Increased temperature and protein oxidation lead to HSP72 mRNA and protein accumulation in the in vivo exercised rat heart
Abstract
Expression of myocardial heat shock protein 72 (HSP72), mediated by its transcription factor, heat shock factor 1 (HSF1), increases following exercise. However, the upstream stimuli governing exercise-induced HSF1 activation and subsequent Hsp72 gene expression in the whole animal remain unclear. Exercise-induced increases in body temperature may promote myocardial radical production, leading to protein oxidation. Conceivably, myocardial protein oxidation during exercise may serve as an important signal to promote nuclear HSF1 migration and activation of Hsp72 expression. Therefore, these experiments tested the hypothesis that prevention of exercise-induced increases in body temperature attenuates cardiac protein oxidation, diminishes HSF1 activation and decreases HSP72 expression in vivo. To test this hypothesis, in vivo exercise-induced changes in body temperature were manipulated by exercising male rats in either cold (4°C) or warm ambient conditions (22°C). Warm exercise increased both body temperature (+3°C) and myocardial protein oxidation, whereas these changes were attenuated by cold exercise. Interestingly, exercise in both conditions did not significantly increase myocardial nuclear localized phosphorylated HSF1. Nonetheless, warm exercise elevated left-ventricular HSP72 mRNA by ninefold and increased myocardial HSP72 protein levels by threefold compared with cold-exercised animals. Collectively, these data indicate that elevated body temperature and myocardial protein oxidation promoted exercise-induced cardiac HSP72 mRNA expression and protein accumulation following in vivo exercise. However, these results suggest that exercise-induced myocardial HSP72 protein accumulation is not a result of nuclear-localized, phosphorylated HSF1, indicating that other transcriptional or post-transcriptional regulatory mechanisms are involved in exercise-induced HSP72 expression.
Cellular adaptation to physiological disturbances is a fundamental requirement for survival. Although the molecular events involved in cellular adaptation to stress continue to be investigated, it is well established that the cellular response to homeostatic perturbations includes a highly ordered set of events that involve rapid changes in gene expression and the synthesis of several heat shock proteins (HSPs) that participate in the adaptation to stress (Morimoto et al. 1996; Pirkkala et al. 2001). Within the HSP70 family of HSPs, the inducible 72 kDa protein (HSP72) is considered to be an important protective molecule.
Numerous studies reveal that cellular upregulation of HSP72 occurs in both heart and skeletal muscle following exposure to a variety of stresses (e.g. heat, oxidative stress and exercise) and that increased cellular levels of HSP72 protect cells from subsequent cellular insults (Latchman, 2001; Morimoto, 2002; Melling et al. 2008). The basis for this protection is mediated by several functional properties of HSP72, including an active participation in the folding of proteins by minimizing incorrect interactions within and between molecules, in maintenance of proteins in their native folded state and in the repair of damaged proteins (Agashe & Hartl, 2000).
There are several mechanisms whereby exercise can induce HSP72 expression in cardiac myocytes and all are linked to disturbances in cellular homeostasis and protein damage, including an increase in tissue temperature and elevated oxidative stress (Khassaf et al. 2001; Morton et al. 2006; Brown et al. 2007). At present, the specific stimuli associated with exercise that promote induction of HSP72 remain unknown. Nonetheless, it is plausible that the combination of elevated cellular temperature and oxidant production could interact to stimulate HSP72 expression during and following exercise (Atalay et al. 2004; Moran et al. 2004). To date, few studies have investigated the combined effect of heat and oxidative stress on the transcriptional activating factor, heat shock factor 1 (HSF1) in the whole animal (Paroo et al. 2002; Watkins et al. 2007). Data from whole animal exercise studies will provide insight in elucidating the mechanisms by which these exercise-related stimuli result in HSP72 upregulation in vivo.
The induction of HSP72 expression in cells is mediated through the interaction of HSF1 with the proximal promoter (heat shock element, HSE) on the Hsp70 gene (Morimoto, 1998; Melling et al. 2007). During unstressed conditions, HSF1 remains a latent monomeric protein in the cytoplasm. Transcriptional activation of HSF1 is a complex multi-step process that involves an oligomerization from inert monomer to an active trimer, acquisition of DNA binding ability, phosphorylation and nuclear localization (Morimoto et al. 1997; Melling et al. 2007). While exercise in warm environments increases body temperature and cardiac HSP72 expression, previous studies reveal that exercise in the cold attenuates the exercise-induced rise in core temperature and prevents the upregulation of myocardial HSP72 expression by an unknown mechanism (Taylor et al. 1999; Hamilton et al. 2001; Quindry et al. 2007). These data suggest that cellular temperature could be a critical variable to mitigate transcriptional competency in myocardial HSP72 expression during exercise.
Recently, we completed a series of in vitro experiments to isolate the impact of both cellular temperature and metabolic stress on HSF1 activation in the isolated working rat heart. Our results indicate that hyperthermia, independent of cardiac workload, promoted an increase in nuclear translocation and phosphorylation of HSF1 in the heart (Staib et al. 2007). Similarly, hyperthermia, independent of workload, resulted in significant increases in cardiac levels of HSP72 mRNA (Staib et al. 2007). Collectively, these data suggest that phosphorylation of nuclear-localized HSF1 and mRNA expression following exercise are linked to elevated heart temperature and are not a direct function of an increase in cardiac metabolic workload. Unfortunately, owing to the short duration of these in vitro experiments, the direct impact of increased heart temperature on HSP72 protein synthesis could not be ascertained. Moreover, the integrated response to aerobic exercise in vivo is more complicated than the conditions posed in the isolated working heart model. Therefore, the present experiments were designed to overcome the limitations of our previous in vitro experiments by investigating the influence of exercise-induced increases in body temperature on myocardial Hsp72 gene expression following exercise in vivo.
The diminished levels of cardiac HSP72 protein observed following exercise in a cold environment in previous experiments could be a result of the attenuation of HSF1 phosphorylation/nuclear translocation, decreased HSP72 mRNA levels, diminished rates of translation or a combination of these factors (Taylor et al. 1999; Hamilton et al. 2001; Quindry et al. 2007). Accordingly, the present study is the first investigation to delineate the in vivo molecular link between exercise-induced increases in body temperature, protein oxidation and the expression of myocardial HSP72. We hypothesized that exercise in a cold environment prevents increases in body temperature and oxidative modification of cardiac proteins. Moreover, we also predicted that the prevention of exercise-induced increases in body temperature and myocardial protein oxidation would be associated with a decrease in HSF1 phosphorylation and migration into the nucleus, resulting in decreased HSP72 expression.
Methods
Animals and experimental design
The University of Florida Animal Care and Use Committee and the Veterans Administration (VA) Animal Care and Use Committee approved all experimental procedures, and the specific guidelines outlined by both the University and VA were followed throughout these experiments. Animals (6 months old) were housed in pairs with access to food and water ad libitum in a room with a 12 h–12 h reverse light–dark cycle maintained throughout the experimental period.
In vivo exercise training
Twenty-eight male Sprague–Dawley rats were randomly selected and assigned to one of three experimental groups, as follows: (1) sedentary control animals (S; n= 10); (2) in vivo exercise trained in a warm environment (22°C; WT; n= 9); and (3) in vivo exercise trained in a cold environment (4°C; CT; n= 9). The objective behind exercising animals in the cold environment was not to expose the animals to stress but rather to improve heat loss and therefore prevent exercise-induced heat gain associated with exercise.
In vivo exercise protocol Animals were habituated to the exercise protocol over the course of 5 days, and then exercised for 60 min day−1 for three consecutive days. The habituation period began with 10 min of running at 30 m min−1, 0% gradient, with 10 min increases in running time until 50 min of running were reached. Following this habituation period, the animals rested for 2 days before beginning the exercise-training programme. The exercise period consisted of three consecutive days of treadmill running (60 min day−1) at 30 m min−1, or approximately at 60–70% of maximal oxygen uptake. This protocol has been shown to induce myocardial HSP72 expression in adult rats (Demirel et al. 2001; Hamilton et al. 2001; Quindry et al. 2007). To standardize handling stress, sedentary control animals were placed directly on a non-moving treadmill for the duration of the 60 min period.
Tissue removal and preparation Within minutes of stress induction, HSP72 mRNA transcripts are present in cells, and peak mRNA levels generally appear within 60 min after the termination of stress and decline during the 3 h period following the stress; protein accumulation reaches its maximum at ∼12 h (Iwaki et al. 1993; Nishizawa et al. 1996; Qian et al. 1998). Specifically, exercise-induced increases in myocardial HSP72 mRNA levels peak at ∼60 min postexercise (Demirel et al. 2001). Hence, based upon these data, animals in the present experiment were killed 60 min postexercise training to capture both HSP72 mRNA peak and HSP72 protein accumulation.
Animals were anaesthetized using 5% isoflurane; once a surgical plane of anaesthesia was reached, hearts were quickly removed, rinsed free of blood and dissected, sectioned into left ventricle (LV), right ventricle (RV) and septum (S) and rapidly frozen in liquid nitrogen prior to storage at −80°C until assay. Samples delineated for RNA isolation and reverse transcriptase-polymerase chain reaction (RT-PCR) were placed in RNAlater (Ambion, Austin, TX, USA) and then stored at −80°C.
Biochemical measures: protein oxidation
Oxidative damage to myocardial protein was determined by measurement of protein carbonyls. Protein carbonyls were assayed using an enzyme immunoassay (Zentech PC Test, Zenith Technology Corp., Dunedin, New Zealand). Samples from the left ventricle (LV) were homogenized in cold 100 mm phosphate buffer (pH 7.4) plus protease inhibitors (2 mm 4-(2-aminoethyl)benzenesulphonyl fluoride hydrochloride, 1 mm EDTA, 130 μm bestatin, 14 μm E-64, 1 μm leupeptin, 0.3 μm aprotinin; Sigma-Aldrich, St Louis, MO, USA) in a 3 ml glass homogenization tube (Sigma-Aldrich). Briefly, all samples, standards and quality controls were normalized to 1.8 mg ml−1. Next, 11 μl of each sample, standard and control solution was incubated in 19 μl of dinitrophenylhydrazine (DNPH) for 45 min at room temperature. Following derivitization with DNPH, 7.5 μl of each sample, standard and control solution was diluted with 1 ml of enzyme-linked immunoassay (EIA) buffer provided with the kit. The manufacturer's instructions were then followed, starting with the section entitled ‘ELISA procedure #3’. All samples, standards and control solutions were run in triplicate.
Biochemical measures: HSF1 nuclear translocation and phosphorylation
Myocardial nuclear and cytosolic fractionation Analysis of HSF1 cellular localization and phosphorylation status was performed as described previously (Yamanaka et al. 2003; Melling et al. 2004, 2006; Shinohara et al. 2004; Staib et al. 2007). Briefly, nuclear and cytosolic extracts were prepared from LVs using NE-PER nuclear and cytoplasmic extraction kit (Pierce, Rockford, IL, USA) with the addition of protease inhibitors (Sigma-Aldrich) and phosphatase inhibitors (Sigma-Aldrich) as directed by the manufacturer. A section of LV was mixed and homogenized on ice in 100 mm phosphate buffer containing 0.05% bovine serum albumin and EDTA (1:10 w/v; pH 7.4) using a glass-on-glass tissue homogenizer. Homogenates were centrifuged at 500g for 4 min at 4°C. The supernatant was removed and incubated for 10 min on ice in 200 μl cytoplasmic extraction reagent (CERI) solution containing 0.75 mm phenylmethylsulphonyl fluoride (PMSF), 2.0 mg ml−1 aprotinin and leupeptin, 20 mm NaF and 2.0 mm Na3VO4. CERII solution (11 μl) was then added, and cytoplasmic extracts were collected by centrifugation at 12 000g for 5 min at 4°C. Nuclear pellets were lysed in 100 μl of nuclear extraction reagent (NER) solution containing 2 mm PMSF, 2.0 mg ml−1 aprotinin and leupeptin. Protein concentrations in the nuclear and cytosolic fractions were determined by methods described by Bradford (Bradford, 1976).
Sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western immunoblotting analysis for HSF1 Nuclear and cytosolic extracts were suspended in Laemmeli sample buffer, denatured, loaded (100 μg protein) and separated on 7.5% polyacrylamide gels along with heat-shocked HeLa cell lysate (LYC-HL101F; Stressgen Biotechnologies, Victoria, BC, Canada) and a molecular weight standard (BioRad, Hercules, CA, USA). Selected samples were treated with 80 U of calf intestine alkaline phosphatase at 37°C for 60 min before electrophoresis (Sigma-Aldrich). The separated proteins were then transferred to nitrocellulose membranes and blocked with 5% non-fat milk in Tris-buffered saline with 0.01% Tween-20 (TBST) for 2 h. Next, membranes were incubated in anti-HSF1 monoclonal primary antibody (SPA-950) overnight at 4°C (Stressgen Biotechnologies). Following incubation in the primary antibody, membranes were washed three times for 10 min each and incubated with a secondary antibody conjugated to horseradish peroxidase for enhanced chemiluminescent detection (Amersham, Piscataway, NJ, USA). Quantification of the bands from the immunoblots was performed using computerized densitometry (KODAK Image Station 4000MM Digital Imaging System, Rochester, NY, USA). To ensure equal transfer and loading of proteins, the Ponceau staining method was used (Sigma-Aldrich), and three non-specific protein bands on the membrane were chosen for densitometric analysis. Given that no differences in protein loading were observed, bands were not normalized to Ponceau-stained bands.
Biochemical measures: isolation of HSP72 RNA and real-time RT-PCR
Total RNA was extracted from the LV as described previously (Staib et al. 2007). The concentration and purity of the extracted RNA were measured spectrophotometrically at absorbances of 260 and 280 nm in 1× Tris-EDTA (TE) buffer (Promega, Madison, WI, USA). The integrity of the extracted total RNA was verified by gel electrophoresis on a 1% agarose ethidium bromide-stained Tris base, boric acid, EDTA tetrasodium salt (TBE) gel. Purified RNA was then stored at −80°C until later assay.
Reverse transcription (RT) was performed using the SuperScript III First-Strand Synthesis System for RT-PCR according to the manufacturer's instructions (Life Technologies, Carlsbad, CA, USA). Reactions were carried out using 5 μg of total RNA and 2.5 μm oligo(dT)20 primers. First strand complementary DNA (cDNA) was subsequently treated with two units of RNase H and stored at −80°C.
Following the addition of 2 μl of GlycoBlue co-precipitant, first strand cDNA was cleaned of RNA and unincorporated nucleotides by treating the cDNA with an RNase cocktail, bringing the sample to a volume of 100 μl with water and applying the sample to a NucAway spin column (Ambion, Austin, TX, USA). Samples were then mixed with phenol:choloroform:I AA (pH 7.9) and the aqueous phase recovered using a 1.5 ml heavy phase lock gel (Eppendorf, Hamburg, Germany). Finally, the cDNA was then precipitated by adding one volume of 5.0 m NH2OAc and two volumes of 100% ethanol and stored at −20°C overnight. Following centrifugation at 13 000g for 20 min (4°C), the cDNA was washed with two volumes of 75% ethanol, centrifuged at 13 000g for 10 min (4°C) and resuspended in 50 μl of 1× TE buffer. The cDNA was subsequently quantified using the Oligreen single-stranded DNA Quantitation Reagent and Kit according to the manufacturer's instructions (Molecular Probes, Eugene, OR, USA).
Primers and probes for HSP72 (assay no. Rn00583013_s1) were obtained from the ABI Assays-on-Demand service and consisted of Taqman 5′ labelled 6-carboxy-fluorescein (FAM) reporters and 3′ non-fluorescent quenchers. Primer and probe sequences from this service are proprietary and, therefore, are not reported. However, the nucleotide sequence surrounding the probe for HSP72 is available and consists of the following: 5′-GAGGAGTTCGTGCACAAGCGGGAGG-3′. Primer and probe sequences, also consisting of Taqman 5′ labelled FAM reporters and 3′ non-fluorescent quenchers, for hypoxanthine guanine phosphoribosyl transferase (HPRT) were obtained from Applied Biosystems (Assays-by-Design) and are: forward, 5′-GTTGGATACAGGCCAGACTTTGT-3′; reverse, 5′-AGTCAAGGGCATATCCAACAACAA-3′; and probe, 5′-ACTTGTCTGGAATTTCA-3′.
Quantitative real-time PCR was performed using the ABI Prism 7700 Sequence Detection System (ABI, Foster City, CA, USA). Each 25 μl PCR reaction, performed in duplicate, contained 1 ng of cDNA template as determined by cDNA quantification described in earlier paragraphs. Gene expression was calculated using the relative standard curve method as described in the ABI User Bulletin #2. Briefly, amplification of the endogenous control was performed to standardize the amount of HPRT added to the reaction. For all experimental samples, the target quantity was determined from the standard curve and then divided by the target quantity of the calibrator. Specifically, the calibrator (i.e. S) becomes the 1× sample, and all other quantities are expressed as an n-fold difference relative to the calibrator. In these experiments, HPRT was selected to normalize the mRNA because the expression of this gene in the LV is not significantly altered by exercise in our laboratory (P > 0.05).
Biochemical measures: HSP72 protein content
Left ventricular HSP72 protein was measured using a commercially available ELISA kit (EKS-700; Stressgen Biotechnologies) specific for recombinant inducible HSP72. Samples from LV delineated for determination of HSP72 protein accumulation were homogenized in the 1× HSP70 extraction reagent provided with the kit plus protease inhibitor cocktail added from Sigma. Total LV protein concentration was determined using the Bradford assay; then samples and standards were assayed in triplicate according to the manufacturer's instructions (Bradford, 1976).
Statistical analysis
Comparisons among experimental groups for each dependent variable were made by one-way factorial ANOVA (group X dependent variable) within each experiment using SPSS version 12.0 (SPSS Inc., Chicago, IL, USA). When significant main effects were observed, Tukey's post hoc analyses were implemented. Significance was established a priori at P < 0.05.
Results
Animal morphological characteristics
Adult male Sprague–Dawley rats (6 months old, body weight 372–433 g) were randomly assigned to one of the three experimental groups. No significant differences existed in animal heart weights (1.05–1.57 g) or the heart weight/body weight ratio (2.82–3.63 mg g−1) among experimental groups.
In vivo short-term exercise: postexercise colonic temperature
The objective of these experiments was to determine the influence of in vivo exercise-induced increases in core temperature on myocardial HSP72 expression. To prevent the rise in core temperature that occurs during 60 min of treadmill exercise, one experimental group exercised for 60 min in a cold room (CT; 4°C) while their warm-trained (WT) counterparts trained in ambient temperatures (22°C). Compared with animals exercised in the cold environment (CT), postexercise core temperatures were significantly higher in the animals exercised in the warm environment (WT; postexercise core temperatures, means ±sem: CT, 38.28 ± 0.11°C; WT, 40.18 ± 0.22°C; P < 0.001). Specifically, all CT animals exhibited less than a 0.6°C elevation in core temperature following the 60 min exercise bout, whereas animals exercised in the warm environment experienced more than 3.2°C increase in body temperature during the 60 min exercise bout. Hence, exercise in the cold environment did not represent cold stress per se, but rather the cold environment improved heat loss and prevented exercise-induced increases in body temperature.
Biochemical measures
Protein oxidation Protein carbonyls, a commonly used marker of protein oxidation, were measured in the LV from animals in both sets of experimental conditions using an enzyme-linked immunosorbent assay (ELISA). Group comparisons revealed that in vivo exercise at room temperature (WT) resulted in significant increases in protein carbonyl levels in the LV versus both the cold exercise (CT) and sedentary control group (S; Fig. 1).
Protein carbonyl content in the left ventricle from in vivo exercise groups as determined by ELISA Values are means +s.e.m. Abbreviations: S, sedentary control; CT, cold trained; and WT, warm trained. The value for WT is significantly higher (*P < 0.05) than for both CT and S.
HSF1 nuclear translocation and phosphorylation To determine the subcellular location and phosphorylation status of HSF1 following exercise, an immunoblot specific to HSF1 was performed. To confirm that HSF1 phosphorylation following exercise occurred, 80 U of alkaline phosphatase was added to nuclear and cytoplasmic sample extracts from warm-exercised animals (WT) and run on a gel containing a positive control (heat-shocked HeLa cell lysates; Stressgen Biotechnologies). The addition of alkaline phosphatase was done to remove phosphate from HSF1. The disappearance of the phosphorylated HSF1 band with alkaline phosphatase treatment confirmed that HSF1 was indeed phosphorylated following the exercise protocol (Fig. 3A).
Exercise-induced nuclear localization and phosphorylation of HSF1 following in vivo exercise in cold (CT) and warm environments (WT) determined by SDS-PAGE and Western immunoblot Heat-shocked HeLa cells, tissue extract from an exercised animal (WT) and alkaline phosphatase (AP) inhibition were used to confirm HSF1 phophorylation postexercise. Note that HSF1 had a similar molecular weight in HeLa cells to that in WT and that these phosphorylated forms of HSF1 were lost with AP treatment. A, a representative blot demonstrating the phosphorylation of HSF1 in the nuclear fraction. B, analysis of phosphorylated nuclear HSF1 expressed as percentage of the total. Values are means +s.e.m. The value for CT is significantly higher (*P < 0.05) than for S.
In vivo exercise did not alter the total nuclear translocation of HSF1 in the heart (Fig. 2; P= 0.64). However, compared with the sedentary group (S), exercise in the cold (CT) resulted in a significant increase in phosphorylated HSF1 within the nuclear compartment (Figs 2B and 3; P < 0.05). Note, however, that no differences existed in nuclear levels of phosphorylated HSF1 between the two exercise groups (Fig. 2BP= 0.40; Fig. 3P= 0.80). Therefore, this observation suggests that the phosphorylation of HSF1 in the nuclear fraction was not proportional to the transcription rates of Hsp72 following in vivo exercise. Finally, no differences (P= 0.46) existed in total cytosolic HSF1 among the experimental groups (Fig. 4).
Exercise-induced nuclear localization and phosphorylation of HSF1 following in vivo exercise in cold (CT) and warm environments (WT) determined by SDS-PAGE and Western immunoblot Analysis of total (A) and phosphorylated nuclear HSF1 from the left ventricle (B). Values are means +s.e.m. The value for CT is significantly higher (*P < 0.05) than for S.
Exercise-induced cytosolic localization of HSF1 following in vivo exercise in cold (CT) and warm environments (WT) determined by SDS-PAGE and Western immunoblot Analysis of cytosolic HSF1 content in the left ventricle from in vivo exercise groups. Values are means +s.e.m. No significant differences were observed among experimental groups.
HSP72 mRNA expression As depicted in Fig. 5, temperature had a powerful effect on the upregulation of myocardial HSP72 mRNA levels. Strikingly, myocardial HSP72 mRNA levels in WT animals were ∼18-fold higher than in S animals and ∼ninefold higher than in CT animals (P < 0.001). Figure 5 also reveals that the upregulation of LV HSP72 mRNA observed in WT was attenuated in CT animals, which were not different from S (P= 0.932).
Real-time quantitative RT-PCR analysis of HSP72 mRNA expression, normalized to HPRT in the left ventricle from in vivo exercise groups Values are means +s.e.m. The value for WT is significantly higher (*P < 0.05) than for both CT and S.
HSP72 protein content Following a similar pattern to mRNA expression, myocardial HSP72 protein accumulation was significantly elevated in WT animals compared with all other experimental groups (Fig. 6). Specifically, following the in vivo exercise training protocol, WT animals exhibited a fourfold increase in cardiac HSP72 protein accumulation above hearts from the S animals and a threefold increase compared with CT animals (P < 0.001). Finally, myocardial HSP72 protein levels did not differ between the S and CT animals (P= 0.084).
HSP72 protein levels in the left ventricle from in vivo exercise groups as determined by ELISA Values are means +s.e.m. The value for WT is significantly higher (*P < 0.05) than for both CT and S.
Discussion
Overview of principal findings
The results of this study provide new insight into the regulation of HSP72 expression in the heart and reveal that elevated body temperature is required for Hsp72 gene expression following in vivo exercise. Further, our results indicate that the exercise-induced increase in body temperature may promote oxidative modification of myocardial proteins. Our data reveal that the acquisition of transcriptional competency for Hsp72 expression in the heart following in vivo exercise is not exclusively due to phosphorylation of nuclear localized HSF1. Indeed, exercise-induced increases in body temperature promote large increases in cardiac HSP72 mRNA levels without a substantial increase in phosphorylated and nuclear localized HSF1. A detailed discussion of these findings follows.
Exercise-mediated temperature elevation and protein oxidation are associated with increased myocardial HSP72 expression
The synthesis of HSP72 is primarily regulated by its transcription factor, HSF1. Evidence indicates that HSF1 activation involves a complex series of regulatory events, including nuclear localization, oligomerization and acquisition of HSE–DNA binding, ultimately resulting in the transcription of Hsp72 (Sarge et al. 1993). However, the cellular signalling component of exercise (i.e. thermal and oxidative stress) responsible for increasing the expression of myocardial Hsp72 remains elusive.
Previously, we and others have shown that exercise in the cold prevents HSP72 protein accumulation, which led us to hypothesize that the increased heart temperature following in vivo exercise is the critical proximal stimulus to initiate Hsp72 transcription and translation (Taylor et al. 1999; Hamilton et al. 2001; Harris & Starnes, 2001; Quindry et al. 2007). In addition to temperature increases, we postulated that the myocardial protein oxidation, secondary to increased core temperature-induced oxidant production, provides an upstream stimulus for HSF1 activation. Our working hypothesis evolved from several lines of evidence, including reports indicating that HSF1 can be activated separately by thermal and oxidative modification of cellular proteins.
The postulate that temperature is the critical upstream HSF1 activator is supported by data indicating that HSF1 can directly and indirectly sense temperature. For example, several studies have shown that HSF1 is directly activated by temperature, since purified HSF1 can bind to the HSE when subjected to heat (Goodson & Sarge, 1995; Farkas et al. 1998; Zhong et al. 1998). Indeed, in the present study, we observed that a 3°C rise in body (and presumably heart) temperature during exercise was followed by increased Hsp72 expression, consistent with the notion that cellular temperature elevation is a critical upstream stimulus that mediates exercise-induced Hsp72 expression.
In contrast, others have argued that elevated temperature is not the primary activator of Hsp72 expression and that the cellular environment is the primary determinant of HSF1 activation (Abravaya et al. 1991; Zhong et al. 1998). These studies report that the temperature at which HSF1 is activated is variable and not an absolute temperature, thereby arguing against HSF1 as a direct sensor of heat stress (Abravaya et al. 1991; Zhong et al. 1998). As such, the possibility exists that heat-induced changes in the cellular environment can directly or indirectly activate HSF1. For example, exposure to heat induces various physiological changes in the organism, including free radical production, the release of stress hormones, changes in tissue oxygenation, thermal unfolding of proteins, mechanical disruption of molecules and an increase in proteolytic activity (Paroo et al. 2002; Ruell et al. 2004; Ogura et al. 2008). These data suggest that the intramolecular mechanism by which HSF1 directly or indirectly senses high temperatures is related to a conformational change in HSF1 such that the monomeric form creates the HSF1 trimer (Paroo et al. 2002; Ruell et al. 2004; Ogura et al. 2008). Therefore, when combined with hyperthermia, oxidative stress within the cellular environment may promote oxidative damage to proteins to initiate HSF1 activation postexercise (Zhong et al. 1998; Ahn & Thiele, 2003; Atalay et al. 2004; Moran et al. 2004). Further, the combination of multiple inducible factors during in vivo exercise may promote a more potent stimulus than any single factor creates alone (Milne & Noble, 2002; Ruell et al. 2004; Brown et al. 2007; Watkins et al. 2007).
Although numerous studies report that heat and oxidative stress directly result in the synthesis of HSP72 in cells in vitro, few studies have investigated the direct effects of heat and oxidative stress on HSF1 activation in the whole animal, as in the present investigation (McDuffee et al. 1997; Zou et al. 1998; Paroo et al. 2002). In the present study, we observed a significant increase in myocardial protein oxidation and Hsp72 expression in the hearts of animals exercised in the warm environment. In contrast, animals exercised in the cold did not exhibit increased cardiac protein oxidation or Hsp72 gene expression. In this regard, hyperthermia may exacerbate oxidative stress in the heart during exercise by increasing oxidant production or by inhibiting the antioxidant defense system (Paroo et al. 2002; Ahn & Thiele, 2003; Atalay et al. 2004; Moran et al. 2004). Our results show that increased temperature and changes in redox status in the heart are associated with increased Hsp72 expression in response to in vivo exercise. Perhaps the combination of high temperature and radical production during in vivo exercise provides a unique sequence of events that lead to Hsp72 expression and HSP72 protein accumulation (Paroo et al. 2002; Ahn & Thiele, 2003; Atalay et al. 2004; Moran et al. 2004).
Temperature-related acquisition of transcriptional competence is not exclusively due to nuclear translocation and phosphorylation of HSF1
Although it is clear that HSP72 synthesis is transcriptionally regulated by HSF1, the mechanisms that relay the various intracellular signals produced during exercise to mediate Hsp72 transcription and translation are unclear. Activation of HSF1 involves a complex, dynamic series of regulatory events, including oligiomerization, nuclear localization, hyperphosphorylation and the acquisition of HSE–DNA binding capacity (Pirkkala et al. 2001; Calderwood, 2005; Guettouche et al. 2005; Melling et al. 2007). However, it appears that HSF1 binding to the HSE is only the first of at least two steps involved in achieving transcriptional competency in a stressed cell. For example, when HSF1 is overexpressed from transfected genes, HSF1 accumulates as DNA-binding trimers; however, these trimers possess only minimal transcription-enhancing activity (Jurivich et al. 1992; Voellmy, 2004). Similarly, compounds such as salicylate, indomethacin, menadione and H2O2 that were only capable of triggering the first step of HSF1 activation fail to prompt HSF1 hyperphosphorylation (Cotto et al. 1996; Xia & Voellmy, 1997). For full transcriptional activation to occur following HSE–DNA binding, evidence suggests that HSF1 trimers undergo an additional, required, stress-induced step such that hyperphosphorylation on specific serine residues occurs to render trimerized HSF1 transcriptionally competent (Jurivich et al. 1992; Xia & Voellmy, 1997; Melling et al. 2004; Calderwood, 2005; Guettouche et al. 2005; Melling et al. 2006).
Indeed, our data reveal that the diminished in vivo exercise stress response observed in normothermic hearts could be due to the failure to achieve transcriptional competence. Note that the myocardial level of nuclear-localized phosphorylated HSF1 in vivo did not differ between animals exercised in warm and cold conditions as they did in our previous in vitro experiments (Staib et al. 2007). These novel findings following in vivo exercise agree with the findings of Melling et al. (2004, 2006), who reported that the phosphorylated form of HSF1 in the nuclear fraction did not increase following in vivo exercise. Importantly, this phosphorylation was also evident in the transcriptionally inert HSF1 (Melling et al. 2004, 2006). Collectively, these data indicate that phosphorylation may regulate the HSF1 DNA-binding activity, but perhaps hyperphosphorylation is required to fully activate heat shock element-driven transcription (Sarge et al. 1993; Cotto et al. 1996; Holmberg et al. 1998; Melling et al. 2004). These studies, combined with the present data, suggest that events downstream of HSE–DNA binding acquisition (i.e. hyperphosphorylation and/or post-transcriptional modification) regulate HSF1 transcriptional competency (Jurivich et al. 1992; Xia & Voellmy, 1997; Melling et al. 2004, 2006, 2008; Calderwood, 2005; Guettouche et al. 2005).
Increased body temperature and protein oxidation promote HSP72 mRNA and protein expression following exercise
Although it is generally believed that the regulation of Hsp72 gene expression is under transcriptional control, other mechanisms may also be involved in the exercise-mediated stimulation of HSP72 mRNA transcription and protein synthesis. Results from the present study indicate that only the warm-trained group exhibited marked increases in HSP72 mRNA and protein levels. However, these events occurred without precursory elevated nuclear translocation and hyperphosphorylation of HSF1. That is, nuclear translocation and phosphorylation of HSF1 was not proportional to the transcriptionally activated HSP72 following in vivo exercise in the warm group compared with cold-exercised animals and sedentary control animals.
The present data corroborate the finding that animals exercising in the cold fail to accumulate HSP72 protein following exercise (Taylor et al. 1999; Hamilton et al. 2001; Harris & Starnes, 2001; Quindry et al. 2007). Collectively, these data argue that differences in HSP72 protein production following cellular changes such as heat or oxidative stress are related to events downstream from promoter activation and may include regulation of mRNA stability and important post-transcriptional regulatory mechanisms that will need to be addressed in future studies (Moseley et al. 1993; Mazroui et al. 2007). Since multiple complex factors are involved with exercise and the mechanisms contributing to HSF1 activation are not well defined, other factors associated with exercise may be involved in initiating transcription and translation of Hsp72 in vivo other than increases in core temperature and protein oxidation (Brown et al. 2007). Therefore, in vivo exercise may pose a unique combination of stimuli which may have a differential effect on RNA stability, translocation and protein stability.
Appendix
Acknowledgements
This work was supported by NIH R01 HL072789 awarded to S.K.P., AHA 0215081B awarded to J.L.S. and the Medical Research Service of the Department of Veterans Affairs.
