Tubular mechanical stretch is the key primary insult in obstructive nephropathy. This review addresses how the renal tubular epithelium senses and responds to mechanical stretch. Using data from renal and nonrenal systems, we describe how sensing of stretch initially occurs via the activation of ion channels and subsequent increases in intracellular calcium levels. Calcium influxes activate a number of adaptive and proinjury responses. Key among these are 1) the activation of Rho, consequent cytoskeletal rearrangements, and downstream increases in focal adhesion assembly; and 2) phospholipase activation and resultant mitogen-activated protein kinase activation. These early signaling events culminate in adaptive cellular coupling to the extracellular matrix, a process termed the cell strengthening response. Direct links can be made between increased expression of genes involved in the development of obstructive nephropathy and initial sensing of mechanical stretch. The review illustrates the repercussions of mechanical stretch as a renal stress stimulus, specific to ureteric obstruction, and provides an insight into how tubular responses to mechanical stretch are ultimately implicated in the development of obstructive nephropathy.
- tubular epithelium
Incidence, Causes, and Consequences of Ureteric Obstruction
Ureteric obstruction (UO) is a common and significant urological diagnosis (lifelong incidence of ∼1/1,000). Evidence from animal models indicates that, even with acute recovery following relief of UO, function may be compromised in the long term by progressive renal fibrosis (45).
Obstruction may be classified according to cause (congenital or acquired), duration (acute or chronic), degree (partial or complete), and side (unilateral or bilateral). There are numerous causes of acquired obstruction: calculi (stones), commonest cause; vesical (bladder) tumor involving one or both ureteric orifices; local extension of prostate/cervical cancer into base of bladder, occluding ureters; ureteric tumor; compression of ureters at pelvic brim by metastatic nodes from prostate/cervical cancer; pregnancy; retroperitoneal fibrosis/tumor; ureteric stricture (postinfection/injury); and iatrogenic.
Causes of Tubular Cell Mechanical Stretch in UO
There are a number of stimuli that cause tubular mechanical stretch in UO. Initially, increases in renal pelvic pressure are important, while in sustained obstruction, the volume and composition of pooled urine also provide a tubular stretch stimulus.
Normal pressures at rest within the renal pelvis and ureter are ∼0–10 cmH2O. Peristaltic pressures for the transport of urine vary between 20 and 60 cmH2O. In acute UO, there is a sudden rise in ureteric and intrarenal pressure. Pressures in excess of 80 cmH2O may be generated. This elevation in pressure is transmitted back to the tubular lumen and is enhanced further by early increments in renal blood flow and glomerular filtration rate (GFR) following local increases in nitric oxide generation (111, 128).
Experimental evidence has shown that stepwise increases in renal pelvic pressure cause a bradykinin-mediated increase in prostaglandin E2 and subsequent N-type calcium (Ca2+) channel-dependent elevations in substance P release from renal sensory neurons (54, 56). This activates a reno-renal reflex characterized by decreased efferent sympathetic activity, causing an enhancement in contralateral kidney urinary flow rate (64). Following a period of sustained unilateral UO (UUO) in rats, artificial elevation of intrapelvic pressure does not induce this reno-renal reflex, possibly due to angiotensin II inhibition of substance P release and/or decreased active substance P secondary to enzymatic degradation (55, 64).
However, this maximal pelvic pressure elevation described above begins to fall. The reasons for the reduction of pressure are threefold in nature. First, there is increasing dilatation of the renal pelvis. Second, we see reduced renal blood flow and GFR (after 1 wk of UUO, GFR in the obstructed kidney is reduced to 20% of preobstruction levels) (58, 123). Finally, altered pyelolymphatic and pyelovenous backflow help to reduce pressure (39, 79). At this point, it is important to consider the pressure/stretch effect of the continued pooling of urine following obstruction.
In acute UUO, the cortical collecting tubules become resistant to arginine vasopressin (AVP), and so cannot produce concentrated urine (35). Aquaporins (AQP) are membrane water channels that play a pivotal role in controlling the water content of cells. AQP1 is most prominent in the proximal tubule and descending thin limb and is critical for urinary concentration. AQP2 is only found in the principal cells of the connecting tubule and collecting duct and is the chief AVP-regulated water channel. AQP3 and AQP4 are both present in the basolateral plasma membrane of collecting duct principal cells and most probably represent exit pathways for water reabsorbed apically via AQP2 (83). A reduced ability to respond to the AVP signal leads to reduced abundance of apical AQP water channels in UUO and is at least partly responsible for the defect in urinary concentration observed (132). Another important contributory factor in the loss of urinary concentration capacity is the UO-induced loss of urea transporter expression, which tends to lead to medullary washout (59).
It is established that sodium handling (and associated water reabsorption) is progressively impaired following graded obstruction, in both the distal and proximal tubules (89). Expression of apical sodium transporters is decreased in the obstructed kidney in UUO in rats (51). Resistance to AVP in UO is also likely to play a part in diminished apical recruitment of sodium channels; mobilization of these channels is demonstrated to be AVP dependent in renal A6 cells (53).
The loss of urinary concentrating capacity leads to the generation of a large volume of pooled urine rich in electrolytes, yet relatively hypotonic (Docherty NG, Fuentes-Calvo I, Quinlan MR, Perez-Barriocanal F, Dillon-Murphy R, Wright E, McGuire BB, Fitzpatrick JM, Watson RW, unpublished observations). Aside from the volumetric stretching effect of large quantities of static urine, its hypotonic nature per se is likely to exert a further wave of swelling-induced mechanical stretch insult, particularly in tubular cells in regions of the nephron routinely exposed to filtrate of a higher osmolality.
The changes in renal pressure and urinary pooling thus described lead to mechanical stretching of the tubular epithelium, which is widely accepted as the key initiating injury underpinning the subsequent development of obstructive nephropathy.
Sensing and Responding to Mechanical Stretch in the Tubular Epithelium
The following section draws on relevant evidence to elucidate the response of the tubular cell to mechanical stretch.
Integrins as mechanosensors and subsequent Ca2+ influx.
Mechanical stresses are imparted to cells through the extracellular matrix (ECM). Mechanical signals propagate from the ECM and converge on cell surface adhesion receptors called integrins, which connect intracellularly to the actin cytoskeleton within focal adhesion units (4). Of particular importance in epithelia is the integrin receptor for the basement membrane protein, fibronectin (93). This, as with all integrins, is a heterodimeric protein consisting of an α- and β-subunit, which mediate adhesion via divalent cation-dependent binding of Arg-Gly-Asp and Pro-His-Ser-Arg-Asn domains (3, 76). The nonimmune cell-related fibronectin receptor consists of the α5β1-subunits (70).
Focal adhesion complexes can contain a large and variable number of proteins. The structural scaffold of the complex consists of the association of the cytoplasmic integrin domains with the actin cytoskeleton, via α-actinin, talin, and vinculin proteins (17). The role of this structure as both an inside-out and outside-in mechanotransducer relies on the association of kinase signaling molecules. Aside from the ubiquitously found focal adhesion kinase, two kinase enzymes that interact with the β1-integrin are of particular interest in mechanosensation in epithelia; these are c-Src and integrin linked kinase (ILK) (14, 36). These are discussed below.
Integrins offer a route for the transfer of a mechanical force across the cell membrane (126). Cells respond to forces applied to integrins by generating a stress-induced strengthening response, consisting of reinforcement of the cellular cytoskeleton (126). This occurs in a Ca2+-dependent manner and aids the ability of the cell to resist deformation and injury (32).
Evidence suggests that integrins can regulate ion channel activity. In endothelial cells, integrins regulate cell spreading by controlling the level of cytosolic Ca2+ (112). Integrins can be both physically and functionally linked to certain ion channels (6). Recent work suggests that stretch-activated ion channels can be found within integrin-focal adhesion units, emphasizing the probable role of Ca2+ influx in the adaptive response to stretch (D. E. Ingber, personal communication).
The role of integrins as cellular mechanosensors/mechanotransducers has been described in a variety of cell types, including bovine capillary endothelial cells, keratinocytes, and human umbilical vein endothelial cells (HUVECs) (52, 68, 108).
Very few studies have investigated integrin expression and function in tubular epithelium during renal injury. However, it is has been shown that, in the obstructed kidneys of mice with UUO, there is an increase in basolateral tubular epithelial integrin α5-subunit expression (130). The authors suggest that this might aid epithelial-mesenchymal transition, but, irrespective of this, it does suggest that integrins play a role in the adaptive response to mechanical stretch.
ILKs are important multidomain focal adhesion proteins. They interact with the cytoplasmic domains of integrins and control actin rearrangements at integrin-adhesion sites (105). ILKs have been shown to partake in the regulation of several integrin-mediated processes, among them changes in cell shape, cell adhesion, and gene expression (131). It is known that, in the zebrafish heart, ILKs are involved in sensing mechanical stretch (8). In another study, the increase in matrix metalloproteinase-2 levels following mechanical stretch of trabecular meshwork cells is ILK dependent (11).
Transforming growth factor (TGF)-β1, a potent profibrotic and proapoptotic agent discussed in more detail later in this review, induces ILK expression in renal tubular epithelial cells, and, of note in the same paper, in vivo ILK is markedly induced in tubular epithelial cells secondary to UO in rats and mice (60). The authors demonstrate that, in the early stages following UUO, ILK induction occurs specifically at the basolateral aspect of the tubular epithelium region, indicating that this phenomenon may be an adaptive response to stretch.
Connective tissue growth factor (CTGF) mediates the downstream fibrogenic activity of TGF-β1, and studies in HK-2 cells (immortalized human proximal tubular epithelial cell line) demonstrate that CTGF induces expression of ILK via extracellular signal-regulated kinases 1 and 2 (ERK1/2) and phosphatidylinositol 3-kinase (63).
As mentioned before, the stress-induced strengthening response of the cell, consisting of integrin-dependent reinforcement of the cellular cytoskeleton (Fig. 1), has been shown to occur in a Ca2+-dependent manner (32, 126). Interestingly, stretch- and swelling-activated cation (SSAC) channels can be found within integrin-containing focal adhesion units, implicating a role of Ca2+ influx in cellular mechanosensing (D. E. Ingber, personal communication). SSAC channels are of fundamental importance in sensing and transducing external mechanical stresses (84). Mechanosensitive channels are ion channels whose activity is contingent on a membrane stress (31). Numerous cell surface ion channels have been identified that become activated when the cell membrane is mechanically tensed (43). A stretch-activated, mechano-gated, nonselective, ATP-sensitive cation channel in the basolateral membrane of the proximal tubule of amphibian kidneys (Ambystoma tigrinum) has previously been described (41). This functions as a Ca2+ entry pathway, and it may be particularly important when intracellular ATP concentration is depleted, as occurs in times of ischemia. Elsewhere, it is known that there is a Ca2+-permeable, stretch-activated cation channel at the apical membrane of the renal proximal tubule of salamander (Necturus) (28). The apical membrane as a site of mechanical stretch-sensing is discussed below. There are notable similarities between this stretch-activated channel and the 22-pS stretch-sensitive channel reported in a cultured cell line derived from opossum kidney (122). In conditions of membrane stretch, such as hypotonicity, the outwardly rectifying potassium current is Ca2+ dependent in cultured rabbit kidney proximal tubule cells (49). Potassium secretion is also Ca2+ dependent in rabbit cortical collecting ducts (61). Mechanical stretch leads to an increase in intracellular Ca2+ in both principal and intercalated cells in the cortical collecting duct of adult female New Zealand White rabbits (62). This is due to coupling between extracellular Ca2+ entry at the basolateral membrane and internal Ca2+ release.
The exact structure of these SSACs is incompletely understood. There is evidence claiming that, in vertebrate cells, the transient receptor potential (TRP) protein C1 (TRPC1) channel forms the stretch-activated cation channel (66). The TRP family represents nonselective cationic channels of six transmembrane spanning subunits and is subdivided into three branches: the TRPC (or short TRP) subfamily, the TRPM (or long TRP) subfamily, and the TRPV (or vanilloid receptor) subfamily. The TRPC subfamily, in common with all family members, is Ca2+ permeable (71, 97).
TRPC1 is an 810-amino acid channel, with a putative pore region found between loops 5 and 6 (129). Additionally, it possesses three or four N-terminal ankyrin-like cytoskeletal association repeats (66, 129). Expression studies indicate that TRPC1 can operate as a homoligomer or a heteroligomer with TRPC-3, -4, and/or -5, as well as polycystin-2 (40, 71, 116, 121). Variation in subunit assembly is proposed to modify mechanosensitivity. In endothelial cells, TRPC1 channel opening has been shown to be dependent on protein kinase Cα-induced phosphorylation, actin polymerization, and RhoA-dependent association of TRPC1 with the inositol 1,4,5-triphosphate (IP3) receptor (1, 69).
Human TRPC1 mRNA has been demonstrated in the kidney (97). TRPC1 has been identified in the basolateral membrane in polarized Madin-Darby canine kidney (MDCK) cells (renal epithelial cells) in culture (7). Another study has further pinpointed TRPC1 channels in the tubules of the cortex and outer medulla; indeed, TRPC1 is essentially restricted to the glomerulus and proximal tubule (33). Strikingly, this paper additionally showed colocalization of TRPC1 with AQP1 in cortical tubular elements, suggesting that TRPC1 can also be found in the apical brush border of the proximal tubule.
TRPC1 is known to interact with calmodulin; indeed, the entry of cations through the TRPC1 channel is feedback inhibited by Ca2+ through its interaction with calmodulin (101).
Another potentially important site of Ca2+ influx in response to stretch in kidney epithelial cells is at the cilium. Cilia are tiny hairlike appendages extending from the cell surface. Motile cilia are not normally expressed by renal tubular cells in the mammalian kidney (85). However, there is evidence that the primary cilium in MDCK cells is, nonetheless, mechanically sensitive and responds to bending (simulating the effect of an increase in apical pressure/stretch) by increasing intracellular Ca2+, via influx through mechanically sensitive channels that appear to reside in the cilium or its base (92). The influx is then followed by Ca2+ release from IP3-sensitive stores. Of significance, TRPC1 has been identified in isolated primary cilia of LLC-PK1 cells (immortalized porcine proximal tubular epithelial cells) (95).
Failure of the mechanosensory property of renal tubular cilia, secondary to inherited disorders of cilia assembly, is implicated in epithelial dedifferentiation, apoptosis, and cystic formation in polycystic kidney disease (88, 133).
Similarly, bending properties of the apical brush-border microvilli have been indicted in glomerulotubular balance. Essentially, brush-border microvilli serve a mechanosensory function in which fluid dynamic torque (flow-rate-induced mechanical bending) is transmitted to the actin cytoskeleton and modulates sodium absorption in kidney proximal tubules by coupling to increased sodium/hydrogen exchanger activity (23, 24).
We can thus reasonably conclude that Ca2+ influxes are established via stretch-activated channels (Fig. 2). The resultant influx of Ca2+ into the cytoplasm appears to induce further intraorganellar Ca2+ release, mediated by ryanodine or IP3 receptors, as demonstrated in New Zealand rabbit cortical collecting duct cells (62).
The rho signaling pathway induces focal adhesion assembly, stress fiber formation, and cytoskeleton strengthening.
When external stresses are applied to integrins, a local intracellular transduction response is activated, culminating in focal adhesion assembly (30). Studies have shown that a cell's ability to strengthen itself in response to an applied force depends on focal adhesion assembly, where the force is applied (27). Force-induced assembly of focal adhesions is mediated by activation of Rho and its downstream targets in fibroblasts (99). Rho regulates the formation of actin stress fibers and assembly of focal adhesions in adherent cells through activation of Rho kinase/Rho-associated coiled-coil forming protein kinase (ROCK), and other effector proteins such as mDia (98, 127). Actin stress fibers end at the aforementioned focal adhesions (12). Therefore, stretch-induced stress fiber organization is contingent on Rho pathway activation (48). Actin stress fiber formation secondary to mechanical force has previously been shown in many disparate cell types, including endothelial cells, cardiac myocytes, and podocytes (25, 102, 135). In renal epithelial cells (podocytes and MDCK cells), administration of a mechanical force to the cell surface by the use of fibronectin-coated ferric beads and exposure of the cells to magnetic force leads to Rho translocation and restructuring of the actin cytoskeleton (139).
Cyclic mechanical strain causes Rho activation in airway smooth muscle cells and in skeletal muscle cells, thereby increasing contractility (117, 138). Thomas et al. (120) proposed that stretch modulates BEAS-2B bronchial epithelial cell function predominantly via the activation of Rho kinases.
The Rho-ROCK system has been shown to play an important role in the pathogenesis of tubulointerstitial fibrosis following UUO in mice; inhibition of the Rho-ROCK pathway and a consequent attenuation of fibrosis is a therapeutic approach attracting attention now (75, 80). However, the relative importance of the cell types targeted by Rho-ROCK inhibition is not clearly defined, with effects on both the renal parenchymal and inflammatory cells likely to contribute to its anti-fibrotic effect. ROCK activity has also been implicated in the progression of fibrosis in other areas, namely in the liver in rats, and also in a murine model of pulmonary fibrosis (78, 115). Specifically, in obstructive nephropathy, interference with Rho-ROCK signaling is likely to alter early adaptive tubular responses to stretch.
In human embryonic kidney cells (HEK-293), an influx of extracellular Ca2+ is required for Rho kinase activation (18). Furthermore, while inhibition of Rho kinase will result in attenuation of the typical morphological cell changes (actin stress fiber formation, etc.), it will not affect intracellular Ca2+ mobilization or ERK activation, highlighting the importance of the Rho pathway downstream of Ca2+.
In fibroblasts, force application directly to the actin cytoskeleton through integrins results in the induction of actin accumulation and the recruitment of actin binding protein 280 (ABP-280), thus fortifying the membrane cortex (32). ABP-280 is a ubiquitous homodimeric actin cross-linking protein that binds to the cytoplasmic tail of integrins (113). Similar to the case with Rho activation, this is via a Ca2+-dependent mechanism, as reflected by the fact that there is no increase in ABP-280 when cells are treated with BAPTA-AM or Ca2+-free buffer. Glogauer et al. (32) suggest that there are far greater levels of cell death in those cells lacking this protein, compared with those cells with it, again emphasizing the role of the actin cytoskeleton in resisting deformation and injury, which is of relevance to tubular cell survival in UO. Importantly, ABP-280 has previously been identified as being expressed in the renal parenchyma (90).
The importance of the integrity of the actin cytoskeleton in regulating Ca2+ influx in response to stretch has been demonstrated in human gastric myocytes, wherein cytochalasin D-mediated interruption of the cytoskeleton prevented Ca2+ influx in response to hypotonic cell swelling (50). The common consensus is that the cytoskeletal architecture is linked to the local level of ECM stiffness, with integrins acting as sensors of any changes in this dynamic (42).
Based on this evidence, we propose that the initial step in detection of the stretch insult by the tubular epithelial cells is a biophysical one, i.e., activation of integrins, Ca2+-dependent activation of Rho, and subsequent changes in the actin cytoskeleton.
Phospholipase and mitogen-activated protein kinase activation.
Numerous cell signaling pathways rely on lipid moieties as signal mediators. Phospholipases are a class of enzymes that convert phospholipids into fatty acids and other lipophilic substances. Four major classes exist, namely A, B, C, and D (PLA, PLB, PLC, and PLD, respectively). PLA1 and PLA2 act on membrane phospholipids to generate arachidonic acid, which is subsequently metabolized by oxygenase enzymes to generate eicosanoid mediators, such as the leukotrienes and prostaglandins. PLB activity combines aspects of both PLA1 and PLA2 functions (46).
PLC releases IP3 and diacylglycerol from phosphatidylinositol biphosphate and is thus of major importance in the release of Ca2+ from intracellular stores and the activation of protein kinase C. The two isoforms of PLD, designated D1 and D2, catalyze hydrolytic release of phosphatidic acid from phosphatidylcholine (46).
Alexander et al. (5) demonstrate that the Ca2+ mobilization via stretch-activated channels in rabbit proximal tubular epithelial cells also triggers PLA2 activation with subsequent arachidonic acid release, in turn initiating signaling through the major mitogen-activated protein (MAP) kinase pathways, principally ERK1/2, which then serve to activate nuclear substrates. The stretch-induced activation of ERK and another of the MAP kinases, p38, is dependent on integrin activation and mobilization of extracellular Ca2+ via cation-selective, Ca2+-permeable, stretch-activated channels (2, 125). In studies involving fluid shear stress, pulling forces, or cyclical stretching of endothelial cells, it has been shown that ERK1/2 activation represents an important part of the signaling cascade (29, 114). Elsewhere, stretch-induced activation of the MAP kinases is considered a key intracellular signaling step in human BEAS-2B cells, bladder smooth muscle cells, and keratinocytes, respectively (52, 57, 86).
The paper of Alexander et al. (5) proposes that stretch- and Ca2+-induced ERK1/2 activation is dependent on phosphorylation of the epidermal growth factor receptor kinase and c-Src activation. In alveolar epithelial cells, mechanical stretch activates ERK1/2 via the epidermal growth factor receptor (16, 106). In HUVECs, it is also known that cyclic stretch activates c-Src downstream of stretch-activated channels, with this effect abolished by extracellular Ca2+ depletion and a blocker of stretch-activated channels (gadolinium) (81). It is known that ERK activation occurs both in in vitro mechanical stretch of tubular cells and in vivo following UO, in which ERK activation is maximal in the most dilated (most stretched) sections of the tubule (5, 67).
Interestingly, in normal rat kidney-52E cells, a rat renal proximal tubular cell line, it is reported that mechanical stretch mediates its proapoptotic effects through the MAP kinase signaling pathways, in this case via c-Jun NH2-terminal kinase/stress-activated protein kinase and p38 (82). The likelihood is that, as in rat cardiac myocytes, tubular cell stretch in UO can rapidly activate a multitude of second-messenger pathways, including phospholipases and downstream MAP kinases (103). Thus we can see that, as Ca2+ levels rise in response to stretch, the MAP kinase pathways are triggered.
Activation of both the MAP kinase and Rho pathways, and reorganization of the actin cytoskeleton, thus represent highly significant steps that occur as a result of the stretch insult associated with UO.
The gap junction as a site of intercellular spread of mechanical stretch-induced increases in intracellular Ca2+.
It is interesting to ask whether there is a concerted response to mechanical stretch in tubular cells occurring through coordinated responses in adjacent cells, i.e., do changes occurring in one cell propagate to another? Interepithelial structure is maintained by cell-cell interactions, involving tight junctions, cadherin-based adherens junctions, and gap junctions, the last of which permits chemical interplay between adjacent cells (94). Gap junctions are found between tubular cells throughout the nephron, as evidenced by connexin immunostaining and RT-PCR of microdissected nephron segments (34, 38). Gap junctions are an important site for wavelike transfer of oscillations in intracellular Ca2+. In airway epithelial cells, mechanical stimulation induces an influx of Ca2+, subsequent Ca2+ release from intracellular stores, and then spread of the signal from the affected cell to the adjacent cells (107). Evidence would suggest that these Ca2+ waves are communicated by the intercellular movement of IP3 via gap junctions between ciliated epithelial cells (107). Boitano et al. (9) also provide evidence that IP3 acts as a cellular messenger mediating transmission through gap junctions between epithelial cells. Felix et al. (26) concur with this mechanism, concluding that stretching of airway epithelial cells leads to intercellular Ca2+ signaling secondary to PLC-synthesized IP3. Thus the gap junction is likely to represent a key mediator of cell-to-cell communication in stretched, renal tubular epithelial cells in UO.
Examples of How Cellular Perception of Mechanical Stretch Leads to the Activation of Downstream Effectors of Renal Injury
We have thus visited the likely mechanisms by which mechanical stretch might be initially sensed and responded to by the tubular epithelium. The remainder of the article will be devoted to using specific examples to illustrate how the sensing of mechanical stretch can be linked to the initiation of responses in the affected kidney that are associated with progressive renal damage in obstructive nephropathy.
Connecting mechanical stretch to generation of tubular cell oxidative stress.
Oxidative stress is heavily implicated in the progression of renal injury (37). Oxidative stress initiated in the renal parenchyma can result in injury and death, which, in turn, promotes the influx of inflammatory cells containing mediators, which, in themselves, accentuate the prooxidant state (e.g., myeloperoxidase from neutrophils/macrophages).
Tubular cell death in response to mechanical stretch occurs via caspase-dependent apoptosis linked to increases in oxidative stress (91, 96). Mechanical stretch of proximal tubular epithelial cells of up to 24 h causes downregulation of the expression of the mRNA for the endogenous antioxidant catalase compared with unstretched control cultures (96). Acatalasemic mice show an increase in tubular apoptosis following UUO, indicating that hydrogen peroxide accumulation contributes to oxidative stress-related cell death in this model (119). Reductions in oxidative stress post-UUO have also been implicated in the prevention of fibrosis in rats and mice treated with angiotensin II receptor antagonists (65, 118).
Looking specifically at potential pathways involved in oxidative stress-induced renal injury, it has been shown in renal epithelial cells that oxidative stress-induced apoptosis is preceded by phosphorylation of ERK (87). This is of note, given the previously described role of Ca2+ mobilization, PLA activation, and phosphorylation of ERK1/2 in response to stretch.
Connecting mechanical stretch to the induction of TGF-β1 and CTGF: major mediators of progressive renal injury in obstructive nephropathy.
A host of cytokines and growth factors contribute significantly to obstruction-induced apoptotic cell death and renal fibrosis (72). TGF-β1 is a potent profibrotic and proapoptotic agent in renal disease (10, 109). This is emphasized in studies showing that TGF-β1 plays a crucial role in the accumulation of ECM during renal fibrosis (137).
TGF-β1 exerts its profibrotic effects principally through the activation of the Smad pathway, transcription factors that are activated by phosphorylation downstream of ligand binding to the TGF-β1 receptor complex (110). Significantly, an absence of Smad 3 in UUO appears to protect mice against tubulointerstitial fibrosis, stressing the importance of this downstream signaling molecule (109). This same paper demonstrates how mechanical stretching of cultured renal epithelial cells induces profibrotic changes following Smad 3-mediated upregulation of TGF-β1.
Renal TGF-β1 mRNA expression increases considerably after the onset of obstruction, initially occurring most strikingly in the stretched tubular cells (47). In renal injury, TGF-β1 is also produced by infiltrating macrophages, a component of the inflammatory response to UO (21, 22). In addition, we know that interstitial fibroblasts synthesize TGF-β1 and that, in vivo in the rat kidney, antisense-mediated silencing of TGF-β1 expression in interstitial fibroblasts significantly reduced interstitial fibrosis following subsequent UO (44, 72). UO (mild to moderate) was also found to increase plasma TGF-β1 concentrations in patients with ureteric calculi (124).
In vitro mechanical stretching of normal rat kidney-52E cells is associated with a significantly increased release of TGF-β1, while antibody-mediated neutralization of TGF-β1 prevents the stretch-induced apoptotic response (73, 74). Interestingly, in osteoblasts, fluid shear stress induced TGF-β1 production, and this effect is significantly inhibited by stretch-activated cation channel blockade (104). This suggests that, in UO, early increases in tubular TGF-β1 expression might be mediated by the stretch-induced increase in cytosolic Ca2+, previously described in the text.
Besides the physical cell strengthening effect of changes in the actin cytoskeleton in response to stretch, it might be proposed that such changes in the cytoskeleton can also have knock-on effects on cell signaling and ultimately gene transcription. Blockade of CTGF following UUO is associated with downregulation of some of the common markers of fibrosis, such as fibronectin (134). Chaqour et al. (13) have demonstrated that, in cyclically stretched bladder smooth muscle cells in vitro and in the detrusor muscle of mechanically overloaded bladder in rats, mechanical stretch-induced changes in actin dynamics ultimately mediated nuclear factor-kB activation and induction of the expression of the murine CTGF homologue, CCN2.
Studies in HUVECs support this contention, given that, in response to shear-stress, Rho-dependent actin polymerization directly leads to an increase in CTGF expression (77). Therefore, early induction of CTGF in UUO may rely directly on the Ca2+-dependent cytoskeletal rearrangements induced downstream of Rho in response to mechanical stretch and serves as a good example as to how early adaptive responses to mechanical stretch in UUO might act as primers for the development of the renal fibrotic process.
Osteopontin: a mechanical stretch-induced proinflammatory gene product.
An array of chemokines mediate inflammatory influx into the renal interstitium in UUO. Osteopontin is a monocyte chemotactic glycoprotein secreted by tubular cells, and its expression is increased as early as 4 h post-UUO, a period coinciding with high retrograde pressure in the kidney (19). Interestingly, osteopontin expression and secretion by the proximal tubule are increased in an angiotensin II, type 1 receptor-dependent manner following cyclic mechanical stretching of freshly isolated rat proximal tubular cells (20). Of further interest concerning the upregulation of this mediator in response to mechanical stretch is the observation that osteopontin expression is increased via intracellular Ca2+ mobilization and activation of MAP kinases in MC3T3-E1 osteoblasts secondary to fluid shear stress (136). Osteopontin is thus a good example of a gene that is directly regulated by mechanical stretch and resultant elevations in intracellular Ca2+ and subsequently acts to provoke an inflammatory response. Therefore, sensing of mechanical stretch and subsequent ion fluxes and activation of cell signaling pathways can be connected to the generation of inflammation of UO.
Mechanical stretching of the tubular epithelium, caused by retrograde pressure shifts and urinary pooling, is regarded as a highly significant step in the progression of obstructive nephropathy.
The review illustrates the repercussions of mechanical stretch as a renal stimulus, specific to UO. We have explored the question of how renal tubular epithelial cells sense stretch, how this is converted into biochemical messages (ion fluxes, cell signaling), and ultimately, the mechanism by which cells adapt to stretch and instigate responses leading to the common features of obstructive nephropathy.
In summary, we suggest that stretch-activated cation channels sense stretch in the first instance, leading to increases in intracellular Ca2+, which, in turn, activates a number of adaptive responses. Key among theses are 1) the activation of Rho, which causes focal adhesion assembly, as well as actin cytoskeletal rearrangements; and 2) phospholipase activation, leading to activation of the MAP kinases. These events combine with other important events, such as activation of TGF-β1 expression and induction of oxidative stress, to elicit the early injury response (Fig. 3).
M. R. Quinlan was supported by a research grant from the British Urological Foundation.
We thank the Department of Clinical Photography in The Mater Misericordiae Hospital for assistance.
- Copyright © 2008 the American Physiological Society