www.elsevier.com/locate/ynimg

NeuroImage 32 (2006) 1195 – 1204

Differential sensitivity of in vivo and ex vivo diffusion tensor imaging

to evolving optic nerve injury in mice with retinal ischemia

Shu-Wei Sun,a,* Hsiao-Fang Liang,a Tuan Q. Le,b Regina C. Armstrong,b,d

Anne H. Cross,c and Sheng-Kwei Song a

aDepartment of Radiology, Washington University School of Medicine, St. Louis, MO 63110, USAbDepartment of Anatomy, Physiology, and Genetics, Uniformed Services University of the Health Sciences, Bethesda, MD 20814-4799, USAcDepartment of Neurology, Washington University School of Medicine, St. Louis, MO 63110, USAdProgram in Neuroscience, Uniformed Services University of the Health Sciences, Bethesda, MD 20814-4799, USA

Received 21 December 2005; revised 22 March 2006; accepted 28 April 2006

Available online 22 June 2006

Decreased axial (kkkk ||) and increased radial (kkkk–) diffusivity have been

shown to reflect axonal and myelin injury respectively. In the present

study, evolving white matter injury within the optic nerves of mice with

retinal ischemia was examined by in vivo and ex vivo measurements of

kkkk || and kkkk–. The results show that at 3 days after retinal ischemia, a

33% decrease in vivo and a 38% decrease ex vivo in kkkk || without change

in kkkk– was observed in the injured optic nerve compared to the control,

suggestive of axonal damage without myelin injury. At 14 days, both in

vivo and ex vivo measured kkkk– increased significantly to 220–240% of

the control level in the injured optic nerve suggestive of myelin damage.

In contrast, the axonal injury that was clearly detected in vivo as a

significantly decreased kkkk || (33% decrease) was not as clearly detected

by ex vivo kkkk || (17% decrease). The current findings suggest that ex vivo

kkkk– is comparable to in vivo kkkk– in detecting myelin injury. However,

the structural changes resulting from axonal damage causing the

decreased in vivo kkkk || may not be preserved ex vivo in the fixed tissues.

Despite the accurate depiction of the pathology using kkkk || and kkkk– in

vivo, the use of ex vivo kkkk || to extrapolate the status of axonal injury in

vivo would require further investigation.

D 2006 Elsevier Inc. All rights reserved.

Introduction

Tissue fixation is a ubiquitous laboratory practice, and it has

become increasingly common to obtain magnetic resonance images

(MRI) from fixed tissues (Ahrens et al., 1998; Bonny and Renou,

2002; Holmes et al., 2000; Hsu et al., 1998; Jacobs et al., 1999;

Mori et al., 1999, 2001; Pfefferbaum et al., 2004; Wedeen et al.,

1053-8119/$ - see front matter D 2006 Elsevier Inc. All rights reserved.

doi:10.1016/j.neuroimage.2006.04.212

* Corresponding author. Biomedical MR Laboratory, Campus Box 8227,

Washington University School of Medicine, 660 S. Euclid Avenue, St.

Louis, MO 63110, USA. Fax: +1 314 362 0526.

E-mail address: richard@bmr.wustl.edu (S.-W. Sun).

Available online on ScienceDirect (www.sciencedirect.com).

2001). Recent applications of diffusion tensor imaging (DTI) to

fixed tissue demonstrated that it is possible to obtain data with high

signal-to-noise ratio (SNR) at high spatial resolution without motion

artifact as compared with in vivo studies (Mori et al., 1999, 2001).

Postmortem brain specimens from patients with Alzheimer’s disease

(Bronge et al., 2002) or multiple sclerosis (Schmierer et al., 2003,

2004) have been examined using MRI to identify neuropathology.

Although increased T2 and decreased magnetization transfer

ratio have shown high correlation to histological abnormalities in

white matter, these MRI measurements do not differentiate the

underlying axonal and myelin damage (Meier et al., 2004). To

improve the specificity of MRI to underlying pathology, the

directional diffusivities derived from DTI were separated to

components describing water movement along (k ||, axial diffusiv-

ity) and across (k–, radial diffusivity) the white matter tract. It has

been demonstrated in living central and peripheral nervous systems

that axonal injury in white matter results in reduced k || whilemyelin damage increases k– (Arfanakis et al., 2002; Beaulieu et

al., 1996; Kim et al., 2006; Song et al., 2002, 2003; Sun et al.,

2006; Thomalla et al., 2004). However, application of directional

diffusivity to fixed specimens has only been reported recently in

studies examining the effect of demyelination on k– (Song et al.,

2005).

Many factors may alter the water diffusion characteristics in

formalin-fixed specimens. For example, decrease in tissue temper-

ature (Sun et al., 2003, 2005), cessation of cellular transport (Le

Bihan, 1995; van der Toorn et al., 1996) and possible change

of cellular membrane permeability (Benga et al., 1986; Gold and

Widnell, 1976; Ormerod et al., 1993; Sehy et al., 2002) in fixed

tissues can lead to changes in the measured DTI parameters. Thus,

careful comparison of in vivo and ex vivo diffusion properties is

crucial in establishing the validity and relevance of MRI of fixed

specimens. Some preliminary findings suggest that intravoxel

anisotropy and inter-voxel coherence of water diffusion are main-

tained in tissue after fixation (Crespigny et al., 2005; Guilfoyle et

al., 2003; Sun et al., 2003, 2005). A 50–70% decrease in mean

S.-W. Sun et al. / NeuroImage 32 (2006) 1195–12041196

apparent diffusion coefficient (ADC) after fixation was also

reported (Sun et al., 2003, 2005). The decrease in ADC is not

uniform across the entire brain. Different regional changes in

diffusion magnitude after fixation result in an altered image

contrast of ex vivo ADC maps. This alteration has resulted in

the loss of sensitivity using ADC to detect an area of acute stroke

after fixation (Sun et al., 2005).

To test the utility of k || and k– as well as to evaluate the general

diffusion characteristics in injured white matter of fixed specimens,

the mouse model of transient retinal ischemia (Kawai et al., 2001;

Rosenbaum et al., 2001; Song et al., 2003) was employed in this

study. The distinct pattern of optic nerve (ON) injury resulting

from retinal ischemia, i.e., acute axonal damage followed by

coexistence of axonal and myelin damage (Adachi et al., 1996;

Song et al., 2003), makes it an ideal model to evaluate both types

of injury. Serial examinations of mice undergoing transient retinal

ischemia have confirmed the previously reported time course of the

injury progression. Two time points were selected representing two

different pathological states based on the in vivo DTI findings.

Histological analyses after ex vivo DTI were performed to verify

the DTI findings.

Methods

Retinal ischemia

Male Swiss Webster mice, 6–8 weeks of age underwent

transient retinal ischemia (Song et al., 2003). Briefly, 100–120 mm

Hg intraocular pressure was applied to the right eye of each mouse

by inserting into the anterior chamber a 32-gauge needle

connecting to a saline reservoir placed above the eye for 1 h.

Reperfusion started immediately after removal of the cannula. The

left eye, which was not cannulated, served as the control.

Eight mice were used to evaluate the temporal evolution of ON

degeneration after retinal ischemia. For each mouse, in vivo DTI

was conducted 1, 3, 5, 7, 14, and 21 days after the retinal ischemia

to assess the extent of axonal and myelin damage (Song et al.,

2003). Based on the longitudinal findings, two time points

representing two different conditions of the injury were examined:

(a) 3 days after retinal ischemia where axonal damage was

observed without myelin damage (Song et al., 2003), and (b) the

coexistence of axonal and myelin damage at 14 days after retinal

ischemia. Sixteen mice undergoing retinal ischemia were divided

into two groups (8 mice each) and examined using in vivo DTI at 3

and 14 days after the surgery followed by ex vivo DTI measure-

ments and histological analyses.

Diffusion tensor imaging

In vivo DTI of live mice

Mice were anesthetized with a mixture of oxygen and

isoflurane (Baxtor Healthcare Corporation, IL, USA) using an

isoflurane vaporizer (D. R. C., Inc, KY, USA). Seven percent

isoflurane was used for induction and 2% for maintenance. The

core body temperature was maintained using warm water

circulating in a pad. After the appropriate anesthesia level was

reached, mice were placed in a custom-made, magnetic-resonance-

compatible stereotaxic device to immobilize the head. A 9-cm

inner diameter Helmoltz coil served as the RF transmitter. A 1.5-

cm outer diameter circular surface coil was placed on top of the

head to serve as the RF receiver. The entire device was put in an

Oxford Instruments 200/330 (4.7 T, 33 cm clear bore) magnet

equipped with a 16-cm inner diameter, actively shielded Oxford

gradient coil (18 G/cm, 200 As rise time). The magnet, gradient

coil, and Techron gradient power supply were interfaced with a

Varian UNITY-INOVA console controlled by a Sun Microsystems

Ultra-60 Sparc workstation. A conventional spin echo imaging

sequence, modified by adding a Stejskal-Tanner diffusion sensi-

tizing gradient pair, was employed for acquisition of the required

series of diffusion-weighted images. The images were acquired

with repetition time (TR) 1.7 s, spin echo time (TE) 50 ms, time

between application of gradient pulses (D) 25 ms, diffusion

gradient on time (y) 8 ms, diffusion sensitizing gradient strength

9.1 G/cm, 4 scans averaged per k space line, slice thickness 0.5

mm (5 slices total), field of view 3 cm, data matrix 256 � 256 (zero

filled to 512 � 512). Diffusion sensitizing gradients were applied

in six orientations: [Gx, Gy, Gz] = [1, 1, 0], [1, 0, 1], [0, 1, 1], [�1,1, 0], [0, �1, 1], and [1, 0, �1]. Two diffusion-sensitizing factors,

or b values, were used: 0 and 0.838 ms/Am2. Each multi-slice,

whole-brain, DTI data set was obtained with an acquisition time of

3 h.

Ex vivo DTI of fixed brains

At the conclusion of in vivo DTI examinations, mice were

euthanized and perfused through the left cardiac ventricle with

phosphate-buffered saline (PBS) followed by 4% paraformalde-

hyde in PBS. The intact brain was excised, placed in 4%

paraformaldehyde/PBS at 4-C for 1 week and then stored in

PBS at 4-C for an extra week before imaging.

Fixed brains were placed in a 1-cm inner diameter solenoid coil

to serve as the RF transmitter and receiver for the MR signal. Data

were acquired in an Oxford Instruments 200/400 (4.7 T, 40 cm

clear bore) magnet equipped with a 10-cm inner diameter, actively

shielded Magnex gradient coil (60 G/cm, 100 As rise time). The

magnet, gradient coil, and Techron gradient power supply were

interfaced with a Varian UNITY-INOVA console controlled by a

Sun Microsystems Ultra-60 Sparc workstation. An electric fan was

used to maintain a constant flow of cool air throughout the entire

experiment to maintain the sample temperature (Sun et al., 2003,

2005). Since the apparent diffusion coefficient (ADC) in a fixed

brain is about 50% lower than it in the live brain (Sun et al., 2003,

2005), D of 45 ms, d of 4 ms, and diffusion sensitizing gradient

strength of 19 G/cm were used to double the diffusion time of in

vivo DTI. In addition, the b value was set to maintain the same b �ADC for the same degree of diffusion weighting for both in vivo

and ex vivo measurements. Therefore, b = 1.805 ms/Am2 was used

for ex vivo DTI. The spatial resolution of ex vivo DTI was

maintained the same as in vivo DTI. Other acquisition parameters

include TR 1 s, TE 63 ms, and 2 scans averaged per k space line to

achieve similar signal-to-noise ratio (SNR) for T2-weighted images

as that of in vivo measurements. Each DTI data set was obtained

with an acquisition time of one hour.

Data analysis

The six independent elements of the diffusion tensor were

calculated from each diffusion-weighted image. The resulting

tensor element maps were used to derive eigenvalues (k1, k2, and

k3) of the diffusion tensor by matrix diagonalization (Basser and

Pierpaoli, 1996). On a pixel-by-pixel basis, quantitative indices,

including axial diffusivity (k ||), radial diffusivity (k–), relative

anisotropy (RA), and trace of the diffusion tensor (Tr), were

S.-W. Sun et al. / NeuroImage 32 (2006) 1195–1204 1197

derived using software written in Matlab (MathWorks, Natick,

MA, USA) defined by the following equations:

Tr ¼ k1 þ k2 þ k3 ð1Þ

k� ¼ k1 ð2Þ

k8 ¼ 0:5� k2 þ k3ð Þ ð3Þ

RA ¼

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffik1 � Tr=3ð Þ2 þ k2 � Tr=3ð Þ2 þ k3 � Tr=3ð Þ2

qffiffiffi3p

Tr=3ð Þ: ð4Þ

Regions of interest (ROI) were selected in left and right ON

(control and injured respectively) based on the image contrast of

DTI indices maps and the images with b = 0 (B0 images or T2-

weighted images, T2WI) referencing to a mouse brain atlas

(Franklin and Paxinos, 1997). The definition of ROI is shown in

Fig. 1. The delineated ROI was also employed for SNR

determinations on B0 images. SNR was defined as the quotient

of mean signal intensity in each ROI and standard deviation of the

background noise divided by 0.66 in consideration of Rayleigh

statistics (Edelstein et al., 1984). Data were presented as mean Tstandard deviation. Two-tailed paired t test was performed to

compare the measurements between control and experimental

groups. Statistical significance was accepted at P < 0.05.

Histological examinations

At the conclusion of ex vivo DTI examinations, 2-mm-thick

coronal tissue blocks were obtained from each brain. The tissues

from four mice each at 3-day and 14-day post-injury were

embedded in OCT for frozen cutting, and the remaining tissues

were embedded in paraffin. For each mouse brain, a notch was

marked on the right hemisphere. Transverse slices were cut starting

from the optic chiasm. The slice matching the ROI selected from

the DTI image located at 1 mm anterior to the optic chiasm. The

transverse sections including both injured (right eye) and the

control ON (left eye) were mounted on the same slide for the

staining.

Fig. 1. Two continuous coronal RA maps are shown with optic nerves marked w

interest (ROI) are marked within the optic nerve circumference as shown in the

selection was based on the confined criteria of all index maps, where RA was use

pixels due to cerebrospinal fluid contamination, and T2WI was used to define the

For examining axonal damage, transverse 15-Am-thick slices

were cut from the OCT embedded blocks. These slices were

immunostained with SMI-32 monoclonal antibody (1:1000 dilution;

Sternberger Monoclonals, Inc., Lutherville, MD), which recognizes

nonphosphorylated neurofilament H epitopes associatedwith axonal

damage (Dandekar et al., 2001; Trapp et al., 1998). SMI-32

immunolabeling was detected with donkey anti-mouse IgG F(ab’)2conjugated with Cy3 (Jackson Immunoresearch, West Grove, PA).

Images of immunostaining results were captured with a Spot 2 CCD

digital camera using Spot Advanced image acquisition software

(Diagnostic Instruments, SterlingHeights,MI) on anOlympus IX-70

microscope. Damaged axons identified by SMI-32 were quantified

by manually counting SMI-32 stained axons on each slide.

For examining myelin damage, transverse 3-Am-thick slices

were cut from paraffin embedded tissue and cleared in xylene. The

primary antibody detecting myelin basic protein (MBP, 1:100;

Zymed Laboratories Inc., South San Francisco, CA) was revealed

by avidin–biotin–peroxidase (Vector Laboratories, Inc. Burlin-

game, CA 94010). Images were captured with a Photometrics CCD

digital camera using MetaMorph image acquisition software

(Universal Imaging Corporation, Downington, PA) on a Nikon

Eclipse 80i microscope. The myelinated axons were counted in

both injured (right) and control ON (left) of each tissue sections by

one of the authors (H. F. L.) in a blinded fashion.

Results

In vivo longitudinal measurements of k|| and k–

Optic nerve degeneration following transient retinal ischemia

was evaluated longitudinally over a 21-day period (Fig. 2). As

previously published (Song et al., 2003), a significant decrease of

k || suggestive of axonal injury was seen at 3 days after retinal

ischemia. The extent of the reduced k || was maintained throughout

the entire time course. In contrast, k– was not affected on day 3

but a statistically significant elevation of k– was seen at day 7 and

reached a plateau by day 14 (Fig. 2), consistent with a previous

report of a shorter time course using the same model (Song et al.,

ithin a white rectangle (a and b). In panel c, white rectangular regions of

expanded view of RA, k ||, k–, and T2-weighted images (T2WI). The ROI

d to differentiate gray and white matter, k || and k– were used to avoid the

lower edge of the ROI to avoid the pixels extending outside of the brain.

Fig. 2. Temporal evolution of optic nerve injury from mice undergoing transient retinal ischemia. The filled symbols represent the measurement from the

injured optic nerve, and the open symbols represent the measurement from the control optic nerve. The k || of injured optic nerve decreased significantly at 3

days after transient ischemia. It remained at this reduced value throughout the course of the experiment. The k– value of injured optic nerve increased and

reached a plateau at 7 days. Decreased Tr was also seen on days 3–7 probably reflecting the decreased k ||. Normalization of Tr is seen from day 7 until the end

of the experiment, resulting from the combined effect of the decreased k || and the increased k–. Relative anisotropy decreased at day 5 and remained at that

level throughout the time course. B0 intensity in the injured optic nerve varied within 10% of the control values. It increased at day 3, renormalized at day 5,

and then slightly decreased at days 7–21. The two time points, 3 and 14 days after transient ischemia, selected for the cross-sectional examination are marked

by the vertical rectangles. *Indicates statistically significant difference ( P < 0.05) seen between the optic nerve from both eyes.

S.-W. Sun et al. / NeuroImage 32 (2006) 1195–12041198

2003). Also consistent with the previous findings, Tr decreased

beginning at 3 days after ischemia and continued to decrease up to

7 days. After 7 days, Tr normalized to its control level (Fig. 2).

Relative anisotropy decreased significantly at day 5 and remained

at this level up to 21 days after ischemia. The intensity of B0

images (the images with b = 0 ms/Am2) from the injured optic

nerve decreased between 7 and 21 days following a slight increase

at 3 days after ischemia.

Ex vivo DTI of fixed tissues

The effects of fixation on the ex vivo DTI parameters were

examined using the control and injured ONs from mice undergoing

retinal ischemia at 3 and 14 days after transient retinal ischemia.

Representative RA, k ||, and k– maps of the ONs obtained from in

vivo and ex vivo DTI at 3 and 14 days were compared. The

contrast between the control and injured ON was qualitatively the

same between in vivo and ex vivo images at 3 days after retinal

ischemia (Fig. 3). However, despite the similar contrast in RA, and

k– maps, a discrepancy in k || between in vivo and ex vivo was

observed at 14 days (Fig. 4). The SNR of control and injured ONs

was comparable between in vivo and ex vivo DTI data sets (Table

1). Thus, the SNR between in vivo and ex vivo measurements did

not contribute to the observed difference.

Comparisons of in vivo and ex vivo DTI at 3 days after ischemia

The group-averaged summary corroborated the qualitative

examinations (Fig. 5). The ex vivo apparent diffusion coefficients

were 50–70% lower than that of in vivo, consistent with previous

findings (Sun et al., 2003, 2005). At 3 days after ischemia, in vivo

k || decreased by 33% in the injured ON compared to that of the

control (1.27 T 0.31 vs. 1.89 T 0.44 Am2/ms, n = 8, P < 0.0001),

while in vivo k– was not changed in the injured ON (0.21 T 0.06

Fig. 3. The RA, k ||, and k–maps were obtained from in vivo and ex vivo DTI of the ON from the same mouse three days after retinal ischemia. Solid and dashed

rectangles indicate the control and injured ON respectively. Decreased k || is seen in the ON from the injured eye in both in vivo (panel a) and ex vivo (panel b) DTI

data. However, at this time point, there is no detectable difference between the injured and the control ON in either RA or k– map in vivo or ex vivo.

S.-W. Sun et al. / NeuroImage 32 (2006) 1195–1204 1199

vs. 0.24 T 0.06 Am2/ms, n = 8, P = 0.22). A 38% decrease of ex

vivo k || was observed in the injured ON compared to that of the

control (0.46 T 0.12 vs. 0.74 T 0.12 Am2/ms, n = 8, P < 0.0001).

The ex vivo k– in the injured ON (0.09 T 0.04, Am2/ms, n = 8)

were not different from values for the control (0.10 T 0.04 Am2/ms,

n = 8, P = 0.57). Thus, both in vivo and ex vivo DTI measurements

detected the axonal injury at 3 days after transient ischemia, and

neither method detected myelin damage at this time point.

Comparisons of in vivo and ex vivo DTI at 14 days after ischemia

Similar to findings at 3 days, the in vivo k || (1.20 T 0.18 Am2/

ms, n = 8) in the injured ON was significantly lower than that of

the control (1.79 T 0.18 Am2/ms, n = 8, P < 0.0001) at 14 days. A

significant elevation of in vivo k– was seen in the injured ON

(0.42 T 0.00 vs. 0.19 T 0.04 Am2/ms, n = 8, P < 0.0001) compared

Fig. 4. The RA, k ||, and k– maps were obtained from in vivo and ex vivo DTI of th

rectangles indicate the control and injured ON respectively. Decreased k || is seen i

elevated k– and decrease RA are seen in both in vivo and ex vivo measurement

to that of the control at 14 days. After fixation, the elevation of ex

vivo k– in the injured ON (0.24 T 0.09 Am2/ms) compared to the

control (0.10 T 0.07 Am2/ms, n = 8, P < 0.0001) was similar to that

observed in vivo. However, the significantly decreased k || seen in

the injured ON in vivo was not detectable in the ex vivo

measurement (0.60 T 0.14 vs. 0.72 T 0.13 Am2/ms as injured vs.

control, n = 8, P = 0.076; Fig. 5b).

Axonal and myelin damage assessed using immunohistochemistry

Axonal and myelin damage was validated using immunohisto-

chemistry after ex vivo DTI. A similar extent of axonal damage,

based on counting SMI-32 immunolabeled axons, was seen in ONs

at 3 and 14 days after retinal ischemia (Fig. 6). This was in

agreement with the in vivo DTI findings in which k || values at 3

and 14 days were significantly lower than the controls. In vivo k ||

e ON from the same mouse 14 days after retinal ischemia. Solid and dashed

n the injured ON in vivo (panel a), but not ex vivo (panel b). However, the

s.

Table 1

Signal-to-noise ratios measured in the control and injured optic nerves

Control Optic Nerve Injured Optic Nerve

Day 3 Day 14 Day 3 Day 14

In vivo 61 T 6.4 62 T 7.6 64 T 7.2 61 T 7.4

Ex vivo 66 T 7.0 63 T 4.7 66 T 4.7 66 T 6.9

Data are presented as mean T standard deviation (n = 8).

S.-W. Sun et al. / NeuroImage 32 (2006) 1195–12041200

correlates with SMI-32 staining (r = �0.91, P < 0.0001) better

than that of ex vivo k || (r = �0.61, P = 0.033; Fig. 6e). MBP

immunostaining indicated that there was no myelin loss at 3 days.

At 14 days, significantly reduced MBP immunostaining suggested

myelin damage in the injured ON (Fig. 7). Both in vivo (r = �0.82,P = 0.013) and ex vivo k– (r = �0.80, P = 0.018) correlated with

the MBP staining (Fig. 7e). This finding agrees with both ex vivo

and in vivo DTI findings in which significant elevation of k–

occurred at 14 days but not at 3 days after ischemia.

Discussion

In this study, the evolution of ON injury was examined both in

vivo and ex vivo using DTI. The feasibility of using k || and k– as

surrogate markers to detect axonal and myelin damage in

paraformaldehyde fixed ON was extensively examined in two

different settings of injury: (a) axonal injury without myelin

damage (decreased k || with normal k– at 3 days) and (b) the

coexistence of both axonal and myelin damage (decreased k || and

increased k– at 14 days). At 3 days after ischemia, a comparable

decrease in ex vivo fixed and in vivo k || was observed (in vivo 33%

decreased; ex vivo 38% decreased) in the injured ON reflecting

axon damage. At 14 days, both in vivo and ex vivo measurements

Fig. 5. In vivo and ex vivo measurements of k || and k– in normal and injured ON

represent the measurements from injured ONs. *Indicates a significant level ( P <

ON. At 3 days, both in vivo and ex vivo k || exhibit a 33–38% decrease in the injure

DTI showed significantly increased k– in the injured ON, compared to that of the

injured ON in vivo (32%) was not detected in the ex vivo DTI.

of k– exhibited a significant increase reaching 220–240% of

control levels in the injured ON reflecting myelin damage.

However, the axonal injury that was clearly detected in vivo as a

significant decrease in k || was not detected as well by ex vivo k || (in

vivo 33% decreased; ex vivo 17% decreased).

A reduced sensitivity of k || to detect axonal damage in fixed

tissue has been suggested recently (Song et al., 2005). In that study,

axonal damage was clearly demonstrated by h-amyloid precursor

protein staining of axons in the corpus callosum from cuprizone

fed mice, but the predicted decline in k || was not observed ex vivo

(Song et al., 2005). In contrast, a recent DTI investigation of live

mice undergoing the same cuprizone treatment showed a signif-

icant decrease of in vivo k || (Sun et al., 2006). The direct

comparison of live and fixed tissues in the present study confirms

the previous observation that ex vivo k || in fixed tissue cannot

reliably be extrapolated to the in vivo situation.

It is not clear why ex vivo k || failed to clearly detect axonal

damage at 14 days after the ischemia while successfully detecting

axonal damage at 3 days. One possible explanation is related to

tissue shrinkage after fixation (Mouritzen Dam, 1979). Shrinkage

caused by fixation might lead to the increase of inter-axonal space

and the increased volume fraction of less-hindered environments of

water movement in that space. Such alteration might be prominent

especially when myelin is also damaged at 14 days after the

ischemia. Since the measured k || includes both intra- and inter-

axonal compartments, the intraaxonal decrease of k || may be offset

by the inter-axonal increase of k || after tissue fixation. This

argument may be supported by the observed 140% elevation of ex

vivo k– in injured optic nerve compared with the 120% elevation

of in vivo k–, a possible indication of the increased inter-axonal

space caused by formalin fixation. Therefore, the sensitivity of k ||

in detecting axonal damage may be reduced by tissue fixation

when myelin is also injured.

s. Open bars represent the measurements from the control ON. Filled bars

0.05) of the difference between control (open bars) and injured (filled bars)

d nerves compared to the control ONs. At 14 days, both in vivo and ex vivo

control from each mouse. However, the significant decrease of k || in 14-day

Fig. 6. Immunostaining to detect nonphosphorylated neurofilaments (SMI-32) of the control and injured ON. Positive staining with SMI-32 can be identified as

diffuse fluorescence positive axons, which are obvious during injury (b and c) but not detectable in the control ON (a). The SMI-32 positive axon counts for

each ON section are summarized in panel d. Both 3-day and 14-day ON sections contain a similar number of SMI-32 immunostained (+) axons indicating

comparable severity of axonal damage at 3- and 14-day injured ONs. (e) The correlation between SMI-32 and k || for the in vivo (solid line and filled symbols)

and ex vivo (dashed line and filled symbols) DTI were examined. Control ON were presented as triangles, and the injured ON were presented as circles. In vivo

k || (r = �0.91, P < 0.0001) correlated better with SMI-32 staining than ex vivok || (r = �0.61, P = 0.033).

S.-W. Sun et al. / NeuroImage 32 (2006) 1195–1204 1201

Another explanation of this discrepancy may result from the

different underlying structural changes of injured axons between 3 and

14 days. Focal misalignment of neurofilaments and accumulation of

organelles leading to axonal swelling may occur acutely and add

hindrances to water molecule movement (Arfanakis et al., 2002;

Christman et al., 1994). Chronically, the build up of cellular debris

from the breakdown of axonal structures may still be reflected as

reduced k || (Frank and Wolburg, 1996; Westerlund et al., 2001). It is

possible that tissue fixation introduces cross-links between proteins

which may reorganize the randomly distributed debris by anchoring it

to the adjacent structures. Thus, although the breakdown of axonal

structure reduces k || in vivo, tissue fixation may create a renewed or-

rrdered environment for water movement along the axonal tract resu-

lting in an increased ex vivo k || offsetting the factors reducing it in vivo.

In contrast, the k– increase ex vivo was comparable to that of in

vivo, consistent with myelin damage. Both in vivo and ex vivo k–

reached 220–240% of control level at 14 days. There was no signi-

ficant change in k– at 3 days after ischemia, reflecting the intact

myelin at that early time point. In a previous report, the well-

characterized cuprizone model of experimental demyelination and

remyelination of corpus callosum in mouse brain was examined by

ex vivo DTI (Song et al., 2005). The extent of increased ex vivo k–

reflected the severity of demyelination observed by histology, and

the subsequent normalization of ex vivo k–was in agreement with

remyelination validated by histology. The available data suggest

that both in vivo and ex vivo measured k– are good indicators of

myelin integrity.

Despite the reduced sensitivity to detect axonal injury, ex vivo

DTI is still needed in many circumstances because the imaging

time can become very long to achieve adequate SNR and

resolution to offset confounding physiological noises of in vivo

DTI. This is of even greater significance when imaging small

subjects, such as the mouse optic nerve. In this study, the ex vivo

image resolution and SNR were maintained equivalent to those of

in vivo. Thus, it is uncertain whether further increased resolution

and better SNR would resolve the observed discrepancy between in

vivo and ex vivo detection of axonal injury.

Both apparent diffusion coefficient and diffusion anisotropy

have also been employed to characterize the extent and progression

of neurological disorders. However, the current findings suggest that

both Tr and RA are not capable of detecting and differentiating

axonal and myelin injury in the white matter. For example, the

axonal injury occurring at 3–7 days after ischemia was reflected by

the decreased Tr. However, normalization of Tr occurred thereafter,

despite the continuing presence of axonal injury of the optic nerve.

Similarly, RA decreased at 5 days after ischemia coinciding with

both axonal and myelin injury. In contrast, decreased k || and

increased k– are highly correlated with the axonal and myelin

injury respectively.

Despite the abundant evidence supporting that decreased k || andincreased k– are potential surrogate markers of axonal and myelin

damage (Arfanakis et al., 2002; Beaulieu et al., 1996; Kim et al.,

2006; Song et al., 2002; Song et al., 2003; Song et al., 2005; Sun et

al., 2006; Thomalla et al., 2004), its general applicability in the

assessment of white matter injury is still inconclusive. Other

pathophysiologies, such as edema and inflammation, may also

affect these DTI parameters. For example, acute edema could cause

diffusion coefficient to decrease (Kuroiwa et al., 2000; Matsumoto

Fig. 7. Immunostaining of myelin basic protein (MBP) in the control and the injured ON. At 3-day after injury (b), the MBP staining is qualitatively similar to

that of the control ON (a). The significantly reduced MBP staining of the 14-day injured ON (c) is apparent comparing to that of the control. Quantitatively,

myelin sheath integrity was evaluated by counting the myelinated axons in the control and the injured ON. The myelinated axon count of the injured ON was

normalized to that of the control (d). The myelinated axon counts of the injured ON at 3 days after injury are comparable ( P > 0.05) to that of the control.

However, significantly reduced number of myelinated axons is seen at 14 days after injury. *Indicates P < 0.05 between injured and control ONs. The

correlation between MBP immunostaining and in vivo (solid line and filled symbols) and ex vivo (dashed line and open symbols) k– was also examined. Both

in vivo (r = �0.82, P = 0.013) and ex vivo k– (r = �0.80, P = 0.018) correlated well with the MBP immunostaining.

S.-W. Sun et al. / NeuroImage 32 (2006) 1195–12041202

et al., 1995; Sevick et al., 1992). Observations made in patients with

multiple sclerosis have suggested that inflammation could result in

elevated water diffusion in the central nervous system (Bakshi et al.,

2005; Miller et al., 1998; Mukherji et al., 2002; Rovaris and Filippi,

2002; Tartaglia and Arnold, 2006). Therefore, further investigations

into the effects of various pathological processes on k || and k–

using animal models with defined pathologies will be crucial in

assessing the power of the proposed diffusion measurements in

clinical settings.

In conclusion, ex vivo DTI of fixed tissues may provide high

quality images with resolution and SNR not easily achievable in

vivo. Better ex vivo imaging measurements may serve to establish

better correlations between MR correlates and the underlying

pathologies, which will be useful in assessing human diseases

when tissue biopsies are not always feasible. It may also be used to

guide the histological analyses. Thus, it is crucial to determine the

correlation between MR parameters in vivo and in fixed tissue ex

vivo. The current findings suggest that the correlation between

increased k– and demyelination is comparable ex vivo and in vivo.

However, the structural changes resulting from axonal damage

causing the decrease in vivo k || may not always be preserved in fixed

tissues.

Acknowledgments

This study was supported in part by the National Multiple

Sclerosis Society (RG 3376, and CA 1012), NIH R01-NS047592,

USUHS grant G170SX, and the Washington University Small

Animal Imaging Resource (WUSAIR) (NIH: R24-CA83060).

AHC was supported by the Manny and Rosalyn Rosenthal–Dr.

John L. Trotter MS Center Chair in Neuroimmunology.

References

Adachi, M., Takahashi, K., Nishikawa, M., Miki, H., Uyama, M., 1996.

High intraocular pressure-induced ischemia and reperfusion injury in

the optic nerve and retina in rats. Graefe’s Arch. Clin. Exp. Ophthalmol.

234, 445–451.

Ahrens, E.T., Laidlaw, D.H., Readhead, C., Brosnan, C.F., Fraser, S.E.,

Jacobs, R.E., 1998. MR microscopy of transgenic mice that spontane-

ously acquire experimental allergic encephalomyelitis. Magn. Reson.

Med. 40, 119–132.

Arfanakis, K., Haughton, V.M., Carew, J.D., Rogers, B.P., Dempsey, R.J.,

Meyerand, M.E., 2002. Diffusion tensor MR imaging in diffuse axonal

injury. Am. J. Neuroradiol. 23, 794–802.

Bakshi, R., Minagar, A., Jaisani, Z., Wolinsky, J.S., 2005. Imaging of

multiple sclerosis: role in neurotherapeutics. NeuroRx 2, 277–303.

Basser, P.J., Pierpaoli, C., 1996. Microstructural and physiological features

of tissues elucidated by quantitative-diffusion-tensor MRI. J. Magn.

Reson., Ser. B 111, 209–219.

Beaulieu, C., Does, M.D., Snyder, R.E., Allen, P.S., 1996. Changes in water

diffusion due to Wallerian degeneration in peripheral nerve. Magn.

Reson. Med. 36, 627–631.

Benga, G., Popescu, O., Borza, V., Pop, V.I., Muresan, A., Mocsy, I., Brain,

A., Wrigglesworth, J.M., 1986. Water permeability in human eryth-

S.-W. Sun et al. / NeuroImage 32 (2006) 1195–1204 1203

rocytes: identification of membrane proteins involved in water

transport. Eur. J. Cell Biol. 41, 252–262.

Bonny, J.M., Renou, J.P., 2002. Water diffusion features as indicators of

muscle structure ex vivo. Magn. Reson. Imaging 20, 395–400.

Bronge, L., Bogdanovic, N., Wahlund, L.O., 2002. Postmortem MRI and

histopathology of white matter changes in Alzheimer brains. A

quantitative, comparative study. Dementia Geriatr. Cognit. Disord. 13,

205–212.

Christman, C.W., Grady, M.S., Walker, S.A., Holloway, K.L., Povlishock,

J.T., 1994. Ultrastructural studies of diffuse axonal injury in humans. J.

Neurotrauma 11, 173–186.

Crespigny, A.J.d., D’Arceuil, H.E., Dai, G.G., Wang, R., Kim, Y.,

Mandeville, J.B., Wedeen, V.J., 2005. Comparison of in vivo and ex

vivo diffusion spectrum imaging (DSI) of rat brain. Proc. Intl. Soc.

Mag. Reson. Med. 13, 583.

Dandekar, A.A., Wu, G.F., Pewe, L., Perlman, S., 2001. Axonal damage is

T cell mediated and occurs concomitantly with demyelination in mice

infected with a neurotropic coronavirus. J. Virol. 75, 6115–6120.

Edelstein, W.A., Bottomley, P.A., Pfeifer, L.M., 1984. A signal-to-noise

calibration procedure for NMR imaging systems. Med. Phys. 11,

180–185.

Frank, M., Wolburg, H., 1996. Cellular reactions at the lesion site after

crushing of the rat optic nerve. Glia 16, 227–240.

Franklin, K.B., Paxinos, G., 1997. The Mouse Brain in Stereotaxic

Coordinates. Academic Press, San Diego.

Gold, G., Widnell, C.C., 1976. Relationship between microsomal mem-

brane permeability and the inhibition of hepatic glucose-6-phosphatase

by pyridoxal phosphate. J. Biol. Chem. 251, 1035–1041.

Guilfoyle, D.N., Helpern, J.A., Lim, K.O., 2003. Diffusion tensor imaging

in fixed brain tissue at 7.0 T. NMR Biomed. 16, 77–81.

Holmes, A.A., Scollan, D.F., Winslow, R.L., 2000. Direct histological

validation of diffusion tensor MRI in formaldehyde-fixed myocardium.

Magn. Reson. Med. 44, 157–161.

Hsu, E.W., Muzikant, A.L., Matulevicius, S.A., Penland, R.C., Henri-

quez, C.S., 1998. Magnetic resonance myocardial fiber-orientation

mapping with direct histological correlation. Am. J. Physiol. 274,

H1627–H1634.

Jacobs, R.E., Ahrens, E.T., Dickinson, M.E., Laidlaw, D., 1999. Towards a

microMRI atlas of mouse development. Comput. Med. Imaging Graph.

23, 15–24.

Kawai, S.I., Vora, S., Das, S., Gachie, E., Becker, B., Neufeld, A.H.,

2001. Modeling of risk factors for the degeneration of retinal ganglion

cells after ischemia/reperfusion in rats: effects of age, caloric

restriction, diabetes, pigmentation, and glaucoma. FASEB J. 15,

1285–1287.

Kim, J.H., Budde, M.D., Liang, H.F., Klein, R.S., Russell, J.H., Cross,

A.H., Song, S.K., 2006. Detecting axon damage in spinal cord from a

mouse model of multiple sclerosis. Neurobiol. Dis. 21, 626–632.

Kuroiwa, T., Nagaoka, T., Miyasaka, N., Akimoto, H., Zhao, F., Yamada, I.,

Ueki, M., Ichinose, S., 2000. Time course of trace of diffusion tensor

[Trace(D)] and histology in brain edema. Acta Neurochir., Suppl. 76,

191–194.

Le Bihan, D., 1995. Molecular diffusion, tissue microdynamics and

microstructure. NMR Biomed. 8, 375–386.

Matsumoto, K., Lo, E.H., Pierce, A.R., Wei, H., Garrido, L., Kowall,

N.W., 1995. Role of vasogenic edema and tissue cavitation in

ischemic evolution on diffusion-weighted imaging: comparison with

multiparameter MR and immunohistochemistry. Am. J. Neuroradiol.

16, 1107–1115.

Meier, D.S., Weiner, H.L., Khoury, S.J., Guttmann, C.R., 2004. Magnetic

resonance imaging surrogates of multiple sclerosis pathology and their

relationship to central nervous system atrophy. J. Neuroimaging 14,

46S–53S.

Miller, D.H., Grossman, R.I., Reingold, S.C., McFarland, H.F., 1998. The

role of magnetic resonance techniques in understanding and managing

multiple sclerosis. Brain 121 (Pt. 1), 3–24.

Mori, S., Crain, B.J., Chacko, V.P., van Zijl, P.C., 1999. Three-dimensional

tracking of axonal projections in the brain by magnetic resonance

imaging. Ann. Neurol. 45, 265–269.

Mori, S., Itoh, R., Zhang, J., Kaufmann, W.E., van Zijl, P.C., Solaiyappan,

M., Yarowsky, P., 2001. Diffusion tensor imaging of the developing

mouse brain. Magn. Reson. Med. 46, 18–23.

Mouritzen Dam, A., 1979. Shrinkage of the brain during histological

procedures with fixation in formaldehyde solutions of different

concentrations. J. Hirnforsch. 20, 115–119.

Mukherji, S.K., Chenevert, T.L., Castillo, M., 2002. Diffusion-weighted

magnetic resonance imaging. J. Neuroophthalmol. 22, 118–122.

Ormerod, M.G., Sun, X.M., Snowden, R.T., Davies, R., Fearnhead, H.,

Cohen, G.M., 1993. Increased membrane permeability of apoptotic

thymocytes: a flow cytometric study. Cytometry 14, 595–602.

Pfefferbaum, A., Sullivan, E.V., Adalsteinsson, E., Garrick, T., Harper, C.,

2004. Postmortem MR imaging of formalin-fixed human brain. Neuro-

Image 21, 1585–1595.

Rosenbaum, D.M., Rosenbaum, P.S., Singh, M., Gupta, G., Gupta, H., Li,

B., Roth, S., 2001. Functional and morphologic comparison of two

methods to produce transient retinal ischemia in the rat. J. Neuro-

ophthalmol. 21, 62–68.

Rovaris, M., Filippi, M., 2002. MR-based technology for in vivo detection,

characterization, and quantification of pathology of relapsing-remitting

multiple sclerosis. J. Rehabil. Res. Dev. 39, 243–259.

Schmierer, K., Scaravilli, F., Barker, G.J., Gordon, R., MacManus, D.G.,

Miller, D.H., 2003. Stereotactic co-registration of magnetic resonance

imaging and histopathology in post-mortem multiple sclerosis brain.

Neuropathol. Appl. Neurobiol. 29, 596–601.

Schmierer, K., Scaravilli, F., Altmann, D.R., Barker, G.J., Miller, D.H.,

2004. Magnetization transfer ratio and myelin in postmortem multiple

sclerosis brain. Ann. Neurol. 56, 407–415.

Sehy, J.V., Banks, A.A., Ackerman, J.J., Neil, J.J., 2002. Importance of

intracellular water apparent diffusion to the measurement of membrane

permeability. Biophys. J. 83, 2856–2863.

Sevick, R.J., Kanda, F., Mintorovitch, J., Arieff, A.I., Kucharczyk, J.,

Tsuruda, J.S., Norman, D., Moseley, M.E., 1992. Cytotoxic brain

edema: assessment with diffusion-weighted MR imaging. Radiology

185, 687–690.

Song, S.K., Sun, S.W., Ramsbottom, M.J., Chang, C., Russell, J., Cross,

A.H., 2002. Dysmyelination revealed through MRI as increased

radial (but unchanged axial) diffusion of water. NeuroImage 17,

1429–1436.

Song, S.K., Sun, S.W., Ju, W.K., Lin, S.J., Cross, A.H., Neufeld, A.H.,

2003. Diffusion tensor imaging detects and differentiates axon and

myelin degeneration in mouse optic nerve after retinal ischemia.

NeuroImage 20, 1714–1722.

Song, S.K., Yoshino, J., Le, T.Q., Lin, S.J., Sun, S.W., Cross,

A.H., Armstrong, R.C., 2005. Demyelination increases radial

diffusivity in corpus callosum of mouse brain. NeuroImage 26,

132–140.

Sun, S.W., Neil, J.J., Song, S.K., 2003. Relative indices of water diffusion

anisotropy are equivalent in live and formalin-fixed mouse brains.

Magn. Reson. Med. 50, 743–748.

Sun, S.W., Neil, J.J., Liang, H.F., He, Y.Y., Schmidt, R.E., Hsu, C.Y., Song,

S.K., 2005. Formalin fixation alters water diffusion coefficient

magnitude but not anisotropy in infarcted brain. Magn. Reson. Med.

53, 1447–1451.

Sun, S.W., Liang, H.F., Trinkaus, K., Cross, A.H., Armstrong, R.C., Song,

S.K., 2006. Noninvasive detection of cuprizone induced axonal damage

and demyelination in the mouse corpus callosum. Magn. Reson. Med.

55, 302–308.

Tartaglia, M.C., Arnold, D.L., 2006. The role of MRS and fMRI in multiple

sclerosis. Adv. Neurol. 98, 185–202.

Thomalla, G., Glauche, V., Koch, M.A., Beaulieu, C., Weiller, C., Rother,

J., 2004. Diffusion tensor imaging detects early Wallerian degenera-

tion of the pyramidal tract after ischemic stroke. NeuroImage 22,

1767–1774.

Trapp, B.D., Peterson, J., Ransohoff, R.M., Rudick, R., Mork, S., Bo, L.,

S.-W. Sun et al. / NeuroImage 32 (2006) 1195–12041204

1998. Axonal transection in the lesions of multiple sclerosis. N. Engl. J.

Med. 338, 278–285.

van der Toorn, A., Sykova, E., Dijkhuizen, R.M., Vorisek, I., Vargova, L.,

Skobisova, E., van Lookeren Campagne, M., Reese, T., Nicolay, K.,

1996. Dynamic changes in water ADC, energy metabolism, extracel-

lular space volume, and tortuosity in neonatal rat brain during global

ischemia. Magn. Reson. Med. 36, 52–60.

Wedeen, V.J., Reese, T.G., Napadow, V.J., Gilbert, R.J., 2001. Demonstra-

tion of primary and secondary muscle fiber architecture of the bovine

tongue by diffusion tensor magnetic resonance imaging. Biophys. J. 80,

1024–1028.

Westerlund, T., Vuorinen, V., Roytta, M., 2001. Same axonal regeneration

rate after different endoneurial response to intraneural glycerol and

phenol injection. Acta Neuropathol. (Berl.) 102, 41–54.