Nonselective NOS inhibition blunts the sweat response to exercise in a warm environment

Exercise-induced thermal stress.

Ten individuals (5 men and 5 women) participated in our exercise studies that examined the impact of NOS inhibition on thermoregulatory control of sweating and skin blood flow during exercise. The participants in this study were on average (means ± SE) 21 ± 1 yr old, were 176.4 ± 2.4 cm tall, and had a body mass of 78.4 ± 6.6 kg.

To establish the required workload for the exercise trial, peak oxygen consumption (V̇o_{2 peak}) was measured with a computer-controlled (Parvo Medics, Sandy, UT) upright cycle ergometer (Excalibur, Lode, The Netherlands). The graded exercise protocol began at a power output of 150 W for male subjects and 100 W for female subjects and increased 20 W every min until the subject was not able to continue despite verbal encouragement. The average V̇o_{2 peak} was 46.9 ± 2.9 ml·min⁻¹·g body mass⁻¹ and peak power output averaged 260 ± 18 W. The power output at 60% of V̇o_{2 peak} averaged 156 ± 11 W.

To ensure proper hydration during the exercise trial each subject ingested a volume of water equivalent to 5 ml/kg body mass during the evening meal the night before testing. On arrival on the morning of the experiment, the subject again hydrated with water (5 ml/kg) and ingested 1,000 mg of aspirin to inhibit prostaglandin (PG) production. Our intent with inhibiting PG production was to limit any potentially confounding action of PGs on sweat gland activity and/or cutaneous blood flow. The ability of PGs to modify sweat gland function is poorly studied, but it has been reported that cultured sweat gland cells have the potential to produce PGs (29) and that PGE₁ and PGE₂ both stimulate sweat secretion in vitro (31). PGs also contribute to changes in skin blood flow during whole body heating in a manner that appears to be additive to the impact of NO but has little impact on the hyperemic response to local heating (25). Each subject wore shorts, athletic shoes, and socks. Male subjects were shirtless, and female subjects wore an athletic bra.

Three intradermal microdialysis probes were inserted into the skin of the dorsal aspect of the forearm. Nonglabrous forearm skin was aseptically prepared for the insertion of three 27 gauge needles. Without anesthesia, the needles were inserted laterally through the dermis at a depth of 1–2 mm and traversing a distance of 2.5 cm. A microdialysis probe was then threaded through the lumen of each needle, and the needle was retracted, leaving the diffusion membrane of the probe beneath the skin. The probes consisted of a luer stub adapter, polyethylene tubing (PE50, PE10), polyimide tubing, and hollow fiber. The hollow fiber had a molecular mass cutoff of 18,000 Da and a length of 2.5 cm. The probe was reinforced with stainless steel wire and gas-sterilized before use. The subjects rested for 2 h to allow the dermis to recover from the insertion trauma. During the first hour, each of the three probes was perfused with sterile saline using a micro-infusion pump (Harvard Apparatus, Boston, MA) at the rate of 5 μl/min. During the second hour, the each probe was perfused with one of the randomly assigned treatments: 0.9% saline, 10 mM l-NMMA, or 10 mM l-NAME. The three sites were spaced ∼4 cm apart in parallel intervals to limit any lateral diffusion of the pharmacological agents from influencing the response at adjacent sites.

After the first hour of recovery from insertion of the microdialysis probes, the subject voided her or his bladder and entered the environmental chamber for equilibration to the environmental conditions of the test chamber (30°C). The subjects acclimatized to the warm ambient temperature for 60 min before the collection of baseline data. Inside the environmental chamber, subjects were seated on a modified Monark cycle ergometer (Monark, Varberg, Sweden) that was retrofitted with a semirecumbent bicycle seat.

During the equilibration period sweat capsules were positioned directly over the microdialysis probes using 3M double-stick circles and thick self-adhesive mole skin. The sweat capsules consisted of a plastic cylinder (enclosing 0.567 cm² of skin), a relative humidity (RH) sensor (model HIH-4000, Honeywell, Morristown, NJ), a thermocouple (CCT series, Omega, Stamford, CT), and an airflow sensor (FMA series, Omega). Before the beginning of this series of experiments, the RH sensors were calibrated against saturated salt solutions (LiCl 11.3% RH, MgCl 32.8% RH, and NaCl 73.3% RH) and dry gas (0% RH). Before each experiment, the RH sensor zero level was check by passing dry gas through the capsule while mounted on a sheet of impermeable plastic. After the last experimental trial, the RH calibration was confirmed using the same saturated salt solutions. Dry nitrogen gas was blown across the surface of the skin under the capsule and directed, via large-gauge vinyl tubing, from the capsule to the flow sensors. The flow across each capsule was controlled independently at ∼100 ml/min. Flow sensors were calibrated against a ruby ball flowmeter (Cole Palmer, Vernon Hills, IL), and the inlet pressure was regulated to 7 psi. Assuming that the gas entering the capsule is completely dry then the calculation of SR (in mg·min⁻¹ cm⁻²⁾ was according to the following equation:

(1)

where SR is sweat rate (in mg·min⁻¹·cm⁻²), P is saturated water vapor tension of gas leaving the capsule (in Torr), AF_cap is air flow through the capsule (in ml/min), R_w is gas constant for water vapor (= 3.464 mmHg·l·g⁻¹·°K⁻¹), A_cap is area of sweat capsule enclosing skin (0.567 cm²), and T_cap is temperature of gas leaving the capsule (in °K; 273 + °C)

We report SR as the absolute value measured during each trial without subtraction of any baseline (preexercise) sweating.

Laser-Doppler flow probes (FlowLab, Moor Instruments) were positioned proximally to the sweat capsules and directly above the embedded hollow fiber. A thin layer of ultrasound gel was applied to the head of the sensor to provide a constant sensor-skin optical interface. Esophageal temperature (T_es) was measured using a thermistor (Takara Thermistors, Tokyo, Japan) placed inside a polyethylene (PE90) sleeve with a rounded epoxy tip. The probe was advanced through the nose and swallowed a distance equal to one-quarter the subjects standing height. A dental suction device was placed into the oral cavity to remove saliva and minimize artifacts in the T_es measurement attributable to swallowing saliva. The calibration of the esophageal thermistor was verified following each trial using a precision mercury thermometer and a fixed-temperature water bath. All instruments were interfaced with a laptop computer (Powerbook G4, Apple, Cupertino, CA) using a 16 channel analog-to-digital converter and Chart Software (AD Instruments, Colorado Springs, CO) sampling at a rate of 40 samples/s.

Surface thermocouples were affixed to the skin at seven skin sites. The temperature at each site was measured relative to 0°C (ice slurry bath) using a thermocouple bridge and an analog-to-digital converter (model 931, Acro Systems) interfaced with an iMac (Apple, Cupertino, CA) and averaged every 15 s (Labview 7.1, National Instruments, Austin, TX). Mean skin temperature (T̄_sk) was calculated from the product of regional area (13) and local relative thermal sensitivity (28) using the weighted average of seven local skin temperatures: forehead (20.5%), chest (11.5%), deltoid (8%), forearm (5.3%), thigh (19%), abdomen (5%), and calf (9.7%), divided by 83%.

While the subject was seated on the recumbent bike, both arms were raised to heart level. The right arm was fitted with an automated brachial sphygmomanometer (model STPD, Colin). The left arm was the location of intradermal microdialysis probes, sweat capsules, and laser-Doppler instrumentation. During the exercise bout, resistance was applied equal to the 60% of the workload measured at V̇o_{2 peak}for 30 min. During baseline and exercise, blood pressure and heart rate (HR) were measured once per minute.

Following the exercise, 28 mM SNP was perfused at 10 μl/min for 30 min through each microdialysis probe to elicit maximal skin blood flow. Blood pressure was also measured to allow calculation of maximum cutaneous vascular conductance (CVC_max).

Sweat gland recruitment experiments.

In the preceding series of experiments we noted a reduction thermoregulatory sweating induced by exercise in a warm environment during NOS inhibition. Local sweat rate is determined by the product of the number of active sweat glands under the sweat capsule and the sweat output per gland. The number of sweat glands activated (per square cm) during exercise in a warm environment was evaluated in this series of experiments (n = 6). Under identical conditions described for the direct measurement of local SR during exercise in a warm environment, the number of active sweat glands was monitored using an iodine-starch technique at two skin sites: saline control and 10 mM l-NMMA-treated site. The skin over the intradermal microdialysis probe was first painted with a 2% iodine (in ethyl alcohol) solution and allowed to dry. A double-stick circle (identical to that used to hold the sweat capsule in place on the skin in the earlier experiments) was located directly over the path of the microdialysis probe. The skin within the circle was then covered with a 65% starch solution in caster oil. Sweat gland activation was detected when a droplet of sweat reached the skin surface, and the reaction of the water with the starch and iodine produced a black “dot” on the skin. The oil-based starch solution delayed dispersion of the sweat droplet and allowed detection of the onset of sweat gland activation and the number of active sweat glands during the first 10–15 min of exercise. Digital color images (1,240 × 1,240-pixel resolution) were collected at 0, 1, 2, 3, 4, 5, 10, 15, and 20 min of exercise. The number of active sweat glands was determined by counting the number of black spots within the 0.567 cm² area of treated skin.

All data were averaged every 15 s. During the local heating experiments blood pressure was stable throughout the measurement period, so we present the skin blood flow data as a percent of maximal skin blood flow achieved during 28 mM SNP perfusion. During the exercise experiments, changes in skin blood flow were evaluated from changes in cutaneous vascular conductance (CVC), calculated by dividing laser Doppler flux (V) by mean arterial blood pressure (MAP; in mmHg) and normalized to the CVC_max achieved during SNP perfusion. MAP was calculated as the sum of one-third systolic blood pressure (SBP) plus two-thirds diastolic blood pressure (DBP). SR was expressed in absolute terms (mg·min^-1·cm⁻²⁾.

The effect of exercise on body temperature was determined by one-way ANOVA for repeated measures looking at the following time points: 0, 1, 2, 3, 4, 5, 10, 15, 20, 25, and 30 min. The effect of NOS inhibition on local SR and CVC as a function of time were evaluated using a two-way (treatment and time) repeated-measures ANOVA. In this analysis, we used data from the following time points: 0, 5, 10, 15, 20, 25, and 30 min of exercise. Post hoc comparisons between treatment groups at any given time period were made using the Tukey's minimum significant difference test. Significance was denoted by a values of P ≤ 0.05.

Local heating.

The skin blood flow response to local skin heating is shown in Fig. 1. Resting skin blood flow at a local skin temperature of 29.5°C was similar at all skin sites and averaged 9 ± 1, 6 ± 1, and 6 ± 1% maximum for saline, l-NAME, and l-NMMA, respectively. The plateau in skin blood flow during the final 2 min of the local skin-heating (39.5°C) period averaged 59 ± 6% maximum in the saline group. The plateau skin blood flow was markedly reduced following NOS inhibition and averaged 16 ± 3 (P < 0.05) and 19 ± 2% maximum (P < 0.05) for l-NAME and l-NMMA, respectively. The reduction in skin blood flow during the plateau phase was similar for 10 mM l-NAME and 10 mM l-NMMA. These data provided evidence that these doses of l-NAME and l-NMMA delivered via intradermal microdialysis produced similar and effective blockade of the NOS system in human skin.

Exercise data.

Table 1 lists the average cardiovascular parameters HR, SBP, DBP, and MAP) associated with rest and exercise. During exercise HR, SBP, and MAP were elevated above resting baseline (P < 0.05). DBP did not differ significantly during exercise compared with rest. From 5 through 30 min of exercise HR or MAP were unchanged, indicating a steady-state condition.

Figure 2 illustrates one min average data for T_es, T̄_sk, and local forearm skin temperature (T_forearm). T_es increased during exercise with a significant rise above rest appearing by 4 min of exercise (P < 0.05) and the final increase in T_es averaging 1.5 ± 0.2°C (P < 0.05). T̄_sk, showed a small but transient decrease at 5 min of exercise (P < 0.05) but remained close to resting levels throughout the remaining exercise time. T̄_sk, averaged 32.2 ± 0.1°C over the entire exercise period. T_forearm was unchanged during exercise, averaging 33.6 ± 0.4°C for the entire exercise bout.

The cutaneous dilator response during exercise in a warm environment is depicted Fig. 3. Resting CVC at a T_forearm of 33.0 ± 0.2°C was higher at the saline site ( 28 ± 8% maximum) compared with the l-NAME (13 ± 3% maximum)- or l-NMMA (14 ± 4% maximum)-treated sites (P < 0,05). In addition, CVC at the saline-treated site was always higher than l-NAME or l-NMMA during exercise (P < 0.05). At the end of exercise, CVC averaged 59 ± 4% maximum at the saline site but only 31 ± 3 at the l-NAME site (P < 0.05 different from saline) and 43 ± 4 at the l-NMMA site (P < 0.05 different from saline). In addition, between 10 and 30 min of exercise, CVC at the l-NMMA-treated site was higher than the l-NAME-treated site (P < 0.05; Fig. 3). If we correct for differences in baseline CVC, we note that the increase in CVC during exercise is attenuated with l-NAME (P < 0.05) but not l-NMMA.

At rest before exercise, local SR was similar at all treatment sites and averaged 0.14 ± 0.5, 0.10 ± 0.4, and 0.11 ± 0.4 mg·min⁻¹· cm⁻² at the saline, l-NMMA, and l-NAME sites, respectively (Fig. 4). Between 10 and 30 min of exercise, local SR at the Saline treated site was higher than either the l-NAME- or l-NMMA-treated sites (P < 0.05). Peak local SR averaged 2.4 ± 0.3, 2.0 ± 0.3, and 1.8 ± 0.2 mg·min⁻¹·cm⁻² at the saline, l-NAME, and l-NMMA sites, respectively. Between 5 and 30 min of exercise, local SR at the l-NAME treated skin site was always higher than SR at the l-NMMA skin site (P < 0.05).

Sweat gland recruitment.

Figure 5 illustrates the time course of sweat gland recruitment during exercise in a warm environment with and without NOS inhibition. The number of active sweat glands per square centimeter increased during the first 10 min of exercise. Images taken beyond 10 min of exercise indicated that sweat droplets began to merge. By 10 min of exercise, the total number of active sweat glands per square centimeter were similar for saline (109 ± 14 sweat glands/cm²) and l-NMMA (111 ± 14 sweat glands/cm²) sites.

Limitations.

The principle aim of this study was to evaluate the impact of NOS inhibition on thermoregulatory sweating during exercise. As such, we began by verifying the efficacy of our NOS inhibitors (10 mM l-NAME and 10 mM l-NMMA). Although we are confident that 10 mM l-NAME and 10 mM l-NMMA are equally effective in blunting NO-mediated dilation during local heat-induced hyperemia (Fig. 2), we cannot be certain that they produce equivalent blocking of NO production during exercise-induced thermoregulatory responses. Furthermore, we cannot be certain that this dose of NOS inhibitor is sufficient to completely block NOS activity in sweat glands. In addition, despite the relatively high doses used in this study for NOS inhibition, we cannot be sure whether there is not a difference in specificity. Although we cannot explain the reason for a difference in the NOS inhibitors, we have clearly shown that 10 mM l-NMMA and 10 mM l-NAME both blunt the local sweat response to exercise in a warm environment. There is a possibility of nonspecific drug interactions in this study. Specifically, Medow et al. (26) have shown an interaction between NOS and cyclooxygenase (COX) inhibitors, such that COX blockade appears to unmask a NO-dependent dilation during cutaneous reactive hyperemia. It is possible that our COX inhibition with systemic aspirin treatment also unmasked a NO-dependent sweating response. However, the interaction between NOS and COX inhibitors is not uniformly seen (24), although it is clearly seen during reactive hyperemia in the skin. In contrast, this interaction between NOS and COX inhibitors is not seen during cutaneous dilation induced by local heating or whole body heating (25). There are no observations of such an interaction between NOS and COX inhibitors impacting sweat gland function.

In conclusion, we have observed that NOS inhibition attenuates the thermoregulatory sweat response to exercise in a warm environment. We conclude that NO has some role in modifying thermoregulatory sweating, specifically by changing sweat gland output. The overall importance of such an interaction has not been determined. However, the magnitude of reduction in local SR by NOS inhibition is as large as that induced by dehydration (11). As such, nitric oxide appears to play an important role in thermoregulatory control of sweating during exercise in a warm environment.