Noninvasive and Three-Dimensional Imaging of CMRO 2 in Rats at 9.4 T: Reproducibility Test and Normothermia/Hypothermia Comparison Study

Abstract

Ability to image cerebral metabolic rate of oxygen (CMRO₂) is essential for studying the fundamental role of oxidative metabolism in brain function and disease. We have demonstrated recently that three-dimensional (3D) CMRO₂ images can be obtained in the rat brain during a 2-min ¹⁷O₂ inhalation using the ¹⁷O MR spectroscopic imaging (MRSI) approach at high field. The feasibility for establishing a completely noninvasive approach for imaging CMRO₂ has also been demonstrated. In this study, we further explored the feasibility of ¹⁷O MRSI approach for performing repeated CMRO₂ measurements within a short period of time and evaluated the reproducibility of the repeated measurements. Subsequently, we applied the ¹⁷O MRSI approach to measure CMRO₂ and cerebral blood flow (CBF) values at two brain temperatures in the α-chloralose anesthetized rat brain at 9.4 T. Finally, we tested the validity of simplified model for noninvasively determining CMRO₂ in normothermic and hypothermic rat brain. The results show (i) an excellent reproducibility among repeated measurements of 3D CMRO₂ images under the same physiologic condition; (ii) a 44% decrease of CMRO₂ across the rat brain at mild hypothermic (32°C) condition as compared with normothermic (37°C) condition; and (iii) a close correlation between CMRO₂ and CBF within a relatively wide physiologic range. This study demonstrates the capability of ¹⁷O MRSI approach for noninvasively imaging CMRO₂ and its changes caused by physiologic perturbation. This approach, thus, should provide a promising neuroimaging modality for studying oxidative metabolism and bioenergetics associated with brain functions and diseases.

Introduction

The cerebral metabolic rate of oxygen utilization (CMRO₂) is tightly associated with brain physiology and pathology (Siesjo, 1978). Determination of CMRO₂ is important for understanding the essential role of oxidative metabolism in brain function and brain diseases. The most straightforward nuclear magnetic resonance (NMR) method for directly measuring CMRO₂ is the use of ¹⁷O NMR technique to determine the dynamic change of metabolic H₂¹⁷O, which is formed in mitochondria through oxygen metabolism of inhaled ¹⁷O₂ (Arai et al, 1990; Fiat and Kang, 1992; Mateescu and Fercu, 1994; Mateescu et al, 1991; Pekar et al, 1991; Zhu et al, 2002). However, the major challenges of ¹⁷O NMR technique are its low intrinsic NMR sensitivity because of the low gyromagnetic ratio of ¹⁷O spins and relatively small change in the metabolic H₂¹⁷O concentration generated during an ¹⁷O₂ inhalation. Recently, we have demonstrated a substantial sensitivity gain for detecting ¹⁷O NMR signal at high/ultrahigh magnetic field (Zhu et al, 2001). This crucial sensitivity improvement at higher field can be attributed to the field-independent relaxivity of the ¹⁷O spins in H₂¹⁷O, which leads to a significant sensitivity gain according to a quadratic power relation with the magnetic field strength (Zhu et al, 2001, 2005). This gain makes it possible to use the in vivo ¹⁷O MR spectroscopic imaging (MRSI) approach for obtaining a three-dimensional (3D) CMRO₂ image in a rat brain at 9.4 T during a 2-mins ¹⁷O₂ inhalation (Chen et al, 2004; Zhu et al, 2002, 2005). Moreover, the feasibility for establishing a completely noninvasive ¹⁷O MRSI approach for imaging CMRO₂ in vivo and the validity of the simplified approach for quantifying CMRO₂ have been demonstrated in the rat brain anesthetized with α-chloralose (Zhang et al, 2004).

One potential and important merit of the ¹⁷O MRSI approach is its ability allowing repeated CMRO₂ measurements during the same experimental session. Because in the brain tissue, the metabolic H₂¹⁷O generated during a brief ¹⁷O₂ inhalation can be washed out quickly and reach a new steady-state concentration within approximately 10 mins after the ceasing of ¹⁷O₂ inhalation (Zhu et al, 2002). This characteristic allows subsequent measurements of CMRO₂, which are particularly valuable for studies aiming at CMRO₂ changes induced by physiologic or pathologic perturbations. However, to fully realize this advantage, the reproducibility of the repeated CMRO₂ measurements has to be assessed and validated. Therefore, the first part of this paper was to evaluate the reproducibility of in vivo ¹⁷O MRSI approach for repeated 3D CMRO₂ measurements in the rat brain under the same physiologic condition.

It is well documented that the basal CMRO₂ is sensitive to the brain temperature (see (Erecinska et al, 2003; Siesjo, 1978) and the references cited therein). However, most studies reported in the literature were based on the global CMRO₂ measurements of entire brain using the Kety—Schmidt method (Kety and Schmidt, 1948a, b ), which was lack of spatial information regarding regional CMRO₂ variation inside the brain. In the second part of this study, we investigated the feasibility of using 3D ¹⁷O MRSI at 9.4 T for imaging and quantifying absolute CMRO₂ values in the rat brain at normal brain temperature (37°C) (i.e., normothermia) and mild hypothermia (32°C) conditions. We also compared CMRO₂ calculation results based on two different models: a completed model requiring both invasive and noninvasive measurements (Zhu et al, 2002), and a simplified model requiring only one noninvasive measurement of dynamic concentration change of metabolic H₂¹⁷O in the brain tissue during a 2-min ¹⁷O₂ inhalation (Zhang et al, 2004). This comparison was performed under both normothermia and hypothermia conditions.

Finally, the measured CMRO₂ results in this study were correlated with the cerebral blood flow (CBF) results, which were experimentally determined via bolus injection of H₂¹⁷O tracer into the brain through the internal carotid artery (Zhu et al, 2001). These data were obtained from α-chloralose anesthetized rat brains under a wide range of physiologic conditions including both normothermia and hypothermia.

Materials and methods

All NMR experiments were conducted on a 9.4 T horizontal animal magnet (Magnex Scientific, Abingdon, UK) interfaced to a Varian INOVA console (Palo Alto, CA, USA). A multinuclear surface-coil probe consisting of a butterfly-shape ¹H coil (400 MHz) and an oval-shape ¹⁷O coil (~1 cm × 2 cm, 54.25 MHz) was used for acquiring anatomic ¹H images and in vivo ¹⁷O MRSI data, respectively. Scout images were acquired using a turbo fast low-angle shot (TurboFLASH) imaging sequence with the following acquisition parameters: repetition time (TR) = 8 msecs; echo time (TE) = 4 msecs, field of view (FOV) = 3 cm × 3 cm, and image matrix size = 128 × 128. The spatial localization of ¹⁷O MRSI was achieved by using the 3D Fourier series window approach (Hendrich et al, 1994; Zhu et al, 2001). All animal surgical procedures and experimental protocols were approved by the Institutional Animal Care and Use Committee of the University of Minnesota.

Reproducibility Study

Five male Sprague—Dawley rats (body weight: 234 ± 23 g) were used for studying the reproducibility of two repeated CMRO₂ measurements in the same rat and experimental session. The rats were anesthetized with ~ 2% (v/v) isoflurane in a mixture of O₂ and N₂O gases (2:3) during surgery. After oral intubations, femoral artery and vein were catheterized for physiologic monitoring and/or blood sampling and chemical administration. Four of the five rats also underwent a surgical procedure for CBF measurements via a bolus injection of H₂¹⁷O directly into the brain through the internal carotid artery (see surgical details in Zhu et al, 2001). After surgery, anesthesia was switched to α-chloralose by a bolus injection (40 mg/kg) and following continuous infusion (~ 20 mg/kg/h) of α-chloralose solution. Blood gases were sampled for monitoring physiologic conditions (pH = 7.39 ± 0.05, pCO₂ = 38 ± 6 mm Hg, pO₂ = 148 ± 13 mm Hg, n = 5). The arterial blood pressure (110 ± 13 mm Hg) and heart rate (410 ± 25/min) were also monitored during the experiment. The rectal temperature of the rats was maintained at ~ 37°C by using a heated circulating water blanket.

For ¹⁷O₂ inhalation experiments, the ¹⁷O-labeled O₂ gas (68.3% enrichment, ISOTEC) was mixed with N₂O gas (~ 2:3) and stored in a cylindrical gas reservoir. The 3D ¹⁷O MRSI acquisitions were performed with the normal gas mixture with nonlabeled O₂ gas first; these natural abundance 3D ¹⁷O MRS images were used as references for quantifying metabolic H₂¹⁷O concentrations in the rat brain during and after ¹⁷O₂ inhalation (Zhu et al, 2002). Then the respiration gas was quickly switched to the ¹⁷O₂-labeled gas mixture while the continuous ¹⁷O MRSI acquisitions were not interfered. After ~ 2 min of ¹⁷O₂ inhalation, the gas line was switched back to the normal gas mixture, and the 3D ¹⁷O MRSI acquisitions were continued for ~ 15 mins after the ceasing of ¹⁷O₂ inhalation. The second ¹⁷O₂ inhalation and CMRO₂ measurement were performed in the same fashion within ~ 30 to 60 min after the completion of the first ¹⁷O₂ inhalation. The CBF measurement was performed after the second ¹⁷O₂ inhalation experiment via a bolus injection of H₂¹⁷O tracer (~0.05 to 0.1 mL, 50% enrichment, ISOTEC) into the brain via an internal carotid artery. The physiologic conditions of the animal during these measurements were kept as close as possible.

The acquisition time for each 3D ¹⁷O-MRSI data set was 11.5 secs with following parameters: TE = 0.4 msecs; TR = 12 msecs (≈ 3 times of the longitudinal relaxation time of ¹⁷O water (Zhu et al, 2001)); FOV = 2.8 × 2.8 × 2.4 cm³; 9 × 9 × 5 phase encodes; voxel size = 105 μL; radio frequency (RF) (hard) pulse length = 50 μs and spectral width = 15 kHz. A 17 × 17 × 9 matrix of free induced delays (FIDs) was generated from the original 9 × 9 × 5 phase encode data for each 3D ¹⁷O MRS image. The FIDs were zero-filled, and a 100-Hz line broadening was used before fast Fourier transformation for signal to noise ratio (SNR) enhancement. The ¹⁷O NMR signals from each voxel were quantified by measuring the H₂¹⁷O resonance peak intensities, which were then converted to absolute H₂¹⁷O concentrations using the natural abundance H₂¹⁷O NMR signal from the same voxel as an internal reference. The H₂¹⁷O concentration was used to calculate CMRO₂ for the ¹⁷O₂ inhalation experiments and/or calculate CBF for the H₂¹⁷O bolus experiments. The CBF values were quantified from the H₂¹⁷O signal decay rate after a bolus injection of H₂¹⁷O tracer (Zhu et al, 2001).

Normothermia and Hypothermia Comparison Study

Five male Sprague—Dawley rats (body weight: 283 ± 25 g) were used for this study. The general surgical procedures were the same as described in the previous section. In addition, three of the five rats underwent a surgical preparation for the CBF measurement; and two of these three rats were also implanted a vascular ¹⁷O RF coil, tuned at the ¹⁷O Larmar frequency and wrapped around the intact carotid artery, allowing simultaneous measurement of arterial input function of H₂¹⁷O (C_a(t)) and the brain H₂¹⁷O (C_b(t)) signal with two ¹⁷O RF coil configuration and two receiver channels (Zhang et al, 2003; Zhu et al, 2002). The physiologic conditions of the rats, including arterial blood pressure, heart rate, inspired and expired gases were recorded throughout the experiment. Blood gases were sampled at different periods of the study. The rat brain temperatures under normothermic and hypothermic conditions were also monitored via a temperature sensor (PhysiTemp Instrument), which was inserted into the inner ear of the rats without poking through the membrane. After surgical preparation and animal settled inside the magnet, the anesthesia was switched to α-chloralose (initial bolus of 40 mg/kg followed by continuous infusion with infusion rate of ~ 25 mg/kg/h). The normothermic and hypothermic conditions were controlled via a rectal temperature probe and a circulating water blanket with heated or refrigerated water.

Under stable normothermic condition (37°C), a short ¹⁷O₂ gas (72.1% or 89.9% enrichment) inhalation (~ 2 min) was performed on each rat, followed by a CBF measurement via bolus injection of H₂¹⁷O tracer (~ 0.05 mL, 50% enrichment) if it was applicable. After these measurements, the water pump for the circulating water blanket were switched to refrigerant mode and allowing the animal temperature to gradually decrease. When the body temperature of the rat reached the desired mild hypothermic (32°C) condition, the water pump was switched back to the heating mode to maintain the temperature at 32°C. This cooling process took approximately one and half hours. When the animal reached a stable hypothermic condition, second ¹⁷O₂ inhalation experiment was performed on each rat. Therefore, the averaged time interval between the two CMRO₂ measurements for this study was longer than 2 h.

The parameters for acquiring 3D ¹⁷O MRSI data were varied slightly among the rats in this group but were kept same for each rat. The acquisition time for each image was 10 to 12 secs with TR of 10 to 12 msecs and TE of 0.35 to 0.4 msecs. The voxel size was ~50 to 100 μL with FOV of 2.2 × 2.2 × 1.8 cm³ to 2.8 × 2.8 × 2.4 cm³ and phase encodes of 9 × 9 × 5. A 50-μs RF hard pulse and 30 kHz spectral width were used for data acquisition, and a 17 × 17 × 9 matrix of FIDs was generated from the original 9 × 9 ×5 phase encode data for each 3D ¹⁷O image. General data process and quantification procedures were the same as described above. The time courses of the metabolic H₂¹⁷O generated in each ¹⁷O MRSI voxel were used to determine the CMRO₂ values and create 3D CMRO₂ images.

Comparison of Complete Model and Simplified Model for CMRO2 Calculation

The dynamic change in the cerebral H₂¹⁷O concentration during an ¹⁷O₂ inhalation is determined by three parallel processes: (i) cerebral oxygen utilization for generating the metabolic H₂¹⁷O in the brain tissue, (ii) blood perfusion resulting in H₂¹⁷O washout from the brain, and (iii) blood flow recirculation bringing the metabolically generated H₂¹⁷O in the entire living body into the brain. The mass balance of the ¹⁷O-labeled H₂¹⁷O concentration in the brain tissue during the period of ¹⁷O₂ gas inhalation can be derived as (Kety and Schmidt, 1948a; Zhang et al, 2004; Zhu et al, 2002)

$$\frac{\mathrm{d}{C}_{\mathrm{b}}(t)}{\mathrm{d}t}-2\alpha f_{1}CMR{O}_2+mCBF(f_2{C}_{\mathrm{a}}(t)-\frac{n{C}_{\mathrm{b}}(t)}{\lambda })$$

(1)

where C_a(t), C_b(t) and C_v(t) are the H₂¹⁷O concentrations in excess of the natural abundance H₂¹⁷O concentration level in the arterial blood, brain tissue and venous blood, respectively, as a function of ¹⁷O₂ inhalation time (t, unit = min); α is the ¹⁷O-enrichment fraction of inhaled ¹⁷O₂ gas; λ is the brain/blood partition coefficient. Two unit conversion factors, f₁ = 1.266 and f₂ = 1.05, are used to achieve consistency of units among all variables used in Eq. (1). The correction factor m was used to account for the fact that water is not completely freely diffusible across the brain—blood barrier. The constant n was a correction factor that accounts for another type of restriction on the permeability specifically for the metabolically generated H₂¹⁷O inside the mitochondria and it can be determined experimentally (Zhu et al, 2005, 2002). In general, C_a(t) is linearly proportional to the inhalation time and can be expressed as C_a(t) = At, in which A is a constant (Zhu et al, 2005, 2002). Solving the linear differential equation of Eq. (1) led to the following equation:

$$CMR{O}_2=\frac{\{[(C_b(t)-(A{f}_2{\lambda }^2/m{n}^{2}CBF)((mnCBF/\lambda )t{\mathrm{e}}^{-(mnCBF/\lambda )t}-1))/(1-e^{-(mnCBF/\lambda )t})]-(A{f}_2\lambda t/n)\}}{(2\alpha \lambda f_1/mnCBF)}$$

(2)

Therefore, the CMRO₂ value at each data point measured at different inhalation time can be calculated using the experimental data of CBF, A, n, C_b(t), and other known constants (f₁, f₂, m, α, and λ) according to Eq. (2). This quantification approach according to Eq. (2) is defined as a complete model, which accounts for all required parameters for precisely determining CMRO₂ (Zhang et al, 2004; Zhu et al, 2005, 2002). However, this approach requires both invasive measurements (e.g., CBF, A and n measurements) and the noninvasive measurement of C_b(t).

In contrast, the invasive measurements could be eliminated by using the simplified model based on the Taylor's and Polynomial Theorems (Zhang et al, 2004). Briefly, the C_b(t) time course can be expressed by a polynomial expansion

$$C_{\mathrm{b}}(t)=a_{1}t+a_2{t}^2+a_3{t}^3+\cdots$$

(3)

In this expansion, the first-order coefficient of a₁ can be used to calculate CMRO₂ using the following equation:

$$CMR{O}_2=\frac{a_1}{2\alpha f_1}$$

(4)

where α and f₁ are known constants. Thus, using the simplified model, only the experimental data of C_b(t) measured by in vivo ¹⁷O MRSI approach is needed and it can be fitted to the polynomial function given by Eq. (3) to calculate a₁, and ultimately CMRO₂ according to Eq. (4). In this study, the polynomial function was approximated by a linear function for a short ¹⁷O₂ inhalation time (Zhang et al, 2004).

For the rats performed with multiple measurements of C_a(t), C_b(t), and CBF, both complete model and simplified model (linear fitting) were used for calculating and comparing CMRO₂ values at both normothermia and hypothermia conditions. The difference of CMRO₂ obtained at these two conditions reflects the change of CMRO₂ induced by temperature change. The CMRO₂ values calculated with completed model and/or simplified model from the same experimental data reflected the accuracy of the simplified model for quantifying CMRO₂ at normal brain temperature and mild hypothermia conditions.

Data and results are reported as means ± standard deviation (s.d.).

Results

In vivo 3D 17O MRSI of Rat Brain

Figure 1 demonstrates an example of 3D ¹⁷O MRS images (~ 11 secs data acquisition and 75 μL voxel size) of natural abundance H₂¹⁷O and corresponding ¹H anatomic images from a representative rat brain acquired at 9.4 T. Five adjacent coronal slices covering the ¹⁷O RF coil-sensitive region of the rat brain are shown for both ¹⁷O and ¹H images. The relative ¹⁷O RF coil size and position (cross-section) on each slice are depicted in the anatomic images. The nonuniform spatial distributions of ¹⁷O NMR signal in the ¹⁷O MRSI data are because of the inhomogeneous RF field (B₁) given by the ¹⁷O surface coil used in the study. Nevertheless, excellent SNR of ¹⁷O signal is evident, especially for the central image voxels.

Figure 1

3D ¹⁷O MRSIs of natural abundance H₂¹⁷O (top row) and corresponding ¹H anatomic images (bottom row) of rat brain at 9.4 T. The ¹⁷O RF surface coil positions and cross-sections are indicated in the images. The parameters for acquiring these 3D ¹⁷O MRSI data were: FOV = 2.5 × 2.5 × 2.2 cm³; voxel size = 75 μL; acquisition time = 11 secs.

Reproducibility Study of Repeated CMRO2 Measurements

Figure 2A shows the stacked plots of H₂¹⁷O spectra from a central voxel of 3D ¹⁷O MRS images acquired before, during, and after two consecutive 2-min ¹⁷O₂ inhalations from a representative rat brain. The excellent sensitivity of ¹⁷O signal and the stability of magnetic resonance imaging scanner are clearly demonstrated in this figure, which are important for detecting the dynamic change of the metabolic H₂¹⁷O content in the rat brain during and after ¹⁷O₂ inhalation. Figure 2B and Figure 2C are the linear fittings of the H₂¹⁷O concentration change in the same voxel during first and second ¹⁷O₂ inhalations, which were used for calculating CMRO₂ (Zhang et al, 2004; Zhu et al, 2002, 2005). The CMRO₂ values in this particular voxel were almost identical between the first and second measurements (2.65 and 2.63 μmol/g/min) under the same physiologic condition.

Figure 2

(A) Stacked plots of H₂¹⁷O spectra from a representative voxel (~ 105 μL, located in the center of ¹⁷O coil-sensitive region near cortex area) of 3D ¹⁷O MRSI data acquired before, during, and after two consecutive 2-min ¹⁷O₂ inhalations in a rat brain. (B), (C) The linear fitting of the metabolic H₂¹⁷O concentration in this voxel during first and second ¹⁷O₂ inhalation for calculating CMRO₂ (2.65 and 2.63 μmol/g/min, respectively).

Figure 3 provides a voxel-by-voxel-based comparison of two CMRO₂ measurements in a representative rat brain under the same physiologic condition. Total 305 voxels located in the ¹⁷O RF coil-sensitive regions (similar as shown in Figure 1) were used for the comparison. The ratios of the first measured CMRO₂ to the second measured CMRO₂ for each voxel were calculated. The linear correlation coefficient (R) between the first and second CMRO₂ measurements shown in Figure 3 was 0.77. The averaged CMRO₂ ratio between these two measurements from all analyzed voxels was closed to unity (1.02 ± 0.09) for this rat.

Figure 3

A voxel-by-voxel-based comparison of two consecutive 3D CMRO₂ measurements in a representative rat brain. CMRO₂ ratio (1st/2nd) = 1.02 ± 0.09 and correlation coefficient = 0.77 (total voxel number = 305).

The same data analysis method was applied for each of the five rats that underwent reproducibility study. The results are summarized in Table 1. The mean ± s.d. values of the first and second measured CMRO₂ values of these five rats were 2.26 ± 0.18 and 2.20 ± 0.14 μmol/gmin, respectively; and the CMRO₂ ratio (1st/2nd) was 1.03 ± 0.05 (n = 5) with an average voxel number of 290 ± 38 (n = 5) per rat used for analysis. In four of the five rats, the differences between two CMRO₂ measurements were well within a 5% margin, and the difference was ~ 10% for the other rat. Paired t-test show that P ≥ 0.25 (n = 5), indicating that the first and second measured CMRO₂ values are statistically identical.

Table 1

CMRO₂ comparisons between two consecutive measurements using in vivo ¹⁷O MRSI approach in rat brain at 9.4T (n = 5; mean ± s.d.)

Rat	First measured CMRO2 (μmol/g/min)	Second measured CMRO2 (μmol/g/min)	Ratio (1st/2nd)
1	2.13	2.19	0.97
2	2.28	2.21	1.03
3	2.35	2.34	1.00
4	2.49	2.26	1.10
5	2.04	1.97	1.04
Mean ± s.d.	2.26 ± 0.18	2.20 ± 0.14	1.03 ± 0.05

CMRO2 Measurements at Normothermia and Hypothermia Conditions

Physiological parameters of the rats under normothermia (37°C) and mild hypothermia (32°C) conditions are summarized in Table 2. While the blood gas data (pH, pCO₂, and pO₂) were similar for the rats at two brain temperatures, the mean arterial blood pressure and the heart rate of the animal decreased significantly under the hypothermia condition.

Table 2

Physiological data of rats at normothermia and hypothermia condition (n = 5; mean ± s.d.)

	pH	pCO2 (mm Hg)	pO2 (mm Hg)	MABP (mm Hg)	Heart rate (beat/min)
Normothermia (37°C)	7.43±0.04	38±5	127±20	100±10	398±22
Hypothermia (32°C)	7.38±0.04	39±8	137±27	86±10	276±24

Figure 4 illustrates the cerebral concentration changes of metabolic H₂¹⁷O in a single voxel taken from the 3D ¹⁷O MRS images as a function of inhalation time during the first ¹⁷O₂ inhalation at 37°C and second ¹⁷O₂ inhalation at 32°C, respectively. The accumulation rate of metabolic H₂¹⁷O during an ¹⁷O₂ inhalation was significantly decreased under the hypothermia condition. The linear fittings to the two time courses of C_b(t) from the same voxel indicate a dramatic CMRO₂ change (1.9 μmol/g/min at 37°C versus 1.1 μmol/g/min at 32°C).

Figure 4

Dynamic changes of metabolic H₂¹⁷O concentration during ¹⁷O₂ inhalation from a representative voxel (~ 105 μL, similar location as in Figure 2) in the rat brain at normal temperature (37°C, full circles) and during mild hypothermia (32°C, open circles). The linear fitting of the data was used for calculating CMRO₂ values at two temperatures (= 1.9 μmol/g/min at 37°C and 1.1 μmol/g/min at 32°C).

We have demonstrated previously that under normal physiologic condition, CMRO₂ calculations based on the complete model, which requires invasive procedures and rigorous measurements of C_b(t), C_a(t), CBF, and n in the same animal, are in good agreement with the CMRO₂ calculation based on the simplified model which requires only one noninvasive measurement of C_b(t) (Zhang et al, 2004; Zhu et al, 2005). In this study, we performed similar measurements in the rat at normal temperature as well as under mild hypothermia to test whether the validity of the simplified model still holds in a wider physiologic range. Figure 5 shows the results of such a study in a representative rat brain in which all parameters involved in the CMRO₂ calculation according to the complete model were measured explicitly (Zhang et al, 2004). The CMRO₂ values calculated with the complete model and simplified model for each individual ¹⁷O MRSI voxels of the rat brain at 37°C and 32°C are presented in Figure 5A. The consistency and tight correlation (R = 0.97) between calculated CMRO₂ values using these two models are clearly evident for both normothermia and hypothermia conditions. The averaged CMRO₂ values of this rat at two different brain temperatures determined with both models are demonstrated in Figure 5B (complete model: CMRO₂ = 1.84 ± 0.29 μmol/g/min at 37°C versus CMRO₂ = 1.10 ± 0.20 μmol/g/min at 32°C; simplified model: CMRO₂ = 1.77 ± 0.23 μmol/g/min at 37°C versus CMRO₂ = 1.13 ± 0.17 μmol/g/min at 32°C). These CMRO₂ values obtained with two models are in excellent agreement at both temperatures without statistic difference. This result reveals that the simplified model based on linear fitting provides a good approximation for the true CMRO₂ values determined by the complete model, and this is valid not only under normal physiologic condition but also in the mild hypothermic rat brain. Using this simplified model, we could obtain 3D CMRO₂ maps of the rat brain in a completely noninvasive way. Figure 6 illustrates an example showing three representative slices of 3D CMRO₂ maps from a rat brain under normothermia (Figure 6A) and hypothermia (Figure 6C) conditions, as well as the corresponding anatomic images (Figure 6B). These images clearly displayed significant reduction of CMRO₂ across the entire brain induced by decreasing brain temperature several degrees. The reduction of the CMRO₂ during hypothermia is consistently observed in other four rats studied. Figure 7 summarizes the CMRO₂ values calculated with the simplified model in the rat brains at normothermia (37°C, CMRO₂ = 1.84 ± 0.09 μmol/g/min) and hypothermia (32°C, CMRO₂ = 1.02 ± 0.09 μmol/g/min) conditions (n = 5). The averaged voxel number used for calculating CMRO₂ for each rat was 249 ± 17 (37°C) and 243 ± 16 (32°C).

Figure 5

(A) Voxel-based CMRO₂ calculation and comparison using the completed model and simplified model from a representative rat (total voxel number used was 224 for 32°C and 254 for 37°C, voxel size = 75 μL). (B) Averaged CMRO₂ values in the same rat brain at normothermia (37°C) and hypothermia (32°C) conditions calculated with the simplified model and completed model, respectively.

Figure 6

Three-dimensional CMRO₂ maps of a representative rat brain at (A) normothermia (37°C) and (C) hypothermia (32°C) condition, as well as their corresponding anatomic images (B).

Figure 7

Measured CMRO₂ values of individual rat calculated with the simplified model at normothermia (37°C) and hypothermia (32°C) conditions. The right bars show the means and s.d.'s of measured CMRO₂ values (n = 5) at normothermia and hypothermia conditions.

Figure 8

Correlation of CBF and CMRO₂ values in the rat brains anesthetized with α-chloralose at brain temperature range of 32°C to 37°C (•). The linear correlation coefficient (R) was 0.97. The open symbols depicted CBF and CMRO₂ data from literature in α-chloralose-anesthetized rat brain measured either directly with micro positron emission tomography ((Δ); Yee et al, 2005, 2006), or indirectly with ¹³C NMR technique ((□); Hyder et al, 2000), or autoradiography ((○) (Nakao et al, 2001) and (*) (Cholet et al, 1997)) techniques.

Correlation Between CMRO2 and CBF

In addition to mapping CMRO₂ in the rat brain, the ¹⁷O MRSI approach was used to determine absolute CBF in a subgroup of rats in both reproducibility study (CBF = 0.72 ± 0.07 mL/g/min (n = 4)) and normothermia/hypothermia comparison study (CBF = 0.56 ± 0.14 (37°C) and 0.27 ± 0.13 (32°C) mL/g/min (n = 3)). Because the physiologic conditions of these rats covered a relatively wide range, it would be interesting to examine the relation between the CMRO₂ and CBF among these rats. Figure 8 summarizes the absolute values of CMRO₂ versus CBF in the α-chloralose-anesthetized rat brains. It shows a strong correlation between CBF and CMRO₂ in the rat brain at temperature ranging from 32°C to 37°C with a linear correlation coefficient of R = 0.97. This tight correlation was observed in the rat brains with the averaged CBF values ranging from 0.2 to 0.8 mL/g/min and averaged CMRO₂ values ranging from 1.0 to 2.5 μmol/g/min, which were experimentally determined in our study. In addition to our results, we also included in this figure the CBF versus CMRO₂ data from the literature reports of rat brains under similar α-chloralose anesthesia. These literature data, particularly CMRO₂ values, were obtained either directly (e.g. measure CMRO₂ with micro positron emission tomography (Yee et al, 2006)) or indirectly (e.g. measure oxidative glucose consumption rate (CMR_{glc, total}) with ¹³C MRS (Hyder et al, 2000) or total glucose consumption rate (CMR_{glc, total}) with autoradiography (Cholet et al, 1997; Nakao et al, 2001)) with various independent technical modalities and they agreed well with our results.

Discussion and conclusion

The superior ¹⁷O NMR sensitivity available at ultrahigh field makes it possible for detecting and imaging small dynamic changes of metabolic H₂¹⁷O during a brief inhalation of ¹⁷O₂, and ultimately, for imaging CMRO₂ in three dimensions in the rat brain (Zhu et al, 2001, 2002). In this study, we further extended our investigation to address one important aspect of ¹⁷O MRSI approach, that is, the reproducibility of CMRO₂ measurements.

Our results show that in the rat brain, especially for those voxels with high SNR as illustrated in Figure 2, the reproducibility of two consecutive CMRO₂ measurements is extreamly high. When considering all the voxels located in the ¹⁷O RF coil-sensitive region as shown in Figure 1, the majority of the voxels present highly reproducible CMRO₂ values when voxel-by-voxel-based analysis was performed. This is evident from the ratio between two repeated CMRO₂ measurements, which is closed to unity in most of the voxels inside the rat brain. Because these repeated measurements were performed in the animal under stable physiologic condition without any perturbation, therefore, we conclude that the in vivo ¹⁷O MRSI approach for repeated CMRO₂ measurements should provide consistent and reproducible outcomes. This capability is essential for studying the oxidative metabolism changes related to perturbations in physiology and brain function, in which at least two measurements are required under control and perturbation (or activation) conditions.

The feasibility of 3D ¹⁷O MRS imaging for determining CMRO₂ changes because of physiologic perturbation was examined in the rat brain at normal brain temperature and under hypothermia condition. Our results show that in the same rat brain anesthetized with α-chloralose, a significant decrease of CMRO₂ was observed across the entire brain when the brain temperature decreased from 37°C to 32°C. The measured CMRO₂ ratio between normothermia and hypothermia conditions was 1.80 ± 0.10 (n = 5). This value is similar to the CMRO₂ reduction observed in the thiopental-anesthetized rat brain when the brain temperature decreased from 38°C to 34°C (Nemoto et al, 1996) and/or calculated from depressed tricarboxylic acid cycle activity in the halothane-anesthetized rat brain when temperature decreased from 37°C to 31°C (Kaibara et al, 1999). If assuming a linear relationship between CMRO₂ and temperature (Hagerdal et al, 1975), we could estimate the Q₁₀ factor from our data resulting in Q₁₀ ≈ 2.2 for the rat brain anesthetized with α-chloralose. This Q₁₀ factor of CMRO₂ is in good agreement with the literature reports where global CMRO₂ measurements based on the Kety—Schmidt method were conducted (Hagerdal et al, 1975; Klementavicius et al, 1996; Michenfelder and Milde, 1991; Michenfelder and Theye, 1968; Nemoto et al, 1996). The actual relation between CMRO₂ and brain temperature as well as their spatial dependence should be obtainable using the 3D ¹⁷O MRS imaging approach; however, this topic is beyond the scope of this study.

In our previous work, we had demonstrated the feasibility to establish a completely noninvasive ¹⁷O NMR approach for imaging CMRO₂ in the rat brain under normal physiology condition (Zhang et al, 2004). In an effort of extending the applicability of this noninvasive CMRO₂ imaging approach to various physiologic conditions and/or in different species, we evaluated the validity of this method in the normothermic and hypothermic rat brains. Our results have clearly demonstrated that the CMRO₂ calculation based on the simplified model (i.e. linear fitting) and using only noninvasively determined brain metabolic H₂¹⁷O concentration could provide a close approximation of the true CMRO₂ values in the rat brain under both normothermic and mild hypothermic conditions. Therefore, the promising results from our previous as well as current studies conducted at ultrahigh field (i.e. 9.4T) provide crucial evidence supporting the notion that noninvasive 3D ¹⁷O MRS imaging of CMRO₂ is not only feasible but also reliable, at least in the rat brain under a wide range of physiologic conditions.

There are only few methods that are able to determine absolute CBF and CMRO₂ values in a living brain without killing animal. Among them, the Kety—Schmidt method (Kety and Schmidt, 1948b) can provide global CMRO₂ and CBF information and has led to much of our knowledge about the circulation and metabolism in animal and human brain under a variety of physiologic or pathologic conditions. This method is considered as the most robust and reliable way to determine CMRO₂ and CBF values without information of their spatial distributions (see (Siesjo, 1978) and references therein). The positron emission tomography approach has been widely applied for imaging CMRO₂ and CBF in humans (Mintun et al, 1984; Ter-Pogossian et al, 1970) by using radioactive ¹⁵O₂ gas and H₂¹⁵O and C¹⁵O tracers. Positron emission tomography can provide information regarding the distribution of the CBF and CMRO₂ values with limited spatial resolution and hence mainly applied for human brain studies. The NMR method capable of directly determining absolute CMRO₂ and CBF values in the animal brain would be the ¹⁷O MRS approach alone (Pekar et al, 1991; Zhu et al, 2002, 2005) or in combination with other MR technique for CBF measurements (e.g. Arai et al, 1998; Pekar et al, 1995; Tailor et al, 2003). At ultrahigh fields such as 9.4 T, we were able to obtain 3D CMRO₂ and CBF images of the rat brain with ~ 50 to 105 μL vexel size as demonstrated in this study. Furthermore, by comparing the CMRO₂ and CBF values obtained from each rat brain at varied physiologic conditions, we were able to extrapolate the relation between CMRO₂ and CBF in the α-chloralose anesthetized rat brain. Our data indicate that CMRO₂ and CBF obey an approximate linear relationship (correlation coefficient: R = 0.97) within the physiologic conditions we studied. When comparing our results with the literature data of CMRO₂ and CBF in the α-chloralose anesthetized rat brains obtained with different technique modalities, we found that although they gave a range of CBF and CMRO₂ values in different studies with various anesthesia level and physiologic condition, these literature data points, however, are coincident with our data and, more importantly, they obey the same correlation relation between CMRO₂ and CBF (see Figure 8). Our results also suggest that the reduction of CBF is comparable with that of CMRO₂ when the brain temperature of the animal decreases from 37°C to 32°C; that means, the coupling between the blood perfusion and oxidative metabolism is not significantly altered in the mild hypothermic rat brain under moderate anesthesia of α-chloralose.

In conclusion, the developed in vivo ¹⁷O MRSI approach at ultrahigh fields is capable of repeated 3D CMRO₂ imaging within the same experimental session; it can provide reproducible CMRO₂ values in the rat brain under the same physiologic condition and can readily determine and distinguish the CMRO₂ changes induced by varying the physiologic states of the animal. The simplified analysis method for calculating CMRO₂ based solely on the noninvasively measured brain metabolic H₂¹⁷O concentration is valid under normal and mild hypothermia condition. Therefore, the same ¹⁷O MRSI approach should be useful for studying other physiologic perturbations such as hypoxia and ischemia, as well as brain activation and brain diseases.