic hepatocellular carcinoma143 contradicting previous studies. A more defined role for cofilin in metastasis may be achieved by examining the entire output from the cofilin pathway which includes inhibitors such as LIMK I and II and stimulators such SSH and CIN phophatase.144,145 Several studies have demonstrated that increased expression of LIMK I contributed to the invasive capacity of prostate and breast cancer cells highlighting LIMK I as a potential malignant biomarker.146,147 Varying the local concentration of cofilin can significantly alters its function with high levels of active cofilin enhancing F-actin stability by stimulating actin nucleation rather than enhancing turnover and lower levels of cofilin favoring actin severing.148 This novel paradigm may also influence the role of the cofilin system in tumorogenesis. Downregulation of multiple HMW tropomyosins (Tm1, 2 and 3) has been detected in virally transformed chicken and rat fibroblasts,149,150 cultured human tumor cells136,151 and in tumor derived patient samples.134 Restoration of Tm2 and Tm3 increased the appearance of distinct microfilament bundles and adhesion proteins with Tm2 expression also restoring anchorage-dependent growth.152 This correlated with reports highlighting a role for Tm1 in anoikis.134,136 Further studies have demonstrated an increased reliance on LMW Tm5NM1 in transformed cells153 with reduced TM5NM1 expression leading to a reduction in melanoma cell motility.154 Together these studies suggest that certain ABPs (such as gelsolin, cofilin and Tm5NM1) have an important role in permitting tumorogenic hallmarks such as enhanced motility and anchorage independent growth. It is now well established that the evasion of apoptosis contributes to the prolonged survival and metastasis of tumor cells.155 Given that the actin cytoskeleton changes dramatically upon cellular transformation investigating apoptosis pathways in tumor cells may further elucidate a direct link between the actin cytoskeleton, apoptosis signaling and tumor cell survival.
De-regulation of apoptosis by oncogenic transformation is partly responsible for the ability of tumor cells to evade normal apoptotic signaling pathways.155 Overexpression of Bcl-2 and other pro-survival proteins has been detected in multiple tumor cell types (androgen independent prostate cancer cells, B pancreatic cancer cells, B cell lymphoma) and correlated with the prolonged survival of B lymphoid tumors.156 Furthermore, synergistic activation of the proto-oncogene cMyc and Bcl-2 accelerated lymphoma tumorogenesis.157 In contrast the loss of pro-apoptotic factors such as Bim and Puma has been correlated with the survival of tumor cells, with re-expression of Bim suppressing the activity of cMyc in leukemic cells158 and Puma re-expression increasing the sensitivity of melanoma cells to apoptosis.159 The generation of specific BH3-only mimetic compounds thus represents a potential anti-tumor therapy that could restore sensitivity of tumor cells to apoptosis. This type of therapy has been proven successful in vivo with the high affinity BH3 mimicking compound ABT-737 triggering Bax/Bak dependent apoptosis in a mouse lymphoma model.160 Indeed many cytotoxic therapies including DNA damaging agents exert their anti-tumor effects by inducing either extrinsic CD95 or Bcl-2 mediated apoptosis.161 However, in a panel of apoptosis resistant tumor cells, the activation of CD95 apoptosis permitted tumorogenic hallmarks such as cell invasion and migration through cellular barriers.162 This highlighted a novel anti-apoptotic role for CD95 in metastasis.163 Furthermore, Rac1 stimulated CD95 activity in developing neurons suggesting that Rho GTPases may regulate the invasive potential of CD95.164 Rac activation was found to be dependent upon association with ezrin suggesting that cellular transformation may convey a tumorogenic role for ezrin in actin-mediated tumor cell invasion.165 Thus ezrin may have a more global role in transducing extracellular signals to the actin cytoskeleton and that transformation alters this signaling pathway to promote tumor cell survival. Given the mounting evidence implicating the actin cytoskeleton in both apoptosis and tumorogenesis, targeting actin filaments represents an attractive anti-cancer therapy. Studies have demonstrated the anti-tumor effects of numerous actin targeting drugs both in vitro and in vivo. The cytochalasins were shown to inhibit cytokine stimulated melanoma cell motility in vitro166 and conferred an anti-proliferative effect in an in vivo mouse melanoma model.167 Jasplakinolide (actin stabilizing drug) has also been shown to possess a potent anti-proliferative capacity in a panel of prostate cancer cells accompanied by distinct apoptotic changes such as multi-nucleated cells and actin filament disruption.168 This study utilized phalloidin as its marker of actin filaments, however phalloidin F-actin binding is out-competed by the binding of Jasplaknolide.62 Therefore the observed actin filament disruption may be incorrect if filaments are already saturated with Jasplakinloide leading to the inhibition of phalloidin binding. More recently, latrunculin A has been demonstrated as an effective anti-tumor agent in both in vitro and in vivo models of gastric cancer.169 Due to the high sequence similarity between all actin isoforms indiscriminate targeting of the global actin filament population has hampered the success of these compounds in pre-clinical trials.170 To circumvent this problem, an ideal approach would be to target a sub-population of actin filaments involved in distinct functions such as cytokinesis or proliferation.171 Given the dramatic changes in ABP expression discussed previously, ABPs could be utilized as novel targets for chemotherapeutic drug design.172 More specifically tumor cells downregulate their HMW tropomyosins and show an increased reliance upon the LMW isoforms such as Tm5NM1/2. A novel strategy would be to target actin filaments containing Tm5NM1 to improve the specificity toward tumor cells.171 Inhibitors of LIMK I have also been postulated as a second target for actin-based chemotherapy because elevated expression of LIMK I was associated with the malignant phenotype.173 Given the high regulatory nature of the ADF/cofilin family and the conflicting expression patterns of cofilin in tumor cells, upstream targeting of LIMK may be a more effective strategy.
In the realm of chemotherapy, multi-drug resistance (MDR) is an emerging issue and changes in the actin cytoskeleton have been identified in MDR-specific cell lines. An altered actin cytoskeleton has been detected in a specific sub-population of MDR osteosarcoma cells with increased resistance correlating to the appearance of cells with distinct actin filament bundles.174 Furthermore, drug resistance in leukemic cells treated with the anti-microtubule drug vincristine produced fragments of actin that were identified via proteomic analysis.175 Reduced expression of the γ-actin isoform was specifically detected in leukemic cells that were resistant to another anti-microtubule drug vinblastine. Consequently a lower level of γ-actin in vinblastine-resistant cells correlated with a worse prognosis in relapse patients diagnosed with acute lymphoblastic leukemia.176 In relation to ABPs, mutational analysis of cofilin revealed the existence of positively charged surfaces that can regulate the activity of the drug transporter PDR1 without compromising mitochondrial function.97 This implicates cofilin as a potential player in the acquisition of multi-drug resistance with human cofilin CFL-2 being identified as a prognostic marker for non-small cell lung cancer drug resistance. Collectively, these reports suggest that changes in the organization of the actin cytoskeleton facilitate the survival of drug resistant cells resulting in a worse prognosis for cancer patients. Thus actin filament changes in MDR-specific tumor cells may represent a point of vulnerability in the actin cytoskeleton that could be treated by drug intervention.
Given the important role of the actin cytoskeleton in cellular homeostasis, it is not surprising that actin also has an important role in apoptosis. In the yeast system, the role of actin and cofilin in sensing oxidative stress has been well established. However in mammalian cells apoptotic mechanisms are more complex and defining a global role for actin in mammalian apoptosis remains challenging. It is clear that actin initiates and mediates mammalian apoptosis via the intrinsic and extrinsic pathways and final degradation of actin filaments amplifies the apoptosis signaling cascade. Actin dynamics is the crucial determinant of whether a cell succumbs to insult or resists with ABPs such as gelsolin, cofilin and tropomyosin conveying an important regulatory function in apoptosis. Given that the actin cytoskeleton significantly changes upon cellular transformation and the correlative changes in actin filament architecture in drug resistant cells, drugs that target MDR tumor-specific actin filaments could be utilized in combination with routine therapies to enhance their effectiveness in patients.
This work was supported by grants from the NH&MRC awarded to Peter W. Gunning and Justine R. Stehn. We acknowledge our major funding body ‘The Kids Cancer Project’ for supporting this work. Justine Stehn is supported by a Kids Cancer Project C4 Fellowship. Melissa Desouza has been awarded an Australian Postgraduate Award. We also thank Shane Whittaker for his assistance in the preparation of this review.
The authors of this review have no affiliation or financial involvement with any organization or entity with a financial interest in the subject material discussed in this review. No writing assistance was utilized in the production of this review.
Previously published online: www.landesbioscience.com/journals/BioArchitecture/article/20975