The temperature change of the fractional dissociation of imidazole (α-imidazole) in resting human lower leg muscles was measured noninvasively using 1H-nuclear magnetic resonance spectroscopy at 3.0 and 1.5 T on five normal male volunteers aged 30.6 ± 10.4 yr (mean ± SD). Using 1H-nuclear magnetic resonance spectroscopy, water, carnosine, and creatine in the muscles could be simultaneously analyzed. Carnosine contains imidazole protons. The chemical shifts of water and carnosine imidazole protons relative to creatine could be used for estimating temperatures and α-imidazole, respectively. Using the chemical shift, the values of temperature in gastrocnemius (Gast) and soleus muscles at ambient temperature (21–25°C) were estimated to be 35.5 ± 0.5 and 37.4 ± 0.6°C (means ± SE), respectively (significantly different; P < 0.01). The estimated values of α-imidazole in these muscles were 0.620 ± 0.007 and 0.630 ± 0.013 (means ± SE), respectively (not significant). Alternation of the surface temperature of the lower leg from 40 to 10°C significantly changed the temperature in Gast (P < 0.0001) from 38.1 ± 0.5 to 28.0 ± 1.2°C, and the α-imidazole in Gast decreased from 0.631 ± 0.003 to 0.580 ± 0.011 (P < 0.05). However, the values of α-imidazole and the temperature in soleus muscles were not significantly affected by this maneuver. These results indicate that the α-imidazole in Gast changed significantly with alternation in muscle temperature (r = 0.877, P < 0.00001), and its change was estimated to be 0.0058/°C.
acid-base homeostasis is important for normal body functions. The imidazole group of histidine can reversibly bind a proton within the physiological pH range and plays a key role in maintaining the physiological pH (5, 25, 28). However, the pH of fluids is known to vary with temperature even in the presence of imidazole (5). It has been reported that the change of pH with temperature is nearly equal to that of the equilibrium constant of imidazole (pKIm), suggesting that the ratio of protonated to unprotonated imidazole is not altered by temperature (5, 25, 28). Because histidine is found at the functional sites of many cellular proteins, a constant charge state for imidazole would preserve protein structure and function as temperature varies (18, 25, 28). Reeves (28) proposed the idea that arterial and intracellular pH are not the primary regulated variables but that the ratio of unprotonated to total protein histidine imidazole (fractional dissociation of imidazole or α-imidazole) is primarily regulated to maintain the several protein functions. The idea is called the “alphastat hypothesis.” The alphastat hypothesis has been evaluated indirectly with the help of the data on pH with changing temperature (16, 28, 34). Direct measurements of α-imidazole with changing temperature, however, are obscure (5, 13), even for poikilothermic animals. Although humans are homeothermic animals, a temperature gradient exists in the body, and the temperature near the surface may be influenced by surrounding conditions. The optimal pH strategy during hypothermic cardiopulmonary bypass and hypothermic treatments of humans remains controversial (29). Therefore, the applicability of the alphastat hypothesis to the human body, which suffers changes in tissue temperature, is an interesting issue (18, 24).
An expression for the fractional dissociation of imidazole can be derived from the following equations: (1) where Im is the imidazole group and brackets denote concentration. α-Imidazole could also be calculated using the nuclear magnetic resonance (NMR) chemical shift of the imidazole proton (13) (2) where δacid is the chemical shift of the peak representing completely protonated imidazole, δbase is the peak position of completely unprotonated imidazole, and δo is the peak under ambient conditions.
Using in vivo and in vitro 1H-NMR spectroscopy, metabolites in the muscles such as creatines (creatine, phosphocreatine, creatinine), cholines (choline, phosphoryl choline, glycerophosphocholine) and carnosine could be simultaneously analyzed (1, 3, 36, 37, 39). Carnosine is abundant in human muscles and contains imidazole protons. The imidazole NMR signal of carnosine is detectable in human muscles in vivo (14, 26, 38). Therefore, we could estimate the α-imidazole of carnosine using 1H-NMR spectroscopy. Hitzig et al. (13) applied this method for intact amphibian skeletal muscle and showed that α-imidazole remained constant with alterations in body temperature. Binzoni et al. (4) examined the change in pH in human muscles with changes in temperature (−0.016 pH units/°C) and suggested the applicability of the alphastat hypothesis. However, direct measurement of α-imidazole in human tissues with changing temperature had hardly been achieved. Thus we tried to measure α-imidazole directly with changing muscle temperature using the idea provided by Hitzig et al. (13).
Several noninvasive thermometers employing NMR techniques have been proposed based on the temperature dependence of NMR parameters, such as relaxation time (30), diffusion coefficient (10), chemical shift (2, 6, 8, 20), or the proton frequency method (11, 15, 19). Reliability with an error range of less than ±1°C is required for application of these techniques to the estimation of the temperature in the human body under physiological conditions (21, 23, 32, 33). Cady et al. (6) and Corbett et al. (8) have reported novel techniques for estimating the absolute temperature of the human brain under physiological conditions. They used the chemical shift difference between the water and N-acethylaspartate as the temperature probe. N-acethylaspartate is one of the endogenous metabolites, and, therefore, noninvasive temperature estimation could be done. We used a similar approach to estimate temperature in human lower leg muscles in which the chemical-shift difference between water and creatine was used as the probe. The temperature change of 1°C causes a shift of water signal of ∼0.01 part per million (ppm) (see results).
The temperature and α-imidazole, therefore, could be estimated by 1H-NMR spectroscopy in the human muscles simultaneously, and the correlation between the temperature and α-imidazole could be assessed.
A part of this study has been reported at the First International Conference on Biomedical Spectroscopy: From Molecules to Men (2002, Cardiff, UK) (38).
MATERIALS AND METHODS
Materials and subjects.
Five volunteers aged 30.6 ± 10.4 yr (mean ± SD) were enrolled in this magnetic resonance spectroscopy study. All of these volunteers were healthy men, and NMR spectra were obtained from their legs at resting supine position. Phantom solutions and excised pig skeletal muscles were also used in this study for calibration of α-imidazole and temperature with NMR spectroscopy. The data for these calibrations are described in results in detail. The Ethics Committee of Iwate Medical University has approved the experimental protocols. NMR measurements for volunteers were performed after obtaining appropriate consent forms signed by the volunteers.
Spectroscopy and data processing in humans.
1H-NMR spectroscopy was performed at 3.0 and 1.5 T (General Electric, SIGNA Horizon LX 3T and LX 1.5 T). The spectra of volunteer's lower leg were obtained from gastrocnemius (Gast) and soleus muscles (Sol), which were identified by MRI. Figure 1 shows an example of lower leg MRI. The magnetic resonance spectra were obtained from blocked A and B regions. These regions indicate Gast and Sol, respectively.
Typical 1H-NMR spectra of a human Gast at 3.0 T are represented in Fig. 2.
The upper spectrum shown in Fig. 2 was obtained under water suppression for estimation of the α-imidazole and the lower one without water suppression for estimation of the temperature. The signal position of the imidazole C-2 proton and that of creatine methyl protons, as shown in the upper spectrum of Fig. 2, were used for the estimation of α-imidazole. The signal position of water and that of creatine methyl protons, as shown in the lower spectrum of Fig. 2, were used for the estimation of temperature. There were no differences in measured chemical shift or calculated temperatures when the two different field strengths were used. The standard deviation of the chemical shift difference was less than ±0.002 ppm at both field strengths.
The total sampling time for obtaining the spectra shown in Fig. 2 was ∼14 min. After the NMR spectra at ambient temperature (21–25°C) were obtained from the lower leg muscles, these muscles were cooled or warmed from the exterior using cooling or warming elements. These elements were commercial cold- and heat-reserving materials and had been cooled to −10°C or warmed to 45°C before use. A towel was inserted between the element and the leg. The NMR spectra were obtained 10–40 min later from the beginning of cooling or warming. The skin surface temperatures during cooling and warming were ∼10 and 40°C, respectively. The changes of skin temperature during measurement were <2°C. Acquisition parameters were as follows: at 3.0 T, 3,000-ms repetition time, 30- to 144-ms echo time, 2-K data size, 2,500-Hz spectral width, 8–256 acquisitions; at 1.5 T, 1,500-ms repetition time, 30- to 288-ms echo time, 2-K data size, 2,500-Hz spectral width, 16–512 acquisitions. The voxel sizes were 25–36 ml for human lower leg muscles and 8 ml for pig skeletal muscles and phantom solutions. The raw data were transferred from a magnetic resonance scanner to a personal computer and processed on the personal computer (zero filling; 2 K, apodization; 1 Hz, fast Fourier transform, and peak fitting). The phase of the spectrum was determined manually. NMR relaxation times (T1 and T2) are generally shorter at 1.5 T than at 3 T, although we did not determine those accurately. Therefore we adopted repetition time = 1,500 and 3,000 ms at 1.5 and 3 T, respectively. Then the accumulation number at 1.5 T was two times that at 3 T. The signal-to-noise ratio of the spectrum obtained at 1.5 T was comparable to that at 3 T.
The NMR spectra were analyzed by an automatic curve-fitting procedure and decomposed into Lorentzian peak components (GRAMS/AI, Galactic Industries, Salem, NH). The fittings were done for three separate regions of 1) imidazole protons of carnosine, 2) water protons, and 3) methyl protons of choline and creatine. In the fitting procedures of regions 1 and 3, a baseline correction was performed. The origin of this baseline is the tail parts of water and lipid signals. As a result of this fitting procedure, the height, full width at half-maximum, position, and area of each peak were determined.
Scheffé's test was used for post hoc multiple comparisons between different pairs of means after one-way ANOVA indicated significant differences. The correlations between the temperature and α-imidazole were examined by Pearson's test.
Calibration for estimation of temperature.
Calibration lines for estimation of the muscle temperature were obtained using 16 excised pig skeletal muscles. The temperature of the muscles was monitored by a copper-constantan thermocouple (absolute accuracy: ±0.1°C). The temperature drift of the pig muscles during measurements was <1.0°C. Although the excised pig muscles had been stored in a refrigerator until use, the energy statuses of the muscles were not controlled. The pH of the pig muscles came to near 6. The change of pH, however, has little effect on the chemical shift difference between water and creatine. Relationships between actual temperature measured by copper-constantan thermocouple and NMR spectra are shown in Fig. 3.
The difference between the chemical shifts for choline or creatine and water showed a linear function of increasing temperature from 20 to 40°C (slope: −0.00971 and −0.00999 ppm/°C; r = −0.996 and −0.996 for choline and creatine, respectively). The slopes of these lines were comparable to those of the temperature dependence of the water signal (12, 27).
Estimation of α-imidazole.
The NMR signals of creatine and carnosine were observed using standard buffer solutions (Wako Pure Chemical Industries) at several temperatures. A small amount (∼1 mM) of dimethylsilapentane sodium sulfonate, creatine, and carnosine was dissolved into buffer solutions. The pH of each buffer solution was 1.68 (oxalate buffer), 4.01 (phthalate buffer), 6.86 (phosphate buffer), 7.41 (phosphate buffer), 9.18 (tetraborate buffer), and 10.01 (carbonate buffer) at 25°C. To confirm the signal position of imidazole, dimethylsilapentane sodium sulfonate, creatine, and carnosine were added to the buffer solutions. The change in pH of the buffer solutions induced by adding these chemicals was <0.03 pH units. The buffer solutions were cooled or warmed to several temperatures, and the spectra were observed. The temperature drifts of the buffers during NMR measurements were <0.5°C. In this experiment, the accuracy and stability of the pH were most important. Therefore, we used the standard buffer solutions, sold to calibrate pH meters, and only a little amount of metabolites was added to the buffers. Although the ionic strengths of the buffer we used and the intracellular fluid of human muscles were not equivalent, the effect of ions to the chemical shift difference between carnosine and creatine is little (26). Therefore, we could estimate α-imidazole of human muscles by using the chemical shift data of carnosine and creatine obtained with buffer solutions. The chemical shift of the imidazole C-2 proton in carnosine, relative to the methyl protons of creatine, changed with the temperature at pH 6.8 and 7.4, as shown in Fig. 4A.
The chemical shift values of the imidazole C-2 proton in carnosine at 20, 30, and 40°C were calculated using the linear regression lines shown in Fig. 4A and were fitted by Eq. 3 (1) (3) where δo equals the chemical shift of the imidazole (C-2 proton) signal relative to creatine (−CH3 signal), and δacid and δbase are the limiting acidic and basic chemical shifts, respectively. The fitting curves are shown in Fig. 4B. The pH of the buffer at each temperature was corrected according to the technical data for the buffer (Wako Pure Chemical Industries). In agreement with the previous reports (1, 9, 26), the limiting chemical shifts were unaffected by temperature. The δacid and δbase were estimated to be 5.533 and 4.625 relative to the creatine signal, respectively. Therefore, if imidazole and creatine signals can be observed by 1H-NMR spectroscopy, α-imidazole can be estimated by Eq. 2 independently of the temperature. The value of pKIm was also estimated by using the fitting curves shown in Fig. 4B to be 7.348–0.0154 T, where T is temperature in degrees Celsius. The values of pKIm and ΔpKIm/ΔT are comparable to those in previous reports (1, 9, 13, 26).
Estimated temperature of volunteers.
Using the calibration line shown in Fig. 3, the temperatures of human muscles could be measured noninvasively. We used the creatine signal as an internal chemical shift reference, since the signal-to-noise ratio of this signal is higher than that for choline and the distortion of the signal caused by the large water signal is minimal. Linear regression analysis yielded δ = 2.0505–0.009992 T, where δ is the water proton signal chemical shift relative to the methyl signal of creatine. The estimated values of temperature in the Gast and Sol are shown in Table 1.
Cooling, ambient, and warming in Table 1 represent the muscle temperature at cooled (10°C), ambient (21–25°C), and warmed conditions (40°C). Temperatures of Gast and Sol at ambient conditions were 35.5 ± 0.5 and 37.4 ± 0.6°C (means ± SE), respectively (significantly different; P < 0.01). These values are comparable to the previously reported values (4, 32, 35). Alternation of the surface temperature of the lower leg from 40 to 10°C significantly changed the temperature in Gast (P < 0.0001) from 38.1 ± 0.5 to 28.0 ± 1.2°C but induced no significant change in Sol (Table 1).
Estimated α-imidazole of volunteers.
The values of α-imidazole in Gast and Sol at ambient temperature were estimated to be 0.620 ± 0.007 and 0.630 ± 0.013 (means ± SE), respectively, using Eq. 2. On lowering the surface temperature of the lower leg from 40 to 10°C, the α-imidazole decreased significantly in Gast (P < 0.05) from 0.631 ± 0.002 to 0.580 ± 0.011, but not in Sol (Table 1). Because the acid and basic chemical shifts of imidazole are unaffected by temperature, we could estimate the dissociation of imidazole (α-imidazole) independently of the temperature.
Correlation between α-imidazole and temperature.
The present study demonstrated that the temperatures and α-imidazole in the lower leg muscles could be estimated simultaneously. Finally, we examined the correlation between the α-imidazole and the temperature in Gast. The α-imidazole in Gast significantly correlated with the temperature (r = 0.877, P < 0.00001) as shown in Fig. 5. The slope calculated was 0.0058/°C.
1H-NMR spectroscopy provided a noninvasive method of estimating the temperature and α-imidazole in human lower leg muscles simultaneously. The 1H-NMR chemical shift differences between the water and creatine of pig skeletal muscles vary in a linear fashion with temperature, and these differences could be used to determine muscle temperature. The temperature-dependent change in chemical shift obtained from pig skeletal muscles was nearly identical to that obtained from water (12, 27). Because we applied the results of pig skeletal muscles to those of human muscles, the measured temperature of human muscle with 1H-NMR might not be absolutely accurate. However, large merit of 1H-NMR spectroscopy is its ability to estimate temperatures at various regions of interest noninvasively. Our results showed that the temperature in Sol was 1.9°C higher than that in Gast at ambient temperature and indicated that the temperature in Gast was significantly affected by surrounding conditions but not in Sol.
Hitzig et al. (13) developed a method for direct measurement of α-imidazole in intact, unanesthetized amphibian skeletal muscles using 1H-NMR spectroscopy and reported that α-imidazole remained constant with alterations in body temperature. Their results support the alphastat hypothesis. We applied their direct method on human lower leg muscles. Our results, however, showed that α-imidazole in human Gast decreased gradually on cooling from the surface of the lower leg and also that the values correlated with the muscle temperature significantly. One of the reasons for this discrepancy may be as follows. Hitzig et al. (13) used water signal as the internal reference standard for the chemical shift of the carnosine peaks. The chemical shift of the water signal changes with a temperature change of −0.01 ppm/°C (12, 27). Therefore, the chemical shift of carnosine in amphibian muscles should change with alterations in body temperature. The change of −0.01 ppm/°C in the carnosine signal corresponds to the change of 0.01/°C in the α-imidazole. This value is about two times larger than that of our results (0.0058/°C). It could be considered that the α-imidazole estimated by us and also by Hitzig et al. changes with alternating muscle temperature. In fact, if the chemical shift is calculated in reference to the water signal, the change in the imidazole chemical shift with temperature is very small, and it may be hard to pick up the change of the imidazole signal in vivo.
The temperature dependence of α-imidazole estimated by us corresponds to −0.005 pH units/°C. This value is one-third of that proposed in the alphastat hypothesis (−0.015 pH units/°C). Therefore, this result may indicate that some buffering systems with small ΔpK/ΔT values contribute in human muscles with changes in muscle temperature (7, 22). Inorganic phosphate is one of the candidates with a small ΔpK/ΔT value, i.e., −0.006 to −0.001 (17, 31). Although concentrations of inorganic phosphate are often very low, inorganic phosphate can sometimes contribute significantly to intracellular buffering. It is considered that the concentration of inorganic phosphate is raised to some extent by hydrolysis of ATP and phosphocreatine (16).
Binzoni et al. (4) reported on the pH in the human Gast over a temperature range from 30 to 36°C using 31P-NMR spectroscopy. They showed that the pH varies, as was suggested by the alphastat hypothesis (−0.016 pH units/°C). Our results were obtained with a wider temperature range, between 23 and 40°C (−0.005 pH units/°C). Our data on α-imidazle on cooling was significantly smaller than those of the other two protocols (Table 1). The values at ambient temperature and on warming were not significantly different (Table 1). Therefore, our results at high temperatures may be consistent with the data of Binzoni et al. The decrease of α-imidazole at lower temperatures may be significant (Fig. 5).
Carnosine is distributed in the cytosol but not in mitochondria or myofibrils (39). Therefore, our results may indicate that the α-imidazole of human muscle cytosol changes with alternating muscle temperature, and, therefore, some buffering systems with small ΔpK/ΔT values other than imidazole could contribute in muscle cytosol on changes in temperature.
In summary, we have here described a method using 1H-NMR spectroscopy for measuring the temperature and α-imidazole in resting human lower leg muscles noninvasively and indicated that α-imidazole in the muscles correlates with muscle temperature.
This work was partly supported by Grants-in-Aid for Advanced Medical Science Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan, and Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 the American Physiological Society