[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 23.23.54.109. Please contact the publisher to request reinstatement.
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Download PDF
Figure 1.
Mean Gray Matter CVR Maps of Patients With Multiple Sclerosis and Controls
Mean Gray Matter CVR Maps of Patients With Multiple Sclerosis and Controls

The mean gray matter cerebrovascular reactivity (CVR) maps of the control group (A) and multiple sclerosis group (B) show the diffuse decrease of CVR in the patients with multiple sclerosis. The color bar shows the range of CVR values (calculated as the percent increase of normocapnic to hypercapnic cerebral blood flow normalized by the change in end-tidal carbon dioxide).

Figure 2.
Comparison of Global and Regional CVR Values for Patients With MS and Controls
Comparison of Global and Regional CVR Values for Patients With MS and Controls

A, Comparison of mean global gray matter (GM) cerebrovascular reactivity (CVR) values (calculated as the percent increase of normocapnic to hypercapnic cerebral blood flow normalized by the change in end-tidal carbon dioxide) between the multiple sclerosis (MS) group and the control group showing a significant decrease in CVR in patients compared with controls. Error bars indicate standard deviation. B, Surface rendering of voxel-based map on Montreal Neurological Institute brain template of significantly decreased CVR in the MS group vs control group (corrected P < .05) is shown in red, as generated by Statistical Parametric Mapping. The 6 surface rendering views shown are anterior/posterior (top row), lateral (middle row), and bottom/top (bottom row).

Figure 3.
GM CVR Correlations With Lesion Volume and Atrophy
GM CVR Correlations With Lesion Volume and Atrophy

Correlation of gray matter (GM) cerebrovascular reactivity (CVR; calculated as the percent increase of normocapnic to hypercapnic cerebral blood flow normalized by the change in end-tidal carbon dioxide) in patients with multiple sclerosis with lesion volume (R = 0.6, P = .004) (A) and with fractional GM (fGM) atrophy index (R = 0.5, P = .03) (B).

Table 1.  
Summary of CBF, ETCO2, and CVR Values Among Patients With MS and Controlsa
Summary of CBF, ETCO2, and CVR Values Among Patients With MS and Controlsa
Table 2.  
Regions With Significantly Decreased CVR in Patients With MSa
Regions With Significantly Decreased CVR in Patients With MSa
1.
Brown  GC, Borutaite  V.  Nitric oxide inhibition of mitochondrial respiration and its role in cell death. Free Radic Biol Med. 2002;33(11):1440-1450.
PubMedArticle
2.
Su  KG, Banker  G, Bourdette  D, Forte  M.  Axonal degeneration in multiple sclerosis: the mitochondrial hypothesis. Curr Neurol Neurosci Rep. 2009;9(5):411-417.
PubMedArticle
3.
Tzourio-Mazoyer  N, Landeau  B, Papathanassiou  D,  et al.  Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain. Neuroimage. 2002;15(1):273-289.
PubMedArticle
4.
Zhang  Y, Brady  M, Smith  S.  Segmentation of brain MR images through a hidden Markov random field model and the expectation-maximization algorithm. IEEE Trans Med Imaging. 2001;20(1):45-57.
PubMedArticle
5.
Lu  H, Golay  X, Pekar  JJ, Van Zijl  PC.  Sustained poststimulus elevation in cerebral oxygen utilization after vascular recovery. J Cereb Blood Flow Metab. 2004;24(7):764-770.
PubMedArticle
6.
Tancredi  FB, Gauthier  CJ, Madjar  C,  et al.  Comparison of pulsed and pseudocontinuous arterial spin-labeling for measuring CO2 -induced cerebrovascular reactivity. J Magn Reson Imaging. 2012;36(2):312-321.
PubMedArticle
7.
Trapp  BD, Stys  PK.  Virtual hypoxia and chronic necrosis of demyelinated axons in multiple sclerosis. Lancet Neurol. 2009;8(3):280-291.
PubMedArticle
8.
Juurlink  BH.  The evidence for hypoperfusion as a factor in multiple sclerosis lesion development. Mult Scler Int. 2013;2013:598093.
PubMed
9.
Wakefield  AJ, More  LJ, Difford  J, McLaughlin  JE.  Immunohistochemical study of vascular injury in acute multiple sclerosis. J Clin Pathol. 1994;47(2):129-133.
PubMedArticle
10.
Wuerfel  J, Bellmann-Strobl  J, Brunecker  P,  et al.  Changes in cerebral perfusion precede plaque formation in multiple sclerosis: a longitudinal perfusion MRI study. Brain. 2004;127(pt 1):111-119.
PubMedArticle
11.
Law  M, Saindane  AM, Ge  Y,  et al.  Microvascular abnormality in relapsing-remitting multiple sclerosis: perfusion MR imaging findings in normal-appearing white matter. Radiology. 2004;231(3):645-652.
PubMedArticle
12.
Ge  Y, Law  M, Johnson  G,  et al.  Dynamic susceptibility contrast perfusion MR imaging of multiple sclerosis lesions: characterizing hemodynamic impairment and inflammatory activity. AJNR Am J Neuroradiol. 2005;26(6):1539-1547.
PubMed
13.
Inglese  M, Park  SJ, Johnson  G,  et al.  Deep gray matter perfusion in multiple sclerosis: dynamic susceptibility contrast perfusion magnetic resonance imaging at 3 T. Arch Neurol. 2007;64(2):196-202.
PubMedArticle
14.
De Keyser  J, Steen  C, Mostert  JP, Koch  MW.  Hypoperfusion of the cerebral white matter in multiple sclerosis: possible mechanisms and pathophysiological significance. J Cereb Blood Flow Metab. 2008;28(10):1645-1651.
PubMedArticle
15.
Bokkers  RP, van Osch  MJ, Klijn  CJ, Kappelle  LJ, Hendrikse  J.  Cerebrovascular reactivity within perfusion territories in patients with an internal carotid artery occlusion. J Neurol Neurosurg Psychiatry. 2011;82(9):1011-1016.
PubMedArticle
16.
Chen  Y, Parrish  TB.  Caffeine dose effect on activation-induced BOLD and CBF responses. Neuroimage. 2009;46(3):577-583.
PubMedArticle
17.
Yezhuvath  US, Uh  J, Cheng  Y,  et al.  Forebrain-dominant deficit in cerebrovascular reactivity in Alzheimer’s disease. Neurobiol Aging. 2012;33(1):75-82.
PubMedArticle
18.
Krainik  A, Hund-Georgiadis  M, Zysset  S, von Cramon  DY.  Regional impairment of cerebrovascular reactivity and BOLD signal in adults after stroke. Stroke. 2005;36(6):1146-1152.
PubMedArticle
19.
Wu  WC, Fernández-Seara  M, Detre  JA, Wehrli  FW, Wang  J.  A theoretical and experimental investigation of the tagging efficiency of pseudocontinuous arterial spin labeling. Magn Reson Med. 2007;58(5):1020-1027.
PubMedArticle
20.
Wang  Z, Aguirre  GK, Rao  H,  et al.  Empirical optimization of ASL data analysis using an ASL data processing toolbox: ASLtbx. Magn Reson Imaging. 2008;26(2):261-269.
PubMedArticle
21.
Chen  Y, Wang  DJ, Detre  JA.  Test-retest reliability of arterial spin labeling with common labeling strategies. J Magn Reson Imaging. 2011;33(4):940-949.
PubMedArticle
22.
Aslan  S, Xu  F, Wang  PL,  et al.  Estimation of labeling efficiency in pseudocontinuous arterial spin labeling. Magn Reson Med. 2010;63(3):765-771.
PubMedArticle
23.
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(3):679-682.
PubMedArticle
24.
Johnson  NA, Jahng  GH, Weiner  MW,  et al.  Pattern of cerebral hypoperfusion in Alzheimer disease and mild cognitive impairment measured with arterial spin-labeling MR imaging: initial experience. Radiology. 2005;234(3):851-859.
PubMedArticle
25.
Mikheev  A, Nevsky  G, Govindan  S, Grossman  R, Rusinek  H.  Fully automatic segmentation of the brain from T1-weighted MRI using Bridge Burner algorithm. J Magn Reson Imaging. 2008;27(6):1235-1241.
PubMedArticle
26.
Smith  SM.  Fast robust automated brain extraction. Hum Brain Mapp. 2002;17(3):143-155.
PubMedArticle
27.
Lancaster  JL, Woldorff  MG, Parsons  LM,  et al.  Automated Talairach atlas labels for functional brain mapping. Hum Brain Mapp. 2000;10(3):120-131.
PubMedArticle
28.
Maldjian  JA, Laurienti  PJ, Kraft  RA, Burdette  JH.  An automated method for neuroanatomic and cytoarchitectonic atlas-based interrogation of fMRI data sets. Neuroimage. 2003;19(3):1233-1239.
PubMedArticle
29.
Acar  G, Idiman  F, Idiman  E, Kirkali  G, Cakmakçi  H, Ozakbaş  S.  Nitric oxide as an activity marker in multiple sclerosis. J Neurol. 2003;250(5):588-592.
PubMedArticle
30.
Sicard  KM, Duong  TQ.  Effects of hypoxia, hyperoxia, and hypercapnia on baseline and stimulus-evoked BOLD, CBF, and CMRO2 in spontaneously breathing animals. Neuroimage. 2005;25(3):850-858.
PubMedArticle
31.
Dai  W, Garcia  D, de Bazelaire  C, Alsop  DC.  Continuous flow-driven inversion for arterial spin labeling using pulsed radio frequency and gradient fields. Magn Reson Med. 2008;60(6):1488-1497.
PubMedArticle
32.
van Gelderen  P, de Zwart  JA, Duyn  JH.  Pittfalls of MRI measurement of white matter perfusion based on arterial spin labeling. Magn Reson Med. 2008;59(4):788-795.
PubMedArticle
33.
Jo  A, Do  H, Jhon  GJ, Suh  M, Lee  Y.  Real-time evaluation of nitric oxide (NO) levels in cortical and hippocampal areas with a nanopore-based electrochemical NO sensor. Neurosci Lett. 2011;498(1):22-25.
PubMedArticle
34.
Fernández-Seara  MA, Techawiboonwong  A, Detre  JA, Wehrli  FW.  MR susceptometry for measuring global brain oxygen extraction. Magn Reson Med. 2006;55(5):967-973.
PubMedArticle
35.
van Osch  MJ, Teeuwisse  WM, van Walderveen  MA, Hendrikse  J, Kies  DA, van Buchem  MA.  Can arterial spin labeling detect white matter perfusion signal? Magn Reson Med. 2009;62(1):165-173.
PubMedArticle
36.
Rashid  W, Parkes  LM, Ingle  GT,  et al.  Abnormalities of cerebral perfusion in multiple sclerosis. J Neurol Neurosurg Psychiatry. 2004;75(9):1288-1293.
PubMedArticle
37.
Lu  H, Xu  F, Rodrigue  KM,  et al.  Alterations in cerebral metabolic rate and blood supply across the adult lifespan. Cereb Cortex. 2011;21(6):1426-1434.
PubMedArticle
38.
Prinster  A, Quarantelli  M, Orefice  G,  et al.  Grey matter loss in relapsing-remitting multiple sclerosis: a voxel-based morphometry study. Neuroimage. 2006;29(3):859-867.
PubMedArticle
39.
Achiron  A, Chapman  J, Tal  S, Bercovich  E, Gil  H, Achiron  A.  Superior temporal gyrus thickness correlates with cognitive performance in multiple sclerosis. Brain Struct Funct. 2013;218(4):943-950.
PubMedArticle
40.
Kollndorfer  K, Krajnik  J, Woitek  R, Freiherr  J, Prayer  D, Schöpf  V.  Altered likelihood of brain activation in attention and working memory networks in patients with multiple sclerosis: an ALE meta-analysis. Neurosci Biobehav Rev. 2013;37(10, pt 2):2699-2708.
PubMedArticle
Original Investigation
October 2014

Impaired Cerebrovascular Reactivity in Multiple Sclerosis

Author Affiliations
  • 1Department of Radiology, Center for Biomedical Imaging, New York University School of Medicine, New York
  • 2Advanced Imaging Research Center, University of Texas Southwestern Medical Center, Dallas
  • 3Department of Neurology, Multiple Sclerosis Comprehensive Care Center, New York University School of Medicine, New York
JAMA Neurol. 2014;71(10):1275-1281. doi:10.1001/jamaneurol.2014.1668
Abstract

Importance  Cerebrovascular reactivity (CVR) is an inherent indicator of the dilatory capacity of cerebral arterioles for a vasomotor stimulus for maintaining a spontaneous and instant increase of cerebral blood flow (CBF) in response to neural activation. The integrity of this mechanism is essential to preserving healthy neurovascular coupling; however, to our knowledge, no studies have investigated whether there are CVR abnormalities in multiple sclerosis (MS).

Objective  To use hypercapnic perfusion magnetic resonance imaging to assess CVR impairment in patients with MS.

Design, Setting, and Participants  A total of 19 healthy volunteers and 19 patients with MS underwent perfusion magnetic resonance imaging based on pseudocontinuous arterial spin labeling to measure CBF at normocapnia (ie, breathing room air) and hypercapnia. The hypercapnia condition is achieved by breathing 5% carbon dioxide gas mixture, which is a potent vasodilator causing an increase of CBF.

Main Outcomes and Measures  Cerebrovascular reactivity was calculated as the percent increase of normocapnic to hypercapnic CBF normalized by the change in end-tidal carbon dioxide, which was recorded during both conditions. Group analysis was performed for regional and global CVR comparison between patients and controls. Regression analysis was also performed between CVR values, lesion load, and brain atrophy measures in patients with MS.

Results  A significant decrease of mean (SD) global gray matter CVR was found in patients with MS (3.56 [0.81]) compared with healthy controls (5.08 [1.56]; P = .001). Voxel-by-voxel analysis showed diffuse reduction of CVR in multiple regions of patients with MS. There was a significant negative correlation between gray matter CVR and lesion volume (R = 0.6, P = .004) and a significant positive correlation between global gray matter CVR and gray matter atrophy index (R = 0.5, P = .03).

Conclusions and Relevance  Our quantitative imaging findings suggest impairment in functional cerebrovascular pathophysiology, by measuring a diffuse decrease in CVR, which may be the underlying cause of neurodegeneration in MS.

Introduction

Multiple sclerosis (MS) is considered a chronic autoimmune inflammatory disease. One of the hallmarks, however, is progressive neurodegeneration, which plays a key role in the advancement of neurological disabilities. Little is known of the link between neuroinflammation and neurodegeneration of the disease. Previous biochemical studies1,2 have shown elevated levels of nitric oxide in patients with MS produced by the activation of nitric oxide synthase secondary to repetitive proinflammatory cascades. Nitric oxide has a wide range of biological functions, one of which is a strong vasodilator that can manipulate cerebral blood flow (CBF), thus mediating neurovascular coupling that is responsible for increased blood supply during transient neural activation.3,4 In healthy volunteers, studies5,6 reported an immense increase of blood flow in the visual cortex in response to visual stimuli (an increase of more than 60%-80% compared with the resting state), indicating that intact neurovascular coupling of the brain is essential for instant and rapid oxygen delivery. However, for patients with MS, chronically high levels of nitric oxide may have detrimental effects on the vascular health of the brain; its overproduction may desensitize endothelial and smooth muscle function (vascular habituation), causing decreased vasodilatory capacity and limited blood supply for neurons that perform demanding tasks; thus, this overproduction may create neuronal activity–induced hypoxia.

Several histopathological studies7,8 have elucidated hypoxia-like tissue injury in patients with MS, which may help us better understand its pathogenetic role in lesion formation and neurodegeneration. This is likely an underlying cause of progressive neurodegeneration secondary to neuroinflammation in MS. Cerebrovascular reactivity (CVR), a useful index of cerebral vascular function, can be used to measure the impaired hemodynamic response to vasodilator stimuli. Although perfusion measurements (eg, CBF and cerebral blood volume) have been reported in numerous studies,914 vasoreactivity has not yet been investigated in patients with MS.

A number of vasoactive stimuli have been explored for CVR assessment, including caffeine, carbon dioxide (CO2), and injection of acetazolamide.1517 CO2 is the most commonly used and most appropriate vasodilator for CVR measurement owing to its practicality, reliability, and ability to be standardized. Using a fixed CO2 gas mixture (5% by volume), previous studies have reported CVR impairment using blood oxygen level–dependent functional magnetic resonance imaging (MRI) for several neurological diseases (eg, Alzheimer disease17 and stroke18). However, blood oxygen level–dependent functional MRI measures CBF changes indirectly because it relies heavily on the blood oxygenation level. The modulation of physiological blood flow can be evaluated using arterial spin labeling (ASL) perfusion MRI, which measures CBF noninvasively by magnetically labeling arterial blood water as an endogenous tracer. A few studies have shown that ASL-based techniques can reliably measure CVR (see, eg, Tancredi et al6).

Our hypothesis is that patients with MS have significant abnormalities in CVR that may be due to the presence of chronically high nitric oxide levels, and such vascular hemodynamic response deficits may be a significant underlying cause of diffuse and progressive neurodegeneration. In the present study, we used a robust and well-validated pseudocontinuous ASL (pCASL)19 perfusion MRI acquisition at normocapnia (room air) and at hypercapnia (5% CO2) conditions to evaluate CVR in patients with MS, compared with healthy controls, and to correlate CVR and brain atrophy measures.

Methods
Participants

Nineteen patients with clinically definite MS (17 with relapsing-remitting MS and 2 with secondary progressive MS), with a mean (SD) age of 42.8 (10.2) years (10 women and 9 men), and 19 healthy control volunteers, with a mean [SD] age of 39.6 [12.9] years (6 women and 13 men), were included in our study. The protocol was approved by the institutional review board and research ethics committee at New York University School of Medicine and NYU Langone Medical Center. After receiving an explanation of the study protocol, all participants provided written informed consent and were financially compensated for participation. Patients and controls with a prior diagnosis of cardiovascular, pulmonary, cerebrovascular, or other neurological disease, as well as active smokers, were excluded. All patients had varied symptoms that included visual disturbances, numbness and tingling, weakness or fatigue, pain or sensitivity to heat, dizziness, bowel or bladder problems, or difficulty thinking clearly. All patients received 1 or more immunosuppressive or immunomodulatory drugs in the 3 years prior to the examination. Among them, 6 patients received natalizumab (Tysabri; Biogen Idec) and glatiramer acetate (Copaxone; Teva Neuroscience), 4 received interferon beta-1a (Rebif [EMD Serono/Pfizer] and Betaseron [Bayer HealthCare]), 3 received fingolimod (Gilenya; Novartis), 2 received interferon beta-1a (Avonex; Biogen Idec), and 1 received rituximab (Rituxan; Genentech/Biogen Idec). None of these medications are known to cause changes in vascular function. All participants were instructed to refrain from drinking coffee for at least 4 hours before the examination. The patients had a mean (SD) disease duration of 10.4 (7.6) years (range, 2.2-24 years) and a mean (SD) Expanded Disability Status Scale score of 2.9 (1.5) (range, 1-6).

Magnetic Resonance Imaging

The MRI data were obtained by use of a 3-T whole-body MR scanner (Siemens Tim Trio; Siemens Healthcare) with a 12-channel head coil, following our standard protocol for intracranial lesion assessment. This includes nonenhanced T1-weighted (repetition time/echo time: 600/14 milliseconds), fluid-attenuated inversion recovery images (repetition time/echo time: 9010/134 milliseconds; inversion time: 2500 milliseconds) and T2-weighted images (repetition time/echo time: 3400/119 milliseconds) covering the whole brain. Postcontrast T1-weighted imaging (repetition time/echo time: 600/14 milliseconds) was acquired to evaluate enhancing lesions. In addition to conventional routine sequences, CBF was measured using pCASL perfusion MRI,19 based on single-shot gradient-echo echo-planar imaging with the following parameters: repetition time/echo time: 3950/17 milliseconds; labeling duration:1470 milliseconds; postlabeling delay: 1230 milliseconds; slice thickness: 5 mm; 32 transverse slices; field of view: 22 cm; 64 × 64 in-plane matrix; and a total of 52 measurements with 26 pairs of label and control images. Labeling was performed at 97 mm below the center-of-imaging volume such that it was approximately perpendicular to the internal carotid and vertebral arteries. We used a generalized autocalibrating, partially parallel acquisition factor of 2 with a scan duration of 3 minutes and 35 seconds.

Two pCASL sequences were recorded, the first sequence under normocapnia, in which the patient breathed room air, and the next sequence under hypercapnia, in which the patient breathed a 5% CO2 gas mixture (5% CO2, 21% oxygen gas, and 74% nitrogen gas) from a Douglas bag. The participants were tightly fitted with a nose clip to restrict breathing to the mouth. During both sequences, end-tidal CO2 (ETCO2) was monitored and measured at a 2-second sampling rate, using the MR-compatible Medrad 9500 MRI Multigas Monitor. After the normocapnia scan, the hypercapnia gas was switched on, and 1-minute, 18-second, T1-weighted anatomical imaging (repetition time/echo time: 320/2.42 milliseconds; voxel resolution size: 0.9 × 0.9 × 1 mm3) with identical imaging orientation, resolution, and slice number as the pCASL was performed for optimal image registration. This time permitted ETCO2 to reach a new steady state, allowing the second pCASL sequence to be run under the hypercapnia condition. Three-dimensional magnetization-prepared rapid acquisition with gradient echo (MPRAGE), high-resolution, T1-weighted imaging was performed for brain tissue segmentation (repetition time/echo time: 2300/2.98 milliseconds; inversion time: 900 milliseconds; voxel resolution size: 1 × 1 × 1 mm3).

Image Processing and Evaluation

The processing of images was conducted using Statistical Parametric Mapping (University College London) software in order to perform motion correction for both pCASL sequences and in order to register the geometrically matched 2-dimensional spin-echo T1-weighted images and 3-dimensional high-resolution MPRAGE anatomical images.20 For both pCASL sequences, the difference between the average labeled image and the control image was corrected for the differences in the acquisition delay time of the slices. The CBF maps were generated for each condition by conducting CBF calibration20,21 usingImage description not available.where ΔM is the difference signal between the control and labeling states; λ = 0.9 g/mL, which is the blood/tissue water partition coefficient; α = 0.86, which is used empirically for the inversion tagging efficiency of pCASL at 3 T19,22; M0 is the equilibrium magnetization of brain tissue determined from a whole-brain region of interest during the control condition, after accounting for T1 relaxation; T1 = 1.6 seconds, which is the conventional T1 constant of blood at 3 T23; w is the postlabeling delay, which is different for individual slices (1.23 seconds + slice acquisition delay)23; and τ is the labeling duration (1.47 seconds in our data).

The resulting maps were subsequently transformed into the Montreal Neurological Institute template of 52 space-masking brain-only regions and were spatially smoothed using a gaussian kernel (full-width at half-maximum = 8 mm). A CBF difference map (hypercapnia − normocapnia) was divided at each voxel by the average gray matter (GM) partial volume–adjusted normocapnia CBF to reflect the CBF percent increase. This was performed by assuming that GM perfusion is 2.5 times white matter (WM) perfusion per unit volume.24 These maps were normalized by individual ETCO2 increases, to adjust for participant-specific variation, resulting in CVR maps. Thus, CVR (calculated as the percent increase of normocapnic to hypercapnic CBF normalized by the change in ETCO2) represents the percent change of CBF per unit of ETCO2 (mm Hg) change. Global GM and WM CBF values were computed by overlaying the tissue mask (defined as >70% probability of being the tissue type) on the normalized CBF maps and calculating the mean value within. These values were used to calculate global CVR according toImage description not available.FireVoxel software (https://files.nyu.edu/hr18/public/)25 was used for semiautomatic MS lesion load quantification, after intensity uniformity correction of the fluid-attenuated inversion recovery images. Lesions were visually inspected to correct errors. The brain was segmented into GM, WM, and cerebrospinal fluid (CSF) regions based on the 3-dimensional, high-resolution, T1-weighted MPRAGE images using the Brain Extraction Tool and the Automated Segmentation Tool from the FMRIB (Functional Magnetic Resonance Imaging of the Brain) Software Library.4,26 Results were visually inspected and manually corrected. The fractional brain parenchymal volume, an index of brain atrophy, was computed as the ratio of brain parenchymal volume (GM and WM) to total intracranial volume (GM, WM, and CSF). Fractional GM (fGM) volume was computed as the ratio of GM volume to total intracranial volume.

Statistical Analysis

The CVR map for each participant was used for voxel-based regional analyses between patients and controls. Group statistics were calculated using the Statistical Parametric Mapping software on voxel-by-voxel bases. A 2-sample t test was used to identify clusters with significantly different CVR values in the MS group compared with the control group, set at a familywise error–corrected threshold of P < .05. Voxels were automatically labeled using the preparcelled Montreal Neurological Institute brain template and the PickAtlas software.27,28 Global CVR values of the patients and controls were compared using a 2-sample t test. For regression analysis, the Pearson correlations between GM CVR and lesion load, fractional brain parenchymal volume (ie, brain atrophy index), Expanded Disability Status Scale score, and disease duration were evaluated for patients with MS.

Results

All participants completed the scan comfortably. Compared with the control group, the MS group showed a significant decrease of global GM CVR (Figure 1). Global GM CBF and mean ETCO2 values are listed in Table 1, with no significant difference between the MS group and the control group shown for either measure at normocapnia or hypercapnia. There was a significantly smaller mean CBF percent increase (31.7%) from normocapnia to hypercapnia in the MS group than in the control group (44.8%) (P = .01). However, the ETCO2 increase was almost identical for both groups (9.2 mm Hg for the MS group vs 8.9 mm Hg for the control group; P = .73). Figure 2 presents the mean (SD) GM CVR values in the control group (5.08 [1.56]) and the MS group (3.56 [0.81]) measured as the percent change of CBF per unit change of ETCO2 (P = .001) (Figure 2A) and the widespread CVR deficit distribution along the cortical surface between the patients with MS and the controls (corrected P < .05) as a result of a voxel-by-voxel comparison of CVR maps (Figure 2B). The mean (SD) global WM CVR was 6.1 (3.0) for the control group and 4.0 (1.2) for the MS group, showing a significant difference (P = .007). Further analysis was restricted to GM because the pCASL signal is less reliable in WM.

Quantitative analysis of the voxel-based group comparison shows a significant reduction in CVR in more than 10% of the GM volume of the temporal, parietal, sublobar, and limbic regions and in less than 10% of the GM volume of the frontal and occipital lobes. Specifically, GM CVR deficits are seen in more than 25% of the superior temporal gyri, supramarginal gyri, rolandic opercula, Heschl gyri, anterior cingulate gyri, and lenticular nuclei. Table 2 summarizes the lobes and contained subregions where more than 10% of GM volume shows a significant decrease in CVR.

The mean (SD) fractional brain parenchymal volume was decreased in the MS group compared with the control group (0.79% [0.04%] vs 0.82% [0.02%]; P = .02), as was the mean (SD) fGM volume (0.41% [0.03%] vs 0.44% [0.02%]; P = .008). The mean (SD) lesion size in the MS group is 14.3 (12.7) cm3. Figure 3 shows the significant negative correlation of GM CVR with lesion volume (R = 0.6, P = .004), as well as the significant positive correlation of GM CVR with the fGM atrophy index (R = 0.5, P = .03). There was no significant correlation between GM CVR and Expanded Disability Status Scale score or disease duration. No enhanced lesions were visible in the postcontrast T1-weighted image.

Discussion

Neurovascular coupling is a complex and important physiologic mechanism during which neuronal activation leads to increased local cortical blood flow resulting from activity-induced vasodilation. This prompt response is critical for maintaining normal brain function by fulfilling instantaneous energy demand, through oxygen and glucose delivery. Such an intrinsic mechanism can be estimated with CVR, which defines the ability to change CBF in response to vasodilatory stimuli. In our study, we found that global GM CVR (by CO2 challenge) significantly decreases and correlates with brain atrophy and lesion volume in patients with MS. The pathophysiological mechanism of such deficits of vascular hemodynamic response in MS may be best explained by chronically high concentrations of nitric oxide (a potent vasodilator causing vascular habituation) secondary to repetitive vascular inflammatory cascades,29 as opposed to other postinflammatory by-products. Over time, inadequate blood flow to neurons due to impaired CVR can cause a chronic state of hypoxia leading to neurodegeneration, especially affecting otherwise healthy neurons. We have provided the first direct evidence that CVR, which is essential for neurovascular coupling, is impaired in MS and may be accountable for neurodegenerative changes often seen in patients with MS.

There is abundant evidence showing higher levels of nitric oxide (and its nitroderivatives in CSF) in patients with MS.29 However, nitric oxide has a very short lifetime and is difficult to use clinically for CVR assessment. Alternatively, mild CO2 mixtures (5% CO2, 21% oxygen gas, and 74% nitrogen gas) are often used as a safe and reliable vasodilator for CVR measures. Increased CO2 tension in blood (referred to as hypercapnia) is known to cause an increase in CBF.30 As performed in our study, comparing CBF at normocapnia (breathing room air) and hypercapnia (breathing 5% CO2) allows for a direct estimation of CVR. Compared with regular continuous or pulsed ASL, the recently developed pCASL sequence is optimized for a better signal to noise ratio, as well as high tagging efficiency within a reasonable short scan time.19,31

A significant drawback of the ASL-type technique is its larger variations of CBF values in WM, which are due to the poor sensitivity and heterogeneous transit time, than in GM.32 Therefore, we do not consider WM CVR changes as reliable, but reduced CVR is expected because it is accepted that, in MS, inflammatory activities with lesion formation are more prominent in WM than in GM. Because nitric oxide is a highly lipophilic and diffusible gas, it permeates biological membranes, reaching distances far beyond where it was generated.33 In light of this, the chronic elevation of nitric oxide is expected to affect the vasculature reserve function found in the GM throughout the brain, which is shown by our finding of a diffuse decrease in CVR. Our results favor the notion that, in MS, atrophy is more prominently seen in GM than in WM because GM has a higher vascular density; thus, the CVR compromise may be more pronounced in GM than in WM. However, with several recent improvements in ASL, including 3-dimensional gradient- and spin-echo readout schemes34 and background suppression techniques,35 a more accurate characterization of WM CVR may be possible in the near future.

In the present study, neither of the CBF conditions show significant differences in the patients with MS compared with controls; however, previous literature has reported both increased and decreased perfusion in patients. More regions are likely associated with underlying inflammatory activities,10,12,36 and regions with decreased perfusion may be due to either decreased neuronal activity (demand) or vasoreactivity. Our study only investigates global mean CBF values because regional CBF is not reliable with current ASL techniques owing to low signal to noise ratio. The CVR of the patients with MS, however, is significantly lower than the CVR of the controls. This is likely driven by the significantly smaller CBF increase in the MS group than in the control group when both groups were exposed to hypercapnia, as opposed to the ETCO2 increase, which is similar in both groups. There is an imbalance in the number of male and female participants in the control group; however, in a study of 152 healthy controls, Lu et al37 show that no sex effects were observed in CVR.

We show a globally diffuse, significant decrease in CVR in the MS group, which can indicate a widespread impairment in blood flow regulation or neurovascular coupling. We hypothesize that such impairment can lead to inefficient blood supply on demand and eventually neuronal dysfunction and death, which supports the increasingly recognized evidence of widespread neurodegeneration in MS. In our study, the GM CVR deficits found in certain cortical areas are consistent with structural changes38 shown in previous studies39,40 and are likely responsible for the common symptoms experienced by patients with MS. For example, decreased GM volume and thickness have been shown in the superior temporal gyrus of patients with MS and have been a suggested marker of cognitive impairment related to attention and information speed processing.38,39 We report that 43.7% of this region has significantly decreased CVR, which could be the underlying cause of GM thinning. Further studies are needed to explore the correlation between symptoms, structural changes, and CVR deficits; however, we provide evidence that CVR may play an important role in initiating these changes that lead to clinical symptoms.

Cerebrovascular reactivity is an important clinical measure because it assesses vascular health and therefore may indicate overall neuronal health. It is a strong predictor of tissue at risk and can guide medical treatment of many disorders, such as stroke.15 For the MS group, the global GM CVR and the fGM volume are significantly correlated. This is to be expected because a decreased fGM volume indicates neurodegeneration, which can be partly due to decreased CVR in patients with MS. This significant correlation is unlikely to be attributed to the partial volume effect of CSF because similar analysis for the control group does not show a significant correlation between CVR and fGM volume.

Conclusions

In conclusion, the direct measure of CVR with hypercapnia perfusion MRI reveals important characteristics of vascular hemodynamic function, and its impairment in patients with MS is usually only detectable in in vivo studies. This technique allows for a better understanding of the underlying mechanism of neurodegeneration and the worsening of symptoms in MS.

Back to top
Article Information

Accepted for Publication: May 15, 2014.

Corresponding Author: Yulin Ge, MD, Department of Radiology, Center for Biomedical Imaging, New York University School of Medicine, 660 First Ave, 4th Floor, New York, NY 10016 (yulin.ge@nyumc.org).

Published Online: August 18, 2014. doi:10.1001/jamaneurol.2014.1668.

Author Contributions: Dr Ge had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Marshall, Lu, Grossman, Ge.

Acquisition, analysis, or interpretation of data: Marshall, Lu, Brisset, Xu, Liu, Herbert, Ge.

Drafting of the manuscript: Marshall, Lu, Brisset, Xu, Liu, Herbert, Ge.

Critical revision of the manuscript for important intellectual content: Marshall, Lu, Grossman, Ge.

Statistical analysis: Marshall, Ge.

Obtained funding: Grossman, Ge.

Administrative, technical, or material support: Marshall, Ge.

Study supervision: Ge.

Conflict of Interest Disclosures: None reported.

Funding/Support: Dr Ge was supported by grants R01 NS076588 and R01 NS029029S1 from the National Institutes of Health and by research grant 4707A1 from the National MS Society. Dr Lu was supported by grants R01 MH084021, R01 NS067015, R01 AG042753, and R21 NS078656 from the National Institutes of Health. Dr Grossman was supported by grant R01 NS029029 from the National Institutes of Health.

Role of the Sponsor: The funding agencies had no role in the design and conduct of the study; collection, management, analysis, or interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Contributions: We would like to thank Jacqueline Smith, Ilya Kister, MD, and Sanjeev Chawla, PhD, from the New York University School of Medicine for helping with data collection. We would also like to thank Danny J. Wang, PhD, from the David Geffen School of Medicine at the University of California, Los Angeles, for providing the pseudocontinuous ASL sequence. These individuals did not receive compensation from the funding sponsors.

References
1.
Brown  GC, Borutaite  V.  Nitric oxide inhibition of mitochondrial respiration and its role in cell death. Free Radic Biol Med. 2002;33(11):1440-1450.
PubMedArticle
2.
Su  KG, Banker  G, Bourdette  D, Forte  M.  Axonal degeneration in multiple sclerosis: the mitochondrial hypothesis. Curr Neurol Neurosci Rep. 2009;9(5):411-417.
PubMedArticle
3.
Tzourio-Mazoyer  N, Landeau  B, Papathanassiou  D,  et al.  Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain. Neuroimage. 2002;15(1):273-289.
PubMedArticle
4.
Zhang  Y, Brady  M, Smith  S.  Segmentation of brain MR images through a hidden Markov random field model and the expectation-maximization algorithm. IEEE Trans Med Imaging. 2001;20(1):45-57.
PubMedArticle
5.
Lu  H, Golay  X, Pekar  JJ, Van Zijl  PC.  Sustained poststimulus elevation in cerebral oxygen utilization after vascular recovery. J Cereb Blood Flow Metab. 2004;24(7):764-770.
PubMedArticle
6.
Tancredi  FB, Gauthier  CJ, Madjar  C,  et al.  Comparison of pulsed and pseudocontinuous arterial spin-labeling for measuring CO2 -induced cerebrovascular reactivity. J Magn Reson Imaging. 2012;36(2):312-321.
PubMedArticle
7.
Trapp  BD, Stys  PK.  Virtual hypoxia and chronic necrosis of demyelinated axons in multiple sclerosis. Lancet Neurol. 2009;8(3):280-291.
PubMedArticle
8.
Juurlink  BH.  The evidence for hypoperfusion as a factor in multiple sclerosis lesion development. Mult Scler Int. 2013;2013:598093.
PubMed
9.
Wakefield  AJ, More  LJ, Difford  J, McLaughlin  JE.  Immunohistochemical study of vascular injury in acute multiple sclerosis. J Clin Pathol. 1994;47(2):129-133.
PubMedArticle
10.
Wuerfel  J, Bellmann-Strobl  J, Brunecker  P,  et al.  Changes in cerebral perfusion precede plaque formation in multiple sclerosis: a longitudinal perfusion MRI study. Brain. 2004;127(pt 1):111-119.
PubMedArticle
11.
Law  M, Saindane  AM, Ge  Y,  et al.  Microvascular abnormality in relapsing-remitting multiple sclerosis: perfusion MR imaging findings in normal-appearing white matter. Radiology. 2004;231(3):645-652.
PubMedArticle
12.
Ge  Y, Law  M, Johnson  G,  et al.  Dynamic susceptibility contrast perfusion MR imaging of multiple sclerosis lesions: characterizing hemodynamic impairment and inflammatory activity. AJNR Am J Neuroradiol. 2005;26(6):1539-1547.
PubMed
13.
Inglese  M, Park  SJ, Johnson  G,  et al.  Deep gray matter perfusion in multiple sclerosis: dynamic susceptibility contrast perfusion magnetic resonance imaging at 3 T. Arch Neurol. 2007;64(2):196-202.
PubMedArticle
14.
De Keyser  J, Steen  C, Mostert  JP, Koch  MW.  Hypoperfusion of the cerebral white matter in multiple sclerosis: possible mechanisms and pathophysiological significance. J Cereb Blood Flow Metab. 2008;28(10):1645-1651.
PubMedArticle
15.
Bokkers  RP, van Osch  MJ, Klijn  CJ, Kappelle  LJ, Hendrikse  J.  Cerebrovascular reactivity within perfusion territories in patients with an internal carotid artery occlusion. J Neurol Neurosurg Psychiatry. 2011;82(9):1011-1016.
PubMedArticle
16.
Chen  Y, Parrish  TB.  Caffeine dose effect on activation-induced BOLD and CBF responses. Neuroimage. 2009;46(3):577-583.
PubMedArticle
17.
Yezhuvath  US, Uh  J, Cheng  Y,  et al.  Forebrain-dominant deficit in cerebrovascular reactivity in Alzheimer’s disease. Neurobiol Aging. 2012;33(1):75-82.
PubMedArticle
18.
Krainik  A, Hund-Georgiadis  M, Zysset  S, von Cramon  DY.  Regional impairment of cerebrovascular reactivity and BOLD signal in adults after stroke. Stroke. 2005;36(6):1146-1152.
PubMedArticle
19.
Wu  WC, Fernández-Seara  M, Detre  JA, Wehrli  FW, Wang  J.  A theoretical and experimental investigation of the tagging efficiency of pseudocontinuous arterial spin labeling. Magn Reson Med. 2007;58(5):1020-1027.
PubMedArticle
20.
Wang  Z, Aguirre  GK, Rao  H,  et al.  Empirical optimization of ASL data analysis using an ASL data processing toolbox: ASLtbx. Magn Reson Imaging. 2008;26(2):261-269.
PubMedArticle
21.
Chen  Y, Wang  DJ, Detre  JA.  Test-retest reliability of arterial spin labeling with common labeling strategies. J Magn Reson Imaging. 2011;33(4):940-949.
PubMedArticle
22.
Aslan  S, Xu  F, Wang  PL,  et al.  Estimation of labeling efficiency in pseudocontinuous arterial spin labeling. Magn Reson Med. 2010;63(3):765-771.
PubMedArticle
23.
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(3):679-682.
PubMedArticle
24.
Johnson  NA, Jahng  GH, Weiner  MW,  et al.  Pattern of cerebral hypoperfusion in Alzheimer disease and mild cognitive impairment measured with arterial spin-labeling MR imaging: initial experience. Radiology. 2005;234(3):851-859.
PubMedArticle
25.
Mikheev  A, Nevsky  G, Govindan  S, Grossman  R, Rusinek  H.  Fully automatic segmentation of the brain from T1-weighted MRI using Bridge Burner algorithm. J Magn Reson Imaging. 2008;27(6):1235-1241.
PubMedArticle
26.
Smith  SM.  Fast robust automated brain extraction. Hum Brain Mapp. 2002;17(3):143-155.
PubMedArticle
27.
Lancaster  JL, Woldorff  MG, Parsons  LM,  et al.  Automated Talairach atlas labels for functional brain mapping. Hum Brain Mapp. 2000;10(3):120-131.
PubMedArticle
28.
Maldjian  JA, Laurienti  PJ, Kraft  RA, Burdette  JH.  An automated method for neuroanatomic and cytoarchitectonic atlas-based interrogation of fMRI data sets. Neuroimage. 2003;19(3):1233-1239.
PubMedArticle
29.
Acar  G, Idiman  F, Idiman  E, Kirkali  G, Cakmakçi  H, Ozakbaş  S.  Nitric oxide as an activity marker in multiple sclerosis. J Neurol. 2003;250(5):588-592.
PubMedArticle
30.
Sicard  KM, Duong  TQ.  Effects of hypoxia, hyperoxia, and hypercapnia on baseline and stimulus-evoked BOLD, CBF, and CMRO2 in spontaneously breathing animals. Neuroimage. 2005;25(3):850-858.
PubMedArticle
31.
Dai  W, Garcia  D, de Bazelaire  C, Alsop  DC.  Continuous flow-driven inversion for arterial spin labeling using pulsed radio frequency and gradient fields. Magn Reson Med. 2008;60(6):1488-1497.
PubMedArticle
32.
van Gelderen  P, de Zwart  JA, Duyn  JH.  Pittfalls of MRI measurement of white matter perfusion based on arterial spin labeling. Magn Reson Med. 2008;59(4):788-795.
PubMedArticle
33.
Jo  A, Do  H, Jhon  GJ, Suh  M, Lee  Y.  Real-time evaluation of nitric oxide (NO) levels in cortical and hippocampal areas with a nanopore-based electrochemical NO sensor. Neurosci Lett. 2011;498(1):22-25.
PubMedArticle
34.
Fernández-Seara  MA, Techawiboonwong  A, Detre  JA, Wehrli  FW.  MR susceptometry for measuring global brain oxygen extraction. Magn Reson Med. 2006;55(5):967-973.
PubMedArticle
35.
van Osch  MJ, Teeuwisse  WM, van Walderveen  MA, Hendrikse  J, Kies  DA, van Buchem  MA.  Can arterial spin labeling detect white matter perfusion signal? Magn Reson Med. 2009;62(1):165-173.
PubMedArticle
36.
Rashid  W, Parkes  LM, Ingle  GT,  et al.  Abnormalities of cerebral perfusion in multiple sclerosis. J Neurol Neurosurg Psychiatry. 2004;75(9):1288-1293.
PubMedArticle
37.
Lu  H, Xu  F, Rodrigue  KM,  et al.  Alterations in cerebral metabolic rate and blood supply across the adult lifespan. Cereb Cortex. 2011;21(6):1426-1434.
PubMedArticle
38.
Prinster  A, Quarantelli  M, Orefice  G,  et al.  Grey matter loss in relapsing-remitting multiple sclerosis: a voxel-based morphometry study. Neuroimage. 2006;29(3):859-867.
PubMedArticle
39.
Achiron  A, Chapman  J, Tal  S, Bercovich  E, Gil  H, Achiron  A.  Superior temporal gyrus thickness correlates with cognitive performance in multiple sclerosis. Brain Struct Funct. 2013;218(4):943-950.
PubMedArticle
40.
Kollndorfer  K, Krajnik  J, Woitek  R, Freiherr  J, Prayer  D, Schöpf  V.  Altered likelihood of brain activation in attention and working memory networks in patients with multiple sclerosis: an ALE meta-analysis. Neurosci Biobehav Rev. 2013;37(10, pt 2):2699-2708.
PubMedArticle
×