The maximal cytoprotective function of the heat shock protein 27 is dependent on heat shock protein 70

Abstract

Endogenous heat shock proteins (HSPs) 70 and 25/27 are induced in renal cells by injury from energy depletion. Transfected over-expression of HSPs 70 or 27 (human analogue of HSP25), provide protection against renal cell injury from ATP deprivation. This study examines whether over-expressed HSP27 depends on induction of endogenous HSPs, in particular HSP70, to afford protection against cell injury. LLC-PK1 cells transfected with HSP27 (27OE cells) were injured by ATP depletion for 2 h and recovered for 4 h in the presence of HSF decoy, HSP70 specific siRNA (siRNA-70) and their respective controls. Injury in the presence of HSF decoy, a synthetic oligonucleotide identical to the heat shock element, the nuclear binding site of HSF, decreased HSP70 induction by 80% without affecting the over-expression of transfected HSP27. The HSP70 stress response was completely ablated in the presence of siRNA-70. Protection against injury, provided by over-expression of HSP27, was reduced by treatment with HSF decoy and abolished by treatment with siRNA-70. Immunoprecipitation studies demonstrated association of HSP27 with actin that was not affected by either treatment with HSF decoy or siRNA. Therefore, HSP27 is dependent on HSP70 to provide its maximal cytoprotective effect, but not for its interaction with actin. This study suggests that, while it has specific action on the cytoskeleton, HSP 25/27 must have coordinated activity with other HSP classes, especially HSP70, to provide the full extent of resistance to injury from energy depletion.

1. Introduction

Kidneys are responsible for maintaining acid–base and fluid balance of the body, and proximal tubules of the nephron play an important role in preserving this stability. Membrane polarity of the renal proximal tubule is essential to maintain body pH and volume equilibrium and its loss leads to disruption in this important kidney function. Ischemia of the kidney causes rapid, duration-dependent, reversible loss of surface membrane polarity in proximal tubule cells [1]. Renal function is disturbed to varying degrees in critically ill patients affected by ischemic injury, which in turn disrupts the delicate fluid and acid–base balance of the body, negatively affecting patient morbidity and mortality [2].

Energy depletion in proximal tubule cells produces cellular changes that replicate those seen in ischemic renal tissue. This tissue culture model is well accepted to study cellular and molecular events in renal cells subjected to ischemic injury [3]. The detachment of Na, K-ATPase from the cytoskeleton following injury is duration-dependent and reversible [1,4]. The loss of Na,K-ATPase anchorage to the cytoskeleton in the basolateral domain of the cell membrane increases Na,K-ATPase detergent solubility. This has led to the routine use of quantitative changes in Na,K-ATPase solubility as a reliable measure of early and reversible injury, the technique used in the present studies [4–8].

Heat shock proteins have been implicated by several groups in modulating the effects of renal injury from ischemia [9–13]. Both 70 kD and 27 kD heat shock proteins have demonstrated protective effects in the cell culture model of reversible renal injury [6,7,9]. The interactions between these two classes of HSPs and their relative roles in protecting renal cells from injury are not well characterized. Defining molecular processes of renal epithelial cell injury and repair affected by these stress proteins may provide therapeutic targets to limit injury or augment cellular repair. The aim of this study was to determine whether HSP27 is dependent on HSP70 in providing tolerance to injury from energy depletion. We further attempt to understand the molecular basis of this relationship.

Porcine kidney proximal tubule (LLC-PK1) cells transfected with HSP27 were subjected to ATP depletion to simulate ischemic kidney injury. Gene silencing techniques [8] that result in suppression of endogenous HSPs, without affecting the level or function of transfected HSPs, were used to modulate the expression of individual classes of HSP. This allows for identifying functional interactions between HSP70 and HSP27 with each other, with actin, and then defining the attendant effect on cell integrity.

2. Methods

2.1. 27OE and VC cell culture

Studies were performed using cultured porcine proximal tubule cells (LLC-PK1) stably transfected with an empty control vector (vector control or VC cells) or a vector containing a gene to over-express human HSP27 (27OE cells). The cells were grown in α-MEM (Cellgro, Mediatech, Herndon, VA) with 10% fetal bovine serum and 1 mg/ml of geneticin at 37 °C in 5% CO₂ as previously described [8]. Cell culture filters (1 μm, BioCoat, BD Biosciences, Bedford, MA) coated with 40 μg of collagen IV (BD Biosciences), were seeded with 2.8 × 10⁵ cells. Studies were conducted 4 days after plating when cells had achieved confluence.

2.2. Cell injury

Acute renal ischemia–reperfusion injury, in vivo, leads to cellular depletion of ATP. This initiates a cascade of events within the renal epithelial cells, the severity of which is dependent on the degree and duration of the insult [1]. Subjecting cultured cells to ATP depletion is an established in vitro model of ischemia–reperfusion injury, producing cellular changes concordant with those found in in vivo injury in the cells [3,4,10,14]. The current studies were performed using this well established in vitro model of sublethal renal cell injury. Four days following plating, confluent 27OE cells or VC cells were injured by incubation in substrate-free media containing the mitochondrial inhibitor antimycin A (0.1 μM) for 2 h and allowed to recover in normal growth media (α-MEM) for 4 h following a wash with PBS as previously described [6–8,14].

2.3. HSF-1 decoy preparation

HSF-1 is the primary transcriptional regulator of rapid elaboration of the inducible stress proteins, HSP70 and HSP25/27. HSF-1 rapidly initiates transcription of HSP70 and HSP25/27 through binding to the heat shock element (HSE). We previously showed that synthesized HSE can be used as HSF-1 decoy to suppress the transcription of HSP70 and HSP25, by 75% and 90% respectively, in LLC-PK1 cells [8]. Oligodeoxynucleotides were synthesized as described previously [8]. Briefly, following the spontaneous annealing of complementary sequences, the free ends were ligated by incubation with 350 mM N-(3-dimethylaminopropyl)-N’-ethylcarbodiimide, 50 mM 2-morpholinoethanesulfonic acid sodium salt and 20 mM magnesium chloride at 4 °C for 60 h. Oligonucleotides were subsequently removed from the reaction mix by ethanol precipitation. This HSF-1 decoy was incubated with cells both during treatment with the injury media and with the recovery media.

2.4. siRNA preparation

The design and synthesis of siRNA used protocols and reagents supplied by Ambion Inc. (Austin, TX, USA). An siRNA previously identified as able to inhibit inducible HSP70 synthesis was transfected into LLC-PK1 cells 24 h prior to injury experiments using an siPort Lipid (Ambion, Austin, TX) based transfection reagent [8].

2.5. Cell harvest

Cells were harvested by scraping into chilled extraction buffer containing 0.1% Triton X-100, 60 mM piperazine-N,N’-bis(2-ethane-sulfonic acid)–pH 6.8, 25 mM HEPES, 10 mM EGTA, and 2 mM magnesium chloride. The cells were either sonicated whole for heat shock protein quantification or centrifuged at 36,000 g for 30 min to separate the Triton-soluble from the insoluble cytoskeleton associated protein fraction [5–8].

2.6. Assessment of protein expression

Sample protein concentration was measured by the BCA protein assay, using BSA as a protein standard (Pierce, USA). Equal amounts of protein (10 μg) were subjected to SDS-PAGE electrophoresis on 4–20% gradient gels (Criterion, Biorad) and transferred onto nitrocellulose membranes. Following blocking of non-specific binding sites with 5% skimmed milk, membranes were incubated for 1 h with monoclonal antibodies directed against inducible HSP70 (SPA-810, Stressgen, B.C., Canada) (dilution 1:5000), HSP25 (SPA-801, Stressgen, B.C., Canada) (dilution 1:2000), HSP27 (SPA-800, Stressgen, B.C., Canada) (dilution 1:5000), β-actin (A5441, Sigma-Aldrich) (dilution 1: 5000) or α1 subunit of Na,K-ATPase [15] (dilution 1:5000) respectively. After repeated washings and incubation with an appropriate species-specific secondary antibody for 1 h, immunoreactive protein was detected with enhanced chemiluminescence and quantified by means of densitometry (Scion Image, USA) as previously described [8]. The linearity of measurements within the experimental range was confirmed by serial dilution and subsequent densitometry.

2.7. Immunoprecipitation

As has been previously described [6], immunoprecipitation was performed following harvesting of the cells in a modified aggregate buffer (0.1% Triton X, 50 mM TRIS (pH 7.4), 50 mM NaCl, 1 mM EDTA, 1 mM EGTA, 2 mM CDTA, 60 mM PIPES, 50 nM PMSF and Complete Protease Inhibitor Cocktail (Roche)) and centrifugation for 10 min at 12,000 g after lysis of cells by 3 cycles of alternating freeze/thaw. Briefly, equal volumes of cell supernatant were adjusted to contain equal quantities of protein and divided into two aliquots, subsequent to preclearing, and incubated for 1 h at 4 °C with 100 μg of antibody (IgG1) directed against either HSP27 (SPA-800, Stressgen, B.C, Canada) or actin (A5441, Sigma-Aldrich); or a normal mouse IgG (SC-2025, Santa Cruz) as a negative control. This was followed by overnight incubation with 20 μl of agarose-bound Protein G at 4 °C. Agarose-bound Protein G was then separated by centrifuging for 30 s at 12,000 g, and the supernatant discarded. The remaining pellets were washed five times with 500 μl of wash buffer, suspended in an equal volume of 2× Western loading buffer and boiled for 3 min prior to SDS electrophoresis on a 4–20% gradient gel as described above.

2.8. Sample size and statistical evaluation

HSP abundance and Na,K-ATPase detachment from the cytoskeleton was measured and compared between VC and 27OE cells under each experimental condition (n = 5 for each condition). Average values were calculated and expressed in terms of the percentage change from the respective level in uninjured, untreated cells. Values are expressed as a mean ± standard error of the mean (SEM). Comparison between experimental groups was made using Student’s t-test. Values were considered significantly different if p<0.05. Immunoprecipitation was performed on a minimum of three occasions to ensure consistency of results.

3. Results

The expression of endogenous heat shock proteins (HSP70 and 25) and transfected heat shock protein (HSP27) was determined in empty vector control cells (VC cells) and in transfected HSP 27 over-expressing cells (27OE cells) in uninjured control cells and in cells injured by ATP depletion (Fig. 1). As expected, VC cells exhibited no HSP27 expression, while 27OE cells had abundant HSP27 with similar expression between control conditions and after injury induced by ATP depletion (Fig. 1, Panel A). Although both VC cells and 27OE cells responded to energy deprivation injury by inducing HSP25 and HSP70, the induction was blunted in 27OE cells compared with VC cells (Fig. 1, Panels B and C).

Fig. 1.

Expression of individual HSPs under uninjured condition (Control) and following recovery for 4 h after 2 h of injury (Injury) in Vector control (VC cells) and HSP27 transfected cells (27OE cells). Shown are the Western blots of cell lysates stained with antibodies directed against transfected human HSP27 (Panel A), endogenous HSP25 (Panel B), endogenous HSP70 (Panel C) and actin (Panel D).

Fig. 2 shows the effect of using the HSF decoy to dampen the stress induction of endogenous HSP70 and HSP25 in 27OE cells. Since the HSF decoy should only affect the expression of endogenous HSPs under the influence of HSF (HSP70 and 25), the transfected HSP27, controlled only by the viral promoter in the plasmid, should not be altered by administration of decoy. As expected, HSF decoy treatment did not change HSP 27 expression in 27OE cells subjected to injury (Fig. 2A, Panel 1). HSP27 expression was not significantly different between uninjured and injured 27OE cells either in the presence or absence of HSF decoy (Fig. 2A, Panel 1 and Fig. 2B, Upper Panel). HSF decoy completely suppressed the induction of HSP25 following injury, and further reduced levels of HSP25 below that found in uninjured control cells (Fig. 2A, Panel 2 and Fig. 2B, Middle Panel). HSF decoy treatment reduced HSP70 induction following injury by 80%, resulting in HSP70 expression at only 5% above that in uninjured control cells (Fig. 2A, Panel 3 and Fig. 2B, Lower Panel). So when 27OE cells were injured in the presence of HSF decoy, HSP70 levels fell substantially but remained above the values seen in uninjured cells; HSP25 abundance was reduced below uninjured levels; and HSP27 expression was unaffected.

Fig. 2.

Effect of HSF decoy on HSP expression, with and without injury by ATP depletion, in 27OE cells. Panel A shows representative Western blots of 27OE cell lysate stained with antibodies directed against HSP27, HSP25, HSP70 and actin respectively. Panel B is a graph of densitometry (mean ± SEM) of Western blots probed for endogenous HSP70 and HSP25 and transfected HSP27, expressed as change from uninjured control (n = 5 for all conditions). * represents p<0.05 between groups.

Fig. 3 shows the effect of gene silencing using siRNA directed against HSP70 (siRNA-70) on expression of endogenous HSP70, endogenous HSP25, and exogenous HSP27 in uninjured 27OE cells and cells injured by energy depletion. As has been described previously in LLCPK cells [8], treatment with siRNA-70 eliminated stress induction of endogenous HSP70 in the 27OE cells (Fig. 3, Panel A). Injuring 27OE cells by ATP depletion in the presence of siRNA-70 did not demonstrate inhibition of HSP25 induction or affect the abundance of HSP27, establishing the specificity of siRNA-70 in blocking HSP70 induction (Fig. 3, Panels B and C).

Fig. 3.

Western blot of 27OE cells stained with antibody directed against HSP70, HSP25, HSP27, and actin. Cells were from uninjured (C) and injured (I) conditions treated either with vehicle or with siRNA against HSP70 (I + siRNA-70).

Detachment of Na,K-ATPase from the cytoskeleton was used as a marker of the severity of cell injury and was compared between VC cells (n = 5) and 27OE cells injured in the presence (n = 5) or absence (n = 5) of HSF decoy and in cells treated with siRNA-70 (Fig. 4). Aspreviously reported, 27OE cells are protected against detachment of Na,K-ATPase from the cytoskeleton following injury [7]. Injured 27OE cells have Na,K-ATPase detachment only 30% above uninjured cells, far less than VC cells that have Na,K-ATPase detachment 110% above uninjured cells (p<0.05). Treatment of 27OE cells with HSF decoy, which dampens HSP70 and HSP25 induction, but has no effect on HSP27 (Fig. 2), partially reversed the protection afforded by HSP27 transfected over-expression. Compared with the mild injury of 30% Na,K-ATPase detachment sustained in 27OE cells, treatment of these same cells with HSF decoy during the insult resulted in greater, intermediate injury of 62% increase in Na,K-ATPase detachment above uninjured VC cells (Fig. 4, p<0.05). Nevertheless, the 27OE cells treated with HSF decoy still had less injury than cells transfected with vector alone (62% increase in Na,K-ATPase detachment in HSF decoy treated injured 27OE cells versus 110% in uninjured VC cells, Fig. 4). Therefore, the maximal preservation of cellular integrity associated with over-expression of HSP27 was significantly reduced by the suppression of injury-related induction of endogenous HSP70 and HSP25.

Fig. 4.

Na,K-ATPase detachment in injured VC cells and 27OE cells injured in the presence or absence of HSF decoy or of siRNA-70. The graph is a summary of detergent extractable, soluble Na,K-ATPase that is detached from the cytoskeleton (mean ± SEM). Each experimental group is expressed as percent change from parallel, uninjured control cells. Below the graph is a representative Western blot of soluble Na,K-ATPase in each of the experimental conditions. n = 5 for all conditions. * represents p<0.05 between groups.

Injury of 27OE cells in the presence of siRNA-70, which abolishes induction of HSP70 without altering stress induction of HSP25 or transfected over-expression of HSP27, increased Na,K-ATPase detachment beyond that seen in 27OE cells and beyond 27OE cells treated with HSF decoy (144% more than uninjured; p<0.01). Therefore, isolated elimination of injury-induced HSP70 expression abolished the cytoprotection provided by over-expression of HSP27 (Fig. 4).

To determine which proteins of interest were associated with the HSP27 complex, immunoprecipitation and co-immunoprecipitation studies were undertaken in 27OE cells (Figs. 5 and 6). Studies were performed in uninjured cells, after 2 h of ATP depletion and after 4 h of subsequent recovery. Immunoprecipitates of HSP27 were studied by Western blot with antibodies against HSP70, Na,K-ATPase and actin. Overlapping detection of HSP27 or actin with the light and heavy chains of immunoglobulin G, which have similar molecular weights, was avoided by employing gradient gel electrophoresis. Fig. 5, Panel A demonstrates that equivalent HSP27 was immunoprecipitated from the cells at each experimental injury and recovery interval (Lanes 1, 2 and 3). HSP27 was not detected in the antibody alone lane (Lane 4) which has the antibody against HSP27 incubated alone (no cell extract) with the Protein G. Actin was detected as a co-precipitant with HSP27 in each experimental setting (Fig. 5, Panel B, Lanes 1, 2 and 3), but not in the antibody alone lane (Fig. 5, Panel B, Lane 4). There was no immunoprecipitation of HSP27 or co-precipitation of actin in the negative control lanes, in which immunoprecipitation was performed with a non-specific antibody in uninjured cells, after 2 h of ATP depletion and after 4 h of subsequent recovery (Fig. 5 Panels A and B, Lanes 6, 7 and 8). Neither HSP70 nor Na,K-ATPase was found in any samples as co-precipitating with HSP27 (data not shown).

Fig. 5.

Immunoprecipitation of HSP27 and co-precipitation of actin in 27OE cells following no injury (Lane 1), injury for 2 h (Lane 2) and injury followed by recovery for 4 h (Lane 3). Incubation of HSP27 antibody alone with Protein G (Lane 4) was used to clarify target proteins from immunoglobulin chains which have molecular weights near target protein. Lanes 5 in both A and B are positive controls. Parallel immunoprecipitation with non-specific normal mouse IgG was performed as a negative control in all three conditions (Lanes 6, 7, and 8). Shown are Western blots stained for HSP27 (Panel A) and Actin (Panel B). These experiments were repeated three times to confirm consistency.

Fig. 6.

Immunoprecipitation of actin and co-precipitation of HSP27 in 27OE cells following no injury (Lane 1), injury for 2 h (Lane 2) and injury followed by recovery for 4 h (Lane 3). Incubation of actin antibody alone with Protein G (Lane 4) was used to clarify target proteins from immunoglobulin chains which have molecular weights near target protein. Lanes 5 in both A and B are positive controls. Parallel immunoprecipitation with non-specific normal mouse IgG was performed as a negative control in all three conditions (Lanes 6, 7, and 8). Shown are Western blots stained for Actin (Panel A) and HSP27 (Panel B). These experiments were repeated three times to confirm consistency.

The association between HSP27 and actin was further studied under the same injury conditions as described for Fig. 5 by immunoprecipitating actin and determining the presence of HSP27 as a co-precipitant (Fig. 6). Panel A demonstrates the amount of actin immunoprecipitated during uninjured (Lane 1), injured (Lane 2) and recovered (Lane 3) conditions. Panel B confirms the co-precipitation of HSP27 with actin in each experimental setting. Just as in the HSP27 immunoprecipitation experiments, neither actin nor HSP27 was detected in samples immunoprecipitated with a non-specific antibody (Lanes 6, 7 and 8).

To determine whether the interaction between HSP27 and actin is affected by the level of HSP70 expression, immunoprecipitation of HSP27 and co-precipitation of actin in 27OE cells were performed in cells treated with HSF decoy or siRNA-70 (Fig. 7). The co-precipitation of actin with the immunoprecipitated HSP27 did not change when the cells were injured in the presence of HSF decoy (Panel A) or siRNA-70 (Panel B). HSF decoy, which inhibits endogenous HSP70 and HSP25 induction during injury, thus does not affect the co-precipitation of actin with HSP27. Similarly, siRNA-70 nullifies the injury-associated induction of endogenous HSP70 without affecting the HSP27–actin interaction. These immunoprecipitation studies, then, show that suppression or ablation of the injury-induced endogenous heat shock protein response does not ablate the association of HSP27 with actin.

Fig. 7.

Immunoprecipitation of HSP27 and co-precipitation of actin in 27OE cells following no injury (Lane 1), injury for 2 h (Lane 2) and injury followed by recovery (Lane 3) in the presence of HSF decoy (Panel A) or siRNA-70 (Panel B). Shown are Western blots stained for actin. These experiments were repeated three times to confirm consistency.

4. Discussion

Resistance to injury by energy depletion, measured as change in detachment of Na,K-ATPase from the cytoskeleton, occurs when either HSP27 or HSP70 is over-expressed in cultured renal tubule cells [6,7]. Even though transfected over-expression of an individual HSP (e.g. HSP27 and HSP70) affords protection to cells, induction of endogenous HSPs might still be an essential event contributing to the cytoprotection observed in HSP over-expressing cells. Whether the cytoprotection produced by over-expression of either HSP70 or HSP27 is from the over-expressed HSP per se, or further requires contribution from endogenous HSP induction, has not previously been determined. The model system developed for the current study enabled assessment of the effect of a single class of heat shock protein on renal tubule cell injury, separated from an inducible stress protein response.

Interaction of HSPs with specific proteins is considered to be integral to their function as chaperone proteins [5–7,16–18]. Specific association between HSP70 and Na,K-ATPase was found in renal cells, and the degree of interaction increased following cell injury [5,6]. The binding of HSP70 to Na,K-ATPase is specific and dynamic and appears to be one fundamental mechanism in the maintenance of renal cell structure [5,6]. Furthermore, HSP27 previously was demonstrated to associate with the cytoskeletal protein actin in renal epithelia [7,19,20]. These protein interactions may be independent and specific to each class of HSP, but they could also be cooperative in protecting against renal cell injury. Our study shows that in in vitro renal ischemia/reperfusion injury, even though the particular function of each HSP may be specific and differ between HSPs, each class of HSP may depend on the other HSP classes to provide the broader effect of cytoprotection.

To determine whether HSP27 function depends on HSP70 induction, we used a culture model of renal proximal tubule cells in which HSP27 is over-expressed by transfection. The HSP27 production is regulated by the promoter within the plasmid and is not influenced by the endogenous regulatory mechanisms within the cell. The endogenous HSPs in these cells, namely HSP70 and HSP25 (porcine analogue of the human HSP27), are controlled by HSF, the endogenous regulator of inducible HSPs. The regulatory difference between endogenous and transfected HSPs allowed us to modulate endogenous HSP expression independent of the transfected HSP27 expression, by using HSF decoy or further protein-specific gene silencing siRNA.

Our study shows that HSP27 transfected cells have the endogenous inducible stress response, albeit to a lesser degree than the vector control cells. Previous studies in LLC-PK1 cells found that injury increased HSP25 by 25% above uninjured control, while HSP70 increased by 80% [21]. The difference in endogenous stress proteinexpression in the HSP27 over-expressing cells might stem either from reduced baseline expression of endogenous HSP70 or HSP25, or negative feedback control of heat shock protein induction following injury [21–23].

Suppressing the endogenous stress response as well as abolishing specific HSP70 induction, by gene silencing using HSF decoy and HSP70 specific siRNA treatment respectively, decreased the resistance to injury afforded by over-expression of HSP27. Furthermore, specific blockade of HSP70 induction eliminated the cytoprotection, measured by detachment of Na,K-ATPase from the cytoskeleton, provided by HSP27 over-expression, despite additional injury-induced expression of the cognate HSP25. We previously found that neither HSF decoy nor siRNA-70 affected cellular ATP levels when compared to the controls that received neither compound, in either uninjured or injured conditions. This indicates that alteration in cellular ATP is not responsible for the decrease or reversal of cytoprotection by treatment with either the HSF decoy or siRNA-70 [8]. Rather, the effect likely is due to diminished cooperative action of the HSPs on disrupted cellular proteins. This study, then, indicates that HSP27 is dependent on the induction of endogenous HSPs to provide protection against disruption of cell architecture by energy deprivation. It further shows that the inducible HSP70 is an essential cofactor to HSP25/27 in delivering beneficial effect.

In this form of renal cell injury, interactions between an individual class of heat shock protein with cytoskeletal or integral membrane proteins may be specific and independent of other classes of HSP, or may require cooperative activity between different HSPs. For example, association of HSP25/27 with specific structural proteins may be dependent on HSP70 chaperone function. To explore this question, we examined the effect of suppressing inducible HSPs on the binding of transfected HSP27 to actin. In the immunoprecipitation studies, HSP27 interacted with actin in each experimental group, uninjured control as well as during injury and recovery. Suppression of inducible, endogenous HSPs did not affect the relationship between HSP27 and actin. This suggests that interaction between HSP27 and actin is independent of the inducible stress response, and in particular inducible HSP70, in injured cells. It appears, then, that the cooperative action of transfected HSP27 and inducible HSP70 is not accomplished through facilitation of HSP27 binding to actin, which might have been anticipated since disruption of the actin based cytoskeleton is a central feature of this cell injury model [3,4]. Rather, HSP27 and inducible HSP70 may be working elsewhere in the cell cooperatively, as co-chaperones to other disrupted cell proteins important to maintaining cell integrity in this form of renal cell injury. Alternately, it may be that each HSP’s individual and specific function is necessary, but not alone sufficient, to prevent breakdown of cell structure. For example, it appears that HSP25/27 class of heat shock proteins are important in maintaining cytoskeletal structure [7,20], while HSP70 supports membrane associated proteins in injured renal epithelia [6]. Although HSP27 is not dependent on induction of HSP70 for its association with actin, stabilization of the membrane associated protein Na,K-ATPase requires HSP70. So, overall protection of cell architecture cannot be achieved by over-expression of HSP27 alone, but requires coordinated induction and action with HSP70.

Taken together our findings suggest that these two classes of HSP can act independently and specifically with particular cell structure proteins, but in so doing cooperate to provide their beneficial effects. Each specific function is equally important to maintain cell integrity, and the lack of one cannot be compensated for by the other. Maximal preservation of cellular integrity and Na,K-ATPase attachment to the cytoskeleton appears to be dependent on a coordinated, but independent interaction of HSP70 with Na,K-ATPase, and HSP25/27 with actin. Thus, the maintenance and restitution of cellular integrity following renal cell injury is a dynamic and complex process.