Abstract
Podocytes exhibit a unique cytoskeletal architecture that is fundamentally linked to their function in maintaining the kidney filtration barrier. The cytoskeleton regulates podocyte shape, structure, stability, slit diaphragm insertion, adhesion, plasticity, and dynamic response to environmental stimuli. Genetic mutations demonstrate that even slight impairment of the podocyte cytoskeletal apparatus results in proteinuria and glomerular disease. Moreover, mechanisms underpinning all acquired glomerular pathologies converge on disruption of the cytoskeleton, suggesting that this subcellular structure could be targeted for therapeutic purposes. This review summarizes our current understanding of the function of the cytoskeleton in podocytes and the associated implications for pathophysiology.
Form Follows Function—Composition and Structure of the Podocyte Cytoskeleton
Podocytes exhibit a complex cellular morphology characterized by a sophisticated cell polarity organization and an extensive network of protrusions which extend from a primary process into numerous secondary projections called foot processes. The latter not only surround glomerular capillaries but also interlink neighboring podocytes via slit diaphragms1–3 (see Figure 1). Because of the anatomic position on the outside of the glomerular capillary, podocytes are exposed to tremendous physical forces, which are required for efficient glomerular filtration. Hence, foot processes are not only involved in the formation of the filtration barrier, but are also essential for increasing the surface, thereby mediating efficient attachment to the glomerular basement membrane.4,5 Retraction and simplification of the podocyte foot process network (termed foot process effacement) is the common feature in all glomerular diseases.6,7 Early work from the 1980s employing immuno-gold–labeled electron microscopy demonstrated that primary and secondary processes have unique cytoskeletal characteristics and components. Primary processes are predominantly on the basis of intermediate filaments (IFs) and microtubules (MT), whereas secondary processes mainly rely on an actin cytoskeleton.8,9
Podocytes rely on a diverse cytoskeletal repertoire to maintain their complex morphology. (A) During development podocytes undergo a dramatic change in morphology, characterized by increasing basal specification ultimately leading to the complex podocyte foot process (FP) network. (B) Schematic depiction of the organization and distribution of the podocyte cytoskeleton: within primary processes mainly IFs and tubulin networks stabilize the morphology (insert, primary processes [PP]). Secondary processes are mainly composed of different actin cytoskeleton structures (insert, FPs, see also C). (C) In the upper panel, FPs are schematically characterized under physiologic conditions: at least two different actin networks are discernible, a central actin bundle within each FP and the cortical actin network beneath the plasma membrane (also connected to adhesion receptors). In the lower panel, FP effacement is depicted as a simplification and retraction of individual FPs. This process is accompanied by a redistribution of the aforementioned actin networks, characterized by the occurrence of a prominent actin accumulation parallel to the glomerular basement membrane (described also as an “actin mat”). Aside from reorganization of the actin network, adhesion receptors are also dissolved resulting in progressive detachment of FPs and podocytes.
The Microtubule Cytoskeleton
Microtubules (MTs) are oriented structures composed of α/β-tubulin subunits, characterized by a fast-growing (plus) end and a slow-growing (minus) pole.10 MTs are highly dynamic structures, involved in a variety of cell biologic processes such as migration and mitosis, as well as cilia formation.11–13 Early studies using in vitro cultured immortalized podocytes suggested that MT-motor proteins CHO/MKLP1 and protein phosphatase 2A (PP2A) are involved in podocyte primary process formation and orientation.14,15 Given the peculiar similarity in terms of cellular morphology between polarized neurons and podocytes (prominent cell body, polarized protrusion network1), it is of special interest that a series of mutations involving certain tubulin isotypes result in neurodevelopmental disorders with a variety of phenotypes, ranging from cortical malformations to degeneration of axons.16–18 Interestingly, in none of those reported genetic syndromes was a kidney or particular podocyte phenotype reported. To date, there is one study employing a mouse model mimicking human TUBB2B mutations (resulting in asymmetric polymicrogyria and epilepsy19) reporting glomerular maturational defects.20 However, this model shows embryonic lethality and ultrastructural studies revealed that despite impaired podocyte maturation (cuboidal shape, occludens type junctions), a podocyte foot process network can be formed. Certainly, more work is needed to clarify the structural effect of the tubulin cytoskeleton for podocyte processes. Aside from the core MT cytoskeleton, there is an array of MT-associated proteins (MAPs) involved in stabilizing, bundling, or severing of MTs. A series of studies focused so far on the expression of different MAPs such as MAP4, MAP3, and lately MAP1B in podocytes.21–23 Although MAP1B showed a high expressional enrichment in podocytes and colocalized with the MT network, the constitutive knockout for MAP1B did not result in any overt glomerular phenotype under physiologic conditions (in contrast to neurons23).
Intermediate Filaments
The second group of cytoskeletal constituents particularly detected within primary processes is the class of intermediate filaments (IFs). In contrast to the actin and MT cytoskeleton, which is composed of single protein type polymers, IFs represent a heterogeneous group organized due to structural homologies into six classes. On the basis of immunohistochemistry studies on various murine experimental models, as well as human samples, it was unequivocally demonstrated that podocytes express the class III type IF vimentin.24,25 Moreover, several studies attempted to characterize alterations in the expression pattern of detected IFs under diseased conditions, showing that the typical muscle cell marker desmin appears to be upregulated upon podocyte stress or damage.26,27 So far, there is only limited evidence directly demonstrating the effect and role of vimentin and desmin for podocyte function, whereas IFs have been demonstrated to regulate the cellular elasticity in isolated glomeruli.28 Surprisingly, knockout models for either desmin or vimentin do not, however, result in any podocyte phenotype.29,30 The pronounced change in the expression pattern of IFs (re-expression of desmin) is generally accepted as a surrogate marker for a phenotypic switch of podocytes under pathologic conditions, often termed as epithelial-mesenchymal transition.31,32 However, one should be cautious to use this term for what rather represents a loss of podocyte differentiation in response to injury.33 Aside from typical mesenchymal IFs (class III), podocytes also express the class VI IF nestin, usually described as a characteristic marker of neuronal stem/progenitor cell populations and also found in several cancer entities.34,35 From studies using neuronal progenitor cells and the nestin knockout model, a cytoprotective, antiapoptotic function was deduced.36,37 This beneficial function is at least partially mediated via an interaction between nestin and the cyclin-dependent kinase 5 (CDK5).36 A similar observation was recently made in podocytes, especially when challenged under high glucose conditions.38–40 In summary, IFs do not appear as predominant stabilizing cytoskeletal components of the podocyte, but might rather be involved in signaling functions or intracellular sorting/transport processes as shown in different other cell types.41
Actin Cytoskeleton
In contrast to the two aforementioned cytoskeletal components, the actin cytoskeleton that constantly undergoes polymerization and severing, shows a rather exclusive localization within secondary processes (FPs).9 Actin filaments are composed of globular actin monomers and the efficient polymerization process involves ATP usage.42 Several polymerizing and severing proteins are known (before excellent review see42) to modulate the composition of the actin cytoskeleton. The ARP2/3 complex plays a particular role, because this molecular machinery efficiently introduces new branches on pre-existing filaments, a process essentially required for the formation of lamellipodia in vitro.43 Early studies using electron microscopy could reveal that even within podocyte secondary processes apparently different actin pools might be present. In the center of each FP an electron dense actin accumulation was detected and interpreted as the correlate of central actin bundles.44,45 Beneath the curved plasma membrane of FPs a dense actin network was described and assumed to represent branched subcortical actin meshwork (Figure 1). On the basis of these observations it was proposed that the central actin bundles might be involved in generating tensional forces, whereas the branched meshwork is essentially required to maintain the complex morphology of podocyte foot processes.45 Foot process effacement as the common pattern of podocyte injury involves active remodeling of the actin cytoskeleton46 often including the formation of a flat actin mat that is thought to provide increased adhesion forces.46 Recent in vitro studies advanced our understanding of the architecture of the actin cytoskeleton.47,48 However, advances in new super resolution microscopy techniques such as stimulated emission depletion or stochastic optical reconstruction microscopy now offer new avenues to fully understand the complexity, composition, and dynamics of the podocyte actin cytoskeleton in vivo.49
Additional insights into the structural and functional relation of FPs have come to light through the study of dramatic changes in podocyte morphology which take place during glomerular development, where cuboidal nonfiltering podocytes differentiate into arborized octopus-like filtering cells50 (Figure 1). In this context, a series of studies demonstrated the requirement of the aPKC/PAR complex for correct podocyte morphogenesis.51–53 The importance of podocyte polarity signaling is also supported by the high level of redundancy of the involved molecules.54–56 In addition, loss-of-function mutations in the gene encoding the polarity protein CRUMBS2 have been shown to result in steroid-resistant nephrotic syndrome.57
From Genes to Function—Lessons from Genetic Syndromes Affecting the Podocyte Cytoskeleton
The list of causal mutations for glomerular diseases has rapidly expanded over the last few years, driven by the development of accurate sequencing platforms. Not surprisingly, cytoskeletal proteins account for a significant proportion of the proteins that are involved in hereditary glomerular diseases (see Figure 2, Table 1). One of the first discovered cytoskeletal linked proteins causing specific podocyte dysfunction was actinin-4 (ACTN4).58 Whereas α-actinin is required to link actin filaments to Z-lines in striated muscles,59 the so-called nonsarcomeric actinins, like actinin-1 and ACTN4, are differentially expressed in nonmuscle cells where they crosslink stress fibers. Interestingly, mutations in ACTN4 cause a wide range of functional impairments ranging from decreased protein stability,60 gain-of-function, subcellular mislocalization, and increased F-actin binding capacity.61 These alterations translate into impaired cytoskeletal dynamics characterized by decreased cellular spreading and attenuation of podocyte migration.62 In line with the recent discovery of mutant ACTN4 forms showing decreased protein stability and protein levels, lack of ACTN4 in mice also results in proteinuria and an FSGS-like phenotype.63 In addition, ACTN4 mutations may also have nonstructural functions such as transcriptional regulation contributing to glomerular disease.64
Interacting signaling cascades determine podocyte cytoskeletal stability and cell-matrix interaction. (A) Various modulators of the podocyte cytoskeleton and their potential therapeutic approach (yellow indicates proteins with a proven genetic disease background; red indicates available drugs or pharmacologic drug classes). (B) The podocyte-specific FA component EPB41L5 titrates actomyosin contractility via recruitment of the small GTPase ARHGEF18, thereby influencing podocyte adhesion and attachment. Actn4, Actinin-4; ANLN, Anilin; EPB41L5, Erythrocyte Membrane Protein Band 4.1 Like 5; FAK, Focal Adhesion Kinase; INF2, Inverted formin 2; ITGα3β1, Integrin α3β1; ROCK, Rho-associated protein kinase.
Cytoskeleton-associated mutations related to SRNS/FSGS phenotypes
The formin protein family member inverted formin 2 (INF2) is another cytoskeletal gene prominently involved in podocyte disease.65 Formin proteins are key regulators of actin polymerization.65 Interestingly, the majority of identified mutations within INF2 are localized in the autoinhibitory domain (diaphanous inhibitory domain), implying increased activity of this protein causes glomerular dysfunction.66 Moreover, a recent study demonstrated that INF2 interacts with the actin-modifying proteins profilin and the capping protein CAPZ, whereas INF2 mutations impaired these interactions.67 Interestingly, INF2 mutations also lead to complex neurologic disease phenotypes such as Charcot–Marie–Tooth neuropathy.68 Because INF2 might also modulate mitochondrial dynamics,69 these observations imply the presence of a diverse range of disease-associated mechanisms that selectively operate in a cell-specific manner.1,4,70
Further genetic studies demonstrated that missense mutations within the actin binding protein anilin segregate with human FSGS.71 Mechanistically, these mutations interfered with the interaction of anilin with CD2AP. CD2AP mutations have also been reported in rare cases of human genetic FSGS72–75 and mouse studies definitively demonstrated the requirement of CD2AP activity in podocytes.72,76 Initially identified as a ligand interacting with T cell CD2 receptor,77 CD2AP supports several functions in podocytes ranging from slit diaphragm signaling78 to glucose transporter trafficking.79 Furthermore, it was shown that CD2AP plays a role in modulating podocyte TGFβ response in order to prevent a proteolytic program that would culminate in decreased levels of essential cytoskeletal components such as dynamin and synaptopodin.80 A similar cellular phenotype was reported in the case of missense mutations of ARHGAP24, a small GTPase modulator which is specifically expressed in podocytes. Recent work showed that these mutations within ARHGAP24 resulted in increased membrane ruffling dynamics, a phenotype which is thought to partially reflect podocyte foot process effacement in vitro.81
Balance Is the Key—Pathophysiologic Concepts of Cytoskeletal Dynamics
Regulatory networks of intrinsic and extrinsic signaling factors tightly control the podocyte cytoskeleton with minor changes shifting the system toward podocyte disease. This was exemplified by mouse models where combined genetic heterozygosity of Cd2ap, Synpo, Fyn proto-oncogene (Fyn), and Neph1 was tested.82 Although heterozygosity of these genes alone did not cause any early-onset phenotype, compound heterozygosity (i.e., combinations of heterozygous mutations) resulted in FSGS-like glomerular disease suggesting synergistic effects of genetic modifiers to reach the threshold of podocyte disease.82
Podocytes, with their complex interdigitating network of foot processes, are accepted as essentially stationary cells, at least under normal physiologic conditions. This concept was recently supported by elegant in vivo two-photon imaging studies employing zebrafish larvae as a model system.83 Also, under prolonged observation no motile phenotype of podocytes was noted. In contrast, experimental studies performed in mice with the same technical approach could visualize a migrating podocyte phenotype under stressed conditions (unilateral ureteral obstruction and adriamycin nephropathy).84 On the basis of in vitro studies, a transition to a motile state has been considered as a pathologic condition to explain the effects of foot process effacement. Treatment of podocytes with nephrotoxic substances such as puromycin aminonucleoside resulted in increased migration.85 Furthermore, more recent studies proposed that rapid oscillations at the leading edge of the cell may reflect cytoskeletal instability, which normally has to be controlled to maintain the complex morphology of podocyte foot processes.81,86 In previous studies the GTPases RhoA, RAC1, and CDC42 were identified as the molecular machinery defining podocyte cytoskeletal stability.87–89 By deleting podocyte GTPases (specifically each GTPase), it was demonstrated that only CDC42 loss resulted in a profound podocyte phenotype, characterized by massive proteinuria and dramatic ultrastructural abnormalities. On the contrary, loss of RhoA and RAC1 did not result in any overt phenotype, at least under physiologic conditions.88,89 However, knowledge is emerging that titration and balancing of those GTPases appears to play a major decisive role in maintaining podocyte integrity, because, for example, either increased or decreased RAC1 activity lead to proteinuria and podocyte foot process effacement.87,90 This line of thought was substantiated by the identification of causative mutations in ARHGDIA in cases of familial nephrotic syndrome.91 ARHGDIA belongs to the Rho GDI (GDP dissociation inhibitor) family, which ensures that Rho family members are kept in their inactive GDP-bound form within the cytosol. Inhibition of RAC1 overactivation in Arhgdia null mice could suppress the albuminuria phenotype,92 once more emphasizing the importance of balanced cytoskeletal dynamics for podocyte integrity.
Don’t Lose Your Grip—Adhesion as a Key Determinant of Podocyte Function in Health and Disease
Podocyte loss is a key feature in the progression of most glomerular diseases. Because apoptosis is very rarely observed in vivo, the critical step toward podocyte loss can be attributed to the failure of podocyte adhesion.5,93 Integrin heterodimeric receptors represent essential adhesion molecules, exhibiting an exquisite tissue and developmental expression profile.94 The importance of integrin receptors was demonstrated by the identification of mutations within ITGα3 resulting in glomerular and skin disease.95,96 Together with a complex array of adaptor and linker molecules, integrin-based adhesion sites form focal adhesions (FAs).97. FAs are composed of a multiprotein complex consisting of different receptor classes, linker molecules, GTPases, kinases, and phosphatases, collectively known as the adhesome.98,99 FAs not only provide physical linkage via connection of cells to the underlying GBM or ECM, but serve also as signaling hubs by integrating physical and biologic cues and thereby influencing a variety of cellular functions ranging from cell migration to metabolism.97 This concept of reciprocal exchange via FAs was termed outside-in/inside-out signaling and extensively studied in the field of podocyte research (for excellent review see 100). Among the different integrin subunit types, ITGβ1 shows the highest expression in podocytes and a series of studies employing animal models demonstrated the critical role for podocyte maintenance and early development.101,102 More recently, integrin receptors have attracted attention as potential druggable targets in glomerular disease. For example, it was suggested that the induction of B7–1 in podocytes results in a diminished activation of ITGβ1 via competition with talin. On the basis of those assumptions, a recent study tried to treat B7-1 positive patients with FSGS with the CTLA4-modulator abatacept.103 In addition to ITGβ1, increased levels of ITGβ3 activation were identified to play a potential pathogenic role in suPAR-mediated proteinuria.104 Through a series of complementary approaches it was shown that either blocking ITGβ3 activity or inhibiting suPAR-ITGβ3 interaction resulted in a protective effect.104 Although our knowledge about basic constituents of the podocyte adhesome is growing, there is still an urgent need to define the exact composition and modification of this complex in order to identify more specific druggable targets and reliable diagnostic techniques. By combining in vivo quantitative MS analysis with complementary in vitro adhesomics, major progress has recently been made.105 This approach allowed for the identification of a highly selectively expressed FA component of podocytes, the FERM-domain protein EPB41L5. Mechanistically, EPB41L5 recruits the small GTPase ARHGEF18 thereby titrating the activity of the actomyosin cytoskeleton and efficient podocyte adhesion (see Figure 2). Given the highly specific expression, EPB41L5 might serve as a new entry point for innovative diagnostics or therapeutic avenues in the future.
The Actin Cytoskeleton as a Therapeutic Target
The common hallmark of podocyte disease is the uniform stress response characterized by retraction and overt simplification of podocyte foot processes. Some investigators interpret this morphologic change as an adaption of podocytes to withstand altered filtration forces and to increase their adhesiveness,7 but this hypothetic concept is still awaiting experimental confirmation. However, it is accepted that changes in actin cytoskeleton architecture might represent the underlying molecular pattern for this tremendous morphogenetic transformation. Consequently, a recent study demonstrated that stabilizing the actin cytoskeleton is of therapeutic value.106 The same authors identified previously the large GTPase dynamin as an essential component of podocyte function and target in proteinuric kidney diseases, cleaved by cathepsin-L.80,107 Dynamin is itself able to induce actin polymerization.80 The application of a small compound, Ringo or BisT23, stabilizes the conformation of dynamin and thereby prevents the degradation of dynamin even in conditions of high levels of cathepsin. The value of this approach was demonstrated by the successful prevention or at least amelioration of proteinuria in toxic as well as genetic models of podocyte damage.108 Nevertheless, the time of treatment in this preclinical study was of course limited and long-term toxicity effects, as well as side effects, on the organism have to be addressed in future studies.
Conclusion and Outlook
Over the last decades our knowledge concerning the structure and function of the podocyte cytoskeleton increased tremendously. Nevertheless, a fundamental issue is how to translate this insight into treatments for glomerular diseases. And finally, we need a more detailed understanding of how the cytoskeleton itself regulates critical cellular processes such as localized transcription, vesicle trafficking, mitochondrial dynamics, and even nuclear transcription. Better understanding of underlying mechanisms could help in the design and rational development of targeted therapeutic approaches.
Disclosures
None.
Acknowledgments
We apologize to all authors whose work was not cited due to space limitations. We appreciate careful reading of the manuscript by Prof. Ketan Patel, University of Reading, UK, and thank all members of the Huber laboratory for their insightful discussions.
This study was supported by the German Research Foundation (DFG): Collaborative Research Centres 1140 (to T.B.H.) and Collaborative Research Centres 992 (to T.B.H.); the Heisenberg program (to T.B.H.), HU 1016/5-1 and HU 1016/8-1 (to T.B.H.); the European Research Council (ERC grant 616891 to T.B.H.); the Horizon 2020–Innovative Medicines Initiative consortium Biomarker Enterprise To Attack Diabetic Kidney Disease (115974, to T.B.H.); the Bundesministerium für Bildung und Forschung–STOP-Focal Segmental Glomerulosclerosis 01GM1518C (to T.B.H.); the Excellence Initiative of the German Federal and State Governments (Center for Biological Signaling Studies to T.B.H. and the Freiburg Institute for Advanced Studies to T.B.H.); the German Society of Nephrology (Deutsche Gesellschaft für Nephrologie to C.S.); the Berta-Ottenstein program, University of Freiburg (to C.S.); and the Else Kröner Fresenius Stiftung, Nierenfunktionsstörungen als Komplikation von Systemerkrankungen (to C.S. and T.B.H.).
Footnotes
Published online ahead of print. Publication date available at www.jasn.org.
- Copyright © 2017 by the American Society of Nephrology