Search tips
Search criteria 


Logo of jcbfmJournal of Cerebral Blood Flow & Metabolism
J Cereb Blood Flow Metab. 2012 March; 32(3): 403–412.
Published online 2012 January 18. doi:  10.1038/jcbfm.2011.191
PMCID: PMC3293125

Characterizing brain oxygen metabolism in patients with multiple sclerosis with T2-relaxation-under-spin-tagging MRI


In this study, venous oxygen saturation and oxygen metabolic changes in multiple sclerosis (MS) patients were assessed using a recently developed T2-relaxation-under-spin-tagging (TRUST) magnetic resonance imaging (MRI), which measures the superior sagittal venous sinus blood oxygenation (Yv) and cerebral metabolic rate of oxygen (CMRO2), an index of global oxygen consumption. Thirty patients with relapsing-remitting MS and 30 age-matched healthy controls were studied using TRUST at 3 T MR. The mean expanded disability status scale (EDSS) of the patients was 2.3 (range, 0 to 5.5). We found significantly increased Yv (P<0.0001) and decreased CMRO2 (P=0.003) in MS patients (mean±s.d.: 65.9%±5.1% and 138.8±35.4 μmol per 100 g per minute) as compared with healthy control subjects (60.2%±4.0% and 180.2±24.8 μmol per 100 g per minute, respectively), implying decrease of oxygen consumption in MS. There was a significant positive correlation between Yv and EDSS and between Yv and lesion load in MS patients (n=30); on the contrary, there was a significant negative correlation between CMRO2 and EDSS and between CMRO2 and lesion load (n=12). There was no correlation between Yv and brain atrophy measures. This study showed preliminary evidence of the potential utility of TRUST in global oxygen metabolism. Our results of significant underutilization of oxygen in MS raise important questions regarding mitochondrial respiratory dysfunction and neurodegeneration of the disease.

Keywords: MRI, multiple sclerosis, neurodegeneration, nitric oxide, oxygen metabolism


Multiple sclerosis (MS) is considered an inflammatory demyelinating disease of the central nervous system. One of the hallmarks in MS is the progressive neurodegeneration that has a key role in the progression of neurologic disabilities (Lisak, 2007; Trapp et al, 1999). Little is known of the link between neuroinflammation and neurodegeneration. Recent biochemical studies suggested a crucial role of nitric oxide (NO) overproduction due to vascular inflammation in neuronal/axonal injury (Brown and Borutaite, 2002; Encinas et al, 2005). The increased NO competitively inhibits the binding of oxygen to mitochondrial respiratory complex and affects ATP synthesis (Brown and Borutaite, 2002; Encinas et al, 2005; Smith and Lassmann, 2002; Su et al, 2009). Unlike a true hypoxia secondary to the blood flow disruption, this phenomenon is called ‘histotoxic' or ‘metabolic hypoxia,' a condition that, though O2 might be available, but cells and tissues are unable to use it (Aboul-Enein and Lassmann, 2005; Trapp and Stys, 2009). The subsequent mitochondrial respiratory dysfunction due to this mechanism may cause progressive neurodegeneration and impairment of the brain function in MS (Aboul-Enein and Lassmann, 2005; Beal, 1992; Su et al, 2009). A better understanding of neural tissue energy metabolism by measuring oxygen utilization in vivo may have potential to pave the way to optimized treatment approach of neurodegenerative processes in MS. Work exploring this mitochondrial hypothesis has been obtained mainly from immunohistopathological studies of postmortem tissue or cerebrospinal fluid (CSF); however, studies addressing oxygen consumption in vivo are clinically needed, yet still rare in MS.

One study by Sun et al (1998), using positron emission tomography (PET), which requires specific tracer such as 15O, has demonstrated that there is cerebral hypometabolism in both gray matter (GM) and white matter (WM) of MS patients. The severity of cerebral hypometabolism was also related to the number of relapses in patients, suggesting that the measurement of oxygen metabolism has a potential role in monitoring disease activity and prognosis. The clinical assessments with PET are often limited by its complicated data corrections, exposure to ionizing radiation and clinical availability. To the best of our knowledge, there are no data available using quantitative magnetic resonance imaging (MRI) to assess the oxygen metabolic changes in MS. In this study, a recently developed T2-relaxation-under-spin-tagging (TRUST) technique (Lu and Ge, 2008; Lu et al, 2008) was used. T2-relaxation-under-spin-tagging imaging was developed to quantitatively estimate venous oxygen saturation via the measurement of pure blood T2-relaxation time in the lower superior sagittal sinus, an accumulative blood pool draining from the most cerebral brain tissues, therefore, providing a global estimation of oxygen consumption. From this technique, Yv, defined as the percentage of oxygenated hemoglobin, is calculated based on the well-established relationship between venous blood T2 and blood oxygenation (Bryant et al, 1990; Gomori et al, 1987; Thulborn et al, 1982; Wright et al, 1991). With whole brain cerebral blood flow (CBF) measurement, it also has potential to provide a global estimation of cerebral metabolic rate of oxygen (CMRO2) or oxygen consumption (Xu et al, 2009).

Since oxygen metabolism is directly linked to cell energy and neuronal activities, we hypothesize that MS has significant oxygen metabolic abnormalities that are tightly linked to the mitochondrial respiratory dysfunction and progressive neurodegeneration (Brown and Borutaite, 2002; Encinas et al, 2005; Smith and Lassmann, 2002; Su et al, 2009). The purpose of the current study was to investigate whether there are detectable global oxygen consumption changes using TRUST MRI in patients with relapsing-remitting MS as compared with age-matched healthy controls. We also evaluated the relationship between cerebral Yv or CMRO2 changes and parenchyma tissue loss, MS lesion load as well as clinical disability scales.

Materials and methods


This study was in compliance with the health insurance portability and accountability act. After receiving an explanation of the study procedure, all the participants provided written informed consent approved by our Institutional Review Board. Thirty patients with clinically definite relapsing-remitting MS (Poser et al, 1983) (11 males and 19 females; mean age, 38.8 years; range, 22 to 60 years) were included in this study. The mean disease duration for the patient group was 4.8 years (range, 8 to 167 months) and the mean expanded disability status scale (EDSS) score was 2.3 (range, 0 to 5.5). Clinical examinations were performed by a neurologist who has over 15 years of experience with MS patients. The patients were recruited with the following inclusion criteria: (1) no pulmonary and cardiovascular morbidity, (2) no cerebrovascular diseases or other clinical neurologic diseases or confirmed on the current MR imaging scans, and (3) no history of respiratory diseases or anemia. Twenty-four patients were on immunomodulating treatment regimens (10 on β-Interferonas, 12 on Glatiramer acetate, and 2 on Natalizumab) and six were untreated at the time of the MRI. None of the patients received Statins in their treatment regimen. There were four patients who had relapses and no one had received steroids within 1 month before MRI. For comparison, 30 healthy volunteers (12 males and 18 females; mean age, 37.0 years; range, 20 to 59 years) were also recruited and they were confirmed to have no brain diseases by history and imaging studies.

Magnetic Resonance Acquisition

Magnetic resonance imaging was performed on a 3.0-T whole body MR scanner (Magnetom Trio, Siemens Medical Solutions, Erlangen, Germany) with an eight-channel array head coil. Since the venous oxygenation may change during sleep, all the subjects were instructed not to fall asleep during the experiments (verified after each session) with finger pulse oximetry monitored during the MRI. The TRUST sequence is similar to proximal inversion with a control for off-resonance effects (PICORE) arterial spin-labeling sequence (Wong et al, 1997), except that the labeling slab is on the venous side and was placed above or caudal to the imaging slice (area of interest) and the venous draining spins are inverted (instead of arterial blood). Figure 1 showed geometric locations of the labeling slab and the image slice in TRUST MRI for venous T2 measurement. The labeling slab thickness was 50 mm and the gap between labeling slab and imaging slice was 25 mm. The spins above the imaging slice, which is placed at the lower level of superior sagittal sinus, are labeled and the venous blood inside the labeling slab will flow downward and enter the imaging slice.

Figure 1
The placement of labeling slab (purple) of T2-relaxation-under-spin-tagging (TRUST) sequence is above the actual imaging slice (green) to obtain draining venous blood signal measured in the lower superior sagittal sinus. Labeling slab thickness 50 mm, ...

T2-relaxation-under-spin-tagging sequence was performed in a transverse plane parallel to the anterior–posterior commissure line and going through the lower superior sagittal sinus (just above the level of confluence of sinuses) for a global estimation of oxygen consumption. For each scan, the sequence begins with a presaturation radio frequency (RF) pulse to suppress the static tissue signal (to improve the signal-to-noise ratio) and is followed by a labeling (or control) RF pulse to magnetically label the venous blood. As shown in Figure 2, by performing a subtraction of labeled and control images, only flowing signal remains in the difference image reflecting pure venous blood. The T2-relaxation time of the blood is estimated by repeating the label/control pairs but with increasing T2-weighting. To minimize the effect of blood outflow on the estimation, a nonselective T2-preparation pulse train is employed for T2-weighting, the duration of which is denoted effective echo time (eTE), instead of the typical multispin-echo acquisition. Therefore, a complete TRUST MRI sequence includes both labeled and control scans acquired at different eTEs for different T2-weightings (Figure 2). The specific imaging parameters of TRUST were as follows: TR (repetition time)/TE (echo time)/TI (inversion time)=8,000/19/1,200 milliseconds, repetition=4, field-of-view=230 mm × 230 mm, matrix=64 × 64, single-shot echo planar imaging, slice thickness=5 mm, four eTEs: 0, 40, 80, and 160 milliseconds, corresponding to 0, 4, 8, and 16 refocusing pulses with an interval (τCPMG) of 10 milliseconds in the T2-preparation. The total acquisition time for TRUST is 4 minutes and 16 seconds.

Figure 2
T2-relaxation-under-spin-tagging (TRUST) magnetic resonance imaging (MRI) images in which the venous blood is either magnetically labeled (Label, middle row) or unlabeled (Control, top row). Each type of image is acquired at four different T2-weightings, ...

Imaging protocols in the current study also consist of the routine MRI studies for MS, including standard 2D turbo spin-echo T2-weighted imaging (TR/TE=5,000/88 milliseconds; matrix, 448 × 364; in-plane resolution, 0.5 × 0.5 mm2), 2D fluid attenuation inversion recovery imaging (TR/TE/TI=9,000/87/2,500 milliseconds; matrix, 256 × 224; in-plane resolution, 0.9 × 0.9 mm2), and contrast-enhanced fast low angle single-shot T1-weighted imaging (TR/TE=354/2.73 milliseconds; matrix, 768 × 640; in-plane resolution, 0.3 × 0.3 mm2). These routine sequences were all acquired with 3 mm thick contiguous axial sections. In addition, a 3D T1-weighted MPRAGE (magnetization-prepared rapid gradient-echo) sequence was performed on all subjects to attain a brain volume measurement. The parameters for 3D MPRAGE were as follows: TR/TE/TI=2,300/2.98/900 milliseconds, field-of-view=256 × 256 mm2, matrix=256 × 240, flip angle=15° 256 slices, 1 mm thick contiguous sagittal sections, voxel size=1 × 1 × 1 mm3. Among 30 patients and 30 healthy controls, 12 patients and 12 controls also had additional CMRO2 protocol performed, which includes global CBF measured with single-slice phase-contrast (PC) MRI with the following parameters: TR/TE/flip angle=39 milliseconds/5.93 milliseconds/20°, voxel size=0.45 × 0.45 × 5 mm3, maximum velocity encoding=80 cm/s. This results in one PC image with scan duration of 2 minutes and 43 seconds. The positioning of the PC scan was based on a time-of-flight angiogram acquired with the following parameters: TR/TE/flip angle=17 milliseconds/4.12 milliseconds/18°, voxel size=1 × 1 × 2.5 mm3, and number of slices=40, one saturation band placed above the imaging slab, imaging duration: 2 minutes and 55 seconds. The phase-contrast imaging slice was placed above the carotid artery bifurcation and oriented perpendicular to the internal carotid and vertebral arteries, allowing simultaneous assessment of the four feeding arteries (Figure 3).

Figure 3
Global cerebral blood flow (CBF) was measured with phase-contrast magnetic resonance imaging (MRI) magnitude (A) and phase (B) images and the positioning of phase-contrast scan was based on a 2D time-of-flight (TOF) angiogram including coronal (C) and ...

All the subjects were asked not to drink caffeine 24 hours before the MRI was performed. We have physiological monitoring of heart rate and blood pressure in 12 patients and controls, the macrovascular hematocrit (Mhct) level from peripheral blood was also acquired in 12 patients during the time of MRI scan.

T2-Relaxation-Under-Spin-Tagging Data Analysis and Processing

All data were processed offline using in-house-written MATLAB (Mathworks, Natick, MA, USA) scripts. The details of data processing procedures for TRUST MRI and estimation of CMRO2 were described previously (Lu and Ge, 2008; Xu et al, 2009). In short, the TRUST data consist of labeled and control images, and each image type is acquired with four different eTEs as shown in Figure 2. The preprocessing includes pairwise subtraction to ensure that only blood signal draining into the superior sagittal sinus is shown and quantified, thus excluding the possibility of measuring other tissues in the venous sinus, such as CSF partial volume, dura matter, or granular tissue.

A thorough theoretical description for the computation of venous sinus blood T2-relaxation time was previously provided (Lu and Ge, 2008). We first define the duration of the T2-preparation sequence as eTE, which is the additional T2-weighting that will contribute to the acquired MR signal with nominal TE. The signal difference of static tissue between labeled and control scans can be canceled out. The blood signal in the control scan is

equation image

And the blood signal in the labeled scan (superior sagittal sinus in this case) is

equation image

Therefore, the measured signal intensity difference (ΔS) is a function of S0, eTE, and a factor C, which is inversely proportional to blood T2:

equation image

in which An external file that holds a picture, illustration, etc.
Object name is jcbfm2011191e4.jpg and C=1/T1b−1/T2b.

The experimental data, as a function of eTE, can then be fitted to a monoexponential function to obtain the exponent C. Note that although C is related to both the T1 and T2 relaxation times of the blood, it is predominantly determined by the term 1/T2b because the term 1/T1b is more than one order smaller. To further minimize the effect of T1b, we used a literature value for this parameter (T1b=1,624 milliseconds) (Lu et al, 2004) and applied the following equation to calculate T2b:

equation image

After motion correction and pairwise subtraction was performed for the control and label images to obtain the difference images of venous blood signal for each eTE (Figure 2), a preliminary ROI (region of interest) covering the lower superior sagittal sinus region was manually drawn. As shown in Figure 2, four voxels in the ROI containing the largest difference signals (which represent the pure venous blood signals) in the image of eTE=0 were selected as the final mask for spatial averaging. Only large venous vessels (e.g., venous sinus) have discernable signal intensities in the difference image because of labeled spins from relatively large flow velocities. The signals from different eTEs were fitted to obtain CPMG (Carr–Purcell–Meiboom–Gill) T2 of the venous blood as described above.

Previous studies showed a direct relationship between blood T2-relaxation time and blood oxygen saturation (Y) (Barth and Moser, 1997; Golay et al, 2001; Wright et al, 1991; Zhao et al, 2007). Validated at various field strengths, an empirical relationship between these parameters was established:

equation image

in which A, B, and C are dependent on magnetic field strength and Mhct. The coefficients used in the present study were based on the results from recent calibration experiments using the TRUST sequence (Lu et al, 2011). Specifically, A=6.80/s, B=0.38/s, and C=60.3/s for an Mhct of 0.42 (Lu et al, 2011). For participants on whom Mhct was measured, the A, B, and C values were based on the actual Mhct value using the relationship described in Lu et al (2011). Based on this relationship, venous oxygenation, Yv, is converted from the venous blood T2.

Cerebral metabolic rate of oxygen is an index indicating the amount of O2 molecules consumed per unit mass tissue per unit time and the whole brain CMRO2 can be calculated using the following equation,

equation image

where CBF is global or whole brain CBF, Ya is oxygen saturation in arterial blood, Yv is oxygen saturation in venous blood, and Ca is the amount of oxygen molecules that a unit volume of blood can carry, and is assumed from hematology literature (833.7 μmol O2/100 mL blood) (Guyton and Hall, 2005). Although both Yv and Ca in equation (6) could vary slightly depending on the assumed Mhct value, their effects on CMRO2 estimation are actually opposite; thus, the dependence of CMRO2 estimation on Mhct is minimal (Xu et al, 2009). We calculated the whole brain CBF based on PC MRI data in a subgroup of MS patients (n=12) and controls (n=12). Among 12 patients, 8 patients were receiving immunomodulating treatment and the rest four were not at the time of MRI. The areas of all four brain feeding arteries (left and right internal carotid arteries) were firstly segmented using a simple threshold (five times the background noise) on magnitude image to create vessel masks. These masks were then applied to phase image (velocity map) to obtain the total blood flow. The whole brain CBF (in mL per 100 g per minute) was finally measured using total blood flow divided by total brain parenchyma (BP) volume (quantified with T1-weighted MPRAGE). Ya was determined by pulse oximetry and arterial vessels deliver blood that usually has an oxygenation level close to unity. Yv was determined using TRUST and calculated using equation (5). Once these three parameters described above are known, CMRO2 can be calculated using the above Fick's principle (equation (6)). Oxygen extraction fraction was defined as (YaYv)/100. The details of CMRO2 quantification using TRUST technique was described in the previous study (Xu et al, 2009).

Brain and Lesion Volume Measurement

The 3D T1-weighted MPRAGE data were used for segmentation of GM, WM, and CSF using VBM8 ( toolbox under Statistical Parametrical Mapping software (SPM-8, Wellcome Department of Cognitive Neurology, London, UK; running on MATLAB R2009a (Mathworks). The method was well described previously (Ashburner and Friston, 2000) with high reproducibility and accuracy. In brief, tissue volumes including GM, WM, and CSF were preliminarily segmented using a fully automated relaxometric (T1) method based on isotropic images with voxel size of 1 × 1 × 1 mm3. The procedures started with image coregistration and reslicing to ensure that all scans were normalized to the standard MNI (Montreal Neurological Institute) template space using mean GM, WM, and CSF templates (Prinster et al, 2010). To increase the tissue specificity in each voxel, the normalized images were then modulated to correct for changes in volume induced by normalization through the process with the voxel intensities being multiplied by the local value in the deformation field from normalization. The resulting modulated images were smoothed with a 12-mm full-width at half maximum (FWHM) isotropic Gaussian filter. Finally, the fractions of GM, WM, CSF, and BP volume relative to the total intracranial volume (the sum of GM, WM, and CSF) were calculated to normalize the head size. In addition, T2 lesion volume and Gd-enhancing lesion number and volume in MS patients were computed using semiautomated in-house software of MIDAS segmentation package (De Santi et al, 2001) based on T2-weighted images. After intensity uniformity correction, the process started by automatic detection of the WM signal intensity, Iwm, in a periventricular ‘seed' region without lesions. Following calculation of these seed-based ‘normal' WM intensity and selection of appropriate thresholds (at or above mean+2.5 s.d.) and apply for all voxels, an abnormal tissue mask is constructed per slice. This process involves three steps: morphological erosion; recursive region growth retaining pixels connected to the ‘seed' and morphological inflation to reverse the effect of erosion. Lesion number and volumes were then obtained from the T2-weighted images and Gd-enhanced T1-weighted images after confirmation and correction by human interventions.

Statistical Analysis

SAS 9.3 (SAS Institute, Cary, NC, USA) was used for all computations. For comparison of quantitative indices (T2, Yv, CMRO2, volumetric measures) between patients and controls, independent-sample unequal-variance t-tests were performed. Associations between Yv or CMRO2 and other end points such as lesion volume, GM, WM, and BP volume fractions as well as EDSS were assessed using Pearson correlations. All reported P values are two-sided with statistical significance defined as P<0.05.


Table 1 shows the comparison of quantitative measurements of venous blood T2 and Yv, oxygen extraction fraction (OEF), brain parenchyma fraction (BPF), global CBF, and CMRO2 in patients and controls. The T2 of the superior sagittal sinus blood was found to be significantly higher (P<0.0001) in the patient group (mean±s.d.: 71.5±10.3 milliseconds) as compared with the control group (mean±s.d.: 60.1 ±6.7 milliseconds). Correspondingly, Yv, which is converted from venous blood T2, was also significantly higher (P<0.0001) among patients (mean±s.d.: 65.9%±5.1%) than among controls (60.2%±4.0%) (Figure 4A), suggesting significantly higher oxygenation level in the venous blood in MS. Oxygen extraction fraction, which was calculated from (YaYv) also showed significant difference in patient group (32.2%±5.4%) compared with control group (37.3%±7.4%). In addition, CMRO2 data acquired in a subgroup of patients (n=12) and controls (n=12) showed significantly lower CMRO2 in patients (mean±s.d.: 138.8±35.4 μmol per minute per 100 g tissue) than healthy controls (180.2±24.8, P=0.003) (Figure 4B), further indicating decreased oxygen consumption in patients. However, there was no significant difference of global brain CBF between these two subgroups (mean±s.d.: 58.5±13.1 versus 61.8±6.0 mL per minute per 100 g tissue; P=0.42) heart rate (mean±s.d.: 64.8±8.4 versus 63.7±9.8, P=0.76), and blood pressure (systolic: 121±7.9 versus 119±10 mm Hg, P=0.48; diastolic: 78.3±4.4 versus 76.5±5.4 mm Hg, P=0.39). The Mhct level measured in MS patients (n=12) was 0.41±0.04 (mean±s.d.), which is within normal range.

Figure 4
Box-plot representation of Yv (n=30) (A) and cerebral metabolic rate of oxygen (CMRO2) (n=12) (B) in control and multiple sclerosis (MS) groups, demonstrating significantly higher Yv (%) (mean±s.d.: 65.9±5.1 versus 60.2±4.0; ...
Table 1
Comparison between MS patients and normal controls: results of independent-sample unequal-variance t-tests

We evaluated the relationships of venous oxygen saturation measures of Yv and CMRO2 with conventional imaging measures including BPF and lesion load as well as clinical measures of EDSS. The results of their correlation coefficient and P values are listed in Table 2. The mean T2 lesion volume in the patient group was 4.01±4.73 cm3 (mean±s.d.). Only seven enhancing lesions were identified in four patients with mean volume of 0.01±0.04 cm3 (mean±s.d.). There was a significant positive correlation between Yv and EDSS (n=30, r=0.54, P=0.002) (Figure 5A) and also between Yv and the total lesion volume (n=30, r=0.40, P=0.03) (Figure 5B). We also found a significant negative correlation between CMRO2 and EDSS (n=12, r=−0.61, P=0.03) (Figure 5C) and between CMRO2 and lesion load (n=12, r=−0.74, P=0.005) (Figure 5D), indicating that patients with higher venous oxygenation level or lower oxygen consumption tend to have higher EDSS and lesion load. There was a significant difference (P=0.03) of BP fraction (BPF) between patients (mean±s.d.: 81.0±3.4) and controls (82.8%±2.4%), implying there is general brain atrophy in MS patients. We did not find any statistically significant correlations between either Yv or CMRO2 with atrophic measures (i.e., BPF), enhancing lesions, or disease duration.

Figure 5
Relationship between Yv and expanded disability status scale (EDSS) (A) and lesion volume (B), and between cerebral metabolic rate of oxygen (CMRO2) and EDSS (C) and lesion volume (D) in patients with multiple sclerosis (MS) patients. A significant positive ...
Table 2
Results of Pearson's correlation coefficient in patients with MS


Using quantitative TRUST MRI, this study defines for the first time a significantly higher oxygenation level of the venous sinus blood (Yv) in MS as a result of considerably reduced oxygen consumption or CMRO2. Our findings are consistent with the findings of the prior PET study (Sun et al, 1998), in which hypometabolism was found in both GM and WM. Because oxygen metabolism is critical for cell energy and brain health, these findings may lead to new insights into the diffuse and progressive neurodegeneration of MS that is likely associated with mitochondrial impairment, which is increasingly being recognized in a number of immunohistochemical studies (Brown and Borutaite, 2002; Encinas et al, 2005; Smith and Lassmann, 2002; Su et al, 2009).

Compared with PET, quantitative MRI techniques for brain oxygen metabolism have many advantages including clinical availability, no radioactive radiation, low cost, and well-established routine use of clinical diagnosis and disease monitoring of MS. The oxygen metabolic measures may provide new probes with functional relevant features concerning the initial causative processes of neuronal cell dysfunction (defective mitochondrial oxidative phosphorylation) before structural damage occurs. T2-relaxation-under-spin-tagging MRI estimates global Yv through its effects on the T2-relaxation time in a large intracranial venous sinus (i.e., superior sagittal sinus). The T2-relaxation time is more accurate than T2* regarding their interrelationship with blood oxygenation (Bryant et al, 1990; Gomori et al, 1987; Thulborn et al, 1982; Wright et al, 1991) with high degree of precision and reliability. The Yv converted from T2 using TRUST has been shown in good agreement with PET studies (Lu and Ge, 2008; Lu et al, 2008). In addition, TRUST employs a spin-labeling approach to isolate the superior sagittal sinus to improve the accuracy of T2-relaxation time measurement from pure venous blood (Lu and Ge, 2008). Unlike conventional T2 mapping, TRUST uses T2-preparation to modulate T2-weighting and also uses composite pulses and RF phase cycling to minimize the effect of RF imperfection on T2 value, thus the conversion from T2 to Yv is more accurate (Lu et al, 2011). Most existing noninvasive MRI techniques for CMRO2 quantification involve special physiological challenges such as hypercapnia or hyperoxia (Chiarelli et al, 2007; Leontiev and Buxton, 2007) and/or exogenous agents (e.g., caffeine). These so-called calibrated functional MRI techniques are sensitive to changes of CMRO2 but not to the baseline values; therefore, they may not be suitable for routine clinical studies.

In our study, the decreased oxygen consumption in MS patients was indicated by the findings of both significantly increased Yv and decreased CMRO2. Cerebral metabolic rate of oxygen is an index signifying the amount of O2 molecules consumed per unit mass tissue per unit time (in μmol per minute per 100 g tissue). The portion that remains in the blood will determine the venous oxygen saturation or Yv. Yv changes depend on the relative changes of both blood supply (i.e., CBF) and oxygen consumption (i.e., CMRO2) (Buxton, 2010) and the increase of CBF without change of CMRO2 could also raise the oxygenation level in the venous blood. Since we did not find statistical difference of global CBF between patients and controls (P=0.42), the significantly increased Yv, thus, is likely caused by reduced oxygen consumption. Although previous perfusion studies using dynamic susceptibility contrast MR imaging have demonstrated regional perfusion abnormalities including increased perfusion in acute lesions and prelesional stage (Broom et al, 2005; Ge et al, 2005; Wuerfel et al, 2004) or decreased perfusion in some chronic lesions (Broom et al, 2005), periventricular WM (Law et al, 2004), and deep GM (Inglese et al, 2007). However, most of perfusion studies were ROI-based and did not provide global CBF information. We have not found significant change of global CBF in MS patients in this study.

The decreased oxygen consumption in MS found in this study are also supported by the results of a previous study using SWI (susceptibility weighted imaging) in MS (Ge et al, 2009), in which the visibility of venous vasculatures in periventricular WM in patients with relapsing-remitting MS was diminished on SWI venography. Susceptibility weighted imaging can image cerebral veins by exploiting the magnetic susceptibility effects from paramagnetic deoxygenated hemoglobin. The signal change of the local venous vasculature visualized on SWI was thought to reflect a result of increased venous oxygenation or correspondingly decreased level of deoxygenated hemoglobin content (Ge et al, 2009). However, TRUST is a more direct and quantitative measure of blood oxygenation than SWI. Both studies showed a significant correlation between these venous oxygenation changes and lesion load (Ge et al, 2009), which was also corroborated by the prior PET study (Sun et al, 1998). This suggests that oxygen underutilization becomes more prominent as lesions increase.

At present, there are no satisfactory explanations for progressive neurodegeneration in MS. It has been proposed that mitochondrial ATP production is compromised in MS (Aboul-Enein and Lassmann, 2005) by increased ambient levels of NO secondary to microvascular inflammation (Encinas et al, 2005; Smith and Lassmann, 2002). The increased NO competitively inhibits the binding of oxygen to respiratory complex (cytochrome C oxidase) and inhibits cell respiratory chain function in mitochondria (Moncada and Bolanos, 2006). Such chronic oxygen starvation of neuronal cells causes ATP depletion, leading to diffuse and progressive neurodegeneration and neuronal dysfunction in MS (Smith and Lassmann, 2002). Since overproduction of NO in MS is mainly driven by lesions, the positive correlation between Yv and lesion load found in this study may substantiate the above mechanisms that increased NO underlies decreased mitochondrial oxygen metabolism. We also found a significant positive correlation between Yv and EDSS and between CMRO2 and EDSS, implying higher venous oxygenation level or lower global oxygen consumption may be related to higher clinical disability. The information gained from noninvasive studies of CMRO2 will have positive implication for in vivo assessment of neuronal cell degeneration and disease progression that is associated with impaired oxygen metabolism in MS.

One challenge from the current data is to answer the question of what is exactly the underlying cause of the changes of Yv and CMRO2 in MS. Although NO hypothesis of competitive inhibition of oxygen uptake in mitochondrial respiratory process was proposed, we could not exclude the contribution from neuronal tissue loss or cell death, which had a decreased oxygen demand and decreased oxygen utilization, based on these imaging data. Nitric oxide is difficult to be measured directly and NO level is mainly measured by its nitroderivatives in CSF or peripheral plasma. Several immunohistochemical studies have shown 1.5 to 3 times higher NO endproducts levels in MS patients compared with controls (Acar et al, 2003; Rejdak et al, 2004) and such increased NO can be diffuse and distal to the lesions (Moncada and Bolanos, 2006; Smith and Lassmann, 2002). Although it is difficult to differentiate these two pathological processes in vivo using our techniques, we note that the CMRO2 reduction in MS patients was about 20% (compared with control subjects), which was much greater than the atrophic change (~2%) or lesion volume (~0.5%). That is, even if the regions with tissue loss or gliotic changes do not use any oxygen, it would still not explain the 20% reduction in oxygen consumption. In addition, we have not found significant correlation between either Yv or CMRO2 and brain atrophy measures in MS patients, which may also underline the NO mechanism (Encinas et al, 2005; Moncada and Bolanos, 2006; Smith and Lassmann, 2002) is a major contributor. Future MRI studies combined with molecular imaging (i.e., 15O-PET) and/or metabolic measure (i.e., N-acetyl aspartate) may help understanding the exact underlying causes of decreased oxygen metabolism in vivo in MS.

There are several limitations in the current study. First, Yv, which is measured in the lower superior sagittal sinus, is not truly underlying the whole brain but rather of the blood drainage from most cerebral cortex. The changes of Yv and CMRO2 based on superior sagittal sinus may underestimate the oxygen changes in the deep WM such as draining blood to the straight sinus. The new version of TRUST sequence (Xu et al, 2009) with positioning of imaging slice at the internal jugular vein has great potential for a true global oxygen consumption estimation in the future studies. Second, not all patients and controls (only 12 out of 30 subjects from each group) had measured heart rate and blood pressure, and no subject had end-tidal CO2 measured. These physiological indices are known to have potential to influence CBF, and hence, oxygen delivery. However, we did not find significant difference of heart rate and blood pressure between the subgroups of patients and controls (n=12). Given careful clinical history monitoring (no pulmonary and cardiovascular diseases) and the findings of no global CBF difference, as well as the pre- and during-scan instruction (asking to be awake and normal breathing), these effects should be minimal. Although the significant changes of Yv and CMRO2 found in patient group is less likely due to such physiological variations, more careful design should be warranted. Third, since T2 calibration also depends on Mhct levels, the T2 difference between the two groups could have been explained by a lower Mhct value in the MS group, compared with the controls. Unfortunately, we only have macrovascular hematocrit (Mhct, from peripheral blood) information on a subgroup of patients (n=12), but not the controls. We have conducted a calculation to see how large would the Mhct difference have to be between the two subject groups to produce the measured difference in T2 if the saturation is exactly the same in the two groups. We found that the difference would have to be 0.08 (from an average normal value of 0.42 to an average value of 0.34 in patients). The Mhct values in the MS group were 0.41±0.04 (mean±s.d.). Thus, Mhct does not appear to explain the T2 differences observed in the present study. To the best of our knowledge, no previous studies have documented abnormal Mhct levels in MS patients and MS is not known to be an anemic disease.

In conclusion, the TRUST method presented in this study provides a sensitive global measurement for quantifying oxygen metabolism in patients with MS. The results of this study highlighted in vivo the impairment of oxygen consumption in MS, which may have a profound impact on the underlying mechanism of progressive neurodegeneration that is associated with neuronal tissue cellular energy failure.


The authors thank Jian Xu from Siemens Medical System for help with sequence installation as well as Sabrina Ouki and Yongxia Zhou for the manuscript preparation.


The authors declare no conflict of interest.


This study was supported by Award Numbers R01NS029029, R01 MH084021, and R01 NS067015 from the National Institute of Health (NIH).


  • Aboul-Enein F, Lassmann H. Mitochondrial damage and histotoxic hypoxia: a pathway of tissue injury in inflammatory brain disease. Acta Neuropathol. 2005;109:49–55. [PubMed]
  • Acar G, Idiman F, Idiman E, Kirkali G, Cakmakci H, Ozakbas S. Nitric oxide as an activity marker in multiple sclerosis. J Neurol. 2003;250:588–592. [PubMed]
  • Ashburner J, Friston KJ. Voxel-based morphometry—the methods. Neuroimage. 2000;11:805–821. [PubMed]
  • Barth M, Moser E. Proton NMR relaxation times of human blood samples at 1.5 T and implications for functional MRI. Cell Mol Biol (Noisy-le-grand) 1997;43:783–791. [PubMed]
  • Beal MF. Does impairment of energy metabolism result in excitotoxic neuronal death in neurodegenerative illnesses. Ann Neurol. 1992;31:119–130. [PubMed]
  • Broom KA, Anthony DC, Blamire AM, Waters S, Styles P, Perry VH, Sibson NR. MRI reveals that early changes in cerebral blood volume precede blood-brain barrier breakdown and overt pathology in MS-like lesions in rat brain. J Cereb Blood Flow Metab. 2005;25:204–216. [PubMed]
  • Brown GC, Borutaite V. Nitric oxide inhibition of mitochondrial respiration and its role in cell death. Free Radic Biol Med. 2002;33:1440–1450. [PubMed]
  • Bryant RG, Marill K, Blackmore C, Francis C. Magnetic relaxation in blood and blood clots. Magn Reson Med. 1990;13:133–144. [PubMed]
  • Buxton RB. Interpreting oxygenation-based neuroimaging signals: the importance and the challenge of understanding brain oxygen metabolism. Front Neuroenergetics. 2010;2:8. [PMC free article] [PubMed]
  • Chiarelli PA, Bulte DP, Piechnik S, Jezzard P. Sources of systematic bias in hypercapnia-calibrated functional MRI estimation of oxygen metabolism. Neuroimage. 2007;34:35–43. [PubMed]
  • De Santi S, de Leon MJ, Rusinek H, Convit A, Tarshish CY, Roche A, Tsui WH, Kandil E, Boppana M, Daisley K, Wang GJ, Schlyer D, Fowler J. Hippocampal formation glucose metabolism and volume losses in MCI and AD. Neurobiol Aging. 2001;22:529–539. [PubMed]
  • Encinas JM, Manganas L, Enikolopov G. Nitric oxide and multiple sclerosis. Curr Neurol Neurosci Rep. 2005;5:232–238. [PubMed]
  • Ge Y, Law M, Johnson G, Herbert J, Babb JS, Mannon LJ, Grossman RI. Dynamic susceptibility contrast perfusion MR imaging of multiple sclerosis lesions: characterizing hemodynamic impairment and inflammatory activity. AJNR Am J Neuroradiol. 2005;26:1539–1547. [PubMed]
  • Ge Y, Zohrabian VM, Osa EO, Xu J, Jaggi H, Herbert J, Haacke EM, Grossman RI. Diminished visibility of cerebral venous vasculature in multiple sclerosis by susceptibility-weighted imaging at 3.0 Tesla. J Magn Reson Imaging. 2009;29:1190–1194. [PMC free article] [PubMed]
  • Golay X, Silvennoinen MJ, Zhou J, Clingman CS, Kauppinen RA, Pekar JJ, van Zij PC. Measurement of tissue oxygen extraction ratios from venous blood T(2): increased precision and validation of principle. Magn Reson Med. 2001;46:282–291. [PubMed]
  • Gomori JM, Grossman RI, Yu-Ip C, Asakura T. NMR relaxation times of blood: dependence on field strength, oxidation state, and cell integrity. J Comput Assist Tomogr. 1987;11:684–690. [PubMed]
  • Guyton AC, Hall JE. 2005. Respiration Testbook of Medical Physiology(Guyton AC, Hall JE, eds),Saunders: Elsevier
  • Inglese M, Park SJ, Johnson G, Babb JS, Miles L, Jaggi H, Herbert J, Grossman RI. Deep gray matter perfusion in multiple sclerosis: dynamic susceptibility contrast perfusion magnetic resonance imaging at 3 T. Arch Neurol. 2007;64:196–202. [PubMed]
  • Law M, Saindane AM, Ge Y, Babb JS, Johnson G, Mannon LJ, Herbert J, Grossman RI. Microvascular abnormality in relapsing-remitting multiple sclerosis: perfusion MR imaging findings in normal-appearing white matter. Radiology. 2004;231:645–652. [PubMed]
  • Leontiev O, Buxton RB. Reproducibility of BOLD, perfusion, and CMRO2 measurements with calibrated-BOLD fMRI. Neuroimage. 2007;35:175–184. [PMC free article] [PubMed]
  • Lisak RP. 2007. Neurodegeneration in multiple sclerosis: defining the problem Neurology 68S5–S12.S12discussion S43–S54 [PubMed]
  • Lu H, Clingman C, Golay X, van Zijl PC. Determining the longitudinal relaxation time (T1) of blood at 3.0 Tesla. Magn Reson Med. 2004;52:679–682. [PubMed]
  • Lu H, Ge Y. Quantitative evaluation of oxygenation in venous vessels using T2-relaxation-under-spin-tagging MRI. Magn Reson Med. 2008;60:357–363. [PMC free article] [PubMed]
  • Lu H, Xu F, Grgac K, Liu P, Qin Q, van Zijl P. Calibration and validation of TRUST MRI for the estimation of cerebral blood oxygenation. Magn Reson Med. 2011;67:42–49. [PMC free article] [PubMed]
  • Lu H, Zhao C, Ge Y, Lewis-Amezcua K. Baseline blood oxygenation modulates response amplitude: physiologic basis for intersubject variations in functional MRI signals. Magn Reson Med. 2008;60:364–372. [PMC free article] [PubMed]
  • Moncada S, Bolanos JP. Nitric oxide, cell bioenergetics and neurodegeneration. J Neurochem. 2006;97:1676–1689. [PubMed]
  • Poser CM, Paty DW, Scheinberg L, McDonald WI, Davis FA, Ebers GC, Johnson KP, Sibley WA, Silberberg DH, Tourtellotte WW. New diagnostic criteria for multiple sclerosis: guidelines for research protocols. Ann Neurol. 1983;13:227–231. [PubMed]
  • Prinster A, Quarantelli M, Lanzillo R, Orefice G, Vacca G, Carotenuto B, Alfano B, Brunetti A, Morra VB, Salvatore M. A voxel-based morphometry study of disease severity correlates in relapsing-remitting multiple sclerosis. Mult Scler. 2010;16:45–54. [PMC free article] [PubMed]
  • Rejdak K, Eikelenboom MJ, Petzold A, Thompson EJ, Stelmasiak Z, Lazeron RH, Barkhof F, Polman CH, Uitdehaag BM, Giovannoni G. CSF nitric oxide metabolites are associated with activity and progression of multiple sclerosis. Neurology. 2004;63:1439–1445. [PubMed]
  • Smith KJ, Lassmann H. The role of nitric oxide in multiple sclerosis. Lancet Neurol. 2002;1:232–241. [PubMed]
  • Su KG, Banker G, Bourdette D, Forte M. Axonal degeneration in multiple sclerosis: the mitochondrial hypothesis. Curr Neurol Neurosci Rep. 2009;9:411–417. [PMC free article] [PubMed]
  • Sun X, Tanaka M, Kondo S, Okamoto K, Hirai S. Clinical significance of reduced cerebral metabolism in multiple sclerosis: a combined PET and MRI study. Ann Nucl Med. 1998;12:89–94. [PubMed]
  • Thulborn KR, Waterton JC, Matthews PM, Radda GK. Oxygenation dependence of the transverse relaxation time of water protons in whole blood at high field. Biochim Biophys Acta. 1982;714:265–270. [PubMed]
  • Trapp BD, Ransohoff R, Rudick R. Axonal pathology in multiple sclerosis: relationship to neurologic disability. Curr Opin Neurol. 1999;12:295–302. [PubMed]
  • Trapp BD, Stys PK. Virtual hypoxia and chronic necrosis of demyelinated axons in multiple sclerosis. Lancet Neurol. 2009;8:280–291. [PubMed]
  • Wong EC, Buxton RB, Frank LR. Implementation of quantitative perfusion imaging techniques for functional brain mapping using pulsed arterial spin labeling. NMR Biomed. 1997;10:237–249. [PubMed]
  • Wright GA, Hu BS, Macovski A. 1991 I.I. Rabi Award. Estimating oxygen saturation of blood in vivo with MR imaging at 1.5 T. J Magn Reson Imaging. 1991;1:275–283. [PubMed]
  • Wuerfel J, Bellmann-Strobl J, Brunecker P, Aktas O, McFarland H, Villringer A, Zipp F. Changes in cerebral perfusion precede plaque formation in multiple sclerosis: a longitudinal perfusion MRI study. Brain. 2004;127:111–119. [PubMed]
  • Xu F, Ge Y, Lu H. Noninvasive quantification of whole-brain cerebral metabolic rate of oxygen (CMRO2) by MRI. Magn Reson Med. 2009;62:141–148. [PMC free article] [PubMed]
  • Zhao JM, Clingman CS, Narvainen MJ, Kauppinen RA, van Zijl PC. Oxygenation and hematocrit dependence of transverse relaxation rates of blood at 3T. Magn Reson Med. 2007;58:592–597. [PubMed]

Articles from Journal of Cerebral Blood Flow & Metabolism are provided here courtesy of SAGE Publications