INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

THE ROLE OF THE ACTIN cytoskeleton in the loss of structure and function that is associated with renal ischemia or ATP depletion has become increasingly evident (19). Under these conditions, marked alterations in the dynamics of actin turnover and assembly result in enhancement of polymerization, disruption of cortical and microvillous cytoskeleton, and diffusion of filamentous actin (F-actin) throughout the renal epithelial cells. During reperfusion or ATP repletion, these alterations are rapidly stabilized and then reversed in sublethally injured cells. The mechanisms by which this remodeling of actin occurs are unknown.

Small stress proteins (sHSPs) include single-copy gene products heat-shock protein 27 (HSP-27) in humans and HSP-25 in rodents (1). Increasing evidence supports a major function of these sHSPs in the regulation of actin dynamics (1, 2, 6, 10-14, 17). Overexpression of HSP-27 prevents microfilament disruption by cytochalasin D (11). Also, HSP-27 demonstrates actin capping activity and enhancement of mitogen-stimulated actin polymerization (12, 17). Since these functions of HSP-25 would be beneficial to the postischemic restructuring of the actin cytoskeleton, we investigated the expression and distribution of HSP-25 in the rat renal cortex after ischemia.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Animal preparation. All experiments were performed on anesthetized male Sprague-Dawley rats weighing 225-300 g as previously described (25). In brief, bilateral renal ischemia was accomplished by selective occlusion of the right renal artery and aorta just proximal to the left renal artery. After 45 min, the clamps were removed, and reperfusion was visually confirmed. After reflow intervals of 15 min, 2 h, or 6 h, the kidneys were rapidly removed. Nonischemic control samples included kidneys obtained immediately after induction of anesthesia as well as after sham operation. Sham-operated animals underwent identical anesthesia, laparotomy, isolation of renal vessels, and harvest procedures as ischemic rats. No difference was observed between sham-operated and immediate postanesthesia kidneys in any of the studies.

Cell protein fractionation. Renal cortex was homogenized in chilled extraction buffer containing 0.1% Triton X-100, 60 mM piperazine-N,N'-bis(2-ethanesulfonic acid), 2 mM trans-1,2-diaminocyclohexane-N,N,N,',N'-tetraacetic acid, 1 mM EDTA, 1 mM ethylene glycol-bis(-aminoethyl ether)-N,N,N',N'-tetraacetic acid, 100 mM NaCl, 0.5 mM phenylmethylsulfonyl fluoride, 0.75 mg/l leupeptin, and 0.1 mM DL-dithiothreitol using a Potter-Elvehjem homogenizer. For cell protein fractionation, the homogenate was centrifuged at 35,000 g for 14 min at 4°C to separate the Triton-soluble from the insoluble protein fraction. The pellet was then resuspended in the same extraction buffer to a volume equivalent to the detergent-soluble extract. Total cortex homogenate, detergent-soluble, and insoluble fractions were stored at 70°C until further analysis.

Western analysis. Protein determinations were performed in duplicate in each subfraction according to Lowry et al. (15) using bovine serum albumin (BSA) as a standard. For induction studies, 100 µg of protein from total cortex homogenate obtained from each experimental animal was used for Western analysis. To assess changes in detergent extractability, equivalent 50-µl aliquots from both the detergent extract and the resuspended pellet were placed in an equivalent volume of 2× sodium dodecyl sulfate (SDS)-sample buffer for protein electrophoresis (9). Protein samples were electrophoresed through 0.1% SDS-12% polyacrylamide gel with 4% stacking gel and electrophoretically transferred to nitrocellulose. Nonspecific binding sites were blocked with 5% nonfat milk, and the membranes were incubated for 1 h with polyclonal antibodies against rodent HSP-25 (StressGen, Victoria, BC, Canada). Detection was with secondary antibodies, reagents and protocols for enhanced chemiluminescence (ECL; Amersham, Arlington Heights, IL). Computerized densitometry of the specific bands on all autoradiograms was performed using IM-4000 image analysis software from Georgia Instruments (Roswell, GA) as previously described (26). The integrated intensities of HSP-25-specific bands from cortical homogenates were determined at each reflow interval and expressed as percent of nonischemic controls. To evaluate alterations of HSP-25 distribution during reflow, the integrated intensity of HSP-25-specific bands in the detergent-soluble and -insoluble fractions were assessed and expressed as the ratio between the insoluble fraction and total cellular HSP-25 (the sum of the densitometric values for the insoluble and soluble fractions from each extraction placed on the same gel).

Immunocytochemical studies. A separate set of animals underwent the same preparation, renal ischemia, and reflow intervals as described above (Animal preparation). The kidneys were perfused in situ, fixed, and stored as previously described (25). Frozen sections of cortex within 100 µm of the capsule were cut on a Leica Cryocut 3000 cryostat at 6 µm thickness, 18°C, and picked up on lysine-coated slides. The slides were soaked in acetone at 20°C for 5 min and air dried. Sections were then rehydrated and blocked with phosphate-buffered saline (PBS), 1% BSA, 1% normal goat serum (NGS) for 1 h at room temperature. The slides were then further blocked using an avidin-biotin blocking kit (Vector), per standard instruction. Tissue was then incubated in a titrated series of primary antibody diluted in PBS with 0.1% BSA and NGS in a humidified chamber for 1 h at room temperature. Sections were then washed three times for 15 min each. Biotinylated secondary antibody (Vector), was diluted 1/100 in PBS and incubated and washed as above. Streptavidin-conjugated fluorescein isothiocyanate (FITC) (Vector) was then applied at 1/100 dilution and washed, again in the same manner as the primary. The slides were then stained with either Texas Red-conjugated phalloidin, to identify F-actin, or Texas Red deoxyribonuclease I (DNase I) diluted 1/300, to identify fragmented actin (Molecular Probes). Negative controls included tissue unexposed to any primary antibody as well as tissue exposed to control immunoglobulin G (Vector). Slides were then cover-slipped using a fade-resisting media, Vectashield (Vector), and edge sealed with Cytoseal Mounting Media. Slides were stored at 4°C until viewing. Proximal tubules were identified in phase microscopy by tubular diameter and cytologic characteristics including cell volume, pattern of basal infoldings and distribution of mitochondria in injured tubules, and the presence of a brush border in control and recovering tubules. Confocal microscopy was performed with a Bio-Rad MRC-600 scanning head attached to a Zeiss Standard upright microscope fitted with a Leitz ×50 water-immersion lens. The scanning head was equipped with simultaneous dual-channel FITC vs. Texas Red filter sets. All microscope settings controlling laser intensity, detection sensitivity, and confocal aperture size were standardized to ensure optimal signal-to-noise ratio for labeled tissue. Negative control, positively labeled normal, and experimental tissue were then viewed under identical conditions. The FITC and Texas Red channels were viewed simultaneously in gray scale, and the data were merged and pseudo-colored with Bio-Rad software merge function. FITC displays as green, Texas Red as red, areas of colocalization display as yellow. Fluorescent signal was undetectable in negative controls under these conditions. Nonconfocal, phase-contrast images of the same tubule segments were also acquired using the Bio-Rad fiber optic transmission attachment.

Statistics. Analysis of variance, with the least significant difference approach and the Dunnett multiple comparison test, were used where appropriate. Values for each reflow interval were compared with the respective control and considered to be significantly different if P < 0.05.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

HSP-25 induction. Figure 1A shows a representative immunoblot of total cortex homogenates isolated at recovery intervals through 24 h after 45 min of ischemia and measured for presence of HSP-25. HSP-25 was constitutively expressed in renal cortex from control animals. Levels of HSP-25 were increased at 2 h reflow after ischemia, peaked by 6 h, and persisted through 24 h. In controls, and at 2, 6, and 24 h reflow, HSP-25 had a duplex appearance with a faint, slower migrating second band demonstrable. At 15 min reflow, only a single band was present. Figure 1B shows the results of densitometry of the HSP-25-specific bands in cortical proteins from six animals at each reflow interval expressed as percent of control. HSP-25 was minimally increased (+20%, not significant) at 15 min reflow, significantly increased (+65%, P < 0.05) at 2 h, and reached a peak level (+260%, P < 0.05) at 6 h. High levels of HSP-25 persisted through 24 h reflow (+240%).

View larger version (16K): [in this window] [in a new window]   View larger version (12K): [in this window] [in a new window]   Fig. 1.   Representative immunoblot (A) and densitometry (B) of heat-shock protein 25 (HSP-25) in total rat renal cortex homogenates in control (C) and during reflow after 45 min ischemia. For densitometry, n = 6 at each interval. HSP-25 is constitutively present in nonischemic control cortex and is minimally increased at 15 min reflow (+20%, not significant). At 2 h, HSP-25 has increased to 165% control and reaches the highest levels at 6 and 24 h with 360% and 340% of baseline, respectively (P < 0.05 for both). Error lines are SE.

HSP-25 detergent extractability. Figure 2A represents the results from immunoblots of Triton X-100-soluble and -insoluble fractions of renal cortex extracted at each reflow interval and compared with control. Figure 2B shows the densitometry results of detergent-insoluble HSP-25 as percent total HSP-25 at a particular interval. Under control conditions, the majority of HSP-25 was recovered in the Triton-soluble supernatant. After 45 min of renal ischemia and 15 min reflow, HSP-25 redistributed from the detergent-soluble supernatant into the insoluble pellet, resulting in an equal distribution between the two fractions. Coordinate with the redistribution of HSP-25 at 15 min reflow was the loss of the slower migrating band. At 2 and 6 h reflow, HSP-25 increased mainly in the supernatant, and its distribution returned to a predominantly detergent-soluble form coordinate with reappearance of the slower band. Densitometry (Fig. 2B) was used to quantify the redistribution of HSP-25 between the cellular fractions during reflow after ischemic injury. Nonischemic controls have only 20% (±6) of total HSP-25 in the insoluble fraction. At 15 min reflow, HSP-25 exhibits a marked shift into the insoluble fraction to reach 50% (±7) of total (P < 0.05). At 2 and 6 h reflow HSP-25 redistributes to the soluble fraction, with only 18% (±4) and 12.5% (±4), respectively, found in the insoluble fraction.

View larger version (48K): [in this window] [in a new window]   View larger version (11K): [in this window] [in a new window]   Fig. 2.   Representative immunoblot (A) and densitometry (B) of HSP-25 distribution between detergent-soluble (S) and -insoluble pellet (P) cellular protein fractions of rat renal cortex. For densitometry results, the portion of insoluble HSP-25 is expressed as percent total HSP-25 present at each reflow interval (n = 6 at each point). In nonischemic controls, the majority of total cellular HSP-25 was recovered in the detergent-soluble fraction (20% insoluble). At 15 min reflow, HSP-25 exhibits a marked shift into the insoluble cytoskeletal fraction (50% of total, * P < 0.05). At 2 and 6 h reflow, while total HSP-25 accumulates progressively, HSP-25 redistributes back from the cytoskeletal pellet into the detergent-soluble fraction (18% and 12.5% insoluble, respectively). Error lines are SE.

Immunolocalization. HSP-25 was readily detectable in glomeruli and to a lesser extent in the peritubule vasculature, from control cortex as previously described (21, 24). At 6 h of reflow, when cortical HSP-25 was increased more than threefold, the pattern of HSP-25 staining in the peritubule vasculature and glomeruli had not changed. There was a qualitative decrease in the fluorescent intensity of HSP-25 in the glomeruli, and no change in the vascular staining at 6 h reflow. HSP-25 was also detectable in control proximal tubules, as previously reported (21), and in proximal tubules from each reflow interval.

HSP-25 immunolocalization within proximal tubular cells compared with DNase-reactive actin and F-actin is represented in Figs. 3 and 4, respectively. The sections of renal cortex displayed are all within 100 µm of the capsule. Phase-contrast images of each tubule segment examined by immunofluorescence are displayed at right of both Figs. 3 and 4. The typical brush border of proximal tubules is present in segments from control and 6 h reflow cortex. The injured tubules at 15 min and 2 h reflow show the usual alterations of apical membrane blebbing and luminal collapse after ischemia while maintaining the cytologic characteristics of proximal tubules including cell volume, pattern of basal infoldings, and distribution of mitochondria.

View larger version (105K): [in this window] [in a new window]   Fig. 3.   Immunolocalization of HSP-25 and DNase-reactive actin (DNase I R'A) within proximal tubules under control conditions and during reflow after ischemia. Individual images for the separately labeled HSP-25 and DNase-reactive actin were merged for the color images (both). HSP-25 displays as green, DNase-reactive actin as red, and areas of colocalization as yellow. In each series, note the phase-contrast image (right) of the tubule examined by immunofluorescence. Bar = 50 µm.

View larger version (108K): [in this window] [in a new window]   Fig. 4.   Immunolocalization of HSP-25 and F-actin within proximal tubules in controls and after renal ischemia. HSP-25 and F-actin images were merged for the combined color images (both). HSP-25 displays as green, F-actin as red, and areas of colocalization as yellow. In each series, note the phase-contrast image (right) of the tubule examined by immunofluorescence. Bar = 50 µm.

The color images in Figs. 3 and 4 demonstrate the combined images of HSP-25 with each form of actin. There was no nuclear translocation of HSP-25 demonstrable at any time point. Under control conditions, HSP-25 was localized predominantly along the apical domain with some diffuse staining in the cytoplasm. As shown in the combined color images ("both") of Figs. 3 and 4, this pattern of HSP-25 localization was intermediate between DNase-reactive actin, which is distributed diffusely throughout the cytoplasm, and F-actin, which was found almost exclusively in the apical domain.

At 15 min of reflow HSP-25 changed to a diffuse pattern throughout the cell with small punctate accumulations in a pattern similar to that of DNase-reactive actin. F-actin remained predominantly in the apical domain. The combined images show that at 15 min reflow HSP-25 localization closely paralleled DNase-reactive actin but differed from F-actin.

At 2 h of reflow, HSP-25 continued to be diffused throughout the cytoplasm with increased accumulations in a coarse granular pattern. DNase-reactive actin remained diffused through the cytoplasm but in coarse accumulations now distinct from HSP-25 (Fig. 3, color image). Microfilamentous actin at 2 h had lost its predominantly apical distribution and had become more diffusely distributed. The combined image in Fig. 4 shows that a portion of F-actin colocalized with HSP-25.

At 6 h reflow, HSP-25 and DNase-reactive actin remained in separate coarse granular accumulations. At this time F-actin reaccumulated in the apical domain. In addition to the persistent granular accumulations, a portion of HSP-25 had returned by 6 h to the subapical distribution seen in controls (Fig. 4).

In summary, the distribution of HSP-25 in control kidney cortex was intermediate between that of DNase-reactive and microfilamentous actin. During early reflow after ischemic injury HSP-25 distribution paralleled DNase-reactive actin, but separated from DNase-reactive actin at later reflow intervals when HSP-25 colocalized, in part, with F-actin. These findings suggest that HSP-25 interacts with both forms of actin during recovery from renal ischemia.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

During the last decade, several modes of action for the sHSPs, including the closely related rodent HSP-25 and human HSP-27, have been proposed. These include protease inhibition, chaperone activity, cross-linkage to other proteins and protection against apoptosis (1, 16). Most evidence, however, points toward a major function of these sHSPs at the level of the actin cytoskeleton under both control and stress conditions (1, 2, 6, 10-14, 17). HSP-25 and -27 have sequence homology and structural features common to the lens protein -crystallin, which is thought to be involved in cytoskeletal organization of the lens, and share the characteristic of -crystallin in forming large aggregates under stressful conditions (1). The chicken homolog of HSP-27 was originally described as an inhibitor of actin polymerization (IAP), and showed actin-binding (capping) properties under control conditions (17). In cells that selectively overexpress HSP-27, the actin-based cytoskeleton was specifically protected against disruption by several stimuli and stresses, including cytochalasin D treatment (1, 6, 10, 12, 13).

Renal ischemia is known to induce disruption and aggregation of the actin cytoskeleton (4, 18, 20, 22, 27). Microvilli are fragmented, cell-cell and cell-substratum junctions dissociate, and cellular polarity is altered. F-actin, which is primarily associated with the circumferential terminal web at the apical pole in intact proximal tubules, is redistributed throughout the cytoplasm. Molitoris and co-workers (18, 20) have shown that redistribution of actin in injured proximal tubule cells is associated with the rapid conversion of soluble G-actin to pelletable F-actin. Golenhofen et al. (4) reported similar results in cultured renal epithelia, finding decreased detergent solubility of actin after cellular ATP depletion. More recently, Nurko et al. (22) demonstrated that actin has enhanced DNase reactivity and requires higher rates of centrifugation to be pelleted from isolated proximal tubules subjected to ATP depletion. Their results suggest that ischemia causes fragmentation of the intact cytoskeleton into smaller but still pelletable actin filaments. Postischemic restructuring of the actin cytoskeleton could be facilitated by specific ischemia related alterations in HSP-25/27 activity. Initially, the actin-binding and capping activity may prevent further fragmentation of the cytoskeleton. Subsequent rephosphorylation and induced expression of HSP-25 may then augment actin polymerization during later recovery (1).

The in vitro actin capping activity of HSP-27 was recently demonstrated to be dependent on the phosphorylation status of HSP-27 (10), and the nonphosphorylated isoform of HSP-25 appears to have a high binding affinity for actin (2). Also, a recent study demonstrated that an interfering, nonphosphorylatable mutant of human HSP-27, which lacks all but the actin-binding function, demonstrates a stress-induced shift into the detergent-insoluble fraction (13). Thus the changes in HSP-25 that have been documented in this study at 15 min of reflow are most consistent with the actin capping properties of this protein, as follows: 1) redistribution from the detergent-soluble to the insoluble fraction, 2) disappearance of the slower second band on gel electrophoresis, and 3) colocalization with DNase-reactive, fragmented, actin. Immediate postischemic reflow is characterized by ATP depletion, which has been shown to result in dephosphorylation of HSP-27 with a concurrent shift into detergent-insoluble cell protein fractions (14) and attachment to the barbed ends of microfilamentous actin.

During later reflow, then, the reappearance of the slower HSP-25 band, redistribution of HSP-25 into the detergent-soluble fraction, and dissociation from DNase-reactive actin would be a function of rephosphorylation of HSP-25. Phosphorylation has been shown to abolish the actin-binding properties of HSP-27, freeing the barbed ends of microfilaments and resulting in polymerization (2, 10). Recently, a stress-activated protein kinase cascade has been described that involves mitogen-activated protein (MAP) kinase p38 and MAPKAP kinase 2 in the ultimate phosphorylation of HSP-27 (8). Although the activation of these kinases involved in HSP-25/27 phosphorylation has yet to be characterized during or following ischemic injury, members of a separate cascade of stress-activated kinases [the stress-activated protein kinases (SAPKs) and extracellular signal-regulated kinase (ERKs)] have been studied in renal ischemia. Ischemia alone did not activate SAPKs or ERKs; however, SAPKs were rapidly activated, within minutes, during reperfusion after ischemia (23). In an analogous manner the kinases responsible for HSP-25/27 phosphorylation may be rapidly activated during reflow. The initial shift of HSP-25 into the insoluble protein fractions caused by ischemia may be reversed by activation of HSP-25/27 kinases, such as MAPKAP kinase 2, resulting in rephosphorylation and a shift back into soluble protein fraction by 2 h reflow. Stress-induced alterations in the phosphorylation state of the sHSPs may then be pivotal in the restructuring of the actin cytoskeleton after ischemia.

HSP-25 protein abundance increased early after renal ischemia, reaching its highest levels, more than three times baseline, by 6 h reflow. The increase in cortical HSP-25 likely reflects changes in the proximal tubule, since glomeruli comprise a relatively small fraction of cortical mass, and the fluorescent intensity of HSP-25 within glomeruli decreased after ischemia. This is consistent with the previously described increase in glomerular HSP-25 fluorescence in a model of puromycin aminonucleoside (PAN)-induced injury (24) where the insult is predominantly glomerular rather than tubular, as in ischemic injury. The course of HSP-25 induction, then, is very similar to the postischemic elaboration of HSP-72 in the kidney (25). In other organ systems the concordance between HSP-27 and HSP-72 has been more variable (5, 7). Cardiac ischemia resulted in increased message for HSP-72 but not HSP-27 (5). In focal brain ischemia, HSP-27 was induced in microglia, but not in ischemic neurons (7). Our finding of a concordant increase in HSP-25 with HSP-72 expression after renal ischemia thus suggests a putative role for these stress proteins to work synergistically in the recovery process. In fact, Ehrnsperger et al. (3) have demonstrated that HSP-25 can function cooperatively with HSP-70 to trap unfolding proteins in a folding competent state. These nonstructural, chaperone functions of sHSPs then may be the primary activity of HSP-25 at 2 and 6 h of reflow when this protein is found in greatest abundance and in an aggregated pattern throughout the cytoplasm. Indeed, these proteins may interact with aggregating proteins to create a reservoir of nonnative proteins that can be reactivated when cellular metabolism has been restored (3). Such a process would substantially aid the recycling of cytoskeletal and other disrupted cellular proteins after ischemia.

In conclusion, this study describes the induction, intracellular localization, and redistribution of HSP-25 in proximal tubule cells after renal ischemia. Combined with the known actin-binding properties of the sHSPs, the postischemic induction and reversible redistribution of HSP-25 between detergent-soluble and -insoluble compartments strongly suggests interactions with the disrupted actin cytoskeleton. Specific interaction between HSP-25/27 and actin during postischemic reorganization of the cytoskeleton is further supported by the temporal and spatial pattern of HSP-25 localization. Future studies will be essential to elucidate whether the sHSPs have a crucial role in the restructuring of sublethally injured epithelial cells.