| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
2003 ARRS President's Award |
1 Department of Radiology, Duke University Medical Center, Box 3808, Durham, NC
27710.
2 Brain Imaging and Analysis Center, Duke University Medical Center, Durham, NC
27710.
3 Department of Pharmacology and Cancer Biology, Duke University Medical Center,
Durham, NC 27710.
4 Serenex, Inc., Durham, NC.
5 Department of Anesthesiology, Duke University Medical Center, Durham, NC
27710.
Received March 12, 2003; accepted after revision July 6, 2004.
Address correspondence to C. D. Lascola.
OBJECTIVE. Gliosis refers to a range of glial cell transformations that vary according to specific brain pathologic states. Disease, however, is not a prerequisite for gliosis because glial reactivity may also be seen in regions of increased physiologic activity. Our study tests the hypothesis that high-field-strength magnetization transfer MRI is a sensitive technique for detecting transient glial reactivity after experimental spreading depression, a relatively benign perturbation unaccompanied by cell injury.
MATERIALS AND METHODS. Unilateral neocortical spreading depression was elicited in mouse cerebral hemispheres and confirmed by transcranial blood flow and extracellular potential measurements. After 3 days, mice were imaged at 4 T using magnetization transfer techniques. Astroglial reactivity was determined immunohistochemically, and protein expression in control and experimental hemispheres was compared using proteomic techniques.
RESULTS. Sixteen ([mean ± SD] ± 3) spreading depressions (n = 10) were recorded in experimental hemispheres. Spreading depression was never observed in control hemispheres. At 3 days, an 8% decrease (p < 0.05, n = 4) in magnetization transfer signal intensity was measured in experimental hemispheres, which was associated with a 37% increase (p < 0.001, n = 4) in the intensity of glial fibrillary acidic protein staining. Proteomic analysis performed 3 days after the induction of spreading depression showed upregulation of at least 56 proteins, including extracellular and intracellular elements.
CONCLUSION. Magnetization transfer at 4.0-T MRI is a sensitive method for detecting glial reactivity and changes in protein expression not associated with cell injury. These results suggest magnetization transfer MRI techniques may have potential for detecting glial reactivity in physiologic processes such as learning and in early disease states.
Reactive gliosis is a hallmark of virtually all disease states in the brain [1]. After numerous forms of injury, glia (astrocytes and microglia) hypertrophy, elaborate cell processes, and sometimes divide [2, 3]. Gliosis does not require cell injury, however, because glial reactivity may accompany physiologic processes such as motor skill learning [4]. The form of glial reactivity depends on underlying disease. Irreversible cell injury, as in infarction, is accompanied by glial proliferation and scar formation [3]. In contrast, increased neuronal activity without cell loss, as in some forms of epilepsy, is often associated with glial hypertrophy but not hyperplasia [5]. Indeed, among the earliest changes seen in Alzheimer's disease, before widespread neuronal loss, is astroglial hypertrophy and process elaboration without proliferation in areas of diffuse amyloid accumulation [6].
Recurrent spreading depression is a non-injurious propagating phenomenon associated with transient loss of membrane potential and electrophysiologic activity in brain cells [7]. After recurrent spreading depression in rats, astrocytes hypertrophy and show an increase in astrocyte-specific glial fibrillary acidic protein immunostaining and content throughout the experimental hemisphere [8, 9]. Recurrent spreading depression in the absence of ischemia does not cause neuronal injury [10]. Therefore, experimental spreading depression is an opportune perturbation to begin studying the phenomenon of glial reactivity in the absence of detectable cell injury.
Magnetization transfer MRI provides a method for augmenting tissue contrast based on exchange of spin magnetization between free protons and those bound to macromolecular proteins [11]. After bound protons are partially saturated, signal from free protons decreases in proportion to the relative amount of water bound to the protein [12]. Magnetization transfer measurements have shown sensitivity in detecting diseases involving edema, loss of cellular structure, and protein mobilization (i.e., proteolysis), as in multiple sclerosis [13, 14], ischemia [15], or neoplastic disease [16]. Our experiments were designed to test whether magnetization transfer is a sensitive method for detecting glial reactivity unaccompanied by injury, in which cells show enhanced cellular structure and increased protein expression.
Materials and Methods
Animal Preparation and Recording
This study was approved by the Duke University Animal Care and Use
Committee. Male C57 black mice (n = 10, 24–27 g) were prepared
for unilateral neocortical spreading depression using modifications of
techniques described by others
[7–9].
Mice were anesthetized by inhalation of isoflurane (5% induction, 1.5% during
surgery, and 1% during cerebral blood flow and electrophysiologic recordings)
in a 50–50 oxygen–nitrogen gas mixture and mechanically
ventilated. Body temperature was monitored with a rectal thermometer and
maintained at a mean ± SD of 37°C ± 1°C using a heating
pad and heat lamp. A 0.8-mm burr hole was drilled in the right posterior
occiput, and recurrent spreading depressions were induced by application of
gel foam saturated in 3M KCl directly to intact dura overlying the right
occipital cortex. Saturated gel foam was refreshed every 15 min so that the
KCl stimulus was effectively constant. Therefore, spreading depressions
occurred at the maximal frequency allowable in mouse cerebral cortex given the
experimental conditions described previously. Spreading depressions were
monitored by measuring changes in cerebral blood flow accompanying cellular
depolarization using a laser Doppler probe placed against the intact skull
over the frontal cortex, 2 mm to the right of the sagittal suture.
Noninvasive laser Doppler measurements were obtained for animals undergoing imaging to exclude the possible confounding effects of injury associated with additional craniotomies and electrode insertion for electrophysiologic recordings. Nonetheless, although blood flow changes described previously were characteristic for spreading depression [17–19], recurrent spreading depressions were also confirmed electrophysiologically in a separate group of mice (n = 3), using extracellular DC microelectrode recordings through small 0.5-mm burr holes over experimental and contralateral frontal neocortexes (Fig. 1A, 1B). Recording electrodes were connected to a low-level amplifier. Spreading depression produced 15- to 20-mV extracellular deflections, in a time course and frequency consistent with spreading depression in other preparations. As expected, spreading depression, which does not cross white matter pathways, was never seen in the contralateral cortex. After 2 hr of spreading depression, recording was stopped, incisions were closed, and animals were extubated. One hour after the righting reflex returned, mice showed normal movement, feeding, and grooming behaviors.
|
|
MRI
Magnetization transfer MRI was performed using a 4.0-T MRI system in a
group of four animals after induction of spreading depression (Fig.
2A,
2B,
2C,
2D). Three days after recovery
from spreading depression, animals were reanesthetized with an intraperitoneal
injection of ketamine and xylazine and placed into the magnet bore. Core
temperature was maintained at 35–37°C using a warm-water jacket. A
customized radiofrequency coil was constructed to acquire coronal slices at a
256 x 256 matrix and 4-cm2 field of view. Eleven 1.5-mm
sections were obtained covering the whole brain. Matrix, field of view, and
slice thickness were the same for all sequences used. Fast spin-echo
T2-weighted images were acquired (TR/TE, 2,000/17; number of excitations, 8;
and echo-train, 8). Magnetization transfer MR images were then obtained with
an offset frequency of 600 Hz (600/13; and number of excitations, 8). Image
intensity in spreading depression neocortex was compared with the
contralateral control neocortex by drawing regions of interest around the left
and right neocortical hemispheres on three consecutive slices beginning at
1.5-mm rostral to the site of KCl application. Because transient glial
reactivity occurs throughout the affected cortex, although somewhat
heterogeneously, regions of interest were drawn to include the entire
neocortical hemispheres (including gray and white matter, separated from
underlying hippocampus) on each side, as with immunohistochemical image
analysis (Fig. 3A,
3B,
3C). Mean image intensity in
the right (experimental) neocortex was compared with that in the left
(control) neocortex. Data were then calculated as the percentage change in
image intensity in the right hemisphere compared with the left. A group of
three animals not undergoing spreading depression were imaged as control
subjects.
|
|
|
|
|
|
|
Immunohistochemistry and Image Analysis
Three days after spreading depression, after MRI was performed, four
animals were reanesthetized with isoflurane and euthanized by transcardial
perfusion fixation with 40 mL of normal saline followed by 40 mL of fixative
(2% paraformaldehyde, 0.075 M of lysine-HCl, 0.010 M of sodium periodate, in
37 mM of phosphate buffer, pH of 6.2). Brains were removed and post-fixed in
the same fixative with 20% sucrose for 24 hr. Brains were then frozen and
stored at –80°C until sectioned for staining.
Thirty-micrometer coronal sections were cut from blocks extending rostrally from the frontal cortex with a cryostat microtome. Astroglial-specific glial fibrillary acidic protein immunohistochemistry was performed with modifications based on prior methods [9]. Cut sections were first floated in phosphate-buffered saline for 20–30 min and then placed into 4% paraformaldehyde in 0.1 phosphate buffer for 20 min. Sections were incubated in phosphatebuffered saline containing 0.3% H2O2 and 0.25% Triton X-100 (Sigma) for 30 min, rinsed in phosphate-buffered saline, and incubated for 1 hr in a blocking solution containing 3% normal goat serum (0.25% Triton X-100 [SIGMA] and 0.001% sodium azide) in phosphate-buffered saline at room temperature. Sections were rinsed, and the primary antiglial fibrillary acidic protein antibody (rabbit anticow, #Z0334, Dako) was applied at a dilution of 1:500 in the phosphate-buffered saline–azide–Triton X-100 solution at 4°C, overnight.
Sections were washed three times in phosphate-buffered saline and incubated with a biotinylated secondary antibody (goat antirabbit, #BA-1000, Vector Laboratories) at a 1:500 dilution. An ABC Elite tertiary detection agent (Vector Laboratories) was applied, and sections were finally incubated in Vector Red chromogen solution (#SK-4800, Vector Laboratories). Sections were air-dried on slides, dehydrated through graded alcohols, and cleared in xylenes before cover-slipping.
Slides were transilluminated using a fiberoptic flat light source, and electronic images were acquired using a 200mm zoom lens attached to a 12-bit charge-coupled camera. The camera was run on a Pentium workstation (NEC Computers International) for Windows (Microsoft), and images were stored on CD-ROMs. MCID-M5 software (Imaging Research) was used to operate the camera. Regions of interest were drawn around the left (control) and right (experimental) neocortical hemispheres, and optical density values based on immunohistochemical staining intensity were obtained using ImageJ software (public domain, Research Services Branch, National Institute of Mental Health). As with imaging, ratios of left-versus-right staining intensity were calculated and compared with left-to-right differences in normal mice.
Proteomics
To evaluate for changes in protein expression between left (control) and
right (experimental) hemispheres, we sacrificed a second group 3 days after
spreading depression, and brains were removed quickly and placed into standard
Ringer's solution at 4°C. Left and right cerebral hemispheres were divided
1.0 mm anterior to the site of KCl application, and the neocortexes were
dissected free from the underlying hippocampus and diencephalon. Neocortical
fragments were then quickly frozen in liquid nitrogen and homogenized in
low-salt buffer (50 mM of hepes, 40 mM of KCl, 60 mM of MgCl2, 0.2%
NP-40, pH of 8.0) and centrifuged at 2,000g for 15 min. Supernatants
were applied to a 0-2 NaCl SMART Mono-Q PC 1.6/5 anion exchange column (GE
Healthcare). Fractionated samples were separated by stretch one-dimensional
sodium dodecyl sulphate-polyacrylamide gel electrophoresis. Protein bands were
developed according to standard techniques
[20–22].
Gels were overlaid and aligned in Photoshop 7.0 (Adobe Systems) for analyses
of changes in protein expression.
The sodium dodecyl sulfate-polyacrylamide gel electrophoresis–separated protein bands were excised, destained, washed, and in-gel digested using porcine trypsin. The resulting peptides were concentrated and desalted on C18 ZipTips (Millipore). The peptides were eluted directly onto a 10 x 10 matrix-assisted laser desorption/ionization (MALDI) target, using 10-mg/mL a-cyano-4-hydroxycinnamic acid in 50% acetonitrile and 0.1% trifluoroacidic acid.
All samples were analyzed using the 4700 Proteomics Analyzer (Applied Biosystems). For each sample, both peptide mass fingerprints and 3 tandem mass spectrometry spectra were acquired in the reflector mode. Data were analyzed using the GPS Explorer software, which is part of the GPS Explorer workstation for the 4700 Proteomics Analyzer (Applied Biosystems). Data were searched against the Swiss-Prot database (Hinxton), using MASCOT (Matrix Science), as a search algorithm.
Results
Spreading Depression
Spreading depressions were elicited in the right neocortexes of
experimental animals. Contralateral neocortexes served as intraanimal controls
in which spreading depression never occurred. Noninvasive laser Doppler blood
flow measurements and surface DC electrode recordings confirmed the presence
of spreading depressions. Application of 3M KCl to intact dura over the right
occipital cortex resulted in 16 ± 3 spreading depressions (range,
13–19 spreading depressions; n = 10) over a 2-hr period.
Spreading depressions were measured in 10 animals using a noninvasive laser
Doppler flowmeter (Fig. 1A,
1B). As seen in the rat
preparation [23], a marked
relative increase in cerebral blood flow accompanied spreading depression,
lasting up to 1 min after recovery of extracellular potential (Fig.
1A,
1B). In three animals in which
concomitant electrophysiologic recordings were performed, extracellular
potential changes were always accompanied by blood flow changes, and blood
flow changes corresponding to spreading depression were never witnessed
without expected electrophysiologic changes.
Magnetization Transfer MRI
Magnetization transfer signal intensity in the experimental cortex was
directly compared with that in the contralateral control cortex. Three days
after spreading depression, magnetization transfer MRI showed a 8.3 ±
3% decrease in signal intensity in the right (experimental) hemisphere
relative to the left (control) (n = 4), representing a statistically
significant difference (p < 0.05) compared with that in normal
animals (mean left-to-right decrease, 0.3 ± 2%; n = 3) (Fig.
2A,
2B,
2C,
2D). Given that the phenomenon
of glial reactivity after unilateral spreading depression occurs throughout
the affected cortex but may be somewhat heterogeneous
[8], regions of interest were
drawn to include each hemisphere in its entirety. This approach was also used
when comparing hemispheres for differences in immunostaining. Three
consecutive slices were analyzed 1.5 mm rostral to the site of KCl application
to avoid the potentially confounding influence of focally increased, rather
than decreased, magnetization transfer signal intensity at the site of the
craniotomy and KCl application. This latter magnetization transfer change was
likely due to focal necrosis in the occipital cortex secondary to toxicity of
direct KCl application.
Immunohistochemistry
Recurrent spreading depressions caused a statistically significant increase
in neocortical glial fibrillary acidic staining. In eight brain sections from
four animals, glial fibrillary acidic protein staining intensity increased by
34% (p < 0.001) in experimental hemispheres compared with those in
controls (Fig. 3A,
3B,
3C). Control animals showed no
significant right-to-left difference (p > 0.5, n = 3).
Higher magnification revealed hypertrophied astroglial cell bodies and
increased numbers of stained cell processes (Fig.
3A,
3B,
3C) in spreading depression
hemispheres.
Proteomic Analysis
Fractionization and sodium dodecyl sulfate–polyacrylamide gel
electrophoresis analysis of homogenates from left and right hemispheres showed
a generalized increase in protein expression in the experimental versus the
contralateral control hemisphere. Figure
4A,
4B shows a comparison of
one-dimensional stretch gels from each cerebral hemisphere in one animal;
Figure 4B shows at least 56
upregulated proteins in the experimental neocortex after spreading depression.
MALDI-TOF-TOF mass spectroscopic analysis of upregulated proteins, combined
with mouse protein and DNA database interrogation, initially identified half
of the proteins (Fig. 4A,
4B). These proteins include
extra- and intracellular structural elements, metabolic enzymes, kinases, and
additional miscellaneous proteins involved in a variety of cellular
processes.
|
|
Discussion
Experimental spreading depression transforms glia into reactive cells such as those seen in neurodegenerative disease, epilepsy, and stroke but does not cause neuronal injury. Defining MRI correlates of spreading depression–induced reactive gliosis could lead to increased sensitivity in the detection of a wide spectrum of brain diseases and could provide a means for monitoring novel therapies. The goal of this project was to provide the initial groundwork for the development of a model of spreading depression–induced reactive gliosis in mouse cerebral cortexes and define its MRI features. Our mouse model of reactive gliosis is intended to serve as a foundation for subsequent experiments combining genomic manipulation with MRI to study mechanisms by which reactive gliosis modulates brain function and regeneration after injury.
After recurrent spreading depression, astrocytes hypertrophy and show a significant increase in astrocyte-specific glial fibrillary acidic protein immunostaining that persists over 2 weeks [8]. Because spreading depressions do not propagate across white matter pathways, the phenomenon (and reactive gliosis) is confined to a single hemisphere. Recurrent spreading depression in the absence of ischemia does not cause detectable neuronal injury [10]. In our experiments, the mouse model was chosen because its genome sequence has recently been completed [24], facilitating proteomic analysis and future studies investigating molecular mechanisms of reactive gliosis.
Our results show an increase in the magnetization transfer effect in the cerebral cortex after experimental spreading depression, as opposed to the decreases in magnetization transfer values that are seen in multiple sclerosis and other disease processes. This effect, represented as a decrease in signal intensity, suggests an increase in the relative fraction of protons bound to macromolecular proteins. Therefore, magnetization transfer signal intensity may reflect an overall increase in protein content in neocortexes undergoing spreading depression. Proteomic analysis 3 days after spreading depression induction confirmed increased protein expression, concomitant with reactive astrocytosis.
Although in our study immunohistochemical staining was targeted to mouse astrocytes, microglial cells are also likely to be reactive after spreading depression, as was seen in the rat [9]. Thus, dynamic changes in both astrocytes and microglia likely contribute to the measured increase in magnetization transfer. Indeed, although morphologic changes in neurons after spreading depression have not been described, spreading depression has been shown to induce changes in expression of some neuronal proteins [25], suggesting that neurons and glia contribute to overall change in protein expression. Several proteins initially characterized by proteomic analysis in these experiments are likely neuronal in origin.
Proteomics refers to the large-scale characterization of the entire protein complement of a cell line, tissue, or organism. The recent rapid growth of proteomics is a direct result of advances made in large-scale nucleotide sequencing of genomic DNA, which facilitates protein identification using advanced microsequencing techniques. The recent completion of the mouse genome greatly aids the identification of spreading depression–induced changes in gene products. Our proteomic analysis represents only a first step in analyzing the multitude of spreading depression–induced changes in the neuroproteome. Using high-throughput MALDI-TOF-TOF mass spectroscopy, we were able to identify 50% of the altered gene products. Ongoing experiments are attempting to sequence the remaining proteins by more conventional electrospray mass spectroscopy. In addition, we are undoubtedly identifying only a small number of possible spreading depression–induced changes in the neuroproteome because of limitations in the initial protein separation technology—protein electrophoresis. Crude cell mixtures are too complex to be resolved in their entirety on single one-dimensional or 2D gels, and the ability to resolve and detect low-copy proteins in particular is limited partly because the most abundant proteins dominate the gel. We are currently addressing these issues by focusing first on meaningful subproteomes isolated by a variety of techniques such as affinity column arrays. By starting the analysis with subproteomes (e.g., membrane-bound, nuclear, or purine-binding proteins), we can resolve low-copy proteins (e.g., signaling molecules) from abundant proteins such as cytoskeletal elements, which are more likely to influence magnetization transfer effects but less likely to have meaningful roles in initiating changes in cell function.
These experiments contribute to the growing awareness that the brain is capable of rapid, dynamic change, not only in diseases but also after comparatively benign stimuli. As we have shown, glia undergo dramatic morphologic change in response to the relatively benign electrophysiologic perturbation of spreading depression. Indeed, the morphologic changes in astrocytes after spreading depression can even be seen in a model of motor skill learning, in which glial cell hypertrophy accompanies increased synaptogenesis [4]. Developing imaging techniques for imaging glial reactivity in the absence of cell injury not only facilitates in vivo study of this phenomenon in physiologic states such as learning but also leads to more sensitive techniques for detecting disease in its earliest stages.
Acknowledgments
We thank Richard P. Kraig for many insightful discussions and helpful suggestions regarding the experiments described in this manuscript.
References
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
