Introduction

The heat shock proteins are a family of highly conserved stress proteins present in the cells of all living organisms. The 72-kDa family member (Hsp72) is the most heat sensitive and highly inducible heat shock protein. Hsp72 is recognized by its primary intracellular function to prevent protein denaturation and to enhance protein recovery in response to hyperthermic stress (Kregel 2002; Gething and Sambrook 1992). Besides this intracellular function, Hsp72 is secreted from both healthy (Lancaster and Febbraio 2005) and necrotic cells (Basu et al. 2000), and can exert an extracellular function, acting as a danger signal or modulator of the immune system (Asea et al. 2000; Campisi et al. 2003; Udono and Srivastava 1993). It seems that the function of extracellular Hsp72 depends on its localization. Surface-bound Hsp72 activates natural killer cells, while extracellular Hsp72 binds to toll-like receptors on antigen-presenting cells and trigger the activation of the immune system to stimulate cytokine and chemokine release (Asea 2008). Asea et al. (2000) for example, demonstrated that extracellular Hsp72 can bind to the surface of human monocytes, resulting in rapid intracellular calcium influx, activation of nuclear factor kappa B, and up-regulation of pro-inflammatory cytokines, including interleukin (IL)-1β, IL-6 and tumor necrosis factor alpha.

In humans, a baseline concentration of extracellular Hsp72 (eHsp72, Hsp72 in plasma/serum) has been detected at low levels in the peripheral circulation of healthy subjects under non-stressed conditions (Pockley et al. 1998). Higher levels of Hsp72 have been reported in the peripheral circulation of severely traumatized patients (Pittet et al. 2002) and children with septic shock (Wheeler et al. 2005). Recently, Ruell et al. (2006) measured higher eHsp72 concentration in runners with serious symptoms of heat illness when compared to asymptomatic runners after a 14-km run. The authors concluded that an increase in eHsp72 concentration in hyperthermic subjects after a long distance run is a useful indicator to distinguish patients who need treatment for heat illness.

A key aspect of heat injury appears to be the rate of heat storage (Flanagan et al. 1995; Herman et al. 1981). In both isolated cell and animal studies, it has been shown that a high rate of heat storage (or heating rate) markedly reduces the chance of survival or increases the degree of thermal injury, independent of final peak core temperature (Flanagan et al. 1995; Burns et al. 1986; Herman et al. 1981). In this regard, we have previously used intracellular Hsp72 (iHsp72) accumulation as a tissue specific biomarker of cellular stress in an animal model and demonstrated that both tissue injury and intracellular Hsp72 (iHsp72) accumulation correlated better with rate of heat storage rather than with maximum core temperature, thermal load, or duration of heat exposure (Flanagan et al. 1995). Thus, the rate of heat storage may be an important consideration in the development of exertional heat stroke.

Although cell and animal studies have shown the importance of the rate of heat storage on heat stress, little information is available in humans. In several occupational scenarios, workers such as firefighters, or military personnel are required to produce a large metabolic heat load in uncompensable heat conditions, resulting in a high rate of heat storage and increased risk of heat illness (Porter 2000). Given our previous results in animals showing a relationship between the rate of heat storage to intracellular Hsp72 accumulation, and given the potential role for eHsp72 in activating the same inflammatory mediators at work in heat injury, we sought to examine the role of heat storage within the physiological ranges seen in exercising humans on eHsp72 appearance in the blood. Therefore, the aim of this study was to compare the effect of two different rates of heat storage on eHsp72 concentration during exercise in which subjects reach the same final core temperature. We hypothesized that with a slower rate of heat storage, the increase in eHsp72 would be less than that seen/observed with a higher rate of heat storage.

Materials and methods

Subjects

Nine (seven males and two females), non-smoking, healthy subjects (mean ± sd, age: 25 ± 3 years, body weight: 75.2 ± 5.3 kg, peak oxygen uptake 54.8 ± 10.1 ml kg−1 min−1, body fat 15 ± 8%) participated in this study. All subjects had less than two positive cardiovascular risk factors and a percent of body fat lower than 15% for men and 30% for women. They also had an above-average aerobic fitness level (peak oxygen uptake >35 ml kg−1 min−1 for women and >40 ml kg−1 min−1 for men) (ACSM 2000). Subjects were excluded from the study if they had a history of heat illness, known diseases or viral infections, or were taking vitamin supplements or other medications that would affect eHsp72, exercise, or thermoregulatory responses. The female subject was taking a monophasic oral contraceptive and all trials were conducted within the 3 weeks of stable hormone intake. The host institution’s human subjects research committee approved the study. Prior to participation, subjects were informed regarding the possible risks and discomforts involved, and their written consent was obtained.

Peak oxygen uptake

Each subject performed a continuous graded treadmill test (Model 966I; Precor, Woodinville, WA, USA) to determine their peak oxygen uptake (VO2 peak). The protocol consisted of 1-min stages with an initial increase in speed of 1.6 km h−1 (1 mph) every minute until a maximal tolerable speed, and then grade was increased in increments of 2%. VO2 peak was defined as the highest 30-s average using breath-by-breath. Equipment calibrations were conducted immediately before each test. The exercise test was carried out in a temperate room (22–24°C dry bulb temperature, 30–40% relative humidity). Breath-by-breath VO2 was measured with a fast response turbine flow transducer (Model S–430; K.L. Engineering, Van Nuys, CA, USA) and custom developed software (LabVIEW; National Instruments, Austin, TX, USA) with O2 and CO2 electronic gas analyzers (Model S-3A and Model CD-3; AEI Technologies, Pittsburgh, PA, USA).

Heat acclimation protocol

Prior to the experimental trials, a period of heat acclimation was performed to reduce the variability among trained and untrained subjects. The heat acclimation protocol consisted of 10 days of walking/running at 56% of their VO2 peak for 100 min (two 50-min bouts of exercise with 15 min of rest between bouts) in a hot, dry environment (42°C and 30% relative humidity).

Heat stress trials

No sooner than 3 days after the last heat acclimation session, the subjects performed two heat stress trials in a hot, dry room (42°C and 30% relative humidity), at a low (LS) or a high rate (HS) of heat storage in a counterbalanced order. Subjects were advised to refrain from vigorous exercise, caffeine, and alcoholic beverages for at least 24 h before each heat stress trial. They were instructed to eat a high carbohydrate meal on the day before the trials (they were given a list of foods high in carbohydrate) and drink 500 ml of water 2 h before arrival to the laboratory.

Upon arrival to the laboratory, an additional 200 ml of water was provided and a urine sample was collected. The subjects were considered hydrated if urine specific gravity (hand held clinical refractometer, model RHC 200ATC) and urine osmolality (Model 3D3; Advanced Instruments, Inc., Norwood, MA, USA) were lower than 1.030 and 600 mOsmol kg−1, respectively. Baseline nude body weights were measured, and subjects were instructed to insert a rectal probe (Steri-probe #491B; Cincinnati Sub-Zero, Cincinnati, OH, USA) 10 cm beyond the anal sphincter. They were instrumented with a heart rate transmitter strap (S810i seriesTM; Polar, USA). Skin thermistors (EUS-U-VL5-0; Grant Instruments, Cambridge, England) were attached to four skin sites: chest, lateral arm, lateral thigh and lateral calf (Yamada et al. 2007). Rectal and skin temperatures were recorded at 15-s intervals using a data logger (Squirrel 2020 series; Grant Instruments, Cambridge, UK). Then, the subjects dressed in military summer fatigues, body armor, and a frame-supported backpack (total gear adjusted to 15% of their body weight). The summer fatigues consisted of a long-sleeved shirt and full-length pants, which partially restricted evaporation (Thermal insulation ∼0.6 clo; Keim et al. 2002). All trials were completed at the same time of day to avoid circadian variation in core temperature and eHsp72 concentration.

Subjects then entered the environmental chamber and stood quietly on the treadmill for 15 min. The environmental chamber was set to 42°C and 30% relative humidity, and it was monitored every 5 min using a portable heat stress monitor (hs-3600 portable heat stress monitor; Quest Technologies Inc, WI, USA). A baseline, non-stasis blood sample (total of 5 ml) was drawn by venipuncture from an antecubital vein without tourniquet in the standing position: 4 ml of blood was allowed to clot for the measurement of serum osmolality and eHsp72 concentration, and 1 ml of heparinized blood was aliquoted for the measurement of hemoglobin (hgb) and hematocrit (hct). Each heat stress trial started with a 5-min warm-up, walking at 5.6 km h−1 (3.5 mph) and 0% grade, and then treadmill speed and grade were increased to 5.4–6.7 km h−1 (3.4–4.2 mph), 0–4% grade, respectively. The combination of speed, grade and gear were chosen to produce a work rate that required 50% of each subject’s VO2 peak. This was confirmed by measuring the O2 consumption at 20 min into the trial using the methods described above. Each subject walked under the above conditions until his/her core temperature reached 38.5°C or heart rate reached 95% of his/her maximum heart rate. Subjects reported thermal scores (where 0 is very cold and 8 is very hot) which were recorded every 10 min (scales modified from Fox and DuBois 1993). A post-exercise non-stasis blood sample was obtained in the standing position from an antecubital vein immediately upon reaching the target core temperature and the sample was aliquoted as outlined above. The heat stress trial was terminated earlier if subject’s experienced symptoms of exertional heat illness (severe nausea, light headedness, ataxia or confusion) or upon subject request.

During the LS trial, subjects wore an “Air Warrior Cooling Vest” which was supplied by the United States Army Research Institute of Environmental Medicine and followed all the procedures above. Body cooling during the LS trial was obtained by perfusing the vest with 15–16°C water chilled by an electrolytic cooler and controlled at a rate of 0.6 l min−1. The heat extraction (watts) was calculated by the difference in water temperature entering and leaving the vest, adjusted for the flow rate. During the HS trial no artificial cooling was provided. The only difference between the two trials was the cooling applied by the cooling vest.

Hemoglobin, hematocrit, serum osmolality

Heparinized blood was used to measure blood hematocrit in triplicate and hemoglobin concentration in duplicate. Microcapillary tubes were filled with heparinized blood, spun in a microcapillary centrifuge for 5 min (Damon, EC Division), and then hct was immediately recorded as a percentage using a hct reader (#2201; International Equipment Company, Needham HTS, MA, USA). For the measurement of hgb, 20 μl of heparinized blood was added to 5 ml of Drabkins solution and absorbance was read at 540 nm (Thermospectronic 401; Milton Roy, Louisville, CO, USA). Whole blood was allowed to clot, and the serum was removed by centrifugation for 10 min, 4,000 rpm at 4°C (Marathon 21 K/BR; Fisher Scientific). Approximately 1 ml of serum was stored at −20°C for eHsp 72 analyses. The remainder was used to measure serum osmolality in duplicate by freezing point depression (Model 3D3; Advanced Instruments Inc., Norwood, MA, USA). Plasma volume change from pre- to post-heat stress trial was calculated using the hct and hgb values and the equation described by Dill and Costill (1974).

eHsp72 concentration

A commercially available enzyme-linked immunosorbant assay (ELISA, EKS-700B; Assay Designs, Ann Arbor, MI, USA) was used to determine serum Hsp72 concentration from pre- and post-exercise blood samples. Serum dilutions of 1:3 and a standard curve ranging from 0.78 to 50 ng ml−1 (r 2 = 0.999) were used to determine Hsp72 concentration. The ELISA sensitivity is reported to be 0.2 ng ml−1, and intra-assay coefficient of variance is 4.7%.

Heat storage calculations

Mean skin temperature (TSkin) was calculated using the Ramanathan (1964) equation:

$$ T_{\text{Skin}} = 0.3\left( {T_{\text{Arm}} + T_{\text{Chest}} } \right) + 0.2\left( {T_{\text{Thigh}} + T_{\text{Calf}} } \right) $$

where, T Arm is the arm temperature, T Chest is the chest temperature, T Thigh is the thigh temperature, and T Calf is the calf temperature.

Mean body temperature (T Body) was calculated using the following equation (Colin et al. 1971):

$$ T_{\text{Body}} = 0.8\left( {T_{\text{Rectal}} } \right) + 0.2\left( {T_{\text{Skin}} } \right) $$

where, T Rectal is rectal temperature and T Skin is the mean skin temperature.

Heat storage (HS) was calculated using the following equation (Adams et al. 1992):

$$ {\text{HS}}\left( {{\text{W\,m}}^{ - 2} } \right) = 0.965\;{\text{BW}}\;\Delta T_{\text{Body}}\; A_{D}^{ - 1} $$

where, 0.965 W h−1°C−1 kg−1 is the specific heat of body tissues, BW is the subject’s body weight in kg, ∆T Body is the change in mean body temperature in degree Celsius, and A D is the subject body surface area in m2 (DuBois and DuBois 1916). The rate of heat storage was calculated by dividing the HS by trial duration in minutes (W m−2 min−1).

Statistical analyses

A two-way (trial × time) analysis of variance with repeated measure design was performed on eHsp72 data (Statistica Inc, version 7, Tulsa, OK, USA). Dependent t-tests were used for all other analyses. Correlation analyses were used to determine the relationship between exercise duration and eHsp72 post-exercise. Based on a priori power analyses and similar studies, nine subjects would result in a 97% probability of detecting significant differences in eHsp72 (Walsh et al. 2001). Figures and data represent means ± SE unless otherwise indicated. The level of statistical significance was set at α < 0.05.

Results

Heat acclimation was confirmed as the subjects were able to exercise longer (58 ± 30 min on day 1 and 115 ± 0 min on day 10, P < 0.001) with lower final core temperatures (38.89 ± 0.23°C on day 1 and 38.31 ± 0.36°C on day 10, P = 0.007) and lower heart rate (178 ± 7 bpm on day 1 and 164 ± 9 bpm on day 10, P = 0.003). A more detailed description of the heat acclimation process is presented elsewhere (Yamada et al. 2007).

During the heat stress trials, seven subjects terminated the protocol due to reaching the target core temperature (38.5°C) and two trials were terminated because the subject reached 95% of his maximum heart rate (Subject 6 = 38.40°C and Subject 10 = 38.0°C). During the LS trials, seven subjects terminated the trial due to reaching the target core temperature, one subject requested to stop before reaching the target core temperature (Subject 10 = 38.31°C, due to heat-related symptoms) and another subject reached 95% of maximum heart rate (Subject 6 = 38.43°C). It is interesting that subject 6 reached 95% of maximum heart rate at the same core temperature on both treatments (38.40 and 38.43°C). The HR recorded at the end of exercise tended to be higher during HS than LS (Table 1, P = 0.06). The thermal scores were higher at the end of HS than LS (7 ± 1 and 6 ± 1, respectively; P = 0.03).

Table 1 Physiological and protocol related comparisons (mean ± sd, n = 9)
Full size table

The rate of heat storage was significantly different between the two treatments (P < 0.001). The HS trial induced a heat storage rate of 1.04 ± 0.29 W m−2 min−1 and the LS rate was 0.54 ± 0.26 W m−2 min−1. Although the rate of heat storage was different between the two trials, neither total heat storage (40.9 ± 8.6 and 36.3 ± 6.4 W m−2 HS and LS, respectively, P = 0.154) nor final core temperature (P = 0.119) were different between treatments (Table 1). As expected, during the LS trial subjects walked a longer distance than during the heat stress trial (Table 1, P = 0.002). This difference was possible because the cooling vest extracted approximately 315 ± 45 W and slowed the rise in core temperature during the LS (Table 1). Figure 1 illustrates heart rate and core temperature response of subject 11 during both trials. The HR achieved at the end of the heat stress trials tended to be higher during the HS trial than the LS trial (Table 1, P = 0.062). In addition, the oxygen consumption was not different between treatments (Table 1, P = 0.927).

Fig. 1
figure 1

a An individual heart rate response during the two heat stress trials. b An individual rectal temperature response during the two heat stress trials. HS high heat storage trial and LS low heat storage trial

Full size image

Figure 2 shows the eHsp72 concentrations pre- and post-heat stress trials. A significant main effect for time was observed (F = 38.3, P < 0.001) revealing that both heat stress treatments induced significantly higher eHsp72 concentrations post-exercise compared to the pre-exercise values. Conversely, a significant main effect was not observed for treatment (LH versus HS; F = 0.35, P = 0.57) or interaction (treatment versus time; F = 0.01, P = 0.99) indicating that the pre- and post-exercise eHsp72 concentrations were not different between the two treatments, even when the concentrations were corrected for changes in plasma volume (F = 0.15, P = 0.71). Of note, one female subject had undetectable concentrations of eHsp72 pre- and post-exercise during both heat stress trials. We did not observe a significant correlation between post-exercise eHsp72 and exercise duration (r = 0.22, P = 0.418; Fig. 3).

Fig. 2
figure 2

eHsp72 (mean ± SE) pre- and post-heat stress trials (n = 9). Asterisk significant difference between pre- and post-heat stress trial (P < 0.05), high heat storage (HS) and low heat storage (LS), respectively

Full size image
Fig. 3
figure 3

Relationship between exercise duration and eHsp72 (n = 8)

Full size image

The subjects were well hydrated before each heat stress trial as measured by their urine specific gravity (1.007 ± 0.005 and 1.007 ± 0.006, HS and LS, respectively) and urine osmolality (274 ± 189 and 276 ± 193 mOsmol l−1, HS and LS, respectively). The plasma volume change tended to be larger during HS than LS (−8.7 ± 5.7 and −4.4 ± 5.5%, HS and LS, respectively, P = 0.067). Final serum osmolality were not different between trials (296 ± 4 and 298 ± 5 mOsmol l−1, HS and LS, respectively, P = 0.110).

Discussion

In the present study, we found that exercising subjects with two different rates of heat storage, but identical final core temperatures show no difference in post-exercise eHsp72 concentration. This finding suggests that rate of heat storage within physiological ranges in exercising humans may not be the determinant of increased eHsp72 in the blood. Instead, eHsp72 may be a function of the core temperature attained.

We hypothesized that the high rate of heat storage would result in higher increase in eHsp72 concentration compared to the low rate of heat storage, even though the subjects reach the same final core temperature and similar total heat storage. Our hypothesis was based on our previous work examining the response of iHsp72 to passive heating in a rat model (Flanagan et al. 1995). This study revealed that iHsp72 accumulation in the liver was related to the rate of heat storage: rats exposed to a high rate of heat storage (0.166°C min−1) showed significantly greater accumulation of iHsp72 in the liver than those exposed to a lower rate of heat storage (0.045°C min−1) during passive heating. This greater accumulation of iHsp72 occurred despite less time in the heat. Other studies, such as those by Herman et al. (1981) also demonstrate that a high rate of heating (0.166°C min−1) markedly reduced survival compared to a low rate of heating (0.026°C min−1), of Chinese hamster ovary cells heated to the same final temperature (42°C).

The rates of heat storage and final temperature attained (Table 1) in the present study are lower than the studies mentioned above, however our subjects had to cope with an exertional heat stress as opposed to passive heating. In this regard, both the morbidity and mortality in exertional heat stroke occur at a lower core temperature than classical or non-exertional heat stroke (Hubbard et al. 1978; Knochel 1989). The mechanisms for this increased severity of exertional versus classical heatstroke may relate to the presence of acidosis in exertional heat stroke or to any number of other factors associated with exercise-induced hyperthermia. The high rate of heat storage in the present study is comparable to the rate of heat storage observed in marathon runners finishing the race with total time of 150 min and final core temperature of approximately 41°C (Maron et al. 1977). Thus, our volunteers had to deal with a rate of heat storage similar to that observed during a sports event, such as a marathon.

Using human peripheral blood mononuclear cells, Lancaster and Febbraio (2005) observed that eHsp72 concentration was proportional to the exposed temperature (43 > 40 > 37°C). Recently, Whitham et al. (2007) observed that under-water exercise with no increase in core temperature induced less eHsp72 than when the same exercise was performed until a final core temperature similar to the present study (38.5°C). Furthermore, Ruell et al. (2006) observed a positive correlation between eHsp72 concentration and final core temperature measured immediately after a 14-km run. The authors also reported that hyperthermic runners with more severe symptoms of heat illness and higher core temperatures had higher eHsp72 concentrations at the end of a 14 km race versus asymptomatic runners. These data indicate that, unlike iHsp72 accumulation and its relationship to heating rate, final temperature may be the primary determinant of eHsp72 concentration.

The exact sources of eHsp72 during exercise in humans are currently unclear. The initial studies conducted by Febbraio et al. (2002) and Lancaster et al. (2004) reported that splanchnic and brain tissue, but not contracting muscles, contribute to the increase of eHsp72 observed during exercise. Interestingly, in another study Febbraio et al. (2004) demonstrated that reducing the metabolic stress through glucose ingestion during prolonged exercise attenuates the increase of eHsp72 in the systemic circulation and also the local release of eHsp72 by the splanchnic region. In the present study, we reduced the heat stress by applying cooling. In accordance with the studies mentioned above (Flanagan et al. 1995; Burns et al. 1986; Herman et al. 1981), we proposed that a lower rate of heat storage would result in a lower heat stress to deep body tissues and this would reduce the release of eHsp72. However, reducing the rate of heat storage did not change the eHsp72 at the end of the exercise.

Finally, the biological function and mechanisms related to eHsp72 are not completely understood, especially when contrasted to iHsp72. In this regard, iHsp72 is associated with increased resistance to endotoxin (Ryan et al. 1992), decreased cytokine production in endotoxin stressed animals (Kluger et al. 1997), and preserved epithelial barrier function during heat stress (Moseley et al. 1994). Recent investigations have suggested that eHsp72 released after cellular stress can activate the innate immune system by a CD-14 pathway. Patients suffering from septic shock (Wheeler et al. 2005) or heat illness (Ruell et al. 2006) have an up-regulation of eHsp72 in the peripheral circulation. Moreover, Aneja et al. (2006) and Kustanova et al. (2006) observed that administration of eHsp72 to cells or animals previously challenged with bacterial endotoxin, reduced the inflammatory response or mortality rate in human monocytes or rats, respectively. It seems that eHsp72 has both immunostimulatory and immunosuppressive potential, and its immunomodulatory function will depend on the context that eHsp72 is present (Pockley et al. 2008). Therefore, we speculate that increases in eHsp72 may act as an anti-inflammatory signal in conditions such as heat stroke where release of Lypopolysaccharide has been proposed to induce a systemic inflammatory response syndrome.

In summary, the present investigation demonstrated that two distinct rates of heat storage induced a similar increase in eHsp72 during exercise in a hot environment in which subjects reached the same final core temperature. This observation suggests that eHsp72 concentration may be a function of core temperature. This may be important in conditions such as heat stroke, where increased eHsp72 may help regulate the systemic inflammatory response to bacterial endotoxins.