Evidence for HSP-mediated cytoskeletal stabilization in mesothelial cells during acute experimental peritoneal dialysis

【摘要】  Low biocompatibility of peritoneal dialysis fluid (PDF) injures mesothelial cells and activates their stress response. In this study, we investigated the role of heat shock proteins (HSP), the main cytoprotective effectors of the stress response, in cytoskeletal stabilization of mesothelial cells in experimental peritoneal dialysis. In cultured human mesothelial cells, cytoskeletal integrity was assessed by detergent extractability of marker proteins following in vitro PDF exposure. Effects of HSP on stabilization of ezrin were evaluated by a conditioning protocol (PDF pretreatment) and repair assay, based on coincubation of cytoskeletal protein fractions with recombinant HSP-72 or HSP-72 antibodies. In the rat model, detachment of mesothelial cells from their peritoneal monolayer during in vivo PDF exposure was assessed with and without overexpression of HSP-72 (by heat conditioning). In vitro, cytoskeletal disruption on sublethal PDF exposure was demonstrated by significantly altered detergent extractability of ezrin and ZO-1. Restoration was associated with significant induction and cytoskeletal redistribution of HSP during recovery. Both the conditioning protocol and in vitro repair assay provided evidence for HSP-72-mediated cytoskeletal stabilization. In the rat model, overexpression of HSP-72 following heat conditioning resulted in significantly reduced detachment of mesothelial cells on in vivo exposure to PDF. Our results establish an essential role of HSP in repair and cytoprotection of cytoskeletal integrity in mesothelial cells following acute in vitro and in vivo exposure to PDF. Repeated exposure to PDF, as is the rule in the clinical setting, may not only cause repeat injury to mesothelial cells but rather represents a kind of inadvertent conditioning treatment.

【关键词】  heat shock proteins cytoskeletal repair stress response cytoprotection

PERITONEAL DIALYSIS FLUIDS (PDF) remove solutes and water from an uremic patient. Several clinical and experimental observations have shown that these PDF are not biocompatible ( 13, 26 ). Studies on clinical outcomes of peritoneal dialysis report a risk of ultrafiltration failure of up to 30% with long-term treatment ( 16 ).

Recent research demonstrated that PDF exposure not only results in cellular injury but also activates an endogenous machinery found in every cell, the so-called stress response ( 2, 5, 34 ). This stress response includes cellular mechanisms that facilitate cellular repair and survival after acute injury ( 8, 27, 29, 31 ). Main effectors of this stress response are the heat-shock proteins (HSP). Induction of HSP was described in mesothelial cells exposed to PDF in in vitro, ex vivo, and in vivo models of peritoneal dialysis ( 1, 2, 5, 34 ). Overexpression of HSP confers cytoprotection in mesothelial cells, improving cellular viability on PDF exposure ( 9 ).

Disruption of the cytoskeletal architecture is regarded as a classic marker of sublethal epithelial cell injury ( 21, 25, 38 ). Interestingly, marked morphological alterations have been described in mesothelial cells in the dialyzed peritoneal cavity, which can be related to such cytoskeletal disruption. Ultrastructurally, the number of microvilli was substantially reduced, intercellular gaps widened, and considerable detachment of mesothelial cells from the basement membrane on exposure to PDF was found (17-20). These observations are of potential clinical relevance as they likely represent the morphological correlate of impaired ultrafiltration, inefficient solute clearance, and significant protein loss.

We and others have recently described clear evidence for HSP-mediated repair during cytoskeletal restoration in the model of renal ischemia ( 3, 4, 6, 10, 11 ). No such data exist for peritoneal dialysis. In this study, we aimed to investigate the role of HSP in cytoskeletal stabilization in mesothelial cells following acute exposure to PDF in in vitro and in vivo models of peritoneal dialysis.

MATERIALS AND METHODS

In Vitro Model of Peritoneal Dialysis

Cell culture. Immortalized human mesothelial cells (Met5A, ATCC CRL-9444) were cultured in M199/MCDB 105 medium (1:1) supplemented with 100 U/ml penicillin, 100 µg/ml streptomycin, and 10% FCS. Cultures were kept in 75-cm 2 tissue culture flasks (Falcon, Becton Dickinson, Oxnard, CA) at 37°C in 5% CO 2 and passaged by regular trypsinization. The medium was changed every 2-3 days. Confluence was reached on average after 6-7 days.

PDF exposure. Confluent cultures were exposed to a standard glucose monomer and acidic lactate-based PDF (Fresenius 2, Bad Homburg, Germany) containing 1.5% anhydrous dextrose at pH 5.5. Exposure times ranged from 30 to 120 min. Depending on the experimental setup, cells were then allowed to recover in regular growth medium for 0-24 h. Control cultures were kept in regular culture media at 37°C and underwent the same "sham-procedures" of media changes; i.e., exposure to PDF was paralleled by exposure to control medium for the times indicated. In selected experiments, cells were exposed to a more biocompatible peritoneal dialysis solution with almost neutral pH and low-glucose degradation products (PD-Bio 1.5%, Gambro, Lund, Sweden) for 4 and 8 h. At the end of each protocol, detergent extractability of cytoskeletal marker proteins, induction of HSP, and cell viability were assessed in parallel cultures as described below. Cells for use in immunofluorescence staining were grown on microscope slides (culture slides, Falcon, BD Bioscience, Franklin Lakes, NJ), using the same medium and atmosphere as mentioned previously.

In vitro conditioning protocol. Confluent mesothelial cells were pretreated for 60 min by standard PDF incubation (classic conditioning) and then allowed to recover in regular culture medium. Twenty-four hours after the pretreatment, cells were subjected to repeat standard PDF exposure and subsequent recovery studies.

Cell viability assessments. Cellular ATP levels of harvested extracts were measured by luciferase assay using a bioluminescent somatic cell assay kit (FLASC, Sigma, St. Louis, MO). Viability of cells was investigated after the indicated exposure times to PDF and 24-h recovery. Dead cells were identified by incapability of standard trypan blue dye exclusion. Cells were stained with trypan blue for 15 min and evaluated in a hemocytometric chamber. Cells capable of dye exclusion (viable) and cells stained with trypan blue (nonviable) were counted, and percent viability was calculated. For lactate dehydrogenase (LDH) analyses, supernatants were removed after the described experimental setup and kept at 4°C until analyzed within 48 h. Measurements were performed in duplicate with a Sigma TOX-7 LDH Kit according to the manufacturer's instructions. LDH efflux is calculated as the percentage of LDH values measured in each negative control experiment.

Triton extraction. Cells were either lysed in buffer A [Na-K-ATPase; containing 0.1% Triton X-100 and (in mM) 60 PIPES, 2 CDTA, 1 EDTA, 1 EGTA, 100 NaC1, and 0.5 PMSF, as well as 0.75 mg/l leupeptin and 0.1 M DTT] or buffer B (all other marker proteins; containing 0.1% Triton X-100 and 300 mM sucrose, 500 mM Tris·HCl, 2 mM EGTA, 200 µM PMSF, and 10 µM leupeptin) (adapted from Ref. 14 ). The homogenate was centrifuged within 10 min at 35,000 g for 15 min at 4°C to separate the Triton-soluble protein fraction from the insoluble cytoskeletal pellet. Protein fractions were saved at -80°C until further Western blot analysis.

In vitro repair assay. This method was adapted from an assay originally designed for postischemic rat renal cortex ( 10, 11 ). In brief, Triton-insoluble protein aliquots (=cytoskeletal pellets) from mesothelial cells were isolated immediately after exposure to PDF ("injured") or after 24 h of recovery in culture medium ("recovered"). For assessment of in vitro repair, aliquots of injured cytoskeletal pellets were incubated in buffer with added recombinant HSP-72 at a concentration of 10 µg/100 µl (Stress Gen Biotechnologies, Victoria, BC) under conditions of blocked (2 mM CDTA) or enhanced ATP hydrolysis (5 mM MgATP). In addition, aliquots of recovered cytoskeletal pellets were incubated in parallel either in pure buffer or in buffer with added anti HSP-72 antibodies at a concentration of 50 µg/100 µl. After coincubations, differential centrifugation was repeated with 35,000 g for 15 min at 4°C. After repelleting, the supernatant represents nontranslocated HSP-72 (or anti-HSP-72 antibody, respectively) plus dissociated cytoskeletal proteins.

In Vivo Model of Peritoneal Dialysis

Acute rat model. The studies were carried out in adult male inbred Sprague-Dawley rats (average weight 310 g). After introduction of anesthesia (100 mg/kg ketamine and xylazine 5 mg/kg im), the animals were placed on a heated small-animal operating table. A sterile catheter was inserted into the peritoneal cavity through a small abdominal midline incision and, according to the protocol, 35 ml of prewarmed (37°C) test fluid (Medium 199 or PDF, CAPD3, FMC) were slowly infused in 45-60 s. In selected experiments, nonprewarmed PDF (room temperature, 20°C) or PDF warmed to 41.5°C were used. The animal was gently moved, a small volume of peritoneal fluid was aspirated, the catheter was withdrawn, and the abdomen was sutured. Animals awoke within 20 min after the procedure and had free access to food and tap water. At appropriate times (4 or 24 h) after the intraperitoneal injection, animals were again anesthetized, the peritoneal cavity was opened, and the peritoneal effluent was collected. In selected animals, material for morphological analysis was obtained.

In vivo conditioning protocols. For heat conditioning, transient hyperthermia of 41.5°C core temperature was induced for 15 min in anesthetized rats by the heated animal table and an infrared warming lamp, followed by 20 h of recovery.

All animals received humane care in compliance with the principles of laboratory animal care formulated by the Institute of Laboratory Animal Resources and the Guide For The Care and Use of Laboratory Animals prepared by the National Academy of Sciences and published by the National Institutes of Health.

Mesothelial cell detachment. For quantification, the peritoneal effluent was collected by gentle aspiration of 2 ml at baseline and after 4 h. Thereafter, animals were killed by cardial puncture and exsanguination, the abdomen was opened by a midline incision, and the complete intraperitoneal fluid was gently collected. The volumes of the collected fluids were recorded, and the total cell count and differential counts were assessed by hand count after giemsa staining and by machine count by a Coulter counter. The total number of detached mesothelial cells was then computed for each rat. This method was adopted from recent studies in human patients ( 7 ) and based on observations of mesothelial cell loss within 10 min following acute oxidative stress in rats ( 22 ).

Staining Procedures

Rhodamine-phalloidin (F-actin) staining. Cells were grown on culture slides and treated as described above. After the experiments, the cells were fixed in 4% paraformaldehyde. For staining, cells were first reconstituted in PBS and exposed to rhodamine-phalloidin (R415, Molecular Probes, Eugene, OR) in a final concentration of 0.5 µM for 30 min in the dark. Cells were washed five times with PBS. Slides were mounted using 50% glycerol/50% PBS and finally sealed with nail polish. Evaluation and microphotographs were done using a Zeiss fluorescence microscope.

Immunofluorescence staining. Cells were grown on culture slides and treated as described above. After the experiments, cells were washed two times with PBS and fixed with 4% paraformaldehyde in PBS, pH 7.4, for 20 min at room temperature. Cells were then permeabilized for 20 min with either 0.1% Triton X-100 or 0.1% saponin in PBS (pH 7.4). Nonspecific binding sites were blocked for 45 min in 5% fetal calf serum and 1% bovine serum albumin in PBS (pH 7.4). Ezrin and ZO-1 were immunodetected by sequential incubation with the respective antibody and a fluorescently labeled second antibody.

For immunofluorescent double labeling of detached mesothelial cells, aliquots of peritoneal effluate were spun onto slides, air-dried, and fixed with fresh cold 4% paraformaldehyde for 20 min. The fixed cells were blocked with 10% goat serum and permeabilized with 0.1% saponin in PBS for 2 min and washed with PBS. After 30-min incubation with the primary antibodies against cytokeratin and HSP-72 at room temperature, slides were rinsed with PBS and bound IgG was visualized using Alexa Fluor 594 goat anti-mouse IgG (Molecular Probes) and Alexa Fluor 488 goat anti-rabbit IgG (Molecular Probes). As a third step, nuclei were stained with 4,6-diamidino-2-phenylindole.

For immunohistochemical studies of resident mesothelial cells lining the peritoneal cavity, median samples from parietal peritoneum were used to prepare tissue blocks after fixation for 24 h at 4°C in 4% paraformaldehyde in PBS. The fixed peritoneal slices were paraffin embedded, and 4-µm sections were cut. After blocking of unspecific binding sites, primary antibodies to cytokeratin or HSP-72 were applied in PBS, 0.5% BSA for a total of 1 h. After repeated washes, the sections were then exposed to the secondary antibody conjugated to horseradish peroxidase in an equivalent procedure to the primary antibody.

Detection, Densitometry, and Statistics

Western blotting. Protein content was determined by Bradford assay (Bio-Rad) and equal amounts of protein samples (5 µg/lane) were separated by standard SDS-PAGE using a Pharmacia Multiphore II unit. Size-fractionated proteins were transferred to PVDF membranes by semidry transfer in a Pharmacia Multiphore II Novablot unit. Membranes were blocked in 5% dry milk in TBS-Tween (10 mM Tris, 150 mM NaCl, 0.05% Tween 20, pH 8.0). Membranes were incubated with the respective primary antibody. Detection was accomplished by incubation with secondary, peroxidase-coupled antibodies (Sigma), and enhanced chemiluminescence (ECL) using ECL Western blotting analysis system and protocols (Renaissance, NEN-Life Science Products, Boston, MA).

Densitometry was performed with image-analysis software (Molecular Analyst software, Bio-Rad). Differential expression and cellular distribution of respective proteins were derived from the ratio of specific immunodensitometric signals at different exposures in the linear range of the protein-signal intensity relationship, normalized to an internal standard, and compared between parallel incubations in each experiment.

Antibodies. Na-K-ATPase (05-369, Upstate Biotechnology), cadherin (A-CAM, C-2667, Sigma), cytokeratin (ZO622, Dako Cytomation), ezrin (3212, Sigma), HSP-27 and HSP-72 (SPA 800, SPA 810, Stressgen), ZO-1 (33-9100, Zymed), and E-cadherin (33-4000, Zymed) were used.

Statistical analysis. Values for treatment conditions were compared with control values using a Wilcoxon signed rank test (SPSS 7.5 for Windows, SPSS, Chicago, IL). The data are expressed as means ± SD. Differences were considered to be significant given a P < 0.05.

RESULTS

Cytoskeletal Disruption in In Vitro Model of Peritoneal Dialysis

The first part of the in vitro experiments demonstrates cytoskeletal disruption in mesothelial cells exposed to standard PDF for 60 min with no apparent morphological alterations, negligible 90% viability. PDF exposure for 120 min caused marked cellular detachment and decreased viability ( Fig. 1 ). PDF exposure for 60 min also resulted in <20% increase of LDH release (106 ± 12% of control) but markedly reduced ATP levels to 20 ± 9% of controls ( P < 0.05) immediately after exposure with fully restored levels at 24-h recovery. Exposure to the more biocompatible peritoneal dialysis solution could be extended up to 8 h without significant increases in LDH release (118 ± 24% of control).

Fig. 1. Effects of peritoneal dialysis fluid (PDF) exposure on cultured human mesothelial cells. Native morphology ( top ) and vital dye staining ( bottom ) of mesothelial cells exposed to PDF for 30 ( left ), 60 ( middle ), and 120 min ( right ) are shown. PDF exposure for 60 min resulted in no morphological changes and preserved viability. Data are representative of at least 3 independent experiments.

As shown in Fig. 2, the actin-based cytoskeleton formed a cortical web in mesothelial cells under control conditions, with well-developed stress fibers. Ezrin showed a distinct punctate staining pattern, and ZO-1 stained the cell-cell contacts. Upon sublethal PDF exposure for 60 min, remarkable disruption of the actin-based cytoskeleton was found, resulting in fragmentation and aggregation of actin filaments. The punctuate ezrin-staining pattern disappeared, and ZO-1 staining demonstrated a slightly more accentuated pattern. Neither of these alterations was fully reversible within 24 h of recovery.

Fig. 2. Sublethal PDF exposure disrupts the cytoskeletal architecture in cultured human mesothelial cells. Phalloidin staining for actin ( top ) and immunofluorescence staining for ezrin ( middle ) and ZO-1 ( bottom ) are shown in mesothelial cells kept under control conditions ( left ) or exposed to PDF for 60 min ( right ). This sublethal PDF exposure resulted in overt cytoskeletal disruption. Data are representative of at least 3 independent experiments.

The results of differential centrifugation with detergent extraction, a more subtle and quantitative marker of the cytoskeletal disruption of mesothelial cells, are given in Fig. 3 A. We found exposure time (="dose")-dependent effects of standard PDF exposure on the disruption of the cytoskeletal anchorage of ezrin, demonstrated by increasing extraction of this protein marker from the Triton-insoluble (cytoskeletal) fraction into the soluble cellular protein fraction. The cytoskeletal association of the additionally tested protein markers cadherin and Na-K-ATPase was differentially susceptible to Triton extraction before and after PDF exposure. Whereas cadherin behaved similarly to ezrin, increases in Na-K-ATPase in the soluble fraction were only seen with 120 min of exposure (=lethal PDF exposure time). In contrast to these markers, ZO-1 markedly shifted into the detergent-insoluble cytoskeletal fraction with PDF exposure. The more biocompatible peritoneal dialysis solution caused comparable redistribution of ezrin after 4 and 8 h of exposure (165 ± 32 and 176 ± 41% of control) as observed at 60 min with the standard PDF (182 ± 42% of control). After 24 h of recovery following sublethal (60 min) PDF exposure, the cellular distribution of all these marker proteins returned to control levels. These data are clearly consistent with reversible disruption of the cytoskeletal integrity in the in vitro model of peritoneal dialysis in mesothelial cells. Coomassie staining of cytoskeletal and noncytoskeletal protein fractions obtained after such sublethal PDF exposure demonstrated no gross alterations of the cellular proteins (data not shown).

Fig. 3. PDF exposure results in time-dependent alterations of detergent extractability of cytoskeletal marker proteins and heat shock proteins (HSP) in cultured human mesothelial cells. Western blot analysis demonstrates that, with increasing duration of PDF exposure, ezrin, Na-K-ATPase, and cadherin become increasingly solubilized from the cytoskeletal pellet (p) into the Triton-soluble supernatant (s), whereas ZO-1 becomes insolubilized in the Met5A cell line ( A ). This altered cellular distribution of cytoskeletal marker proteins is associated with induction and shifting of HSP-72 and HSP-27 into the cytoskeletal pellet ( B ). As demonstrated on the far right, these alterations are fully reversible within 24 h of recovery ( A ) after sublethal 60-min PDF exposure, associated with marked HSP induction ( B ). Data are representative of at least 3 independent experiments.

Cytoskeletal Restoration and HSP in In Vitro Model of Peritoneal Dialysis

As a next step, we searched for induction and cellular distribution of HSP during cytoskeletal disruption and repair of mesothelial cells following in vitro exposure to PDF. Under control conditions, we found constitutive expression of HSP-27 and low amounts of HSP-72, both predominantly Triton soluble ( Fig. 3 B ). With increasing PDF exposure, HSP-27 and HSP-72 were both induced and increasingly shifted into the cytoskeletal fraction. At 24 h of recovery, HSP-27 and HSP-72 showed potently elevated total cellular amounts with a predominant redistribution into the noncytoskeletal (=Triton-soluble) protein fraction.

Densitometric analysis, as given in Fig. 4 for ezrin, ZO-1, and HSP-72, demonstrates the statistical significance of the effects of PDF exposure on Triton extractability of these marker proteins. Moreover, it also becomes evident that their stabilization during recovery from PDF exposure was at least temporally and spatially associated with induction and redistribution of HSP. These findings support the hypothesis that HSP interact with disrupted elements of the cytoskeleton and participate in its restoration.

Fig. 4. Restoration of cellular distribution of cytoskeletal marker proteins is initiated by a cytoskeletal shift of HSP in cultured human mesothelial cells on PDF exposure. This densitometric analysis demonstrates that the disruption of the marker proteins ezrin (solubilized) and ZO-1 (insolubilized) is reversed following a transient shift of HSP-72 into the cytoskeletal pellet during recovery from sublethal (60-min) PDF exposure. Data are derived from 6 independent experiments. *Statistically significant differences of marker proteins at 60 min vs. control values ( P < 0.05).

To confirm the role of the mesothelial stress response, we manipulated HSP expression in mesothelial cells by a classic conditioning protocol. As shown in Fig. 5, pretreatment by 60-min PDF exposure resulted in high cellular levels of HSP-72 after 24-h recovery. On reexposure to PDF, HSP-72 readily shifted into the cytoskeletal cellular protein fraction. HSP overexpression did not protect against disruption of the cytoskeletal anchorage of ezrin immediately after repeat PDF exposure but resulted in significantly enhanced restoration of its cytoskeletal redistribution within 4 h of recovery.

Fig. 5. Cytoprotection in cultured human mesothelial cells pretreated by nonlethal PDF exposure. A : Western blots for Triton-soluble (=disrupted) ezrin and pelleted (=cytoskeletal) HSP-72 are shown on top for nonpretreated and PDF-pretreated mesothelial cells under control conditions, immediately following (repeat) exposure to PDF, and after 4 h of recovery. B : densitometric analysis of differential stabilization of ezrin between nonpretreated (open bar) and pretreated (filled bar) mesothelial cells during the first 4 h of recovery. Conditioning by sublethal pretreatment with PDF resulted in markedly elevated cytoskeletal HSP levels and significantly enhanced stabilization of ezrin during recovery from (repeat) exposure to PDF. Data are derived from 6 independent experiments.

To provide more direct evidence that HSP-72 is responsible for stabilization of ezrin in the cytoskeletal fraction of mesothelial cells during recovery from sublethal PDF exposure, we utilized an in vitro repair assay. Coincubation of cytoskeletal proteins, isolated immediately after PDF exposure (=injured) or after 24 h-recovery (=recovered) was performed either with recombinant HSP-72 or with anti-HSP-72 antibodies, respectively. As shown in Fig. 6, addition of recombinant HSP-72 under conditions of enhanced ATP hydrolysis resulted in stabilization of ezrin in the injured cytoskeletal protein fraction, whereas anti-HSP-72 antibodies caused increased Triton extractability of this membrane protein from the recovered cytoskeletal protein fraction. These data are consistent with a specific role of HSP-72 in the in vitro repair of disrupted ezrin in mesothelial cells.

Fig. 6. Evidence of HSP-mediated cytoskeletal repair in human mesothelial cells during recovery from sublethal PDF exposure. In an in vitro repair assay, effects of HSP-72 on the cytoskeletal anchorage of ezrin were analyzed following coincubation of "injured" cytoskeletal pellet ("PDF" = harvested immediately after PDF exposure) with recombinant HSP-72 ( left ) and of "recovered" pellet ("PDF+Recovery" = harvested 24 h after PDF exposure) with anti-HSP-72 antibody ( right ). Enhancing exogenously added HSP-72 function (on enhanced ATP hydrolysis) resulted in reduced detergent extractability, whereas blocking endogenously upregulated HSP-72 (with antibody) resulted in increased detergent-soluble ezrin. Data are derived from 6 independent experiments. Both changes were statistically significant (* P < 0.05).

Mesothelial Cell Detachment and HSP in Rat In Vivo Model of Peritoneal Dialysis

To define an in vivo correlate to cytoskeletal damage in mesothelial cells following acute exposure to PDF, we quantified mesothelial cell detachment from their peritoneal monolayer in the rat model of peritoneal dialysis. When PDF was intraperitoneally injected, only few mesothelial cells ( 1% of the total cells in the effluent) were identifiable at baseline. As shown in Fig. 7 A, this number significantly increased within a 4-h dwell and then returned to control levels after 1 day. To discriminate for effects of PDF vs. the fluid injection per se, we treated rats either with a control solution or with standard PDF and compared the increase in mesothelial cell counts in effluents between both groups. As shown in Fig. 7 B, PDF resulted in a significantly higher increased detachment of mesothelial cells. There was no significant effect of PDF temperature on detachment of mesothelial cells (20°C: 62 ± 24, 37°C: 94 ± 46, and 41.5°C: 124 ± 43 mesothelial cells, P = 0.43).

Fig. 7. Mesothelial cell detachment in the rat model of peritoneal dialysis. Box (25th and 75th), whiskers (10th and 90th percentile) and median plot of mesothelial cell counts in PDF effluent collected during an intraperitoneal dwell are shown. In vivo PDF exposure resulted in transiently significant increase in mesothelial cell detachment ( A ). At 4 h of intraperitoneal dwell, PDF exposure caused significantly higher mean numbers of mesothelial cell detachment than exposure to control culture medium ( B ). Data are obtained in 6 rats/group in 3 independent experiments.

Next, we analyzed the "natural course" of HSP-72 expression in the detached and resident rat mesothelial cell populations following acute in vivo exposure to PDF. As shown in Fig. 8, at 4-h dwell overexpression of HSP-72 was specifically found in those mesothelial cells that were detached on in vivo exposure to PDF but not in mesothelial cells that remained resident in the peritoneal monolayer until the end of the dwell.

Fig. 8. In vivo PDF exposure induces HSP-72 expression in detached but not in resident rat peritoneal mesothelial cells. Top and middle : coimmunofluorescence staining (yellow) for cytokeratin (green) and HSP-72 (red) in cytospins of rat peritoneal cells in PDF effluents at 4-h dwell following in vivo PDF exposure. Light microscopy ( top left ) and 4,6-diamidino-2-phenylindole stain (DAPI; middle right ) show the absence of these proteins in other peritoneal cell types. Bottom : negativity for HSP-72 in resident mesothelial cells lining the peritoneal cavity following a 4-h dwell with PDF. Data are representative of 5 independent experiments.

Finally, we investigated the effects of manipulating HSP-72 expression on detachment of mesothelial cells in the in vivo rat model of peritoneal dialysis. The mesothelial stress response was induced before PDF exposure by transient hyperthermia, and effects on HSP induction and mesothelial cell detachment were assessed. As shown in Fig. 9, this heat conditioning protocol resulted in marked overexpression of HSP-72 and significantly reduced mesothelial cell detachment following in vivo PDF exposure during a 4-h dwell.

Fig. 9. Effects of heat conditioning on mesothelial HSP-72 expression and detachment following acute in vivo PDF exposure in the rat model of peritoneal dialysis. Top : heat-induced HSP-72 overexpression is demonstrated in Western blots of protein extracts of resident mesothelial cells harvested from untreated and conditioned rats by trypsin washout of the peritoneal cavity. Bottom : plot (means ± SD) of mesothelial cell counts in PDF effluent collected after 4 h of intraperitoneal dwell. Heat conditioning resulted in significantly reduced mesothelial cell detachment following in vivo PDF exposure. Data are obtained in 4 rats/group in 2 independent experiments.

DISCUSSION

The results of this study establish an essential role of HSP in repair and cytoprotection of cytoskeletal integrity in mesothelial cells following acute in vitro and in vivo exposure to PDF.

In the first part of this study, sublethal PDF exposure was evaluated for its ability to induce cytoskeletal disruption in mesothelial cells in vitro. Ezrin, cadherin, ZO-1, and Na-K-ATPase are essential for an intact cellular architecture ( 14, 15, 23, 24, 32 ). Based on the known effects of PDF exposure on ultrastructural morphology in mesothelial cells, such as reduced numbers of microvilli, widened intercellular gaps, cellular detachment, and loss of cellular polarity, these proteins were particularly attractive markers for PDF-induced cytoskeletal disruption. In addition to classic morphological assessment of cytoskeletal integrity, this study primarily used test systems that we and others had previously established as markers for acute nonlethal injury in renal tubule cells ( 6, 10, 11, 28, 33, 37, 40 ). The detergent Triton X-100 solubilizes cell membranes, and the cellular pool of proteins can be fractionated into an insoluble pellet (=cytoskeletal-associated fraction) and a soluble supernatant (=noncytoskeletal, dissociated fraction) by simple differential centrifugation. The distribution of such marker proteins between pellet and supernatant has been shown to accurately and reproducibly quantify cytoskeletal disruption on nonlethal cellular injury ( 28, 37 ).

Cytoskeletal findings in mesothelial cells exposed to PDF were comparable to observations in polarized renal tubular epithelia following sublethal injury. Prolonged exposure to standard PDF resulted in severe loss of cell-cell and cell-matrix contact of mesothelial cells and caused cellular detachment. Shorter exposures, or improved biocompatibility, resulted in disruption of the actin-based cytoskeleton and increased Triton extractability of the marker proteins with otherwise apparently unchanged morphology. Effects of acute nonlethal exposure to PDF can thus be quantified by dose-dependant, reversible disruption of the cytoskeletal anchorage of marker proteins in mesothelial cells. These results confirm that methods based on detergent extractability of membrane proteins may be transferred to the in vitro model of peritoneal dialysis.

In the following parts of this study, cytoskeletal restoration was investigated in the in vitro model of peritoneal dialysis, focusing on the concept of HSP-mediated repair. Both HSP-27 and HSP-72 showed distinct temporal patterns of postexposure cellular expression that can be readily related to the recycling of injured proteins. The finding of a shift of these HSP into the Triton-insoluble pellet also strongly suggests association with the disrupted cytoskeleton during the cellular reorganization process ( 3, 4, 6, 10, 11 ). The prevalence of such cellular recycling mechanisms in cellular repair of the peritoneal mesothelium is further supported by previous reports that overall de novo expression of proteins is reduced in these cells on exposure to PDF ( 13, 26 ). These results therefore suggest that HSP represent attractive candidates to interact with disrupted elements of the cytoskeleton and participate in its restoration in mesothelial cells during recovery from exposure to PDF. These observations, however, do not prove direct functional interactions and provide only indirect evidence for HSP-mediated repair.

As a next step, we therefore manipulated HSP expression in mesothelial cells and analyzed the effects of this intervention on cytoskeletal integrity on in vitro exposure to PDF. It is typical for HSP-mediated processes that pretreatment with a nonlethal dose of cellular stress results in survival of a subsequent otherwise lethal dose of the same or other type of injury ( 12 ). The finding of an increased resistance to repeat cellular injury is called cytoprotection, and the treatment resulting in this cytoprotection is termed "conditioning" ( 29, 31 ). Assuming functional interactions between cytoskeletal proteins and HSP in mesothelial cells during recovery from PDF exposure, increased HSP levels should improve cytoskeletal integrity ( 4, 10, 11 ). For these experiments, ezrin was chosen as the cytoskeletal marker protein, because ezrin was rapidly solubilized on sublethal PDF exposure and because loss of microvilli can be regarded as a hallmark of cytoskeletal architecture disruption following PDF exposure.

Our results demonstrated that conditioning by prior nonlethal PDF exposure indeed improved cytoskeletal stabilization as assessed by detergent extractability of ezrin following repeat PDF exposure in the in vitro model of peritoneal dialysis. Interestingly, the conditioning treatment did not affect Triton extractability of ezrin immediately after PDF exposure but resulted in a significantly more rapid stabilization during the subsequent recovery period. This finding suggests that the increasing abundance of cytoskeletal-associated HSP in the conditioned mesothelial cell (caused by both cellular redistribution and induction) is rather mediating enhanced postexposure repair than mere nonspecific resistance against cellular injury.

These results are in good agreement with the known chaperoning effects of HSP, but they still provide only indirect evidence for HSP-mediated repair of mesothelial cells on exposure to PDF. Incubation of denatured proteins such as luciferase or citrate synthase with purified HSP or with total cellular (reticulocyte) HSP-rich lysates has resulted in significant in vitro repair, readily quantifiable by gain of function (reviewed in Ref. 29 ). Therefore, we adapted a simple in vitro repair system analyzing functional interactions between HSP and disrupted elements of the cytoskeleton that were recently established in subcellular fractions of ischemic renal cortex. This assay is based on the ability of molecular chaperones to bind to hydrophobic sites of denatured proteins and subsequently undergo conformational changes, resulting in refolding, stabilization, and eventually release of reconfirmed proteins ( 4, 10, 11 ).

Using this functional assay, we found more direct evidence for the role of HSP-72 during repair of the cytoskeleton of mesothelial cells after nonlethal injury. Cellular distribution of ezrin was clearly affected by HSP-72 abundance and/or function. Coincubation with recombinant HSP-72 resulted in ATP-dependent stabilization of the cytoskeletal protein fractions from injured mesothelial cells, obtained immediately after PDF exposure. On the other hand, cytoskeletal redistribution of the membrane proteins during recovery could be abolished on coincubation with blocking antibodies against HSP. These results demonstrate that exogenously added as well as endogenously upregulated HSP interacted with substrates in the cytoskeletal fraction of mesothelial cells. These findings are consistent with HSP-mediated repair of disrupted proteins involved in the cytoskeletal organization of mesothelial cells ( 10, 11 ). Based on their known function, HSP-25 and/or HSP-90 might have been effective to convert the disrupted cytoskeletal elements to a "folding-competent" state, which can subsequently be refolded by HSP-72 ( 11 ).

Taken together, our results in the in vitro model of peritoneal dialysis suggest a novel and essential role of HSP in the repair of cytoskeletal integrity of mesothelial cells. First, markers of transient cytoskeletal disruption were defined in mesothelial cells on nonlethal cell injury following PDF exposure. Restoration of cytoskeletal integrity was then correlated with induction and cellular redistribution of HSP in that system. Finally, using conditioning protocols and novel repair assays, direct evidence for HSP-mediated cytoskeletal repair was delineated in mesothelial cells recovering from acute sublethal PDF exposure in vitro.

In the peritoneal cavity, however, rapid physicochemical equilibration of PDF and interactions between different peritoneal cell types will cause considerably increased complexity of both injury- and repair-related cellular processes ( 34, 36 ). In the next set of experiments, we therefore performed studies in intact animals to clarify the potential biological relevance of our in vitro findings.

Previously, we have described strong induction of HSP in the rat model of in vivo exposure to PDF ( 2 ). In that report, mesothelial cells were harvested by intraperitoneal trypsin digestion, precluding any differentiation between adherent and detached cell populations. In the present study, we separately investigated these mesothelial cell populations in PDF effluents and peritoneal tissue in the acute rat model. Weakening of cell-cell and cell-basal contact with subsequent cellular detachment is well accepted as a classic marker in several other models of sublethal cellular injury causing cytoskeletal disruption ( 21, 28, 37 ). In ischemic renal failure, this phenomenon is made responsible for renal tubular obstruction and urinary backleak ( 21 ). However, cellular detachment in the kidney cannot be directly evaluated, whereas our data suggest that simple assessment of mesothelial cell counts in PDF effluent might be used as a direct marker of sublethal mesothelial cell injury in the in vivo model of peritoneal dialysis.

When PDF was intraperitoneally injected, the number of detached mesothelial cells significantly increased within a 4-h dwell. Exposure to PDF resulted in a significantly higher count than exposure to control medium fluid. This is in good agreement with clinical observations, since the loss of mesothelial cells from the peritoneal surface is traditionally regarded as one of the hallmarks of PDF toxicity (17-20). Interestingly, our findings also provide strong evidence that the vast majority of the detached mesothelial cells are indeed viable but sublethally injured, as their protective cellular mechanisms were strongly induced ( 8, 35 ). Overexpression of HSP-72 was specifically found in those mesothelial cells that were detached following in vivo exposure to PDF but not in mesothelial cells that remained resident in the peritoneal monolayer until the end of the dwell. As a next step, we hypothesized that manipulation of HSP expression in the resident population before PDF injection might induce cytoprotection against in vivo exposure of the peritoneum to PDF.

Indeed, pretreatment by transient hyperthermia resulted not only in upregulation of HSP-72 but also in significantly reduced mesothelial cell detachment. These results are in good agreement with many studies that showed that organisms and tissues which overexpress HSP-72 are more resistant to cellular injury ( 12, 29, 31 ). Recently, transfection of mesothelial cells with HSP-72 resulted in protection of cell viability against subsequent PDF exposure in the cell culture model ( 9 ). In vivo, we found stabilization of the cytoskeletal integrity in the rat model of renal ischemia during repeat injury, following a classic conditioning protocol: at high cellular abundance of HSP-72, repeat renal ischemia resulted in neither increased Triton extractability nor altered cellular localization of cytoskeletal marker proteins in proximal tubular cells, in contrast to marked disruption after single ischemia at basal levels of HSP ( 4 ). Repeated exposure to PDF, as is the rule in the clinical setting, may therefore not only cause repeat injury to mesothelial cells but rather represents a kind of inadvertent conditioning treatment counteracting acute PDF toxicity. Future studies are needed to assess the effects of differential HSP expression on markers of more chronic peritoneal injury ( 39 ).

Taken together, our in vivo data clearly suggest the biological relevance of HSP-mediated cytoskeletal stabilization in mesothelial cells during peritoneal dialysis. First, mesothelial cell detachment could be related to PDF exposure. Next, HSP expression was found to be induced in detached cells but remain low in the resident mesothelial cell population. Finally, overexpression of HSP in that resident population resulted in significant reduction of mesothelial cell detachment following acute in vivo PDF exposure in the rat model.

Although HSP-mediated cytoprotection has long since been described as a potential therapeutic concept for acute cellular injury, its clinical application is seriously hampered because this concept strictly depends on planned "time course and dosages" of cellular insults ( 27, 29, 31 ). The majority of clinically relevant injuries, however, are complex and occur rather unpredictably with regard to both time course and dose. In peritoneal dialysis, with its repeated (and thoroughly predictably timed and dosed) exposure of mesothelial cells to PDF, therapeutic manipulations of HSP expression represents a unique opportunity for transferring cytoprotection from bench to bedside in future studies in more chronic models ( 30 ).

ACKNOWLEDGMENTS

C. Aufricht gratefully acknowledges the support of the Else Kroner Fresenius Foundation. We are deeply grateful to Ingrid Raab and Katalin Bojarszky-Nagy for excellent technical help and assistance.

【参考文献】
  Arbeiter K, Bidmon B, Endemann M, Bender TO, Eickelberg O, Ruffingshofer D, Mueller T, Regele H, Herkner K, Aufricht C. Peritoneal dialysate fluid composition determines heat shock protein expression patterns in human mesothelial cells. Kidney Int 60: 1930-1937, 2001.

Arbeiter K, Bidmon B, Endemann M, Ruffingshofer D, Mueller T, Regele H, Eickelberg O, Aufricht C. Induction of mesothelial HSP-72 upon in vivo exposure to peritoneal dialysis fluid. Perit Dial Int 23: 499-501, 2003.

Aufricht C, Ardito T, Thulin G, Kashgarian M, Siegel NJ, Van Why SK. Heat-shock protein 25 induction and redistribution during actin reorganization after renal ischemia. Am J Physiol Renal Physiol 274: F215-F222, 1998.

Aufricht C, Bidmon B, Ruffingshofer D, Regele H, Herkner K, Siegel NJ, Kashgarian M, Van Why SK. Ischemic conditioning prevents Na,K-ATPase dissociation from the cytoskeletal cellular fraction after repeat renal ischemia in rats. Pediatr Res 51: 722-727, 2002.

Aufricht C, Endemann M, Bidmon B, Arbeiter K, Mueller T, Regele H, Herkner K, Eickelberg O. Peritoneal dialysis fluids induce the stress response in human mesothelial cells. Perit Dial Int 21: 85-88, 2001.

Aufricht C, Lu E, Thulin G, Kashgarian M, Siegel NJ, Van Why SK. ATP releases HSP-72 from protein aggregates after renal ischemia. Am J Physiol Renal Physiol 274: F268-F274, 1998.

Bajo MA, Selgas R, Castro MA, del Peso G, Diaz C, Sanchez-Tomero JA, Fernandez de Castro M, Alvarez V, Corbi A. Icodextrin effluent leads to a greater proliferation than glucose effluent of human mesothelial cells studied ex vivo. Perit Dial Int 20: 742-747, 2000.

Beere HM, Green DR. Stress management-heat shock protein-70 and the regulation of apoptosis. Trends Cell Biol 11: 6-10, 2001.

Bidmon B, Endemann M, Arbeiter K, Ruffingshofer Regele H D, Herkner K, Eickelberg O, Aufricht C. Overexpression of HSP-72 confers cytoprotection in experimental peritoneal dialysis. Kidney Int 66: 2300-2307, 2004.

Bidmon B, Endemann M, Mueller T, Arbeiter K, Herkner K, Aufricht C. Heat shock protein-70 repairs tubule cell structure after renal ischemia. Kidney Int 58: 2400-2407, 2000.

Bidmon B, Endemann M, Mueller T, Arbeiter K, Herkner K, Aufricht C. HSP-25 and HSP-90 stabilize Na,K-ATPase in cytoskeletal fractions of ischemic rat renal cortex. Kidney Int 62: 1620-1627, 2002.

Borkan SC, Wang YH, Lieberthal W, Burke PR, Schwartz JH. Heat stress ameliorates ATP depletion-induced sublethal injury in mouse proximal tubule cells. Am J Physiol Renal Physiol 272: F347-F355, 1997.

Breborowicz A, Oreopoulos DG. Biocompatibility of peritoneal dialysis solutions. Am J Kidney Dis 27: 738-743, 1996.

Chen J, Cohn JA, Mandel LJ. Dephosphorylation of ezrin as an early event in renal microvillar breakdown and anoxic injury. Proc Natl Acad Sci USA 92: 7495-7499, 1995.

Contreras RG, Shoshani L, Flores-Maldonado C, Lazaro A, Monroy AO, Roldan ML, Fiorentino R, Cereijido M. E-cadherin and tight junctions between epithelial cells of different animal species. Pflügers Arch 444: 467-475, 2002.

Davies SJ, Phillips L, Griffiths AM, Russell LH, Naish PF, Russell GI. What really happens to people on long-term peritoneal dialysis? Kidney Int 54: 2207-2217, 1998.

Di Paolo N, Garosi G, Petrini G, Monaci G. Morphological and morphometric changes in mesothelial cells during peritoneal dialysis in the rabbit. Nephron 74: 594-599, 1996.

Di Paolo N, Sacchi G. Atlas of peritoneal histology. Perit Dial Int 20: S5-S96, 2000.

Di Paolo N, Sacchi G, DeMia M, Gaggiotti E, Capotondo L, Rossi P, Bernini M, Pucci AM, Ibba L, Sabatelli P. Morphology of the peritoneal membrane during continuous ambulatory peritoneal dialysis. Nephron 44: 204-211, 1986.

Dobbie JW. New concepts in molecular biology and ultrastructural pathology of the peritoneum: their significance for peritoneal dialysis. Am J Kidney Dis 15: 97-109, 1990.

Fish EM, Molitoris BA. Alterations in epithelial polarity and the pathogenesis of disease states. N Engl J Med 330: 1580-1588, 1994.

Gotloib L, Wajsbrot V, Cuperman Y, Shostak A. Acute oxidative stress induces peritoneal hyperpermeability, mesothelial loss, and fibrosis. J Lab Clin Med 143: 31-40, 2004.

Granes F, Urena JM, Rocamora N, Vilaro S. Ezrin links syndecan-2 to the cytoskeleton. J Cell Sci 113: 1267-1276, 2000.

Grune T, Reinheckel T, North JA, Li R, Bescos PB, Shringarpure R, Davies KJ. Ezrin turnover and cell shape changes catalyzed by proteasome in oxidatively stressed cells. FASEB J 16: 1602-1610, 2002.

Ito T, Yorioka N, Yamamoto M, Kataoka K, Yamakido M. Effect of glucose on intercellular junctions of cultured human peritoneal mesothelial cells. J Am Soc Nephrol 11: 1969-1979, 2000.

Jorres A, Topley N, Gahl GM. Biocompatibility of peritoneal dialysis fluids. Int J Artif Organs 15: 79-83, 1992.

Minowada G, Welch WJ. Clinical implications of the stress response. J Clin Invest 95: 3-12, 1995.

Molitoris BA, Geerdes A, McIntosh JR. Dissociation and redistribution of Na +,K + -ATPase from its surface membrane actin cytoskeletal complex during cellular ATP depletion. J Clin Invest 88: 462-469, 1991.

Morimoto RI, Tissieres A, Georgopoulos C. The Biology of Heat Shock Proteins and Molecular Chaperons. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press, 1994, p. 1-594.

Morimoto RI, Santoro MG. Stress-inducible responses and heat shock proteins: new pharmacologic targets for cytoprotection. Nat Biotechnol 16: 833-838, 1998.

Nover L. (Editor). Heat Shock Response. Boca Raton, FL: CRC, 1991.

Nusrat A, Parkos CA, Verkade P, Foley CS, Liang TW, Innis-Whitehouse W, Eastburn KK, Madara JL. Tight junctions are membrane microdomains. J Cell Sci 113: 1771-1781, 2000.

Price VR, Reed CA, Lieberthal W, Schwartz JH. ATP depletion of tubular cells causes dissociation of the zonula adherens and nuclear translocation of beta-catenin and LEF-1. J Am Soc Nephrol 13: 1152-1161, 2002.

Ruffingshofer D, Endemann M, Arbeiter K, Bidmon B, Mueller T, Herkner K, Aufricht C. Induction of heat shock protein 72 in mesothelial cells exposed to peritoneal dialysate effluent. Perit Dial Int 23: 74-77, 2003.

Samali A, Holmberg CI, Sistonen L, Orrenius S. Thermotolerance and cell death are distinct cellular responses to stress: dependence on heat shock proteins. FEBS Lett 461: 306-310, 1999.

Topley N. What is the ideal technique for testing the biocompatibility of peritoneal dialysis solutions? Perit Dial Int 15: 205-209, 1995.

Tsukamoto T, Nigam SK. Tight junction proteins form large complexes and associate with the cytoskeleton in an ATP depletion model for reversible junction assembly. J Biol Chem 272: 16133-16139, 1997.

Van Why SK, Kim S, Geibel J, Seebach FA, Kashgarian M, Siegel NJ. Thresholds for cellular disruption and activation of the stress response in renal epithelia. Am J Physiol Renal Physiol 277: F227-F234, 1999.

Williams JD, Craig KJ, Topley N, Von Ruhland C, Fallon M, Newman GR, Mackenzie RK, Williams GT, Peritoneal Biopsy Study Group. Morphologic changes in the peritoneal membrane of patients with renal disease. J Am Soc Nephrol 13: 470-479, 2002.

Zimmerhackl LB, Momm F, Wiegele G, Brandis M. Cadmium is more toxic to LLC-PK 1 cells than to MDCK cells acting on the cadherin-catenin complex. Am J Physiol Renal Physiol 275: F143-F153, 1998.

作者单位：1 Department of Pediatrics, Univ. Kinderklinik Wien, 2 Division of Biomedical Research, and 3 Department of Pathology, Vienna Medical School, Vienna, Austria