Retinal Preconditioning and the Induction of Heat-Shock Protein 27

Abstract

purpose. Brief periods of ischemia have been shown to protect the retina from potentially damaging periods of ischemia. This phenomenon has been termed ischemic preconditioning or ischemic tolerance. In the present study the cellular changes in levels of heat shock protein (Hsp)27, -70, and -90 mRNA and expression of Hsp in the rat retina associated with ischemic preconditioning were evaluated.

methods. Unilateral retinal ischemia was created in Long-Evans and Sprague-Dawley rats for 5 minutes. Rats were then left for 1 hour to 7 days, to allow the retina to reperfuse. Retinas were dissected, the mRNA and protein isolated, and Northern and Western blot analyses conducted to detect changes in expression of Hsp27, -70, and -90. Immunohistochemical studies were used to identify retinal regions where Hsp changes occurred. Selected animals were subjected to a second ischemic event, 60 minutes in duration, to correlate the changes in expression of Hsp with functional protection of the retina from ischemic injury.

results. In control and sham-treated animals retinal Hsp27, -70, and -90 mRNAs were detectable. Five hours after retinal preconditioning, levels of Hsp27 mRNA were elevated above control levels, and 24 hours later, mRNA levels increased 200% over basal levels. Hsp27 expression remained elevated for up to 72 hours and then began to return to control levels. Hsp27 protein levels were increased by 200% over basal levels 24 hours after retinal preconditioning, remained at this level for 72 hours, and then returned to control levels. In contrast, no consistent change in Hsp70 or -90 mRNA or protein levels was observed during the course of the study. Immunohistochemical studies demonstrated that the increase in expression of Hsp27 was localized to neuronal and non-neuronal cells in the inner layers of the retina. Electroretinography studies demonstrated a strong correlation between the protection of retinal function from ischemic injury and the expression of Hsp27.

conclusions. These results provide evidence that the induction of Hsp27 is a gene-specific event associated with ischemic preconditioning in the retina. This increase in expression of Hsp27 occurs in both neuronal and non-neuronal retinal cells, and appears to be one component of the neuroprotective events induced by ischemic preconditioning in the retina.

In the heart and brain, studies have provided evidence that short periods of ischemia can protect cells from future severe ischemic injury. This has been termed ischemic preconditioning or ischemic tolerance. The development of ischemic tolerance in many tissues is accompanied by the induction of heat shock proteins (Hsps), ¹ a family of stress-activated proteins that participate in protein folding, repair, and degeneration. ² Hsps range in molecular weight from 10 to 170 kDa. Although these proteins were initially identified in Drosophila in response to heat stress, they are found in all cells, and their expression is upregulated in response to different forms of cellular stress, including ischemia. The primary physiological function of Hsps is to act as molecular chaperones that assist in protein folding. Other functions include participation in activation of glucocorticoid receptors, polymerization of actin, and transfer of proteins to lysosomes. ² In addition, overexpression of specific Hsps has been shown to prevent apoptotic cell death. ^{3 4 5} The induction of Hsps involves both transcriptional activation and translational changes. Studies have provided evidence that depletion of adenosine triphosphate (ATP) and the induction of the mitogen-activated protein (MAP) kinase pathway play significant roles in the activation of specific Hsps. ¹  

In the retina, a short preconditioning stimulus (i.e., 5 minutes of ischemia) can produce significant protection from future ischemic events up to 72 hours after the preconditioning stimulus, and this protective action can be blocked by inhibition of protein synthesis. ^{6 7} Results presented in this report demonstrate that a brief preconditioning ischemic event produced a selective upregulation of expression of Hsp27 in the inner retinal layers, and this upregulation correlated strongly with protection of the retina after a preconditioning stimulus. 

Materials and Methods

Animals were handled in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and the study protocol was approved by the Animal Care Committees at the Medical University of South Carolina and the University of Chicago. Long-Evans and Sprague-Dawley rats (150–200 g) were given free access to food and water under a normal 12-hour light-dark cycle. To induce retinal ischemia, rats were anesthetized with ketamine and xylazine (50 mg/kg and 6 mg/kg, intraperitoneally). The cornea was then anesthetized with 1 drop of 0.5% proparacaine (Alcon, Fort Worth, TX). The conjunctiva was resected, the optic nerve exposed, and 4-0 silk sutures placed around the nerve and associated blood vessels. The suture was then passed though a short length of polyethylene tubing (PE-200; Clay Adams, Parsippany, NJ), and ischemia was produced by clamping the suture-tubing tightly together. At the end of the ischemia, the suture-tubing was unclamped and the suture removed. The absence of blood flow and subsequent reperfusion were monitored by means of direct ophthalmoscopy. For sham treatments, surgical procedures were identical with those performed in the ischemia-treated rat, but the suture and tubing were not clamped together. The opposite eye of each animal served as the control. After a brief period (5 minutes) of retinal ischemia preconditioning, the rats were left for 1, 5, 24, 72, 120, and 168 hours, to allow the retina to reperfuse. Animals were killed by an overdose of pentobarbital, and the eyes were enucleated and the retinas dissected. 

Northern Blot Analysis

For each treatment, retinas of four eyes were isolated, pooled, and homogenized in extraction reagent (Trizol; Gibco, Grand Island, NY) at 4°C. Total RNA was isolated from the retina by chloroform extraction and precipitated by isopropyl alcohol. The RNA pellet was then resuspended in sterile, diethylpyrocarbonate (DEPC)-treated H₂O and stored at −70°C. RNA concentrations were determined by absorbance at 260 nm. Fifteen micrograms of total RNA from each sample was used for Northern blot analysis. RNAs were resolved on 1.2% agarose gel and transferred overnight to a nylon membrane (Amersham Biosciences, Arlington Heights, IL). After cross-linking, the blots were prehybridized (Ultrahyb solution; Ambion, Austin, TX) for 30 minutes at 42°C. Blots were then incubated overnight at 42°C with cDNA probes for ³²P-labled-Hsp27, -70, or -90. After hybridization, blots were washed twice (5 minutes each) with 2× SSC, 0.1% SDS in DEPC-H₂O, and followed by two 15-minute washes with 0.1× SSC and 0.1% SDS. Blots were then exposed to x-ray film and bands quantified on computer by densitometry (image-analysis software by Scion, Frederick, MD). To normalize differences in loading, the blots were stripped and probed with a ³²P-18s-rRNA probe. Multiple isoforms of both Hsp70 and -90 have been identified in the rat. ^{8 9} Based on sequence homology (>98%), the cDNA probe for Hsp70 detects Hsp70.1 and -70.2 (GenBank accession numbers X77207, X77208, respectively; http://www.ncbi.nlm.nih.gov/Genbank; provided in the public domain by the National Center for Biotechnology Information, Bethesda, MD), and the cDNA probe for Hsp90 detects Hsp-90β (GenBank accession number S45392). 

Western Blot Analysis

For each treatment, retinas were isolated and lysed by adding 1 mL lysis buffer (50 mM β-glycerophosphate, 20 mM EGTA, 15 mM MgCl₂, 1 mM NaVO₄, 1 mM dithiothreitol (DTT), 0.1% NP-40, and 1 μg/mM of protease inhibitor cocktail (Roche Molecular Biochemicals, Indianapolis, IN). Total retina lysate was then transferred to microcentrifuge tubes and sonicated for 5 seconds. A small aliquot of the supernatant of each sample was removed for protein assay, and SDS running buffer was added to the remaining fraction. Samples were heated for 5 minutes at 95°C and then stored at −80°C. Sample protein concentrations were determined with a protein assay kit (Bio-Rad, Hercules, CA). Equivalent amounts of protein were loaded onto 12% SDS polyacrylamide gel, and the proteins were separated according to molecular weight by standard SDS-PAGE protocols and transferred to nitrocellulose filters. Levels of Hsp27, -70, or -90 were then determined by immunoblot analysis. Blots were probed with one of the following antibodies: Hsp27 (SPA-801, anti-Hsp25, 1:2000; StressGen, Victoria, British Columbia, Canada), Hsp70 (SPA-810, 1:1000; StressGen); and Hsp90 (H38220, 1:1000; Transduction Laboratories, Lexington, KY). Bands were visualized by the addition of peroxidase-conjugated anti-rabbit secondary antibody (1 μg/mL), or anti-mouse horseradish peroxidase (HP)-conjugated secondary antibodies, and enhanced chemiluminescence reagents (ECL; Amersham, Buckinghamshire, UK). Blots were then stripped by incubation in stripping buffer (62.5 mM Tris, [pH 6.7]) 100 mM β-mercaptoethanol, 2% SDS) for 30 minutes at 50°C. The level of β-actin was then determined by immunoblot analysis with anti-β-actin monoclonal antibodies and chemiluminescence reagents (Amersham). Band densities were quantified with image-analysis software (Scion) and the level of Hsp normalized for differences in loading by using the β-actin protein band intensities. Based on product literature and personal communications with companies, primary antibodies can detect the following proteins in the rat: anti-Hsp25 (StressGen) Hsp25/27; anti-Hsp70 (StressGen) Hsp70.1, -70.2, and -70.3; and anti-Hsp90 (Transduction Laboratories) Hsp90α and -90β. 

Immunohistochemistry

Rats were killed by an overdose of pentobarbital and transcardially perfused with 4% paraformaldehyde in 0.1 M PBS. Eyes were then enucleated, fixed for 1 hour in 4% paraformaldehyde in 0.1 M PBS at 4°C and bisected at the ora serrata. The anterior segment was removed, and the fixation continued for 3 hours. Eyes were then washed three times in PBS, cryopreserved by placing the tissues in a solution of 30% sucrose overnight, and placed in optimal cutting temperature (OCT) mounting medium and frozen. Parasagittal sections (10 μm) were cut, and the sections mounted on gelatin-coated slides. The monoclonal antibody G-A-5 to glial fibrillary acidic protein (GFAP; 1:400) from Sigma (St. Louis, MO) and the anti-mouse IgG conjugated to Texas red (1:50) were used to visualize retinal GFAP. The polyclonal antibody SPA-801 (1:200; Stress Gen) and the anti-rabbit IgG conjugated to FITC were used to visualize retinal Hsp27. Sections were viewed and photographed with a microscope (Axioplan; Carl Zeiss, Thornwood, NY). 

Electroretinography

To quantitate postischemic functional recovery, electroretinography (ERG) was performed in selected animals. The ERG was recorded as previously described. ⁷ In brief, Sprague-Dawley rats were dark adapted overnight, and surgical preparation and ERG recordings were performed under dim red light. Pupillary dilation was maintained with topical 0.5% tropicamide (Alcon) and 0.2% cyclopentolate and 1% phenylephrine HCl (Cyclomidril; Alcon). Platinum needle electroencephalographic electrodes (Grass-Telefactor, Inc., West Warwick, RI) were bent and placed in contact longitudinally with the corneal surfaces bilaterally, to avoid contact between the sharp needle tip and the cornea. A reference electrode was clipped on the tongue, and a grounding electrode was placed subcutaneously in the animal’s back. Corneal electrical responses to 10-μs white-light flashes were delivered by a Ganzfeld stimulator (Nicolet, Madison, WI) and recorded (Spirit 486 System; Nicolet). 

In selected sham-treated and preconditioned rats, electroretinogram b-wave data were obtained before a 60-minute period of ischemia (baseline), during 60 minutes of ischemia, and at 1, 24, 72, and 168 hours after ischemia ended. Data at each time point were collected, and the results averaged for every three flashes, delivered at least 2 minutes apart. Flashes at a single maximal intensity were used, because previous studies had indicated the accuracy and adequacy of this approach. ⁷ The ERG b-wave amplitude at each time point after ischemia was measured and reported as a percentage of the baseline, nonischemic b-wave amplitude. 

Results

As shown in Figure 1 , Northern blot analysis demonstrated that Hsp27, -70, and -90 are expressed in normal rat retinas. Twenty-four hours after preconditioning stimulus, no change in expression of Hsp70 or -90 mRNA was evident in the retinas of these animals. However, Hsp27 expression exhibited a 200% increase over basal levels. Figure 2 shows the time course for the change in mRNA for Hsp27, -70, and -90, after 5 minutes of preconditioning. Increases in Hsp27 mRNA were noted as early as 5 hours after the preconditioning event and peaked by 24 hours. This elevation in Hsp27 mRNA levels remained evident for more than 72 hours after preconditioning and began to return to control levels by 120 hours. No consistent change in Hsp27 mRNA was observed in retinas from sham-treated eyes. When compared with contralateral control retinas or retinas from sham-treated animals, no consistent changes in Hsp70 or -90 were observed during the course of the study. 

Figure 3 shows representative Western blots from retinas 24 hours after sham or ischemic preconditioning procedures. No consistent changes in Hsp70 or -90 proteins were evident in the retinas of these animals 24 hours after preconditioning. However, a 200% increase in Hsp27 protein compared with the sham-treated control was detected in these animals. Figure 4 shows the time course for the change in protein levels for Hsp27, -70, and -90, after 5 minutes of preconditioning. Increases in Hsp27 protein levels were noted by 24 hours and remained elevated for up to 72 hours after the preconditioning event. By 120 hours, Hsp27 levels began to decline, and by 168 hours Hsp27 protein levels were only slightly elevated. When compared with the sham control, no consistent change in the levels of Hsp70 or -90 protein were observed. 

Localization of Hsp27 in sham-treated and preconditioned retinas was evaluated by immunohistochemical analyses (Fig. 5) . In animals undergoing sham treatment 72 hours before isolation of the retina, staining for Hsp27 was observed only in the endothelial cells of inner retinal vessels and the choroid. In retinas that had received a preconditioning stimulus 72 hours before retinal isolation, expression of Hsp27 was observed in the ganglion cell and inner plexiform layers of the retina. Hsp27 was expressed in both GFAP-positive and -negative cells in this layer. The expression of Hsp27 in the inner plexiform layer appeared to be associated with the dendritic projections from neuronal cell bodies in the ganglion cell layer. Light staining of the cell bodies of the inner nuclear layer was also observed. 

Figure 6 shows the correlation between the increase in retinal expression of Hsp27 protein with the protection of retinal function from a second 60-minute ischemic injury after retinal preconditioning. The increase in Hsp27 and the maintenance of an ERG b-wave was not observed until 24 hours after the preconditioning event. Maximal Hsp27 protein expression and retinal protection were observed at 24 and 72 hours after preconditioning. At times beyond 72 hours, both the expression of Hsp27 and the protective action induced by preconditioning were no longer evident. 

Discussion

Retinal ischemia is thought to contribute to the etiology of several retinal diseases, including diabetic retinopathy and glaucoma. Although studies have identified many of the pathologic steps leading to ischemic retinal degeneration, no effective therapies have been developed to date. Original studies of the heart have demonstrated that brief periods of ischemia protected the tissue from subsequent ischemic injury. ¹⁰ This effect has been termed ischemic preconditioning or ischemic tolerance. Recent studies of the eye have shown that ischemic preconditioning can also be demonstrated in the retina. ⁷ Studies have shown that the activation of specific adenosine receptors and the opening of ATP-sensitive potassium channels may be involved in the initiation of this response. ^{11 12} Although this response can be inhibited by blocking protein synthesis, ⁶ there is little information regarding the changes in expression of protein that accompany this response. 

The present study was designed to evaluate whether changes in specific Hsps contribute to the development of ischemic preconditioning. Hsps are expressed in all organisms and function as molecular chaperones that regulate the synthesis, translocation, and activity of a number of proteins. ² In general, Hsps are a family of proteins ranging in molecular weight from 10 to 170 kDa. Several studies in the heart, brain, kidney, and many other tissues, have shown that Hsps are upregulated during periods of metabolic stress, including preconditioning. In the normal retina, Hsp27, -60, -70, and -90; Hsc70; and αβ-crystallin have been identified. ^{13 14 15 16 17}  

Our data confirm that Hsp27, -70, and -90 are constitutively expressed in normal retina. After a 5-minute preconditioning ischemic stimulus, no change in Hsp70 or -90 were observed. Hyperthermic upregulation of endogenous Hsp70 or the intravitreous injection of this protein has been shown to protect the retina from light damage in vivo. ^{18 19} Other studies have demonstrated that prolonged ischemia (60–90 minutes) can induce Hsp70 expression in the retina. ²⁰ Hence, the absence of any increase in expression of Hsp70 was unexpected. However, recent studies evaluating preconditioning in the brain have shown that the induction of Hsp70 proteins is not tightly correlated with the time course of ischemic preconditioning ²¹ indicating that the induction of Hsp70 may not be related to the neuroprotective action after preconditioning. Although our studies support the idea that Hsp70 and -90 are not involved in the development of retinal preconditioning, it is possible that other protein isoforms or closely related proteins (e.g., Grp75, -78) are altered by preconditioning. 

Studies investigating the normal expression of Hsp27 in the mammalian retina have produced contradictory results. Studies in rats by Strunnikova et al., ²² have indicated that Hsp27 is strongly expressed in retinal ganglion cells, photoreceptor outer segments, and retinal pigment epithelium, whereas work by Dean and Tytell ¹⁵ showed only weak expression of Hsp27 in the retinal ganglion cells, the inner plexiform layer, the inner nuclear layer, photoreceptors, and retinal vessels. In the present study, basal expression above background levels was only evident in cells of the choroidal and retinal vasculature (Fig. 5) . However, 5 hours after the preconditioning stimulus, there was a significant increase in the expression of Hsp27 mRNA, and by 24 hours a 200% increase in Hsp27 protein over basal levels was evident. Immunohistochemical studies demonstrated that this increase in Hsp27 protein was localized in the GFAP-positive and -negative cells in the retinal ganglion cell layer. Punctate staining of the inner plexiform layer and light staining of the cell bodies in the inner nuclear layer were also observed. These data are consistent with previous results in glaucomatous human eyes obtained at autopsy, showing expression of Hsp27 in the ganglion cell layer. ²³ Taken together, these results support the idea that retinal stress leads to an increased expression of Hsp27 in the inner retina, and that this enhanced expression is most pronounced in the neuronal and non-neuronal cells in the ganglion cell layer. 

Previous studies evaluating changes in expression of GFAP have shown that ischemic periods of 30 to 60 minutes can increase expression of GFAP in Müller cells. ^{24 25 26} In our studies, brief (5-minute) ischemic preconditioning did not produce any detectable increase in Müller cell expression of GFAP. Our results support the idea that large increases in Müller cell expression of GFAP are not necessary for retinal preconditioning. 

Figure 6 shows the correlation between the neuroprotective action and the changes in expression of Hsp27 after ischemic preconditioning in the retina. The strong correlation between the expression of Hsp27 and the neuroprotective action of ischemic preconditioning provide evidence that ischemic preconditioning in the retina involves an increase in expression of Hsp27. Recent studies have shown that Hsp27 can prevent apoptotic cell death by inhibiting the action of cytochrome c when released from mitochondria after cellular stress. ^{5 27} Work by Yokoyama et al., ²⁸ has shown that electroporation of Hsp27 can prevent ischemic injury. Taken together, the results in these studies support the idea that the expression of Hsp27 is an important component in the development of ischemic tolerance in the retina. In addition, recent results from this laboratory have shown that the overexpression of Hsp27 in the rat retinal ganglion cell line (RGC-5) increases cell viability after hypoxic injury (Crosson CE, unpublished data, 2001). 

In summary, these results demonstrate that retinal preconditioning in rats induces the expression of Hsp27. This increase in expression of Hsp27 was confined to the inner retinal layer. Functional studies demonstrated a strong correlation between the expression of Hsp27 and the neuroprotective action of ischemic preconditioning in the retina. Thus, an increase in expression of Hsp27 appears to be one component of the neuroprotective events induced by ischemic preconditioning in the retina. 

 Supported in part by National Eye Institute Grants EY09741 (CEC), EY10343 (SR), and EY12600 (J-XM); the South Carolina Lions; the Retina Research Foundation, Houston, Texas; and Research to Prevent Blindness.

 Submitted for publication March 7, 2002; revised August 8, 2002; accepted September 20, 2002.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Northern blot analysis of induction of Hsp mRNAs after sham (Sh) or ischemic preconditioning (PC) procedures in the retina. Retinas from contralateral (C) control eyes are shown for comparison.

Figure 2.

 

Time-dependent changes in expression of Hsp mRNA after sham or ischemic preconditioning in the retina. Data represent average band densities as a percentage of control levels (n = 2): (A) Hsp27, (B) Hsp70, and (C) Hsp90. To correct for differences in loading, individual values were normalized to 18s rRNA.

Figure 3.

 

Western blot analysis of Hsp induced by sham (Sh) or ischemic preconditioning (PC) procedures in the retina. Retinas from contralateral (C) control eyes are shown for comparison.

Figure 4.

 

Time-dependent changes in expression of Hsp after sham or ischemic preconditioning procedures in the retina. Data represent average band densities as a percentage of control levels (n = 2): (A) Hsp27, (B) Hsp70, and (C) Hsp90. To correct for differences in loading, individual values were normalized to β-actin.

Figure 5.

 

Double-label immunofluorescence of rat retina 72 hours after a sham treatment (A) or ischemic preconditioning (B). In retinas of sham-treated eyes, GFAP (red) was expressed in the astrocytes located in the ganglion cell layer and surrounding the inner retinal vessels. Expression of Hsp27 (green) was observed only in retinal vascular cells of the inner retinal vessels and the choroid. In retinas from preconditioned eyes, expression of GFAP was again noted only in the astrocytes located in the ganglion cell layer. However, Hsp27 was expressed in the astrocytes (coexpression with GFAP is yellow) and neuronal cell bodies in the ganglion cell layer. In addition, expression of Hsp27 was identified as punctuate staining throughout the inner plexiform layer. This punctuate staining was observed to be associated with dendritic projections from the neuronal cells in the retina. GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; ONL, outer nuclear layer; OS, outer segment. Scale bars, 50 μm.

Figure 6.

 

Correlation between expression of Hsp27 and protection of retinal function induced by ischemic preconditioning of the retina. All animals were subjected to an initial 5-minute period of ischemia and allowed to recover for 1, 5, 24, 72, or 168 hours during the induction period. Rats were then killed and retinal Hsp27 determined, or the animals were subjected to a second period of ischemia for 60 minutes. Rats subjected to the 60-minute ischemic insult were allowed to recover for an additional 7 days, after which retinal function was estimated by measurement of the ERG b-wave. Changes in retinal expression of Hsp27 and recovery of the ERG b-wave were plotted for the different induction times (n = 2 for Hsp27 measurements; and n = 5 for ERG measurements).

The authors thank Luanna Bartholomew for critical reading of and comments on the manuscript.