Extracellular Calpains Increase Tubular Epithelial Cell Mobility

Calpains are intracellular Ca2+-dependent cysteine proteases that are released in the extracellular milieu by tubular epithelial cells following renal ischemia. Here we show that externalized calpains increase epithelial cell mobility and thus are critical for tubule repair. In vitro, exposure of human tubular epithelial cells (HK-2 cells) to μ-calpain limited their adhesion to extracellular matrix and increased their mobility. Calpains acted primarily by promoting the cleavage of fibronectin, thus preventing fibronectin binding to the integrin αvβ3. Analyzing downstream integrin effects, we found that the cyclic AMP-dependent protein kinase A pathway was activated in response to αvβ3 disengagement and was essential for calpain-mediated increase in HK-2 cell mobility. In a murine model of ischemic acute renal failure, injection of a fragment of calpastatin, which specifically blocked calpain activity in extracellular milieu, markedly delayed tubule repair, increasing functional and histological lesions after 24 and 48 h of reperfusion. These findings suggest that externalized calpains are critical for tubule repair process in acute renal failure.

Calpains are intracellular Ca 2ϩ -dependent cysteine proteases (1). The major isozymes, calpain 1 or -calpain and calpain 2 or m-calpain, are distributed ubiquitously and activated in vitro by micromolar and millimolar concentrations of Ca 2ϩ , respectively. They are heterodimers composed of a ϳ80-kDa catalytic subunit (encoded by CAPN1 and CAPN2 forand m-calpain, respectively) and a common ϳ30-kDa regulatory subunit (encoded by CAPN4). Binding of Ca 2ϩ toor m-calpain induces the release of constraints imposed by domain interactions and results in a two-stage activation process: first, the release of ϳ30-kDa regulatory subunit; and second, the rearrangement of the active site cleft in ϳ80-kDa catalytic subunit (2). Calpain activity is tightly controlled by calpastatin, a specific endogenous inhibitor, which contains four equivalent inhibitory domains (1). By conducting limited proteolysis of intracellular substrates, calpain activity has been shown to be critical for a great diversity of cellular responses. They include rearrangement of cytoskeletal linkages to the plasma membrane during cell adhesion and mobility, modification of molecules in signal transduction pathways, degradation of enzymes controlling the cell cycle, and activation of proteolytic cascades leading to cell apoptosis or necrosis (1,3,4).
Recently, several groups showed that calpains may be released from cells into the extracellular environment and thus may have an extracellular role. Like other intracellular enzymes, calpains may indeed leak out from injured and dying cells such as hepatocytes exposed to toxic chemicals (5). The release of intracellular calpains from blood mononuclear cells (6), osteoblasts (7), chondrocytes (8), and parathyroid cells (9) is not due to cell death, but rather to a nonclassical pathway of secretion, which involves in certain cells the shedding of membrane vesicles (7,9). In an extracellular Ca 2ϩ -rich environment, activated calpains trigger plasma membrane proteins of the neighboring cells and extracellular matrix proteins.
The proximal straight tubule in the outer medulla of the kidney is particularly susceptible to ischemia/reperfusion injury, which remains the leading cause of acute renal failure (ARF) 3 (10,11). Damages to this segment are characterized initially by the disruption of tight junctions that control cell polarity (12,13). The loss of cell polarity is responsible for the redistribution of integrin subunits from the basolateral to the apical membrane, contributing to the shedding of cells into the tubule lumen. With more sustained ischemia/reperfusion, epithelial cells of the proximal tubule undergo necrosis or apoptosis (14). Calpains are considered as a key mediator of this death. Their activation results from both a rise in cytosolic Ca 2ϩ through endoplasmic reticulum Ca 2ϩ release (15) and a caspasedependent decrease in calpastatin activity (16). Upon activation, calpains hydrolyze the cytoskeleton-associated paxillin, talin, and vinculin, thus contributing to increased plasma membrane permeability and cell death (17). Epithelial cells that do not die participate in the regeneration of tubular epithelium and the restoration of renal function (11). They migrate into areas denuded by exfoliation, where they dedifferentiate, proliferate, and differentiate again (18). Because calpains leak out from dead tubular epithelial cells (19), the question arises as to * This work was supported in part by the Institut National de la Santé et de la Recherche Mé dicale (INSERM) and the Faculté de Mé decine Saint-Antoine. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  whether externalized calpains play a role in this repair process. Thus, our study has focused on exploring the effects of extracellular calpains on tubular epithelial cells. By using an in vitro approach, we demonstrated that extracellular calpains increased the mobility of tubular epithelial cells by promoting the cleavage of fibronectin, the disengagement of ␣ v ␤ 3 integrin, and thereby the activation of cyclic AMP signaling pathway. By using a model of ischemic ARF, we showed that extracellular calpains are indeed critical for tubule repair process.
Cell Cultures-Human proximal tubular epithelial cells (HK-2 cell line; American Type Culture Collection) were cultured at 37°C under a 5% CO 2 and 95% air atmosphere in a serum-free keratinocyte medium supplemented with human recombinant epidermal growth factor and pituitary bovine extract (Invitrogen). After 24 h, the medium was replaced with KRH medium (KCl 5 mM, NaCl 115 mM, KH 2 PO 4 1 mM, MgSO 4 1.2 mM, Hepes 25 mM, pH 7.4) supplemented with CaCl 2 2 mM, and HK-2 cells were exposed to -calpain or vehicle alone. Cell viability and cell proliferation were assessed by measuring the exclusion of trypan blue and the uptake of [ 3 H]thymidine, respectively (20).
To assess the role of extracellular matrix proteins, multiwell slides were coated with human fibronectin, collagen IV, and laminin (50 g/ml) overnight at 4°C. Cells were allowed to spread on coated wells, with or without blocking anti-␣ v ␤ 3 antibody (15 g/ml), for 24 h before exposure to calpain.
Video Microscopy and Migration Assay-For video microscopy experiments, HK-2 cells were grown on chambered coverglass system (Lab-Tek, Nunc) for 24 h before exposure to calpain. Sequential phase contrast images were taken by confocal microscopy at 10-s interval over 30 -90 min using a Leica TCS laser scanning confocal microscope (Lasertechnik).
Migration assays were performed using 8-m pore Transwell inserts (Costar). HK-2 cells (100,000) were added to the upper chamber of inserts in a serum-free keratinocyte medium supplemented with human recombinant epidermal growth factor and pituitary bovine extract. After 24 h, the medium was replaced with KRH medium supplemented with CaCl 2 2 mM before HK-2 cells were exposed to -calpain and allowed to migrate from the upper to the lower chamber for 6 h at 37°C.
Then, cells remaining on the upper surface of the Transwell inserts were removed with an absorbent tip, and cells that had migrated on the lower side of the inserts were fixed with ethanol, stained using hematoxylin/eosin, and counted.
Immunofluorescence Microscopy-HK-2 cells were allowed to spread on chamber slides (Lab-Tek II, Nunc) for 24 h before exposure to -calpain. Immunofluorescence staining of actin cytoskeleton, ␤ 1 , ␤ 3 , ␣ v ␤ 3 integrins, and fibronectin was performed on cells fixed and permeabilized with cold methanol for 10 min. Slides were incubated with appropriate dilutions of specific primary (1/200) and secondary (1/400) FITC-or TRITC-labeled antibodies. Confocal microscopy was performed using a Leica TCS laser scanning confocal microscope (Lasertechnik).
Immunoprecipitation and Western Blot Analysis-For immunoprecipitation of ␣ v ␤ 3 -fibronectin complexes, cell lysates were prepared by scraping HK-2 cells into an ice-cold protease inhibitory buffer. The lysate was centrifuged (4000 ϫ g, 4°C for 30 min). A portion of the supernatants was reserved for protein determination, and protein concentration in supernatant samples was adjusted with the protease inhibitory buffer. These samples (100 l) were incubated for 20 h with A/G agarose coupled to anti-␣ v ␤ 3 integrin antibody before the complexes were washed extensively and solubilized in SDS sample buffer with reduction.
For Western blot analysis, proteins that had been immunoprecipitated or extracted from HK-2 cells, extracellular milieu of HK-2 cells, and mouse urine samples were separated by electrophoresis on Novex BisTris 7.5% gels using XCell SureLock TM Mini-Cell (NuPAGE, Invitrogen) as described by the manufacturer and transferred onto nitrocellulose membrane (Immobilon-P, Millipore) prior to detection of integrin or extracellular matrix protein with a specific primary antibody (dilution 1/200) and a peroxidase-labeled anti-IgG secondary antibody (dilution 1/4000). Thereafter, the membrane was developed with the ECL plus detection reagent (Amersham Biosciences). Image J software was used to analyze the density of the bands in arbitrary densitometry units.
Calpain Assay-The calpain-like activity was determined in intact cells as described previously (21).
Determination of Intracellular Cyclic AMP Concentrations-HK-2 cells were exposed to -calpain in KRH medium supplemented with 2 mM CaCl 2 and 0.1 mM 3-isobutyl-1-methylxanthine. After 10 min, cyclic AMP was extracted with an ice-cold ethanol/formic acid mixture (85:5, v/v) and quantified using a radioimmunoassay ([ 125 I]cAMP assay, Amersham Biosciences) according to the manufacturer's protocol. Values were normalized to the protein concentration using a Bradford procedure and expressed as fmol/g protein.
Induction of Ischemic Acute Renal Failure-The studies were conducted by following established guidelines for animal care, and all protocols were approved by the Institut National de la Santé et de la Recherche Médicale. C57BL/6 mice were anesthetized by intraperitoneal administration of avertin (Sigma) and subjected to bilateral flank incisions as described previously (22). Both renal pedicles were cross-clamped for 25 min. This time was chosen to obtain a reproducible acute renal failure, while minimizing animal mortality. After clamp removal, the kidneys were inspected for restoration of blood flow, the surgical wounds were sutured, and animals were allowed to recover. Finally, to maintain fluid balance and volume status, mice were given 0.5 ml of warm saline intraperitoneally. After 24, 48, or 72 h of reperfusion, they were reanesthetized, blood and urine samples were collected, and kidneys were removed for morphological analyses.
Assessment of Renal Function and Histology-Samples of serum were collected from all the mice to measure both creatinine and blood urea nitrogen (BUN) using an autoanalyzer.
For light microscopy, kidneys were fixed in 4% paraformaldehyde and processed for paraffin embedding. Sections of 3-m thickness were made and stained with Masson trichromic solution. Tubular injury was scored by estimating the percentage of tubules in the cortex or the outer medulla that showed epithelial necrosis or had luminal necrotic debris and tubular dilatation as follows: 0, none; 1, Ͻ10%; 2, 10 -25%; 3, 25-50%; 4, 50 -75%, and 5, Ͼ75%) (22). All evaluations were made on 10 fields/section and 10 sections/ kidney by two different blinded observers.
Statistical Analysis-Results are expressed as mean Ϯ S.E. Comparisons between groups of values were made with the Student's t test for unrelated groups. A difference between groups of p Ͻ 0.05 was considered significant.

Extracellular -Calpain Modulates the Behavior of Proximal
Tubular Epithelial Cells-To assess initially the impact of extracellular calpains on tubular epithelial cell viability, proliferation, and morphology, we exposed HK-2 cells to purified -calpain. Concentrations of -calpain ranged between 0.1 and 2 g/ml, as such concentrations are reached in the extracellular medium of activated cells (6). Under these conditions, -calpain did not affect HK-2 cell viability and growth, as judged by trypan blue exclusion and [ 3 H]thymidine uptake, respectively (Fig. 1A). By contrast, exposure of HK-2 cells to -calpain resulted in a marked alteration of their morphology. Within 30 min, they adopted a more rounded mor-FIGURE 1. Extracellular -calpain modulates the behavior of proximal tubular epithelial cells. A, HK-2 cells were exposed to the indicated concentrations of -calpain. [ 3 H]Thymidine uptake (left) was measured after 24 h, and trypan blue exclusion (right) was measured both immediately (filled circles) and after 24-h exposure to 5 mM H 2 O 2 (empty circles), oxygen-derived free radicals being known to dramatically reduce epithelial survival in ischemic ARF. Values represent the mean Ϯ S.E. of three independent experiments. B, sequential phase contrast images of HK-2 cells were taken immediately before and 15, 30, and 45 min after the addition of 1 g/ml -calpain to the incubation medium. C, HK-2 cells were seeded into a 24-well chemotaxis chamber and exposed to the indicated concentrations of -calpain. After 6 h, migration of cells was assessed by counting cells that traversed and spread on the lower surface of the filter. Values represent the mean Ϯ S.E. of three independent experiments (*, p Ͻ 0.05). D, actin fibers (red) and integrin ␤ 1 (green) were visualized by confocal microscopy after staining with rhodamine-conjugated phalloidin and anti-␤ 1 antibody/FITC-conjugated secondary antibody, respectively. Scale bars, 20 m. E, calnexin (red) and integrin ␤ 1 (green) were visualized in control HK-2 cells by confocal microscopy. Scale bars, 20 m. phology and migrated before aggregating into heaps scattered over the monolayer (Fig. 1B and Video 1). An assay using chemotaxis chambers demonstrated further that extracellular -calpain promoted HK-2 cell migration in a concentrationdependent manner (Fig. 1C). The response to -calpain was specific because 1) the addition of a serine protease (thrombin up to 10 units/ml) or a matrix metalloprotease-9 up to 2 g/ml had no effect on HK-2 cell morphology and migration; and 2) the addition of cell-impermeable calpastatin (5 g/ml) completely prevented the response to -calpain (data not shown).
Because cell migration requires a redistribution of integrincytoskeleton linkages, we analyzed both the organization of actin cytoskeleton and the expression of integrins (Fig. 1D). The most important integrins in the proximal tubule are ␣ 3 ␤ 1 , ␣ 6 ␤ 1 , and ␣ v ␤ 3 (23). In control cells, the actin cytoskeleton was characterized by randomly oriented cytoplasmic stress fibers, and the ␤ 1 integrin was expressed at basolateral membrane and in a large intracellular pool, where it colocalized with the endoplasmic reticulum-resident protein calnexin (Fig. 1E). After -calpain addition to the culture medium, actin bundles were reinforced at the cell periphery, and ␤ 1 integrin molecules were redistributed from the endoplasmic reticulum to the plasma membrane. Similar results were shown by analyzing ␤ 3 integrin (data not shown). Together, these results demonstrate that extracellular -calpain induces cell migration and that this effect is associated with a redistribution of integrin-cytoskeleton linkages.
Extracellular -Calpain Targets Fibronectin in Extracellular Matrix of Proximal Tubular Epithelial Cells-Intracellular calpain activity is involved in the redistribution of integrin-cytoskeleton linkages during cell migration (24). Thus, we considered initially the possibility that extracellular -calpain enters the cells and/or affects intracellular calpain activity. In fact, measurements of calpain activity showed that -calpain, being added to the extracellular medium either continuously or transiently, did not affect calpain activity of HK-2 cells (Fig. 2A). Further, pre-exposure of HK-2 cells to cell-permeable calpastatin peptide (5 g/ml) did not affect their response to extracellular calpain (data not shown), excluding again a role for intracellular calpain activity. Because integrins are a known substrate for calpains (3), we next examined whether extracellular -calpain would affect HK-2 cell spreading and mobility by cleaving the extracellular domain of integrins. Western blot analysis of whole cell extracts established that HK-2 cells express ␣ 3 , ␣ v , ␤ 1 , and ␤ 3 integrin subunits (Fig. 2B). Exposure of these cells to -calpain did not result in any detectable integrin cleavage.
Thus, we finally examined whether integrin ligands in the extracellular matrix would be a preferential target for extracellular -calpain. To this aim, HK-2 cells were cultured on surfaces coated with or without definite extracellular matrix proteins including fibronectin, laminin, and collagen type IV (Fig.  2C). After 24 h of culture, HK-2 cells attached equally to either surface. Interestingly, cells attached to fibronectin adopted a more rounded morphology and migrated in response to -calpain exposure, whereas cells attached to laminin or collagen type IV did not. These data suggested that in our experimental FIGURE 3. Fibronectin cleavage by extracellular -calpain is responsible for the disengagement of ␣ v ␤ 3 integrin. A, HK-2 cells were exposed for 45 min to the indicated concentrations of -calpain. Associations between ␣ v ␤ 3 and fibronectin were determined by immunoprecipitation with anti-␣ v ␤ 3 antibody and probing with anti-fibronectin antibody. Relative binding levels of fibronectin to ␣ v ␤ 3 integrin were determined by densitometry (*, p Ͻ 0.05; n ϭ 4). B, fibronectin (red) and ␣ v ␤ 3 integrin (green) were visualized by confocal microscopy. Scale bars, 20 m. C, HK-2 cells were cultured in the presence of anti-␣ v ␤ 3 blocking antibody for 24 h. Phase contrast images of HK-2 cells were taken immediately before (upper panel ) and 45 min after the addition of 1 g/ml -calpain to the incubation medium (lower panel ).
conditions, fibronectin was a major substrate for -calpain. Western blot analysis confirmed this hypothesis. In the extracellular milieu of HK-2 cells exposed to -calpain both the monoclonal and the polyclonal anti-fibronectin antibodies reacted mainly with a fibronectin fragment of ϳ140 kDa (Fig.  2D). This breakdown product of 250-kDa fibronectin was undetectable in the extracellular milieu of control HK-2 cells. By comparison, collagen type IV was not released, and laminin ␤-1 was released without cleavage in the extracellular milieu of both control and -calpain-treated HK-2 cells (data not shown). Thus, fibronectin would be the main target of -calpain in extracellular matrix in vivo, other targets such as proteoglycans (8) not being excluded.
Fibronectin Cleavage by Extracellular -Calpain Is Responsible for the Disengagement of ␣ v ␤ 3 Integrin-The ␣ v ␤ 3 integrin is the main fibronectin receptor in proximal tubular epithelial cells (25). Thus, to determine whether fibronectin cleavage interfered with fibronectin binding to ␣ v ␤ 3 , fibronectin-␣ v ␤ 3 complexes were analyzed by both immunoprecipitation and confocal microscopy. Immunoprecipitation experiments demonstrated the presence of fibronectin-␣ v ␤ 3 complexes in HK-2 cells, which was limited significantly upon cell exposure to -calpain (Fig. 3A). Double immunolabeling of HK-2 cells showed that ␣ v ␤ 3 localized mainly to basolateral membrane, in close contact with fibronectin fibrils in extracellular matrix (Fig. 3B). Exposure of HK-2 cells to -calpain resulted initially in a marked fragmentation of fibronectin fibrils and later on in a notable accumulation of fibronectin in the cytoplasm, whereas ␣ v ␤ 3 was still expressed at the basolateral membrane (Fig. 3B). Thus, these results suggest the disappearance of fibronectin-␣ v ␤ 3 complexes in HK-2 cells exposed to -calpain.
To determine further the involvement of ␣ v ␤ 3 integrin in response to -calpain, HK-2 cells were cultured in the presence of anti-␣ v ␤ 3 blocking antibody. Anti-␣ v ␤ 3 -treated cells attached less to extracellular matrix and showed no morphology change upon exposure to -calpain, i.e. no migration and aggregation into heaps (Fig. 3C). Collectively, these results indicate that fibronectin cleavage by -calpain is responsible for ␣ v ␤ 3 disengagement, which is essential to modify HK-2 cell spreading and mobility.
Unligated ␣ v ␤ 3 Integrin Induces PKA-dependent Mobility of Proximal Tubular Epithelial Cells-Integrin ligation is a potent regulator of PKA, and in turn, PKA pathway regulates actinbased migration of cells (26,27), including tubular epithelial cells (28,29). Thus, we tested the hypothesis that the cyclic AMP/PKA pathway would become activated following calpaindependent disengagement of ␣ v ␤ 3 and was involved in HK-2 cell migration. Fig. 4A shows that exposure of HK-2 cells to -calpain for 10 min produced a concentration-dependent increase in intracellular cyclic AMP levels. Blocking cyclic AMP accumulation and PKA activity by using the cell-permeable adenylyl cyclase inhibitor 2Ј,5Ј-dideoxyadenosine (data not shown) and the PKA inhibitor H-89 ( Fig. 4B and Video 2), respectively, completely prevented -calpain-induced HK-2 cell migration. By contrast, H-89 did not prevent -calpain-dependent fibronectin cleavage (Fig. 4C). This response was specific, as blocking PKB/Akt and protein kinase C activities by using 1-L-6-hydroxymethyl-chiro-inositol 2-(R)-2-O-methyl-3-O-octadecylcarbonate and calphostin C, respectively, was ineffective in preventing -calpain-induced HK-2 cell migration (data not shown). Our results indicate that ␣ v ␤ 3 binding to fibronectin would suppress PKA activity in HK-2 cells, and conversely, ␣ v ␤ 3 disengagement would restore this activity, thereby inducing cell migration.
Externalized Calpains Play a Role in Tubular Repair after Renal Ischemia-Tubular epithelial cell migration is an essential feature of tubular repair. Indeed, epithelial cells surviving ischemia migrate over denuded areas of basement membrane before proliferating to replace lost cells (30). Because we demonstrated in vitro the role of extracellular calpains in tubular epithelial cell migration, we investigated in vivo their involvement in tubular repair after renal ischemia. To this aim, calpastatin domain I, a very specific non-cell-permeable calpain inhibitor, was administered intraperitoneally to mice, just after induction of ischemic ARF. This small macromolecule (14-kDa molecular mass) would cross the glomerular capillary wall readily and thus reach the tubular lumen. In vehicle-treated mice subjected to bilateral renal pedicle clamping for 25 min, serum creatinine levels increased after 24 h of reperfusion, stabilized at 48 h, and decreased progressively thereafter (Fig. 5A). BUN demonstrated a similar pattern. In mice given calpastatin, FIGURE 4. Unligated ␣ v ␤ 3 integrin induces PKA-dependent mobility of proximal tubular epithelial cells. A, HK-2 cells were exposed for 10 min to the indicated concentrations of -calpain before cyclic AMP accumulation was measured by radioimmunoassay. Values represent the mean Ϯ S.E. of three independent experiments (*, p Ͻ 0.05). B, HK-2 cells were exposed for 10 min to 10 M H-89, a specific PKA inhibitor. Phase contrast images of HK-2 cells were taken immediately before (left) and 45 min after the addition of 1 g/ml -calpain to the incubation medium (right). C, Western blot analysis of fibronectin cleavage and release into the extracellular milieu of HK-2 cells exposed to the indicated concentrations of -calpain in the absence or the presence of H-89.
the ischemia/reperfusion-dependent rise in serum creatinine and BUN was more severe and markedly prolonged (Fig. 5A). Remarkably, this worsening of renal dysfunction was associated with a significant increase in tubular injury (including necrosis, tubular dilation, and sloughing of epithelial cells) in the cortex and the outer medulla as assessed by semiquantitative analysis (Fig. 5B).
Extracellular -calpain degraded fibronectin in extracellular matrix of HK-2 cells (Fig. 2D). Thus, we next examined whether calpastatin administration would prevent fibronectin degradation by externalized calpains in postischemic kidney. Immunofluorescence studies demonstrated a faint fibronectin staining at the basal surface of tubular epithelial cells after 24 and 48 h of reperfusion (Fig. 6A). There was a marked increase in this staining in mice given calpastatin. Fibronectin was not detectable by immunoprecipitation and Western blot analysis in urine of control mice (Fig. 6B). After 24 h of reperfusion, a ϳ30-kDa fibronectin fragment became detectable, suggesting a release of the ϳ140-kDa fibronectin fragment from tubule extracellular matrix and its subsequent degradation by proteases in the urinary tract. Densitometric analysis showed that this expression decreased significantly in mice given calpastatin (8.2 Ϯ 1.5 versus 19.5 Ϯ 3.7 densitometry units, p Ͻ 0.05, n ϭ 7). Therefore, externalized calpains seem to play a key role in kidney repair after an ischemic insult. This protective effect is accompanied by a degradation of extracellular matrix fibronectin.

DISCUSSION
Here, we have provided evidence that -calpain released from necrotic epithelial cells in ARF would play an essential role in tubule repair. Specifically, in vitro exposure of tubular epithelial cells to extracellular -calpain reduced their adhesion to fibronectin in extracellular matrix and thereby increased their mobility, which was critical to repair. Effective concentrations of calpain (ϳ1 M) were presumably reached in the fluid of proximal tubules 24 h after induction of ischemic ARF, as estimated by measuring calpain activity in urine and taking tubule fluid concentration process along the nephron into account (ϳ4 M).
One of the hallmarks of ischemic ARF is tubular cell necrosis, a damage leading to calpain externalization. Previous studies have demonstrated that externalized calpains cleave extracellular proteins, including the latent form of transforming growth factor-␤ (31) and the extracellular matrix components, vitronectin and fibronectin (32,33). Consistent with this latter report, we found that extracellular -calpain hydrolyzes fibronectin. In the kidney, fibronectin is localized in interstitial matrix, being secreted in particular by proximal tubular epithelial cells. As a consequence of ischemia-reperfusion injury, interstitial matrix proteins such as fibronectin are exposed (18) and thus become a new substrate for externalized calpains. Our data indicate that fibronectin hydrolysis by -calpain results in the appearance of fibronectin breakdown products in the extracellular milieu both in vitro and in vivo. This process could help explain the observation that immunoreactive fibronectin is abundantly expressed in tubular lumen 3-24 h postischemia (18). We measured the size of fibronectin breakdown products by Western blot analysis. Both the monoclonal and the polyclonal anti-fibronectin antibodies reacted mainly with a fragment of ϳ140 kDa (Fig. 2D). Because the polyclonal antibody is directed against the COOH terminus of fibronectin molecule, the ϳ140-kDa fragment contains this domain and thus results from the cleavage of fibronectin molecule at a site located between type III domains 7 and 10. This region is involved in interaction with several integrins (e.g. ␣ v ␤ 3 ) (34) so that its proteolytic cleavage would impede fibronectin binding to such integrins.
In the kidney, fibronectin interacts with ␣ v ␤ 3 and weakly or not with ␣ 3 ␤ 1 at the surface of proximal tubular epithelial cells (25,35). We indeed found a direct interaction of fibronectin with ␣ v ␤ 3 integrin by coimmunoprecipitation and colocalization experiments. This binding was markedly reduced in HK-2 cells exposed to -calpain, providing strong evidence for the hypothesis that fibronectin degradation by extracellular calpains causes ␣ v ␤ 3 disengagement. A recent work by Akimov et al. (36) reported that tissue transglutaminase interacts directly with fibronectin in extracellular matrix and ␣ v ␤ 3 integrin on the cell surface, thereby potentiating integrin-mediated cell adhesion and outside-in signal transduction. As calpains are known to cleave tissue transglutaminase (37), they could regulate ␣ v ␤ 3 downstream signaling by hydrolyzing tissue transglutaminase in addition to fibronectin.
Integrin ligation activates downstream signaling pathways that are involved in the control of cell migration. We found that ␣ v ␤ 3 disengagement in response to fibronectin hydrolysis activates cyclic AMP signaling pathway. Similarly, recent studies in endothelial cells have demonstrated that impeding ␣ 5 ␤ 1 binding to extracellular matrix ligand activates PKA (38). The mechanisms whereby integrin ␣ v ␤ 3 disengagement activates cyclic AMP signaling pathway are unknown. The increase in cyclic AMP resulted probably from the activation of adenylyl cyclase, given that our measurements were performed in the presence of the phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine. Furthermore, there have been reports that ␣ v ␤ 3 integrin and integrin-associated protein (CD47) may combine to form a functional seven-transmembrane complex, which associates with G i protein and,once activated, decreases cyclic AMP (39). Thus, ␣ v ␤ 3 -CD47-G i complexes could be involved in the control of adenylyl cyclase activity by ␣ v ␤ 3 engagement/disengagement. However, further work is needed to assess the precise role of these complexes in the cell response to extracellular calpain.
Cyclic AMP signaling pathway is known to exert both negative and positive effects on cell migration (27). The current observation that ␣ v ␤ 3 disengagement is associated with the accumulation of cyclic AMP in HK-2 cells and that inhibiting cyclic AMP signaling pathway impedes HK-2 cell migration provides strong evidence for the hypothesis that cyclic AMP signaling pathway plays a role in stimulating cell mobility in response to extracellular -calpain. Previous studies have also demonstrated that cyclic AMP signaling pathway stimulates tubular epithelial cell migration in vitro (28,29). In addition, cyclic AMP signaling pathway has been implicated in the protection of these cells against the cytotoxic effect of ischemia and intracellular ATP depletion (40). Because adenosine is released from injured or dying cells and induces both cyclic AMP-dependent survival and migration of tubular epithelial cells (28,29), there are at least two potential mechanisms (i.e. calpain externalization and adenosine release) whereby necrosis of epithelial cells could speed up repair of tubules in ARF.
In the present study, the concept that externalized calpains may accelerate tubule repair in ischemic ARF is supported by the demonstration of persistent lesions of tubules after administration of calpastatin domain I. The decreased level of fibronectin breakdown product in urine of mice given calpastatin indicates clearly that calpastatin reached tubule lumens and effectively targeted externalized calpains. By contrast, a recent report shows that two cell-permeable calpain inhibitors, PD 150606 and E-64, reduce the renal dysfunction and injury caused by ischemia reperfusion (41). Altogether, this is good evidence that extracellular calpains triggered by calpastatin domain I contribute to tubule repair, partly by inducing epithelial cell migration, whereas intracellular calpains triggered by PD 150606 and E-64 participate in tubule injury, partly by increasing oxidative stress. To our knowledge, only one other study has examined the role of externalized calpains in repair FIGURE 6. Calpastatin administration prevents fibronectin degradation by externalized calpains in postischemic kidney. Mice received normal saline or calpastatin domain I (6 mg/kg) intraperitoneally immediately prior to and once a day during the reperfusion period. They were subjected to 25 min of renal ischemia followed by the indicated periods of reperfusion. A, fibronectin (red) and cell nuclei (blue) were visualized by fluorescence microscopy. Representative kidney section of untreated (left) and calpastatin domain I-treated (right) mice after 48-h reperfusion is shown (ϫ200). B, ϳ30-kDa fibronectin cleavage product was detected in urine of mice after 24-h reperfusion by electrophoretic separation of anti-fibronectin immunoprecipitations and immunoblotting with anti-fibronectin antibody. Double arrowheads indicate IgG heavy chain that was recognized by the secondary antibody.
process. Mehendale and Limaye (5) reported that calpains released from necrotic hepatocytes hydrolyze proteins in the plasma membrane of neighboring cells, leading to progression of injury rather than to repair. One hypothesis to explain the differences between this result and the present study is that calpain concentrations reached in the extracellular milieu would be markedly higher in liver exposed to acute toxic insult than in ischemic kidney.
The major conclusion of our work is that calpains released from necrotic cells in ischemic ARF trigger repair process. This effect is likely secondary to fibronectin cleavage, ␣ v ␤ 3 integrin disengagement, and cyclic AMP signaling pathway activation. Thus, our study highlights the importance of extracellular calpains and intracellular cyclic AMP as targets for therapeutic intervention in ARF. It also proposes a novel evidence that necrotic cells may directly trigger the tissue repair response.