Magnetic Resonance in Medicine 45:159 –163 (2001)
In Vivo Magnetization Transfer Measurements of
Experimental Spinal Cord Injury in the Rat
Paula J. Gareau,1 Lynne C. Weaver,2 and Gregory A. Dekaban3*
Magnetization transfer (MT) imaging techniques were implemented to study a clip compression model of spinal cord injury
(SCI) in the rat. The purpose of this study was to determine if the
magnetization transfer ratio (MTR) could be used to classify the
stage and severity of SCI. Two clip compression injuries were
studied: mild SCI and severe SCI. MTRs were determined for
gray matter (GM) and white matter (WM) regions and the
GM–WM contrast was determined on days 1 and 7 following
surgery. Despite differences in pathologic features of mild and
severe SCI, the GM–WM contrast did not allow discrimination
between the two degrees of severity of SCI. WM MTR allowed
differentiation of mild and severe SCI on day 1. These preliminary results suggest that WM MTR may provide an indication of the severity of injury in SCI. Magn Reson Med 45:
159 –163, 2001. © 2001 Wiley-Liss, Inc.
Key words: magnetization transfer ratio; spinal cord injury;
magnetic resonance imaging; gray matter–white matter contrast
Traumatic spinal cord injury (SCI), a common event in
North America, not only leads to a loss in mobility and
sensation, but can also result in serious disturbances in
normal organ function, cardiovascular control, and
chronic pain (1). An important current research in this
area will be the development of effective ways of monitoring the progress of a particular treatment in vivo in a
noninvasive manner over a period of time. While qualitative assessments are useful, methods to quantify changes
that occur in vivo in the injured area of the cord are
critical. MRI is the preferred modality for visualizing spinal cord injuries. In animal models of SCI, monitoring
disease progression by MRI can be linked to measurements
of spinal cord pathology.
Several studies have shown that MRI is a valuable
method for in vivo assessment of intact spinal cord and
spinal cord pathology in animal models of SCI (2–7). Only
a few groups have used MRI in vivo to study the time
course of pathological changes in rats following SCI (8 –
12). In vivo MRI of small animals, and in particular of the
spinal cord, is challenging. A major drawback is often the
long imaging times required to collect high-resolution images with suitable SNR. The use of high magnetic field
strengths and custom built gradient and RF coils offsets
some of the difficulties. Implantable RF coils have been
1
Imaging Research Laboratories, The John P. Robarts Research Institute,
London, Ontario, Canada.
2
Neurodegeneration Group, The John P. Robarts Research Institute, London,
Ontario, Canada.
3
Gene Therapy and Molecular Virology, The John P. Robarts Research Institute, London, Ontario, Canada.
*Correspondence to: Gregory A. Dekaban, Gene Therapy and Molecular
Virology, The John P. Robarts Research Institute, P.O. Box 5015, 100 Perth
Drive, London, Ontario N6A 5K8, Canada. E-mail: dekaban@rri.on.ca
Received 22 February 2000; revised 2 June 2000; accepted 20 July 2000.
© 2001 Wiley-Liss, Inc.
used to obtain high resolution, high SNR in vivo images of
rat spinal cord (8,9,13). With this solution, Narayana et al.
(9) successfully assessed gray matter-to-white matter (GM–
WM) contrast to characterize SCI in the rat. The results of
their study suggest that a loss of GM–WM contrast is
closely related to loss of cholesterol throughout cord tissue
and subsequent edema.
Alterations in the lipid components of nervous tissue,
such as cholesterol, can also be measured using magnetization transfer (MT) (14). MT contrast is obtained by applying off-resonance RF irradiation designed to preferentially saturate the immobile, or bound, protons of macromolecules. This saturates the energy level of the bound
protons, inducing energy transfer to the mobile, or free
water, protons and consequently reducing their signal intensity. The relative difference in signal intensity (SI) with
and without magnetization transfer can be easily and reliably measured as the magnetization transfer ratio (MTR)
(15).
MT has not previously been applied to study SCI in
vivo, although Berman et al. (16) showed that, in fixed
tissue from cord-injured rats, MT histogram analysis was
highly predictive of injury level and pathological events.
The goal of our study was to determine if MTRs could be
used to classify the severity and stage of SCI in a rat model
of SCI. We hypothesize that the pathology of SCI would
lead to a reduced MTR in spinal cord WM that would vary
depending on the severity and stage of injury, thus allowing the characterization of SCI in a more quantitative manner. As a corollary, we predicted that the MTR would
provide a more specific indication of spinal cord pathology than the determination of GM–WM contrast.
METHODS
Rat SCI Model
SCI was induced in Wistar rats (200 –250g) by the clip
compression method of Rivlin and Tator (17), which employs a spring-loaded modified aneurysm clip calibrated
to close with a specific force. Two clip compression injuries were used in these studies: a mild SCI induced by a
20 g compression clip (n ! 4) and a severe SCI induced by
a 50 g compression clip (n ! 5). All rats were given
preanesthesia medications of 3.5 mg/kg diazepam ip and
0.05 mg/kg atropine sc. The rats were then anesthetized by
induction with 4% halothane in 100% oxygen and maintained on 1–2% halothane throughout the surgical procedures. A dorsal laminectomy was performed and a clip
compressed the spinal cord for 1 min at the T3 segment.
For the purposes of comparison, one rat was subjected to
only the dorsal laminectomy without the clip compression
and five intact rats served as controls.
159
160
Gareau et al.
FIG. 1. Representative axial SE images of normal rat spinal cord, obtained with (a) and without (b) MT saturation pulses. ROIs (black circles), used to make MTR and contrast measurements, are indicated in the gray matter (a)
and white matter (b).
Imaging Protocol
Imaging was performed at 4.0 T on a Varian/Siemens
UNITY INOVA using a custom-built solenoid RF coil
(5 cm diameter, 4 cm length). Rats were placed in the
supine position with the epicenter of injury centered in
the RF coil. Anesthesia was induced with 2% isoflurane in
oxygen and maintained during the MRI experiment with
1% isoflurane administered into a Plexiglas chamber that
housed the RF coil. All SCI animals were scanned on days
1 and 7 following surgery. A multislice proton densityweighted spin echo (SE) sequence was used to identify the
axial spinal cord image slice to be used for subsequent MT
experiments.
MT imaging was performed with a proton densityweighted SE single-slice acquisition. The single-slice
mode was chosen to avoid incidental MT effects. The
imaging parameters were as follows: repetition time (TR)
of 5 sec, echo time (TE) of 20 msec, slice thickness of
3 mm, 5 cm field of view, and image matrix of 256 "
256. The in-plane spatial resolution was 195 #m2. The
characteristics of the saturating pulse train (pulse shape,
pulse width, average power (B1), frequency offset, number
of saturation pulses, and time between pulses) were tested
and optimized previously (18). A distilled water phantom
was included in the field of view as a reference to estimate
the direct saturation of the pulse train. The optimized
parameters were: MT saturation produced by a train of
30 15 ms Gaussian RF pulses with average B1 of 4.2 #T,
1.5 kHz off resonance. The time between pulses was
3.75 ms, resulting in a duty cycle of 80% for the saturating
RF pulse train. The mean square saturating power (Psat)
and effective flip angle ($sat), calculated as described by
Berry et al. (19), were 20.1 #T2 and 961°, respectively. The
amount of direct saturation, estimated from SI measurements in the distilled water phantom, was found to be less
than 7% for the optimized MT sequence.
Image Analysis
The MTR was calculated as:
MTR ! %1 " M s/M 0 # 100%
where M0 is the mean intensity of pixels within a region of
interest (ROI) on an image without saturation and Ms is the
mean intensity of pixels corresponding to the same ROI on
an image with saturation. ROIs were measured in GM and
WM in the spinal cord. An average MTR was determined
for each group. The GM–WM contrast was measured in the
proton density-weighted SE images obtained without saturation and calculated as the ratio of the average SI of GM
and WM in the spinal cord image slice. For SCI rats,
measurements were obtained in spinal cord sections rostral to the injury epicenter. Data are presented as the mean
plus or minus the standard error. Statistical analysis consisted of one-way ANOVA with the Student-NewmanKeuls method for multiple comparisons. Statistical significance was defined with P & 0.05.
Histopathology
After MR imaging on day 7 all rats were anesthetized with
urethane (3 g/kg ip) and perfused transcardially with
250 ml of oxygenated tissue culture medium at pH
7.4 followed by 500 ml of 4% formaldehyde fixative in
0.1 M phosphate buffer (pH 7.4). The thoracic spinal cords
were removed. Following overnight fixation, the spinal
cords were cut into horizontal sections (50 #m) on a cryostat, mounted on slides, and stained with cresyl violet
using standard procedures.
RESULTS
Magnetization Transfer Ratios
Representative axial SE images of intact rat spinal cord,
obtained with and without saturation pulses are shown in
Fig. 1a,b. The mean SNR in the pre-MT images was 61.6 for
WM and 72 for GM. The in-plane spatial resolution for all
images was 195 #m2. The average MTR for intact spinal
cord WM and GM were 45.6% ' 3.21% and 40.4% '
2.29%, respectively. The MTRs for all groups are presented in Fig. 2. The average WM MTRs measured on day
1 and 7, after mild SCI, were not significantly different
from those in intact cords.
After severe SCI, the average WM MTR measured at day
1 was significantly reduced in comparison to values in
intact cords or after mild SCI on day 1. In addition, the
WM MTRs measured at day 7 were significantly greater
than those at day 1. The average MTR for GM in all SCI rats
was not significantly different from that in intact rats.
GM–WM Contrast Measurements
The average GM–WM contrast measured in intact rat spinal cord was 1.35 ' 0.08. Proton density-weighted spinal
cord images for mild SCI and severe SCI are shown in Fig.
MT of Rat Spinal Cord Injury
FIG. 2. Average MTRs for mild and severe SCI animals measured
on days 1 and 7 after spinal cord injury. Error bars represent the
standard error of the mean. Symbols: * ! statistically significant
difference from the control value; † ! statistically significant difference from mild SCI on day 1; ● ! statistically significant difference
from severe SCI on day 1.
3. A loss in the GM–WM contrast, as compared to the
intact spinal cord (Fig. 1b) is apparent in all of the images.
The spinal cord images in the severe SCI group typically
demonstrated gross abnormalities in structure that were
not evident in the images of cord after mild SCI. For both
groups, and at each time point, the GM–WM contrast was
significantly less in injured cords than that measured in
intact spinal cord. However, no statistically significant
differences were found between groups or between measurements made on day 1 or day 7. These results are
presented in Fig. 4.
Histopathology
In Fig. 5, representative cresyl violet-stained spinal cord
sections are presented for a laminectomy control rat (5a,b),
161
FIG. 4. Average GM–WM contrast measurements for mild and severe SCI animals measured on days 1 and 7 after spinal cord injury.
Error bars represent the standard error of the mean. Asterisk !
statistically significant difference from the control value.
a mild SCI rat (5c,d) and a severe SCI rat (5e,f). For each
example a low-magnification view of the section, with the
corresponding MR image inset, and a high-magnification
view of an outlined region of interest are shown. Normal
gross morphology and cellular architecture was observed
in the laminectomy control spinal cord sections (5a,b).
Neurons appeared to be of normal size and morphology
(arrows, 5b). The MR images of the laminectomy control
spinal cord appear similar to the images obtained from
intact rats. After mild SCI, localized regions of intense
inflammation were visible under low magnification. The
inflammatory infiltrate was most intense in the GM and
appeared to expand into the WM. Under high-power magnification, the infiltrate can be seen to consist mostly of
macrophages. The absence of staining in a portion of the
cord WM (arrow) may represent the loss of cellular integ-
FIG. 3. Typical proton density-weighted spinal cord images after SCI. Mild SCI, day 1 (a); mild SCI, day 7 (b); severe SCI, day 1 (c); severe
SCI, day 7 (d).
162
Gareau et al.
FIG. 5. Representative cresyl-violet stained
spinal cord sections obtained from: (a,b) a
laminectomy control rat, (c,d) a mild SCI rat,
and (e,f) a severe SCI rat. The low magnification view of each spinal cord section in
the top row is accompanied by an inset of
the corresponding spinal cord MR image
that has been smoothed and cropped of
tissue surrounding the cord. In each of
these views a box is drawn in a region of
intense inflammation, for the SCI examples,
and in a corresponding anatomical location
for the laminectomy control example. In c a
localized region of pallor in the cresyl violetstained cord section (arrow) appears to be
represented in the analogous MR image (arrow). High-magnification views for each box
are shown in b, d, and f. The arrow in b
depicts a normal neuron, while the arrows in
d and f indicate neurons with abnormal morphology.
rity in this region. Severe SCI was characterized by cavitation and vacuolization of GM especially, which contributed to a total loss of the normal gross architecture and
GM–WM discrimination in the histologic section. Inflammation after severe SCI was more widespread than after
mild SCI. The pathologic features of severe SCI are reflected in the MR image in which compression of the
spinal cord is visible as well as a complete loss of GM–WM
contrast. In both the mild and severe SCI, neurons in areas
of inflammation appeared shrunken in size and were more
rounded.
DISCUSSION
SCI has not previously been studied in vivo using MTR. In
these experiments, significant reductions in the mean WM
MTR were measured in the rat spinal cord on day 1 after
severe SCI. By day 7 the average MTR for this group had
increased by approximately 80% and was no longer statistically different from the average MTR for intact spinal
cord WM. The average WM MTR for mild SCI at day 1 or
day 7 was not different from the control value.
After the initial injury the cord undergoes sequential
pathologic changes, including edema, inflammation, hemorrhage, and necrosis followed by infarction and demyelination. The severity of these pathologic features is much
greater after severe SCI than after mild SCI (20). This was
confirmed in our histological examination and is a possible explanation for reduced WM MTRs in severe SCI but
not in mild SCI. The increase in the WM MTR over time
after severe SCI may signify changes in the characteristics
of the pathophysiology between days 1 and 7. After the
initial injury, edema and hemorrhage diminish as inflammatory infiltrates arise and gliosis and cell death occur. It
is important to note that although the MTR increased from
day 1 to day 7 after severe SCI, approaching values for
intact cord, this change was not reflected in the MR images
of the spinal cord and the GM–WM contrast was not different from day 1 to day 7.
Of additional interest is the relative insensitivity of MTR
to inflammatory changes in the cord GM. In our previous
MT studies, using experimental allergic encephalomyelitis
brain of guinea pigs, we found that the MTR was sensitive
to inflammatory-related changes to the structure of normal-appearing WM (18). Within regions of intense inflammation in normal-appearing GM, however, the MTR was
unchanged. The composition and structure of WM and GM
govern their relative MT effects and may also determine
the extent to which inflammatory processes modulate the
degree of intermolecular interaction between macromolecules and water.
The GM–WM contrast measurement was abnormal for
all SCI subgroups. We did not observe any transition in the
GM–WM contrast over the 7-day examination period. This
is in agreement with Narayana et al. (9), who reported that
a temporal recovery of the GM–WM contrast occurs only
in spinal cord sections caudal to the epicenter of the injury. Clip compression causes injury to spinal cord tissue
both caudal and rostral to the epicenter of the injury. Since
our primary interest was the characterization of SCI using
MT, and not the recovery of GM–WM contrast in SCI, we
chose to examine SCI in sections that were located approximately 2 mm rostral to the site of injury. At this position
we were able to acquire high-quality images without the
need for respiratory gating to compensate for motion-generated image artifact, which we commonly observed in
images caudal to the injury.
Despite the differences in the pathophysiologic features
of mild and severe SCI, the proton density-weighted
GM–WM contrast did not allow discrimination between
the two degrees of severity of SCI. On the other hand, using
WM MTR allowed differentiation of mild and severe SCI at
day 1. Because the MT behavior of many general pathologic processes overlap, additional studies are required to
elucidate the specific underlying pathology related to the
reduction in the MTR. GM–WM contrast was only examined in proton-density weighted SE images. This type of
contrast ratio, measured from other images, for example,
MT of Rat Spinal Cord Injury
T1-, T2-, or T *2-weighted images, might also provide a
measure of the extent of injury in SCI. These preliminary
results suggest that WM MTR may provide an indication of
pathology in SCI. More advanced MR methods, perhaps
the determination of T1sat and MT rate constants, may
generate additional information as to the state of the injured cord tissue in vivo.
ACKNOWLEDGMENT
The authors thank Mrs. Carmen Simedrea for assistance.
REFERENCES
1. Papadopoulos SM. Spinal cord injury. Curr Opin Neurol Neurosurg
1992;55:554 –557.
2. Ohta K, Fujimura Y, Nakamura M, Wantanabe M, Yato Y. Experimental
study on MRI evaluation of the course of cervical spinal cord injury.
Spinal Cord 1999;37:580 –584.
3. Fraidakis M, Klason T, Cheng H, Olson L, Spenger C. High-resolution MRI
of intact and transected rat spinal cord. Exp Neurol 1998;153:299 –312.
4. Wamil AW, Wamil BD, Hellerqvist CG. CM101-mediated recovery of
walking ability in adult mice paralyzed by spinal cord injury. Proc Natl
Acad Sci USA 1998;95:13188 –13193.
5. Takahashi T, Suto Y, Kato S, Ohama E. Experimental dorsal compression of cat spinal cord: correlation of magnetic resonance signal intensity with spinal cord evoked potentials and morphology. Spine 1996;
21:166 –173.
6. LeMay DR, Fechner KP, Zelenock GB, Sun JK, Zee CS, D’Alecy LG.
High resolution magnetic resonance imaging of the rat spinal cord.
Neurol Res 1996;18:471– 474.
7. Guizar-Sahagun G, Rivera F, Babinski E, Berlanga E, Madrazo M,
Franco-Bourland R, Grijalva I, Gonzalez J, Contreras B, Madrazo I.
Magnetic resonance imaging of the normal and chronically injured
adult rat spinal cord in vivo. Neuroradiology 1994;36:448 – 452.
8. Weirich SD, Cotler HB, Narayana PA Hazle JD, Jackson EF, Coupe KJ,
McDonald CL, Langford LA, Harris JH Jr. Histopathologic correlation of
magnetic resonance imaging signal patterns in a spinal cord injury
model. Spine 1990;15:630 – 638.
163
9. Narayana PA, Abbe R, Liu S-J, Johnston D. Does loss of gray- and
white-matter contrast in injured spinal cord signify secondary injury?
In vivo longitudinal MRI studies. Magn Reson Med 1999;41:315–320.
10. Falconer JC, Narayana PA, Bhattacharjee MB, Lui S-J. Quantitative MRI
of spinal cord injury in a rat model. Magn Reson Med 1994;32:484 –
491.
11. Fukuoka M, Matsui N, Otsuka T, Murakami M, Seo Y. Magnetic resonance imaging of experimental subacute spinal cord compression.
Spine 1998;15:1540 –1549.
12. Runge VM, Wells JW, Baldwin SA, Scheff SW, Blades DA. Evaluation
of the temporal evolution of acute spinal cord injury. Invest Radiol
1997;32:105–110.
13. Ford J, Hackney D, Joesph P, Phelan M, Aslop D, Tabor S, Hand C,
Markowitz R, Black P. Method for in vivo high resolution MRI of rat
spinal cord. Magn Reson Med 1994;31:218 –233.
14. Wolff SD, Balaban RS. Magnetization transfer contrast (MTC) and tissue water proton relaxation in vivo. Magn Reson Med 1989;10:135–
144.
15. Dousset V, Grossman RI, Ramer KN, Schnall MD, Young LH, GonzalezScarano G, Lavi E. Cohen JA. Experimental allergic encephalomyelitis
and multiple sclerosis: lesion characterization with magnetization
transfer imaging. Radiology 1992;182:483– 491.
16. Berman JI, Hackney DB, Ford JC, Lavi E, Ulrich AM, McGowan JC. MT
histogram analysis of experimental spinal cord injury. In: Proceedings
of the 7th Annual Meeting of the ISMRM, Philadelphia, 1999, #992.
17. Rivlin AS, Tator CH. Effect of duration of acute spinal cord compression in a new acute cord injury model in the rat. Surg Neurol 1978;10:
38 – 43.
18. Gareau PJ, Rutt BK, Karlik SJ, Mitchell JR. Magnetization transfer and
multi-component T2 relaxation measurements with histopathologic
correlation in an experimental model of MS. J Magn Reson Imaging,
2000;11:586 –595.
19. Berry I, Barker G, Barkhof F, Campi A, Dousset V, Franconi JM, Gass A,
Schreiber W, Miller DH, Tofts PS. A multicenter measurement of magnetization transfer ratio in normal white matter. J Magn Reson Imaging
1999;9:441– 446.
20. Fehlings MG, Tator CH. The relationships among the severity of spinal
cord injury, residual neurological function, axon counts, and counts of
retrogradely labeled neurons after experimental spinal cord injury. Exp
Neurol 1995;132:220 –228.