TRPC6 channels and their binding partners in podocytes: role in glomerular filtration and pathophysiology

Abstract

Loss or dysfunction of podocytes is a major cause of glomerular kidney disease. Several genetic forms of glomerular disease are caused by mutations in genes that encode structural elements of the slit diaphragm or the underlying cytoskeleton of podocyte foot processes. The recent discovery that gain-of-function mutations in Ca²⁺-permeable canonical transient receptor potential-6 channels (TRPC6) underlie a subset of familial forms of focal segmental glomerulosclerosis (FSGS) has focused attention on the basic cellular physiology of podocytes. Several recent studies have examined the role of Ca²⁺ dynamics in normal podocyte function and their possible contributions to glomerular disease. This review summarizes the properties of TRPC6 and related channels, focusing on their permeation and gating properties, the nature of mutations associated with familial FSGS, and the role of TRPC channels in podocyte cell biology as well as in glomerular pathophysiology. TRPC6 interacts with several proteins in podocytes, including essential slit diaphragm proteins and mechanosensitive large-conductance Ca²⁺-activated K⁺ channels. The signaling dynamics controlling ion channel function and localization in podocytes appear to be quite complex.

the discovery in 2005 that mutations in canonical transient receptor potential-6 (TRPC6) channels can cause familial focal segmental glomerulosclerosis (FSGS) (92, 128) brought the analysis of ion channels in podocytes to the forefront of experimental nephrology. Ion channels rarely if ever operate in isolation; in most systems, a number of different channels operate in positive and negative feedback loops to regulate various aspects of cell physiology. TRPC6 channels in podocytes interact with a number of different proteins, including scaffolding molecules, signaling proteins, cytoskeletal elements, and other types of ion channels. Here, we review literature on TRPC6 channels and their binding partners, which include mechanosensitive large-conductance Ca²⁺-activated K⁺ channels (BK_Ca channels), as well as a number of other proteins essential for the normal function of the glomerular slit diaphragm. We also review studies showing that TRPC6 channels in podocytes may affect the cytoskeleton, gene expression, and ultimately glomerular filtration. We begin with a brief overview of podocyte biology.

Role of Podocytes in the Regulation of Glomerular Filtration

The initial steps in renal function entail ultrafiltration of blood in the glomerulus, which functions as a sieve that is permeable to water and small solutes, but that largely excludes macromolecules on the basis of size and charge from the urinary spaces in the lumen of Bowman's capsule. The glomerular filtration barrier is composed of several elements, including the highly fenestrated capillary endothelium, the glomerular basement membrane (GBM) that covers the outer surface of the capillary, and a population of specialized cells called podocytes. In addition, mesangial cells play a role in the process of glomerular filtration by regulating the overall mechanical responses of glomeruli to changes in hydrostatic pressure (23, 30).

Podocytes are highly differentiated cells that extend primary processes that ramify and terminate in specialized structures called foot processes. The podocyte foot processes attach to the GBM in a regularly spaced interdigitating pattern such that a series of filtration slits are formed between adjacent foot processes. These slits are bridged by specialized cell junctions known as slit diaphragms (SDs), whose precise structure and location are essential for normal glomerular filtration (92, 96, 98, 124). Additional specialized contacts between podocyte foot processes and the GBM are also crucial for normal glomerular function, as changes in the pattern of these contacts and in the structure of the foot processes, the most common form of which is known as podocyte effacement, lead to altered glomerular function and proteinuria (107). Podocyte cell bodies and primary processes also contact the GBM by a series of structures called anchoring processes that are structurally similar to foot processes (20, 81) and which give rise to a distinct subpodocyte space (SPS). The SPS in turn maintains contact with the rest of the urinary space in Bowman's capsule through small pores that are few in number (82).

The discovery of glomerular SDs suggested an intuitively simple model of glomerular function in which filtration occurs through a series of increasingly refined sieves. In this model, capillary fenestrations and the GBM prevent permeation of cells and large macromolecules, thereby acting as a type of “prefilter,” whereas the SDs represent the final size-selective barrier to the permeation of macromolecules by excluding proteins with Stokes-Einstein radii comparable to that of albumin (96). In its most basic form, this model implies a static and passive role for podocytes. The underlying glomerular capillary has unusually high pressure compared with other capillary beds, and it has been proposed that podocytes generate force to dynamically counteract distension of the capillary wall that would otherwise lead to protein leakage (64–66). Podocytes were originally predicted to be contractile based on the arrangement of F-actin, myosin, and α-actinin in foot processes (16, 45, 65, 66). There is now direct evidence that podocytes contract in response to a variety of mechanical stimuli and fluid stresses in vitro (21, 24).

Recent ultrastructural observations on the SPS have required some revision to the simple picture of a three-element glomerular filtration barrier. The SPS is defined as the space formed between the GBM and the cell body and primary processes of the podocyte. It covers ∼60% of the glomerular filtration surface, and its boundaries define a highly convoluted pathway that eventually feeds, via a marked constriction called the SPS exit pore, into larger compartments that feed the primary filtrate into the tubules (81, 82, 99). In effect, the SPS adds additional layers with their own resistances and compliances to the overall process of glomerular filtration. Measurements of the dimensions of the SPS have suggested that the SPS is highly restrictive to fluid flow (82), a prediction that has been confirmed by in vivo imaging (99). The hydrodynamic properties of the SPS provide a mechanism for the podocyte to rapidly regulate the overall resistance of the glomerular filtration barrier by changing the volume, pressure, and compliance of this space and by controlling the resistance of the SPS exit pore (82). An important feature of this model is that the highly restrictive SPS could allow for transient flow reversals across the glomerular filtration barrier, which could serve to clean the filtration barrier of clogging proteins (82, 99). The SPS also provides a mechanism for the podocyte to sense and respond to changes in pressure within the urinary space, or changes in glomerular filtration rate, and to make these parameters highly sensitive to any dynamic behavior of the podocyte itself, i.e., the entire podocyte, not just the foot processes (99). By analogy to other cell types, such as neuronal dendritic spines, it is reasonable to suppose that dynamic responses in podocytes and their foot processes are mediated by cytoskeletal elements and Ca²⁺-dependent processes (21, 44, 78, 93). The discovery that mutations in TRPC6 channels can cause familial FSGS (94, 128) underscored the potential importance of Ca²⁺ dynamics for podocyte function and the dynamics of glomerular filtration and opened an entirely new line of investigation.

TRPC Channel Family

TRPC channels comprise a family of nonselective Ca²⁺-permeable cation channels that are widely expressed in vertebrate tissues. There is a general consensus that members of the TRPC family become active in response to signal transduction cascades associated with stimulation of phospholipase C (PLC). These types of cascades lead to hydrolysis and local membrane depletion of phosphatidylinositol-4,5-bisphosphate (PIP₂), thereby liberating diacylglycerol (DAG) and inositol 1,4,5-trisphosphate (IP₃) (116). While all members of the TRPC family are thought to play a role in PLC-mediated signaling cascades, different members of the TRPC family may contribute to these cascades in different ways that are not well understood. TRPC channels are nonselective cation channels, and therefore their activation results in cell depolarization. The fact that TRPC channels are finitely permeable to Ca²⁺ means that their activation may be sufficient, at least under certain conditions, to modulate a host of secondary cascades, resulting in integration or amplification of cell signals.

There are seven functional members of the TRPC channel family in mice (TRPC1–7), and six in humans, in which TRPC2 is a pseudogene (119). The TRPC channels are members of a larger channel family known as TRP channels that have a broad role in chemo- and mechanosensation and in allowing cells to sense changes in their local environment (19). Based on amino acid sequence homologies, functional human TRPC channels fall into three groups: 1) TRPC1; 2) TRPC4 and TRPC5; and 3) TRPC3, TRPC6, and TRPC7. TRPC channels can form heteromers and homomers and thereby have the potential to form a very large number of unique ion channels (70).

Minimally functional TRPC channels are tetrameric proteins composed of subunits that contain a membrane domain with six transmembrane segments, a pore-forming region between the membrane segments 5 and 6, and with NH₂ and COOH termini facing the cytosol. This arrangement is shown schematically in Fig. 1. A variable number of ankyrin repeats, four in the case of TRPC6, are present near the NH₂ terminus and are followed by an NH₂-terminal coiled-coil domain. All of the TRPC channels have a highly conserved 25-residue TRP domain immediately distal to the transmembrane segments, which includes a motif known as TRP box 1, followed by a proline-rich sequence known as TRP box 2. The TRP domain is followed by a calmodulin and IP₃ receptor-binding region (CIRB region) and a COOH-terminal coiled-coil domain (119). Additional calmodulin-binding domains have been identified in both the NH₂ and COOH termini of some members of this family, including TRPC6 (134). Ankyrin repeats, coiled-coil domains, and proline-rich domains often play a role in protein-protein interactions, and the domain structure of TRPC channels raises the possibility of extensive associations with other proteins and cytoplasmic factors. Interactions between members of the TRPC family, including formation of homomeric complexes, are known to involve both the NH₂ and COOH termini (70).

Fig. 1.

A: domain structure of transient receptor potential-6 (TRPC6) channels. Boxes indicate human mutations known to cause focal segmental glomerular sclerosis (FSGS). All of the known disease-causing mutations are located in the cytoplasmic NH₂ and COOH terminals of TRPC6. B: immunogold localization of TRPC6, TRPC1, and TRPC5 to podocyte foot processes (arrows). The micrograph of TRPC6 expression also shows that it is present along a primary process.

Activation and Modulation of TRPC Channels

The mechanisms underlying activation of TRPC channels are complex and incompletely understood. Moreover, their gating properties depend on whether the channels are part of a heteromeric complex containing other TRPC members, scaffolding proteins, or even members of other ion channel families. For these reasons, studies of TRPC channels carried out in heterologous expression systems need to be interpreted with caution. Under certain conditions, TRPC1, TRPC4, and TRPC5 can become active as a result of Ca²⁺ mobilization from intracellular stores mediated by IP₃ (129). Moreover, knockdown of TRPC4 reduces store-operated Ca²⁺ current in mouse mesangial cells (122). By contrast, TRPC3, TRPC6, and TRPC7 are not activated by IP₃ in heterologous expression systems and instead become active after application of DAG analogs in excised inside-out patches (29, 36). On this basis. it has been suggested that TRPC6 channels are components of signaling cascades that are delimited to the plasma membrane and that the gating of these channels is independent of Ca²⁺ store depletion (12, 36). There is also evidence that TRPC6 channels are inhibited by PIP₂ and that local depletion of this phospholipid may play a role in regulating TRPC6 gating (69).

Recent findings on TRPC-interacting proteins suggest that the gating properties of TRPC family members can vary in different cell types depending on whether certain interacting proteins are expressed. For example, depletion of intracellular calcium stores failed to activate murine TRPC4 and -5 channels expressed in HEK293 cells (100). TRPC1, TRPC2, TRPC4, and TRPC5 are capable of interacting with a protein known as stromal-interacting molecule 1 (STIM1) (129, 131) that can sense the Ca²⁺ content of intracellular stores (73, 97). STIM1 binding to TRPC1, TRPC4, and TRPC5 can cause them to function as Ca²⁺ store-operated channels, possibly by conferring Ca²⁺ sensitivity to their gating, or by simply tethering TRPC channels to the Ca²⁺ release zone on endosomes (28). STIM1 does not bind to TRPC3 or TRPC6, but it has the potential to alter their behavior indirectly, by promoting their heteromerization with TRPC4 and TRPC5 (131). This raises the possibility that differences in STIM1 expression can underlie markedly different gating behaviors of TRPC channels in different cell types. It is not known whether STIM1 is expressed in podocytes, but such a discovery would be quite significant.

As already noted, the CIRB region of TRPC3 channels can interact physically with IP₃ receptors inside cells (62), and it has been suggested that TRPC3 channels can be gated by a conformational coupling mechanism (61). Therefore, whether an endogenously expressed TRPC channel operates as a Ca²⁺ store-operated or a receptor-operated plasma membrane channel may also depend on its own level of protein expression in a cell. Gating of TRPC3 by mobilization of Ca²⁺ stores generally occurs only at high expression levels (62, 71, 118) and may also depend on other proteins that regulate their subcellular localization. In podocytes, the role of a TRPC channel may also depend on whether the channels are expressed in foot processes or in the cell body, as endoplasmic reticulum is only found in the later compartment (51).

The potential functional variability that results from formation of TRPC heteromers and from interactions with other proteins complicates interpretation of pharmacological studies of the TRPC channel family. This is because drugs that are selective for a particular TRPC channel overexpressed in a heterologous expression system may have quite different specificities for native TRPC channels expressed in tissues. At this time it is safest to conclude that there is no pharmacological agent that can function as a selective activator or inhibitor of TRPC channels, either collectively or individually. Even experiments with small interference (si) RNA-mediated knockdown and gene knockout have to be interpreted with caution, as there can be compensatory (and even overcompensatory) increases in expression of other members of the family (14, 105).

These considerations are relevant to podocytes because multiple members of the TRPC family are expressed in these cells. TRPC3 has been detected in immortalized mouse podocyte cell lines (53). In addition, TRPC1, TRPC2, and TRPC5 were observed in mouse glomeruli in a pattern consistent with expression in podocytes (94). In Fig. 1 we show immunogold labeling of TRPC6, TRPC5, and TRPC1 in mouse foot processes. It is also possible to see TRPC6 expression in the primary processes. Therefore, it is possible that at least some of the TRPC6 channels in podocytes are components of a heteromeric complex and the TRPC6 subunits may have different functions in different cellular compartments.

There are several other mechanisms for TRPC regulation that may be important in podocytes. In general, these operate over longer time scales. TRPC6 channels normally show very low levels of constitutive activity, a feature that depends on its dual glycosylation at sites on extracellular loops (13). Changes in glycosylation could affect functions regulated by TRPC6 (14). In addition, TRPC6 channels are regulated by direct phosphorylation, for example by Src-family tyrosine kinases, which causes an increase in DAG-evoked TRPC6 activation in excised patches (35). TRPC6 can also be phosphorylated by cAMP-dependent protein kinases (31), although the functional significance of this is as yet unknown. The functional activity of TRPC channels (26), including TRPC6 (7), can be regulated by rapid insertion and removal from the plasma membrane. Given this, it is interesting that TRPC5 and TRPC6 coimmunoprecipitate with dynamin (25), which plays a role in membrane protein trafficking and cytoskeletal regulation in many cell types, including podocytes (106). Finally, there are reports that TRPC6 gating is mechanosensitive owing to effects of membrane curvature on the gating of the channels (17, 112). This mode of gating is not seen in every expression system and may therefore depend on other interacting proteins (27).

TRPC6 Channel Permeation

TRPC channels are neither efficient nor selective Ca²⁺ channels. The most extensive analyses of TRPC permeation have been carried out on TRPC6 channels. When heterologously expressed in HEK293 cells, TRPC6 forms cation channels that can be activated by DAG analogs. These channels show marked outward and inward rectification with physiological ionic gradients (22, 47, 108). Increasing external Ca²⁺ concentration causes some reduction in current through TRPC6 channels (22, 47, 108), especially at more depolarized membrane potentials, at least in part because of pore blockade. This blockade can be accurately described with a standard single-site, two-barrier pore model (22). In this model, TRPC6 channels have a single binding site for cations located ∼85% of the distance across the transmembrane electric field. This site has a twofold higher affinity for Ca²⁺ than it does for monovalent cations, and occupation of this site by a permeant ion results in pore blockade (22). More importantly, as a result of Ca²⁺ absorption to this site, membrane depolarization markedly reduces the ratio of Ca²⁺ permeability to that of Na⁺ (the P_Ca/P_Na ratio), effectively causing TRPC6 to function as a monovalent cation channel. In other words, with membrane depolarization there is a reduction in current, but more importantly, the contribution of Ca²⁺ influx to the current that remains becomes negligible. This effect can be observed directly when Ca²⁺ imaging is combined with perforated patch recordings in HEK293 cells (22).

The limited Ca²⁺ permeability has important implications for TRPC6 channels as regulators of Ca²⁺ dynamics in different cell types. For example, in excitable cells that express voltage-activated Ca²⁺ channels (Ca_v channels), the main role of TRPC6 may be to simply evoke sufficient depolarization to cause activation of Ca_v, leading to robust Ca²⁺ influx (111). For this purpose, it is immaterial whether TRPC6 channels have significant permeability to Ca²⁺. Membrane depolarization by itself is sufficient to trigger another source of Ca²⁺ influx. However, in nonexcitable cells such as podocytes that lack Ca_v channels (39), TRPC6 would be unlikely to be a source of significant Ca²⁺ influx unless there is a mechanism to prevent the depolarization that occurs as a result of TRPC6 activation. As will be discussed further below, BK_Ca channels, which interact and colocalize with TRPC6 channels, may serve such a function in podocytes (53). It is also worth noting that the permeation model formulated by Estacion and coworkers (22) precisely captures the inward and outward rectification that is observed in TRPC6 channels, including endogenously expressed TRPC6 channels (47).

In addition to pore blockade, Ca²⁺ produces complex effects on the gating properties of TRPC channels. For example, Ca²⁺-calmodulin appears to facilitate or accelerate activation of TRPC5 channels (85) and TRPC6 channels (4, 108). Cytosolic Ca²⁺ concentrations at the high end of physiological cause inhibition of TRPC6 channels independently of pore blockade, which may entail channel phosphorylation by protein kinase C (108). This provides a potential mechanism for negative feedback. It also raises an interesting possibility for signaling cross talk, as it is possible that Ca²⁺ influx or mobilization through one type of channel could modulate signaling mediated by TRPC6 channels.

Mutations in TRPC6 Gene and Glomerular Disease

The importance of ion channels for podocyte function was underscored by the discovery that the TRPC6 mutations shown in Fig. 1 can cause inherited glomerular disease such as FSGS, which generally entails effects on podocytes at early stages (94, 128). TRPC6 mutations accounted for 5 of 71 families with FSGS in the study by Reiser and colleagues (94). The ethnic backgrounds and age range at disease presentation of patients in that study were broad, ranging from 16–61 yr. This relatively late onset stands in marked contrast to several other genetic glomerulopathies that manifest in infancy, such as congenital nephrotic syndrome of the Finnish type (52), steroid-resistant nephrotic syndrome of the podocin type (5), or Wilms' tumor protein 1 (WT1)-related diseases (83). The early onset of those diseases can be readily explained by the disruption or loss of essential structural components of the SD. The question remains as to why the onset of kidney disease in patients with TRPC6 mutations occurs at a relatively advanced age. One possible explanation is that mutations in TRPC6 may produce subtle changes in intracellular function that lead to irreversibly altered cell behavior only over time and possibly after the accumulation of other renal insults. These could include elevated glomerular capillary pressure, ischemia, metabolic stresses such as diabetes, or serum factors, or the presence of other mutations. Similar explanations have been offered for the adult onset and dominantly inherited forms of FSGS associated with mutations in α-actinin-4 (49) and CD2AP (42).

Interestingly, patients with TRPC6-related FSGS do not generally present with any other pathological phenotype. This is surprising because TRPC6 is expressed in many other tissues, for example, in smooth muscle, where it is thought to contribute to the regulation of vascular tone (14, 126). It is possible that the glomerular phenotype is the consequence of a unique role of TRPC6 in the organization of the filtration apparatus or an unusual susceptibility of this structure to even subtle changes in Ca²⁺ dynamics or cytoskeletal regulation. TRPC6 is not unique in this respect, as mutations in the genes encoding α-actinin-4 and CD2AP result primarily in glomerular kidney disease even though they are widely expressed and associated with a variety of cellular functions (42, 49).

All of the TRPC6 mutations associated to date with FSGS map to the intracellular NH₂-terminal and COOH-terminal domains of the channel protein, and most are associated with some detectable increase in activity compared with wild-type in heterologous expression systems. These include six NH₂-terminal missense mutations (G109S, P112Q, M132T, N125S, N143S, and S270T) that are located in or near the various NH₂-terminal ankyrin repeats (94, 128). These mutations may compromise the ability of TRPC6 to oligomerize with other TRPC subunits (101) or alter the location or steady-state level of channel expression at the cell surface (125). Among these, the P112Q mutation described in a large family from New Zealand exhibits increased surface expression as assessed by cell-surface biotinylation assays as well as an increase in angiotensin II-mediated Ca²⁺ influx in heterologous expression systems (128). NH₂-terminal mutations can also affect TRPC6 gating properties, as the N143S and S270T mutations result in channels with an increased mean open-time compared with wild-type (32). The NH₂-terminal mutation with the largest effect on channel function observed to date is M132T. This mutation gives rise to maximal currents in heterologous expression systems three- to fivefold larger than those of wild-type channels, as well as delayed inactivation in the presence of continuous agonist stimulation (32). Interestingly, the M132T mutation gives rise to FSGS disease in children as early as 9 yr of age, suggesting that the age of onset may be related to the magnitude of the gain of function with respect to integral Ca²⁺ influx over time.

At the COOH terminus of TRPC6, the mutation L780P is located near the Trp box 1 motif that is conserved in all TRPC channels, whereas four additional mutations (K874*, Q889K, R895C, and E897K) map to a predicted coiled-coil domain located further downstream. While no study has specifically investigated the importance of the COOH-terminal coiled-coil motif in TRPC channel gating, one study suggests that the COOH-terminal tails of TRPC4 and TRPC6 participate in channel oligomerization (70). Importantly, the R895C and E897K mutations both resulted in TRPC6 channels with higher maximal current amplitudes in heterologous expression systems, effectively resulting in a net gain in function. It is possible that these are caused by increased trafficking to the cell surface.

Role of TRPC6 in Acquired Forms of Proteinuric Kidney Disease

The discovery of TRPC6 mutations provided a molecular basis for understanding certain hereditary forms of proteinuric glomerular disease. A recent study showed that wild-type TRPC6 was upregulated in a subset of acquired human proteinuric kidney diseases as well as in experimental models of acquired glomerular disease (78). Of particular interest was the induction of TRPC6 detected in patients with membranous glomerulonephritis (78). In membranous nephropathy, subepithelial immune deposits form in situ as a result of circulating antibodies against as yet unidentified antigens (46). The formation of immune deposits is associated with complement activation, leading to assembly of C5b-9 membrane attack complexes on podocyte plasma membranes (90). It has also been reported that podocyte damage mediated by the complement C5b-9 complex is associated with an activation of PLC and an increase in intracellular free Ca²⁺ (11). This raises the interesting question of whether podocyte TRPC6 channels are regulated by complement and whether this contributes to the pathophysiology of membranous disease. Podocyte TRPC6 channels are also upregulated in response to angiotensin II (133) and as a result of increased generation of reactive oxygen species during puromycin aminonucleoside nephrosis (123), as well as in neonatal nephrin^−/− mice (94). With respect to angiotensin II, siRNA knockdown of TRPC6 reduces angiotensin II-mediated Ca²⁺ transients in podocyte cell lines (133), and this neurohormone causes an increase in cytosolic Ca²⁺ concentrations in primary cultures of podocytes through stimulation of Ca²⁺ influx and mobilization of intracellular Ca²⁺ stores (33). Thus sustained excessive Ca²⁺ influx through TRPC6 channels may be a common pathway underlying a host of glomerulodegenerative conditions that originate in podocytes, although it is possible that TRPC6 dysfunction in other glomerular cell types could contribute to the overall pathology. Some of these changes presumably entail changes in gene expression. In this regard, TRPC6 appears to be a component of a Ca²⁺-dependent regulatory loop that drives NFAT-dependent transcription in podocytes (102) and that may also cause activation of NF-κB signaling (133). Over time, alterations in intracellular Ca²⁺ dynamics may result in irreversible damage and loss of podocytes (107) or podocyte dedifferentiation that ultimately leads to glomerulosclerosis (1). It is also possible that sustained disruptions of TRPC6 signaling could lead to podocyte detachment from the GBM owing to changes in cytoskeletal dynamics of the foot processes, altered integrin function, or altered susceptibility to fluid sheer stress (107).

Interaction of TRPC6 with Podocin and Nephrin

TRPC6 is expressed in podocyte foot processes in the vicinity of the SD as well as in the cell body and throughout the major processes (94) (Fig. 1). TRPC6 interacts with nephrin and podocin, at least in cultured podocytes (41, 43, 94), and these interactions may be central to the physiological function of these channels at the SD.

Podocin is a member of the stomatin gene family, and its homolog Mec-2 contributes to mechanosensation in Caenorhabditis elegans (37). Podocin has been shown to regulate the activity of TRPC6 in a cholesterol-dependent manner, and Huber and coworkers (43) have proposed that podocin binds sterols and alters the local lipid environment surrounding the channel molecule, thereby facilitating the ability of TRPC6 to respond directly to deformation of the plasma membrane. In heterologous expression systems, TRPC6 has also been reported to respond to pressure stimuli in membrane patches even when PLC is inhibited, and the authors of that study suggest a common biophysical basis for membrane stretch activation and activation of TRPC6 by DAG based on inhibition of both processes by the spider toxin GsMTx-4 (112). Podocin may cause local depletion of membrane sterols, resulting in local changes in membrane fluidity, thereby lowering energy barriers for conformational changes in the channel molecule caused by membrane curvature.

TRPC6 interactions with nephrin near the SD may underlie regulation of gating and trafficking of these channels (38, 120). Functional interactions between these proteins may serve to monitor the integrity of the filtration apparatus in podocytes to sense mechanical stimuli and to trigger Ca²⁺ signaling cascades that could alter cytoskeletal dynamics (64–66). For example, engagement of the ectodomains of nephrin results in localized activation of Src family tyrosine kinases (120). TRPC6 channels are substrates for Src family kinases such as Fyn, and tyrosine phosphorylation by these types of kinases causes an increase in TRPC6 activity (35), which would in turn cause alterations of cytoskeletal dynamics. Moreover, the absence of nephrin as shown in neonatal nephrin^−/− mice led to increased expression of TRPC6 in podocytes compared with podocytes of wild-type littermates (94).

Biochemical and Biophysical Properties of BK_Ca Channels

BK_Ca channels interact with TRPC6 channels and other SD proteins and may play an important role in podocyte signaling (53–55, 58, 95). Functional BK_Ca channels were first reported in human podocyte cell lines (79) and were later described in mouse podocyte cell lines (53–55, 58, 95). Slo1 proteins can also be detected by immunofluorescence and immunoblotting in human (79) and mouse (58) glomeruli. In mice, BK_Ca channels colocalize with synaptopodin, suggesting a podocyte pattern of expression (58). As we will describe further below, the BK_Ca channels of podocytes are mechanosensitive (79).

Functional BK_Ca channels are composed of four pore-forming subunits, often referred to as Slo1 proteins, which are encoded by the KCNMA1 gene (also known as Slo1 and KCa1.1). BK_Ca channels are expressed in a remarkably wide range of tissues, and in several cell types in kidney, including arteriolar smooth muscle and the distal nephrons, as well as in podocytes (Fig. 2). Open BK_Ca channels always have a very large unitary conductance (180–220 pS in symmetrical 150 mM KCl and ∼105 pS with physiological ionic gradients). This means that activation of a relatively small number of them can have profound effects on cell physiology, especially in electrically compact cells. Each vertebrate Slo1 protein has seven membrane-spanning domains with an extracellular NH₂ terminus and a large globular cytoplasmic COOH-terminal domain that comprises about half of each Slo1 subunit.

Fig. 2.

Large-conductance Ca²⁺-activated K⁺ (BK_Ca) channels in podocytes. A: topography of the pore-forming subunits of BK_Ca channels (Slo1 proteins) showing Ca²⁺-binding domains in the large cytoplasmic COOH terminal. Also shown is a noncanonical SH3 domain that mediates stretch-sensitive gating by interactions with actin-binding proteins. B: colocalization of Slo1 (red) and synaptopodin (green) in a mouse glomerulus revealed by double-label immunofluorescence and a merged image. Areas of colocalization indicate expression in podocytes. Slo1 also appears to be expressed in the mesangium, where there is no synaptopodin. C: model for coordinated activation of TRPC6 and BK_Ca. Although Ca²⁺ can permeate TRPC6, it also causes pore blockade and with depolarization TRPC6 behaves as a monovalent cation channel. However, as a result of physical interactions and resulting colocalization, Ca²⁺ influx through TRPC6 can cause activation of BK_Ca. This may prevent membrane depolarization, thereby providing a driving force that can sustain Ca²⁺ infux through TRPC6.

The gating of BK_Ca channels is markedly voltage dependent, and Ca²⁺ binding to multiple allosterically interacting domains in the cytoplasmic COOH terminus of Slo1 proteins shifts activation into the physiological range of membrane potentials (130, 132). These Ca²⁺-activation domains include the so-called “calcium bowl” (103) and a pair of “regulator of potassium conductance” (RCK) domains (2) that are separated by a flexible hinge region that allows for their conformational coupling (60, 91). The domain structure is summarized schematically in Fig. 2A.

The mammalian KCNMA1 gene is expressed in a large number of splice variants with markedly different voltage and Ca²⁺ sensitivities and gating kinetics, although all variants have high unitary conductance. This gene has 35 exons and is subject to alternative splicing at no fewer than 7 different sites (3). More than one BK_Ca channel splice variant can exist in a single cell type (55, 57), and it is therefore possible that some native BK_Ca channels are heteromeric. This has important consequences for the regulation of both gating and trafficking to the cell surface (9, 57, 76).

A large number of factors determine the overall activity of BK_Ca channels in cells. These include 1) gating properties, especially voltage and Ca²⁺ sensitivities, which can be acutely altered by posttranslational modification and by interactions with other proteins; 2) the extent to which BK_Ca channels are either trafficked into the plasma membrane or trapped in intracellular compartments; 3) variations in KCNMA1 transcript levels and alternative splicing; and 4) the nature and proximity of Ca²⁺ sources that are usually needed to drive BK_Ca gating. A great deal of literature on the biochemical and biophysical factors controlling the gating of BK_Ca channels has been recently reviewed (10). Here, it is important to note that robust activation of BK_Ca channels usually requires cell depolarization and micromolar concentrations of free Ca²⁺ in the cytosolic microenvironment of the channels, far above the concentrations normally found in bulk cytosol. From the laws of diffusion, this implies that BK_Ca channels must in the majority of cases be closely colocalized to their Ca²⁺ sources, generally within nanometers of Ca²⁺-permeable pores, to see concentrations sufficient for substantial activation (80). In addition, reversible phosphorylation of various residues within the large intracellular COOH terminus of BK_Ca channels can have complex effects on their gating properties. For example, proto-oncogenic tyrosine kinases such as Src and Fyn can enhance the activity of BK_Ca channels when Ca²⁺ is present (72), which suggests that they could be modulated coordinately with nephrin and TRPC6 in podocytes (35, 38, 120). Changes in BK_Ca gating provide a means to integrate several other cellular pathways including cellular REDOX potential (113) and hypoxia (77, 127) that may be important in the context of kidney pathophysiology.

BK_Ca gating is typically sensitive to mechanical stimuli, a process that in some cell types requires interactions with actin filaments subjacent to the plasma membrane (18, 89). This type of gating has been described in some detail in immortalized human podocyte cell lines (79). Membrane stretch-sensitive gating is mediated at least in part by binding of adaptor proteins such as cortactin and Grb2 (and possibly many others) to a highly conserved noncanonical Src-homology 3 domain that can tether plasma membrane channels to subjacent actin filaments (114, 115). This proline-rich site is located directly between the two RCK domains in the large cytoplasmic COOH terminal of the Slo1 protein. It is possible that mechanical forces cause displacement of this hinge region, leading to perturbation of the allosteric coupling between the two RCK domains, thereby altering BK_Ca activation. Perturbing these types of interactions with actin may also underlie actions of certain neurohumoral signals on BK_Ca gating (84). BK_Ca channels can potentially interact with actin by at least three other mechanisms: these include binding to filamin A (56), interactions with caveolins (6), and by direct interaction with an actin-binding domain present in a more downstream region of the COOH terminus of Slo1 that is present in all vertebrate variants (135). It is not known whether these latter interactions play a role in regulating BK_Ca gating, but they clearly play a role in BK_Ca trafficking to the cell surface (56, 135). Thus interactions between BK_Ca channels and actin filaments are complex and are likely to have multiple effects on the overall activity of BK_Ca channels.

The properties of BK_Ca channels are markedly altered by auxiliary β-subunits that do not form functional pores by themselves, but which alter the gating properties of associated Slo1 proteins (75). There are four distinct vertebrate genes that encode BK_Ca channel β-subunits (β1–4). All share a basic membrane topology consisting of two transmembrane segments connected by an extracellular loop domain and with cytoplasmic NH₂ and COOH termini (86). The presence of β-subunits can cause complex effects on the Ca²⁺ and voltage sensitivity of BK_Ca gating, as well as the access of the channel pore to scorpion toxin inhibitors such as charybdotoxin and iberiotoxin. Mouse and human podocytes express β4-subunits (55, 79). These subunits cause marked slowing of BK_Ca activation at all membrane potentials, which is one of the hallmarks of macroscopic BK_Ca currents in podocytes (55). β4-Subunits also cause reduced channel activation at low cytosolic Ca²⁺ concentrations, but markedly enhanced activation when cytosolic Ca²⁺ is elevated (121). In other words, β4-subunits cause BK_Ca channels to function as a robust on-off switch, rather than to become active in a graded fashion. Furthermore, the slow kinetics conferred by the presence of β4-subunits suggests that these channels would not normally respond to transient changes in membrane potential. The presence of β4 is also reported to cause a reduction or loss of sensitivity to blockade of BK_Ca channels by charybdotoxin and iberiotoxin (48).

Properties of BK_Ca Channels in Podocytes

Slo1 immunostaining of mouse glomeruli is consistent with expression of BK_Ca channels in podocytes and mesangial cells (58) (Fig. 2), and Slo1 subunits and β4-subunits are also expressed in mouse and human podocyte cell lines (55, 79). Functional BK_Ca channels with slow kinetics are readily detected in single-channel recordings from inside-out patches excised from immortalized podocytes (79). Podocyte BK_Ca channels with unusually slow activation kinetics can also be detected in whole-cell recordings and are most easily observed by including micromolar Ca²⁺ in the recording pipettes and evoking currents by application of depolarizing voltage steps (53–55) (Fig. 2B). These outward currents are not seen when recording electrodes contain 10 mM EGTA to buffer the free Ca²⁺ to much lower concentrations (55). Currents with similar properties can be detected in podocytes dissociated acutely from rat glomeruli. The macroscopic BK_Ca currents in mouse podocyte cell lines do not inactivate in the face of continued depolarization. Podocyte BK_Ca channels are only partially inhibited by iberiotoxin (55), but they are completely blocked by the selective BK_Ca inhibitors paxilline (55) and penetrem A (79) as well as by nonselective K⁺ channel inhibitors such as tetraethylammonium (79).

The BK_Ca channels in human podocyte cell lines become active in response to membrane stretch (79). Thus the probability of BK_Ca opening in cell-attached patches is increased an average of fourfold by applying modest negative pressure (−20 cmH₂O) to recording pipettes. A similar effect is observed by placing podocytes in a bath solution with a tonicity that is 70% of control, which would be expected to cause swelling and stretch of the podocyte plasma membrane (79). These effects were seen in patches that were substantially depolarized from the resting membrane potential, and it is not known whether this reflects a direct effect of membrane stretch on BK_Ca gating or whether it is an indirect effect mediated by a stretch-dependent increase in local Ca²⁺ concentration. We have observed that acute depolymerization of actin with cytochalasin-D does not affect the amplitude or kinetics of macroscopic BK_Ca channels when intracellular Ca²⁺ is clamped at 5 μM (58), which suggests that the stretch sensitivity in these cells may not be mediated by actin interactions. Several alternative mechanisms are available in podocytes. For example, Huber and colleagues (43) demonstrated that coexpression of the cholesterol-binding protein podocin in heterologous systems can increase the activity of TRPC6 channels and that this effect appears to be attenuated by prior cholesterol depletion of cells and requires the cholesterol-binding domain of podocin. Given this, it is interesting to note that cholesterol depletion increases activity of BK_Ca channels in a variety of systems (8, 68, 110), most likely by changing the local fluidity of the surrounding membrane and by altering association of the channels with lipid rafts. It will be important to determine whether podocin affects BK_Ca gating and whether it mediates stretch sensitivity of BK_Ca function in podocytes.

Interactions Between TRPC6 and BK_Ca Channels in Podocytes

Activation of BK_Ca channels at physiological membrane potentials usually requires micromolar concentrations of free Ca²⁺ or even higher. It is rare for bulk cytoplasmic concentrations of Ca²⁺ to reach these levels except in extreme pathological situations. However, nanodomains immediately surrounding the pores of Ca²⁺-permeable channels can exceed this concentration (80). For this reason, the functional coupling of Slo1 channels with Ca²⁺-permeable channels usually requires very close colocalization. There is literature on excitable cells showing BK_Ca channels bind directly to voltage-activated Ca²⁺ channel proteins, which ensures the necessary colocalization for functional coupling of their gating (74, 136). However, podocytes are not excitable cells and do not express Ca_v channels (39). Instead, coimmunoprecipitation and glutathione-S-transferase pull-down assays have shown that Slo1 subunits can bind to TRPC6 channels (53). This occurs with endogenously expressed channels of mouse podocyte cell lines, as well as with channels, and domains of channels, expressed in HEK293T cells. These interactions may allow for sufficient colocalization to allow Ca²⁺ influx through TRPC6 channels to cause a coordinate activation of BK_Ca channels. It is also interesting that coexpression of TRPC6 with certain Slo1 splice variants increases surface expression of the latter in heterologous expression systems, whereas siRNA knockdown of TRPC6 reduces surface expression of endogenous BK_Ca channels in mouse podocyte cell lines (53). Somewhat surprisingly, this does not occur with TRPC3, which is also able to bind to BK_Ca channels. These results raise the possibility that at least a subset of TRPC6 and BK_Ca channels traffics as a complex in podocytes.

The available evidence suggests that TRPC6 channels represent a physiologically important source of Ca²⁺ influx in podocytes (133). For this reason, it is important to consider how podocytes maintain conditions that allow TRPC6 channels to transport Ca²⁺. As already discussed, depolarization tends to cause TRPC6 to function as a nonselective monovalent cation channel, owing to voltage-dependent pore blockade by Ca²⁺, which causes large effects on the P_Na/P_Ca ratio (22). Conversely, hyperpolarization allows TRPC6 to transport Ca²⁺ into the cell more efficiently both by increasing the driving force and by reducing pore blockade (22). Activation of TRPC6 will also cause cell depolarization unless there is an opposing mechanism to prevent this from happening. A physical interaction between BK_Ca and TRPC6 channels on the cell could provide such a mechanism by keeping BK_Ca channels within the Ca²⁺ microdomain that surrounds the pores of active TRPC6 channels. Such colocalization would allow for their coordinated activation in space and time. The resulting Ca²⁺ activation of BK_Ca would oppose cell depolarization caused by TRPC6 activation and thereby maintain the conditions needed for efficient Ca²⁺ influx into the cell. In this model, cell surface BK_Ca channels colocalized with TRPC6 provide positive feedback to Ca²⁺ influx in nonexcitable cells such as podocytes. By contrast, they have long been known to provide negative feedback to Ca²⁺ influx in excitable cells that express Ca_v channels that colocalize with BK_Ca (74). The difference is a simple consequence of the fact that TRPC6 and Ca_v channels have an effectively opposite dependence on membrane potential, through voltage effects on gating in the case of Ca_v (74) and voltage effects on Ca²⁺ permeability in the case of TRPC6 (22).

Another possible role of BK_Ca as part of the TRPC6 complex is to increase the mechanosensitivity of the overall complex. The extent to which native TRPC6 channels in podocytes are stretch sensitive is not known, although their interaction and colocalization with podocin in situ makes this a very real possibility (43). BK_Ca channels are demonstrably stretch sensitive in podocyte cell lines (79), which could provide a mechanism to fine-tune Ca²⁺ influx during normal glomerular function. This is because BK_Ca activation would increase the driving force for Ca²⁺ influx through TRPC6 channels.

Are there other mechanisms that could contribute to coordinated activation or deactivation of TRPC6 and BK_Ca? Src family protein kinases have been reported to increase the activity of both TRPC6 (35) and BK_Ca channels (72). In podocytes, Src activation can occur as a result of engagement of the nephrin ectodomain (120), which raises the possibility that both of these channels are components of a nephrin-dependent outside-in signaling cascade. It is also possible that certain signals have opposing effects on these two channel molecules, PIP₂ being one possibility (69, 117). This could provide a mechanism to terminate or fine-tune the Ca²⁺ signaling mechanism, but to date this has not been investigated in podocytes.

Interaction of BK_Ca Channels with Other Podocyte Proteins

The COOH-terminal domains of BK_Ca channels interact with >100 different proteins (50), but the functional significance has only been established for a small subset of proteins. We have already noted that TRPC6 channels interact with the COOH-terminal domains of Slo1 proteins in heterologous expression systems and in podocyte cell lines (53). Other SD proteins that interact with BK_Ca channels include nephrin (55), Neph1 (54), MAGI-1 (95), and synaptopodin (58). None of these protein interactions directly regulate the gating of BK_Ca channels in HEK293T cells, but all appear to play a role in regulating the steady-state surface expression of functional BK_Ca channels. Thus siRNA-mediated knockdown of nephrin, Neph1, or synaptopodin reduced the number of functional surface BK_Ca channels observed by electrophysiology or immunochemistry in mouse podocyte cell lines (54, 55 58). Conversely, coexpression of either nephrin or synaptopodin with Slo1 increases surface expression of BK_Ca channels in HEK293T cells compared with what was observed when Slo1 is expressed by itself (55, 58). The situation with Neph1 is interesting because the functional consequences of its interaction with BK_Ca channels depend on the cell type examined. Thus coexpression of Neph1 with Slo1 proteins in HEK293T cells suppressed expression of functional BK_Ca channels on the cell surface (54). An effect in the same direction was also seen in parasympathetic neurons, which express Neph1 endogenously. In those cells, siRNA knockdown of Neph1 caused an increase in steady-state surface expression of Slo1 and an increase in macroscopic Ca²⁺-dependent K⁺ current (54). It is not known why Neph1 produces different effects on BK_Ca channels in different cell types. MAGI-1 is a scaffolding protein expressed in podocyte SDs that also interacts with nephrin (34), synaptopodin (87), and α-actinin (87). Coexpression of MAGI-1 suppresses expression of BK_Ca channels on the surface of HEK293T cells (95). Conversely, siRNA knockdown of MAGI-1 expression in podocytes causes an increase in functional surface expression of BK_Ca channels (95).

These data indicate that protein interactions in SD domains regulate trafficking of ion channels and possibly other proteins into the plasma membrane. It is possible that signaling cascades differentially alter the stability of certain of these protein complexes, thereby affecting the steady-state surface expression of ion channels.

Conclusion

The discovery that mutations in TRPC6 channels cause familial FSGS has made analysis of ion channels in podocytes an important subject in renal physiology and experimental nephrology. Many questions remain unanswered, including the mechanisms whereby enhanced TRPC6 function leads to podocyte dysfunction and glomerular disease, the signaling cascades that control gating of podocyte ion channels, the role of multichannel complexes in controlling overall Ca²⁺ flux through TRPC6 channels, and the role of other members of the TRPC family and other types of cation channels in podocyte function.

One question that has not been addressed is what terminates Ca²⁺ signaling in podocytes. It seems likely that podocytes contain a number of systems for sequestering and buffering Ca²⁺ in podocytes, and it is quite possible that Ca²⁺-dependent cascades lead to termination of DAG-dependent TRPC6 activation through negative feedback. It is possible that these cascades eventually cause suppression of the gating of TRPC6 and BK_Ca channels, and possibly, over longer time scales, their expression on the cell surface. Some of the interactions between TRPC6 channels and other proteins are shown schematically in Fig. 3, which also shows a few of the many unresolved questions. The available data already suggest reciprocal interactions between podocyte ion channels and the cytoskeleton. The net effect of Ca²⁺ influx controlled by TRPC6 appears to include regulation of podocyte actin dynamics, possibly to allow SDs to adapt to mechanically adapt to changes in their immediate environment. Interactions with the cytoskeleton may in turn regulate the gating of podocyte ion channels by contributing to their mechanosensitivity, or by controlling their steady-state expression at the SD and at other locations on the cell surface. Ultimately, it is likely that sustained changes in podocyte ion channel function lead to Ca²⁺-dependent changes in gene expression, cell metabolism, and alterations in the efficiency of the glomerular filtration barrier that can lead over time to maintenance or loss of the entire nephron.

Fig. 3.

Models of TRPC6 and BK_Ca function in podocytes. These channels form a complex with each other in the plasma membrane, which ensures that Ca²⁺ influx through TRPC6 can cause coordinated activation of BK_Ca. This serves to prevent membrane depolarization and thereby maintain the Ca²⁺ permeability of TRPC6. In foot processes (left), these channels also interact with nephrin, podocin (P) and with the underlying actin cytoskeleton, which allows them to function as mechanosensors and to allow the resulting Ca²⁺ influx to modulate nephrin signaling and cytoskeletal dynamics acting in part through small GTPases and proteins such as Nck and CD2AP (shown schematically as a black circle). BK_Ca channels are also mechanosensitive through interactions with the cytoskeleton mediated by cortactin (blue circle) or other actin-binding proteins. TRPC6 channels can become active in response to G protein-mediated signaling cascades associated with PLC activation, for example, during angiotensin signaling. The entire slit diaphragm complex may be also regulated by tyrosine kinases such as Fyn. Other TRPC family members are expressed in podocytes, and their functions are unknown. In the cell body (right), they may form heteromeric complexes with TRPC6 and in this way function as Ca²⁺-store operated channels in G protein signaling cascades associated with PLC activation, possibly in complex with other proteins such as STIM1 and the 1,4,5-trisphosphate (IP₃) receptor. Influx of Ca²⁺ in the cell body can regulate gene expression in podocytes. Excessive Ca²⁺ influx through TRPC6 channels as a result of mutations or other pathological events can ultimately lead to loss of podocytes through apoptosis as well as detachment from the glomerular basement membrane (GBM). Foot process effacement and podocyte hypertrophy may represent an adaptive response of the remaining podocytes to maintain coverage of the GBM.

GRANTS

This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants RO1 DK82529 (to S. E. Dryer) and RO1 DK73495 (to J. Reiser).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.