Pflügers Archiv - European Journal of Physiology

, Volume 460, Issue 3, pp 571–581

Canonical TRP channels and mechanotransduction: from physiology to disease states

Authors

  • Amanda Patel
    • IPMC-CNRSUniversité de Nice Sophia Antipolis
  • Reza Sharif-Naeini
    • IPMC-CNRSUniversité de Nice Sophia Antipolis
  • Joost R. H. Folgering
    • IPMC-CNRSUniversité de Nice Sophia Antipolis
  • Delphine Bichet
    • IPMC-CNRSUniversité de Nice Sophia Antipolis
  • Fabrice Duprat
    • IPMC-CNRSUniversité de Nice Sophia Antipolis
    • IPMC-CNRSUniversité de Nice Sophia Antipolis
Invited Review

DOI: 10.1007/s00424-010-0847-8

Cite this article as:
Patel, A., Sharif-Naeini, R., Folgering, J.R.H. et al. Pflugers Arch - Eur J Physiol (2010) 460: 571. doi:10.1007/s00424-010-0847-8

Abstract

Mechano-gated ion channels play a key physiological role in cardiac, arterial, and skeletal myocytes. For instance, opening of the non-selective stretch-activated cation channels in smooth muscle cells is involved in the pressure-dependent myogenic constriction of resistance arteries. These channels are also implicated in major pathologies, including cardiac hypertrophy or Duchenne muscular dystrophy. Seminal work in prokaryotes and invertebrates highlighted the role of transient receptor potential (TRP) channels in mechanosensory transduction. In mammals, recent findings have shown that the canonical TRPC1 and TRPC6 channels are key players in muscle mechanotransduction. In the present review, we will focus on the functional properties of TRPC1 and TRPC6 channels, on their mechano-gating, regulation by interacting cytoskeletal and scaffolding proteins, physiological role and implication in associated diseases.

Keywords

Cation channelMechano-electrical transductionMechanoreceptorMechanosensitive channelTransient receptor potential

Abbreviations

SACs

Stretch-activated cation channels

MscL

Bacterial mechano-sensitive large conductance channel

TRP

Transient receptor potential channels

TRPC

Canonical TRP channel

FSGS

Familial focal segmental glomerulosclerosis

DMD

Duchenne muscular dystrophy

SOCs

Store-operated ion channels

ER

Endoplasmic reticulum

STIM1

Stromal interacting molecule 1

Orai proteins

The pore-forming components of CRAC channels

RNA

Ribonucleic acid

ROCs

Receptor-operated channels

DAG

Diacylglycerol

GsMTx-4

Grammostola spatulata toxin inhibiting SACs

AT1R

Angiotensin II type 1 receptor

Ang II

Angiotensin II

G protein

GTP-binding protein

PLC

Phospholipase C

GPCR

G protein-coupled receptor

TMD

Transmembrane domain

M5R

Muscarinic type 5 receptor

H1R

Histamine type 1 receptor

ETAR

Endothelin receptor

V1AR

Vasopressin type 1 receptor

A7R5

Rat vascular smooth muscle cell line

HETE

Hydroxyeicosatetraenoic acid

MR

Myogenic response

VSMCs

Vascular smooth muscle cells

MEF

Mechanoelectric feedback

TREK channels

Mechano-gated K2P channels

OAG

1-oleoyl-2-acetyl-sn-glycerol

MHC

Myosin heavy chain

NFAT

Nuclear factor of activated T cells

TAC

Transverse aortic constriction

ATG

Angiotensinogen

CAV

Caveolins

eNOS

Endothelial nitric-oxide synthase

FlnA

Filamin A

Introduction

Cellular mechanotransduction is a fundamental process allowing conversion of mechanical forces into an electrical and/or biochemical signal [16]. Opening of non-selective stretch-activated cation channels (SACs) plays a central role in the cellular response to pressure. SACs were initially identified by the Sachs' group using cell-attached patch clamp recordings of chick skeletal myocytes, some 25 years ago (Fig. 1) [7, 8]. The bacterial MscL channel was the first SAC to be cloned and reconstituted into artificial bilayers by the Kung laboratory [9]. Gating of SACs can be direct with activating tension coming from the lipid bilayer itself or from the tethered extracellular matrix and/or cytoskeleton (Fig. 2). An alternative indirect model involves another primary mechanosensor(s), such as a membrane receptor and associated second messenger(s) (Fig. 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs00424-010-0847-8/MediaObjects/424_2010_847_Fig1_HTML.gif
Fig. 1

Functional properties of stretch-activated ion channels (SACs). Non-selective SACs are characterized by a single-channel conductance in the range of 30 pS and are permeable to sodium, potassium, and calcium, resulting in a reversal potential close to 0 mV. Substitution of sodium with a non-permeant cation such as choline or NMDG shifts the current to voltage relationship in the hyperpolarizing direction (dashed line; left panel). Opening of non-selective SACs depolarizes cells such as arterial myocytes with a resting membrane potential of about −50 mV. These channels are blocked by Gd3+ and the tarantula venom peptide GsMTx-4. Interestingly, GsMTx-4 inhibits atrial fibrillation induced by atrial distension and protects mdx skeletal muscle against cellular damages associated with stretching. The open channel probability of SACs gradually rises with increasing stretch intensity (indicated next to the current traces recorded in the cell attached patch configuration at −80 mV from an isolated mesenteric arterial myocyte) (right panel). Adapted from [69]

https://static-content.springer.com/image/art%3A10.1007%2Fs00424-010-0847-8/MediaObjects/424_2010_847_Fig2_HTML.gif
Fig. 2

Models for stretch-activated ion channels (SACs) gating. SACs are inserted in the lipid bilayer of the plasma membrane. Activating force can be the tension in the surrounding lipids inducing a conformational change of the channel complex (bilayer model), as previously described for the bacterial SAC MscL. Alternatively, tension can be transmitted from either the extracellular matrix or the cytoskeletal elements tethered to the channel subunits (tether model). This model may be at play for the mechanosensory channel involved in hearing and touch. An indirect mechanism for SAC gating involves another primary mechanosensor (for instance AT1R), which might be also tethered to the extracellular matrix or cytoskeleton and could involve the release of a second messenger (for instance DAG) resulting ultimately in the opening of a TRP channel such as the ROC TRPC6. An indirect model of SAC gating is proposed to be at play in cardiac mechanotransduction and hypertrophy. Modified from [6]

Subsequently, genetic screens in yeast, Caenorhabditis elegans, Drosophila, and Danio rerio, indicated that transient receptor potential (TRP) channel subunits are involved in mechanical sensing [1016]. Recently, it was claimed that TRPC1 and TRPC6 may encode the mammalian SACs [17, 18]. TRP channels have six transmembrane segments (TMS), a pore (P) domain between TMS 5 and 6 and both the amino and carboxy terminal domains facing the cytosol [4, 19, 20]. The mammalian TRP channel family (28 members) is subdivided into six structural subfamilies: canonical TRP channel (TRPC; seven members), TRPM (eight members), TRPV (six members), TRPML (three members), transient receptor potential polycystic (TRPP; three members), and finally TRPA1 (one member). The additional TRPN subfamily (one member) only concerns lower animal organisms including Drosophila [21]. TRP subunits form tetramers and may heteromultimerize. Their permeation ranges from cationic non-selective to highly Ca2+ selective [4, 19, 20]. TRPC1 and TRPC6 are non-selective calcium permeant channels which have been recently implicated in muscle mechanotransduction [2225]. Furthermore, these channels are involved in inherited and acquired pathologies including familial focal segmental glomerulosclerosis, Duchenne muscular dystrophy (DMD), and cardiac hypertrophy [24, 2632].

The canonical TRPC1 and TRPC6 channels are candidate SACs

TRPC1 is widely expressed and present in the myocytes of the heart, the arteries, and the skeletal muscle [33]. Numerous studies indicate that TRPC1 is a non-selective “store-operated ion channel” (SOC) involved in the calcium entry following calcium depletion of the endoplasmic reticulum (ER). The ER calcium sensor STIM1 (stromal interacting molecule) and the Orai proteins (the pore-forming components of CRAC channels that mediate store-operated calcium entry in T lymphocytes and other hematopoetic cells) interact with TRPC1 [34]. TRPC1 has also been proposed to be directly opened by membrane stretch [17]. Using reconstitution of solubilized Xenopus oocyte (which express high levels of native SACs) membrane proteins into artificial liposomes, it was claimed that TRPC1 underlies the vertebrate SAC [17]. Heterologous expression of hTRPC1 results in about a 10-fold increase in SAC density, whereas injection of a TRPC1-specific antisense RNA abolishes endogenous SAC activity in Xenopus oocytes [17]. Moreover, transfection of hTRPC1 into CHO cells also significantly increases SAC activity [17].

TRPC6 is another calcium-permeable non-selective cation channel which is widely expressed, including in the cardiovascular system [35]. TRPC6 is activated in response to PIP2 hydrolysis by stimulation of various plasma membrane receptors, thus called a “receptor-operated channel” (ROC). TRPC6 is directly activated by diacylglycerol (DAG), independently of protein kinase C [36]. TRPC6 assembles into either homo or heterotetramers with TRPC3 and TRPC7 [35]. TRPC6 also interacts with a number of adaptors including cytoskeletal proteins (for review see [35]). TRPC6 is located in the glomerular slit diaphragm, where it interacts with podocin [37]. In humans, gain of function mutations of TRPC6 lead to kidney focal and segmental glomerulosclerosis (FSGS) [26, 27]. Recent evidence indicates that TRPC6 may be a sensor of mechanically and osmotically induced membrane stretch [18]. TRPC6 is inhibited by the spider toxin GsMTx-4, a blocker of SACs [18, 38, 39] (although also see [40]). Membrane tension has been proposed to directly gate TRPC6 [18] (Fig. 2). DAG (opener) and GsMTx-4 (blocker) may differentially affect membrane curvature and thus modulate TRPC6 activity [18, 20].

While TRPC6 expression at the plasma membrane is well established, TRPC1 is mostly localized in the ER when expressed in transiently transfected cells [41]. However, there is a possibility that TRPC1 (or TRPC6) may require to assemble as an heterotetramer (such as TRPC1/4/5) to function as a mechanosensitive ion channel activated by membrane stretch at the plasma membrane. Such a possibility needs to be considered when comparing endogenous channels (native currents) with expression systems.

Are TRPC1 and TRPC6 directly gated by force?

We and others failed to confirm the direct mechanosensitivity of both TRPC1 and TRPC6 in transfected mammalian cells [23, 40, 41]. Indeed, mock transfected COS-7 and CHO cells show endogenous SAC activity comparable to the one observed in TRPC1 or TRPC6-transfected cells [17, 41]. TRPC6 channel activity could be elicited by DAG, indicating that the channel is indeed present and active at the plasma membrane, but not reactive to stretch [23, 40, 41]. These findings indicate that TRPC1 and TRPC6 per se are not mechanosensitive and that membrane stretch does not primarily gate these channels [23, 40, 41].

How may these apparently opposite results be reconciled? How can we explain the mechanosensitivity of TRPC1 and TRPC6 observed by some, but not by others (including our own group)? One possibility is that stretch sensitivity is indirect and may require an additional component(s) which may be either present or absent in different cell types (Fig. 2).

GPCR requirement for TRPC6 mechano-sensitivity

TRPC6 was shown to become mechano-sensitive when co-expressed with the angiotensin II type 1 receptor AT1R, even in the absence of the ligand Angiotensin II (Ang II) [23]. TRPC6 activation by stretch or swelling is blocked by PLC inhibition (although also see [18]) or by GDPβs, which impairs G protein activation [23, 42]. It was concluded from this study that membrane stretch causes an agonist-independent conformational change of AT1R thereby activating the downstream Gq/11/PLC/DAG/TRPC6 signaling cascade [23] (Fig. 3).
https://static-content.springer.com/image/art%3A10.1007%2Fs00424-010-0847-8/MediaObjects/424_2010_847_Fig3_HTML.gif
Fig. 3

Cartoon illustrating the role of canonical transient receptor potential channel (TRPC)3/6/7 in cardiac hypertrophy. The AT1R stimulated by membrane stretch and/or by its ligand AngII, leads to PIP2 hydrolysis (with possible synergism). Subsequently, DAG directly activates TRPC3/6/7 channels at the plasma membrane. TRPC3 has also been shown to physically interact with the IP3 receptors. Activation of calcineurin by Ca2+-calmodulin dephosphorylates NFAT which then translocates into the nucleus and turns on pro-hypertrophic genes, including TRPC6. TRPC channels are localized in caveolae and interact with caveolins. Disruption of caveolae by CAV silencing or by cholesterol buffering impairs mechanotransduction (adapted from [98])

The AT1R G protein-coupled receptor (GPCR) is made of seven transmembrane helices with an extracellular N-terminal domain and a cytoplasmic C terminus. Mechanically activated receptors (independently of Ang II) adopt an active conformation, allowing for productive G protein coupling [23, 43, 44]. Cell stretch induces an anticlockwise rotation and a shift of transmembrane domain (TMD) 7 into the ligand-binding pocket [43]. As an inverse agonist, candesartan suppresses the stretch-induced helical movement of TMD7 through the binding of its carboxyl group to specific residues of the receptor [43]. Importantly, other Gq/11-coupled receptors including, the muscarinic M5R, histamine H1R, endothelin ETAR and vasopressin V1AR can substitute for AT1R, whereas Gs-coupled receptors such as the β2-adrenergic receptor as well as tyrosine kinase receptors do not link mechanical stimuli to TRPC6 activation [23]. Remarkably, increasing the AT1R density in mechanically unresponsive A7R5 cells renders these cells mechanosensitive [23]. The other DAG-sensitive TRPC channels TRPC3 or TRPC7 can replace TRPC6 and be activated by osmotic stimuli when co-expressed with AT1R [23]. In the absence of AT1R overexpression, there is no activation of TRPC6 by membrane stretch (although see [18]), suggesting that endogenous GPCRs do not underlie stretch-induced TRPC6 activation [23, 40, 41]. A more recent study further demonstrates a synergistic effect of receptor activation and mechanical stimulation on TRPC6 channels activity which is mediated by the phospholipase A2/ω-hydroxylase/20-HETE pathway [40] (Fig. 3). These results confirm that TRPC6 is not intrinsically mechanosensitive and its activation by stretch is dependent on ligand-induced receptor activation. Whether the possible mechanosensitivity of G-proteins/phospholipases is involved in the activation of TRPC channels by membrane stretch has not yet been directly tested.

Differences in the type of native receptor(s), level of expression or release of endogenous ligand(s) may thus explain the discrepancies between the various studies on TRPCs mechanosensitivity [17, 18, 23, 40, 41].

TRPC6, unlike TRPC1, plays a functional role in the arterial myogenic response

Resistance arteries with a small diameter constrict in response to increasing intraluminal pressure [45]. This autoregulation called the myogenic response (MR) is required to maintain the blood flow constant. The myogenic tone is responsible for protection against hypertension-induced injury [45]. The pressure-dependent MR which is intrinsic to the arterial myocytes involves several molecular pathways, including focal adhesions, ion channels and the cytoskeleton [45, 46]. An initial response to increased pressure is a depolarization as a consequence of SACs opening (amplified by the secondary recruitment of voltage-dependent L-type calcium channels), which precedes muscle contraction [22, 47, 48].

TRPC1−/− mice are viable, healthy and fertile, and express normal levels of the other TRPC channels, demonstrating an absence of genetic compensation [49]. The pressure threshold for MR and the degree of vasoconstriction are identical between WT and TRPC1−/− mice [49]. Additionally, in smooth muscle cells derived from cerebral arteries of WT or TRPC1−/− mice, cation currents induced by hypotonic swelling or positive pipette pressure are comparable [49]. Therefore, these results indicate that TRPC1 is not an obligatory component of the SAC complex in vascular smooth muscle cells [49].

TRPC6-/- mice are characterized by high blood pressure, increased contractility of isolated aortic rings in response to agonists, and enhanced MR in cerebral arteries [50]. However, the level of TRPC3 expression is much higher in arterial myocytes from TRPC6−/− mice. These smooth muscle cells are more depolarized and show a higher basal cation entry [50]. Elimination of this current by TRPC3-specific siRNA indicates that TRPC3 channels are up-regulated in TRPC6−/−mice, leading to vasoconstriction [50]. To avoid this compensatory gene remodeling, a TRPC6 antisense oligonucleotide strategy has been used in isolated cultured arteries. TRPC6 knock down strongly attenuates pressure-induced myocyte depolarization, MR and swelling-induced current [22]. The myogenic tone of cerebral and renal arteries is also profoundly diminished by the AT1R inverse agonist losartan [23]. These findings are in line with the proposed role of Gq/11-coupled receptors as sensors of membrane stretch in VSM cells inducing TRPC6 channel opening, cell depolarization, calcium influx and myogenic contraction [22, 23].

TRPC1 and TRPC6 are both involved in cardiac mechanotransduction and ventricular hypertrophy

The heart, which is linked to force generation, is in turn sensitive to both external and internal mechanical forces. At the extreme, the heart can be stopped by an external impact applied to the chest that does not cause tissue damage, a phenomenon first described more than 100 years ago as commotio cordis (for review see [51]). Internal feedback such as a heart-rate response to lung filling and the observation that catheterization can induce ectopic beats when the catheter-tip approaches the endocardium further underscores the importance of mechanoelectric feedback (MEF) in cardiac function (for review see [5254]). Both hyperpolarizing K+-selective channels (encoded by the TREK K2P channels) and depolarizing non-selective SACs are implicated in cardiac MEF [55, 56]. Stretch depolarizes the resting membrane potential of ventricular myocytes, accelerates the initial action potential repolarization phase and retards the later repolarization phase [57]. Moreover, stretched myocytes can generate pacemaker-like depolarizations leading to extra systoles [57].

Remarkably, antibodies against TRPC6 prevent stretch activation of the voltage-independent, Gd3+ and GsMTx-4-sensitive non-selective cation conductance of isolated cardiomyocytes [25]. This mechano-sensitive current, which can also be activated by the TRPC6 openers 1-oleoyl-2-acetyl-sn-glycerol (OAG) or flufenamate, is absent from cells whose T-tubules have been removed [25] (Fig. 3). These findings indicate that in ventricular cardiac cells, TRPC6, localized in T tubules, is involved in the stretch-sensitive depolarizing current.

When ventricular cardiac myocytes are subjected to mechanical overload (as for instance in the case of hypertension), they respond in the short term by adaptive physiological and beneficial hypertrophic growth and remodeling [5860]. The increase in wall thickness aims at restoring normal wall stress according to Laplace’s law. Mechanical stress imposed on cardiomyocytes by haemodynamic overload is considered to be one of the primary events triggering the hypertrophic response [5860]. In the long term, this adaptive mechanism becomes pathogenic leading ultimately to heart failure and sudden death. The hypertrophic cardiomyocytes expand in size and increase their protein synthesis. They also re-express a fetal genetic program, including a lower expression of α-MHC and an up-regulation of β-MHC, with a consequent decrease in cardiac contractility [60]. The increase in intracellular calcium plays a key role in the development of cardiac hypertrophy. Activation of the calcium-dependent serine-threonine phosphatase calcineurin dephosphorylates and induces the translocation of the nuclear factor of activated T cells (NFAT) transcription factors to the nucleus. Activation of NFAT results in the re-expression of fetal cardiac genes [5860]. Recent evidence demonstrates that TRPC channels are involved in the increase in intracellular calcium during cardiac hypertrophy [2832] (Fig. 3).

TRPC6 (and TRPC3) expression is increased in hypertrophic and failing hearts [28]. The presence of two conserved NFAT consensus sites in the TRPC6 promoter is responsible for this upregulation [28]. Moreover, overexpression of TRPC6 or TRPC3 in the heart of transgenic mice results in increased responsiveness to transverse aortic constriction (TAC) and susceptibility to heart failure [28]. Dominant negative TRPC3, TRPC6 or TRPC4 blocking the activity of the TRPC3/6/7 or TRPC1/4/5 channel complexes in the heart blunted cardiac hypertrophy induced by pressure-overload [61]. Moreover, a TRPC3 channel inhibitor (Pyr3) suppresses stretch-induced cardiomyocyte hypertrophy further demonstrating the clinical importance of TRPC channels in the cardiovascular system and indicating that specific inhibition of TRPC channels may emerge as a powerful cardiac anti-hypertrophic strategy [62]. All together, these findings indicate that TRPC channels play a major role in the pathological cardiac remodeling [28, 30, 31] (Fig. 3). Since TRPC3 and TRPC6 associate (and are activated by stretch when co-expressed with AT1R) [35], a heteromeric channel might therefore be at play in cardiac hypertrophy. Although TRPC7 also associate with TRPC3 and TRPC6 channels, it is rather involved in AngII-induced cardiomyocyte apoptosis [63].

TRPC1 is overexpressed in TAC-induced (pressure overload) hypertrophy [64]. Moreover, TRPC1-/- mice subjected to TAC or chronic AngII stimulation show a blunted cardiac hypertrophy response with a modest increase in cardiac mass and preserved contractility as compared to WT mice [32]. Cardiomyocytes isolated from TRPC1-/- mice fail to show an increase in the markers of cardiac remodeling in response to cyclic stretch. Treating the WT cells with the AT1R inverse agonist losartan also suppresses the hypertrophic response [32]. Moreover, phosphorylated NFATC3 (inactivated) is increased in the heart from TAC-operated TRPC1-/- mice [32]. Chronic pressure overload (TAC) induces the expression of a non-selective linear whole cell current in WT cardiomyocytes. The amplitude of this current is significantly decreased in TRPC1−/−mice [32]. On the contrary, Ang II increases this current by 2-fold, an effect which again is lost in the TRPC1−/− cells. However, the OAG-induced current as well as the store depletion-induced calcium entry are comparable in WT and TRPC1−/− myocytes [32]. Cell swelling, used as a way to mimic stretch, increases the non-selective cationic current in WT, but not in TRPC1-/- cells. Moreover, losartan as well as the phospholipase C inhibitor U-73122 decrease this swelling-activated current in WT cells. This current is also blocked by the SACs blocker GsMTx-4. All together, these findings demonstrate that stretch-activation of TRPC1 via the AT1R is responsible for calcium entry which participates in cardiac hypertrophic signaling [32]. Importantly, TRPC1 is unlikely to form a functional homotetrameric channel when expressed by itself, but rather may participate to heterotetramers together with TRPC4 and TRPC5 (for review see [33]). The cardiac phenotype of the TRPC1 knock out mice may thus be related to the impairment of these heteromultimeric complexes [32].

Remarkably, the AT1R inverse agonist candesartan reverses cardiac hypertrophy induced by TAC in both control and angiotensinogen knock out mice [44]. Thus, mechanical stress can induce cardiac hypertrophy in vivo through the AT1R in the absence of AngII. All together these findings show that AT1R, TRPC1 and TRPC3/6/7 signaling are part of a central mechanism involved in the development of cardiac hypertrophy. While TRPC6 opening has been linked to DAG stimulation, the mechanism of TRPC1 opening via the AT1R upon mechanical stimulation has not yet been determined [32].

Interaction of cytoskeletal and scaffolding proteins with TRPC1 and TRPC6

The gating of SACs critically depends on the mechanical properties of the lipid bilayer in which they are inserted [2, 65, 66]. Lipid composition and interaction with the cytoskeleton and extracellular matrix, which strengthen the membrane, both greatly influence SAC activity. For instance, increase in SAC activity observed in dystrophic myocytes correlates with the destabilization of the cortical actin cytoskeleton and consequent alteration of membrane surface tension [67, 68]. Disruption of the cytoskeleton is thought to increase the radius of membrane curvature (the law of Laplace: T = Pr/2), thereby increasing tension in the bilayer for a given pressure stimulus and thus increasing SAC activity [69] (see later Fig. 5).

Cytoskeleton interactions with the membrane are modulated by a multitude of membrane associated proteins and cytosolic cross-linking proteins. DMD is a degenerative muscle disease associated with dilated cardiomyopathy, caused by the lack of dystrophin [70, 71]. Dystrophin is a membrane-associated protein linking the intracellular cytoskeleton to the extracellular matrix. TRPC1, which is increased in the heart and skeletal muscles from mdx mice, interacts with dystrophin and α1-syntrophin [72]. Remarkably, TRPC1 silencing in skeletal myoblasts inhibits SAC activity [73, 74]. Increased calcium entry through a TRPC1-encoded SAC was proposed to contribute to cell damage in mdx hearts and skeletal muscle [73, 75]. Importantly, pharmacological blockade of SACs with GstMx4 protects skeletal myofibers from stretching-induced cell damage and reduces intracellular calcium in cardiomyocytes from aged mdx mice [24, 76].

A similar calcium overload has been demonstrated in the Homer 1 knock out myotubes, a scaffolding protein interacting with TRPC1 [73] (Fig. 4). Mice lacking Homer 1 exhibit a myopathy with decreased muscle fiber cross-sectional area and decreased skeletal muscle force generation [73]. The enhanced cation influx is suppressed by gene silencing of TRPC1 [73]. These results show that Homer 1, which is diminished in mouse models of DMD, tunes the mechanosensitivity of TRPC1 in skeletal myocytes [73]. Homer 1 has been proposed to bridge TRPC1 to the Z-disk/costamere [73]. The Z disk anchors actin thin filaments, the site of various signaling molecules and is also important for force transmission and mechanotransduction. When Homer 1 is present, TRPC1 would be open by membrane stretch (while closed at rest), but when it is absent, TRPC1 would be constitutively active (and not anymore mechano-gated). The resulting basal calcium influx would activate calpains leading to cytoskeleton proteolysis and myopathy [73] (Fig. 4). All together these data indicate that Homer 1 is a key regulator of TRPC1 mechano-gating.
https://static-content.springer.com/image/art%3A10.1007%2Fs00424-010-0847-8/MediaObjects/424_2010_847_Fig4_HTML.gif
Fig. 4

Homer-1 regulates canonical transient receptor potential channel (TRPC)1 mechano-gating. Homer proteins dimerize through coiled-coil domains and link TRPC1 to the costamere. The costameres are sub-sarcolemmal protein assemblies circumferentially aligned with the Z-disk of peripheral myofibrils. They physically couple force-generating sarcomeres with the sarcolemma in striated muscle cells. In response to stretch, SACs (i.e. TRPC1) open and allow calcium influx which activates downstream calcium signaling pathways. When Homer proteins are absent, TRPC1 is not gated anymore by stretch and instead becomes constitutively active. Constitutive calcium influx in the absence of Homer (or dystrophin in DMD patients and mdx mice) activates calpains, leading to cytoskeleton degradation and myopathy. Modified from [73]

TRPC1 also interacts with caveolins (CAV), the scaffolding proteins of caveolae [77, 78] (Fig. 3). Caveolae are invaginations of the cell membrane formed by the dynamic clustering of sphingolipids and cholesterol and function as platforms for protein attachment, trafficking and signaling [79]. Caveolae are particularly abundant in specific cell types including smooth muscle and endothelial cells, where they play a central role in mechanotransduction. Indeed, Cav1−/− mice show altered flow-dependent dilation and remodeling [80]. eNOS activation is lost in Cav1−/− mice (i.e., in the absence of caveolae), but is rescued upon endothelial CAV1 re-expression [80]. The stretch-induced smooth muscle cell proliferation is also impaired in the absence of CAV1 [81]. CAV1 and CAV3 bind to TRPC1 and promote its assembly into channels, surface delivery, localization in caveolae and regulate its activity [81] (Fig. 3). Accordingly, cholesterol depletion results in SACs inhibition [74]. Caveolins are most likely important not only for TRPC1 but also probably for other TRPC channels and have been clearly recognized as a major determinant of arterial myogenic responses [82].

The role of the cytoskeleton and scaffolding proteins in the regulation of SACs activity needs to be considered with caution as these proteins are obviously involved in a multitude of functions which might also indirectly affect channel behavior through for instance trafficking or control of signal transduction complexes.

TRPC1 forms a heteromultimeric channel together with the flow sensitive polycystin TRPP2 channel

TRPP subunits are abundant in the kidney, but are also found in other tissues at high levels, including the cardiovascular system [83]. TRPP1 (PKD1, PC1) is a large (∼460 kDa) transmembrane glycoproteins with an extended N-terminal extracellular domain, 11 TMS and a short intracellular C-terminal domain [83, 84]. TRPP2 (PKD2, PC2) is a member of the TRP channel family (see above) [85]. The TRPP1/TRPP2 complex is expressed at the surface of the primary cilium in kidney epithelial, nodal (TRPP1 is absent in these cells) and endothelial cells [8691]. An increase in intracellular Ca2+ concentration in the primary cilium is induced by shear stress and this response is lost in cells lacking TRPP1 or TRPP2 [8688, 92].

Loss of function mutations in the PKD1 and PKD2 polycystin genes are responsible for autosomal dominant polycystic kidney disease (ADPKD), the most frequent monogenic kidney disease (1/1,000) (for recent reviews [93, 94]). ADPKD is characterized by fluid-filled cysts in the kidneys, the pancreas and the liver. A cardiovascular phenotype is also associated with ADPKD. Intracranial aneurysms and arterial hypertension are among the most frequent vascular complications of this multisystem disease [83].

We recently demonstrated that TRPP2 inhibits SACs in a variety of cell types, including arterial myocytes [69] (Fig. 5). This specific effect, which is absent with other TRP channels such as TRPC1, TRPC6 or TRPV4, is antagonized by TRPP1, suggesting that the TRPP1/TRPP2 dosage tunes SACs opening. Moreover, knocking out Pkd1 in smooth muscle cells impairs SAC activity and the MR, an effect reversed by knocking down Pkd2. TRPP2 interacts with the actin cross-linking protein filamin A (FlnA) which is critically required for SAC regulation [69] (Fig. 5). FlnA makes the actin cytoskeleton stiffer and indirectly modulates SACs activity (for review [95]).
https://static-content.springer.com/image/art%3A10.1007%2Fs00424-010-0847-8/MediaObjects/424_2010_847_Fig5_HTML.gif
Fig. 5

The TRPP1/TRPP2 ratio tunes SACs activity. TRPP2 which is present either at the plasma membrane or in the endoplasmic membrane (ER) interacts with the actin cross-linking protein filamin A (FlnA). This interaction is proposed to reinforce the cortical cytoskeleton and yields to a convoluted bilayer (the upholstery model). SACs are activated by lateral tension in the membrane. According to the law of Laplace, the tension (T) at a given pressure (P) applied at the back of the patch pipette (insert) is proportional to the radius (r) of membrane curvature. When the membrane is convoluted by the cortical cytoskeleton, r is small and thus T is reduced at a given P. By contrast, in the absence of TRPP2 or FlnA or when the actin cytoskeleton is disrupted either mechanically or chemically with for instance latrunculin A, the membrane is flatten, r is increased resulting in a higher activating T at a given P. In these conditions, SAC activity is enhanced. Adapted from and for further details see [69]

Inactivation of either TRPP1 or TRPP2 impairs flow sensing by the primary cilium, while it is the TRPP1/TRPP2 dosage which tunes SACs gating [69, 86, 88].

Interestingly, TRPP2 interacts with TRPC1 to form a channel that is activated in response to GPCR activation and shows a pattern of single-channel conductance, amiloride sensitivity and ion permeability distinct from that of TRPP2 or TRPC1 alone [96, 97]. However, our own results indicate that the heteromeric channel TRPP2/TRPC1 is not directly activated by stretch [69]. Furthermore, TRPC1 does not affect the inhibitory effect of TRPP2 on SACs.

Conclusions

These recent findings indicate that TRPC1, a SOC and TRPC6, a ROC, are involved in muscle mechanotransduction. However, direct activation of these channels by membrane tension has been ruled out. Rather, indirect gating mechanisms via GPCRs (including the AT1R), acting as the primary mechanosensors, are at play. This mechanosensitive molecular pathway has been implicated in the myogenic arterial response to intraluminal pressure (TRPC6) and in cardiac hypertrophy (TRPC1 and TRPC3/6/7). Importantly, TRPC1 interacts with STIM/orai proteins contributing to SOCs [34]. Whether this complex is also involved in mechanotransduction remains to be determined. Moreover, TRPC1 is involved in the myocyte damages associated with DMD. TRPCs interact with several cytoskeletal and scaffolding proteins including dystrophin and α1-syntrophin, CAV1/3 and Homer 1, which regulate channel activity.

Whether these channels are involved, either directly or indirectly, in other important mechanosensory functions such as hearing or touch will require further investigation. The definitive subunit composition of a cardiac mechanosensitive cation channel thus still remains enigmatic. The development of specific blockers for SACs (i.e., TRPC channels) may represent a valuable pharmacological option for the treatment of cardiac hypertrophy [62].

Acknowledgements

We are grateful to the ANR 2005 cardiovasculaire-obésité-diabète, to the ANR 2008 du gène à la physiopathologie, to the Association for information and research on genetic kidney disease France, to the Fondation del Duca, to the Human Frontier Science Program Organization-long term fellowship, to the Fondation de la recherche médicale, to the Fondation de France, to the Fondation de recherche sur l’hypertension artérielle, to the Fédération pour la recherche sur le cerveau, to Société Générale AM, to the Université of Nice Sophia Antipolis and to the CNRS for financial support. We are grateful to Dr. Sophie Demolombe for critical reading of this manuscript.

Copyright information

© Springer-Verlag 2010