Sustained reperfusion after blockade of glycoprotein-receptor-Ib in focal cerebral ischemia: an MRI study at 17.6 Tesla.

Sustained Reperfusion after Blockade of Glycoprotein-
Receptor-Ib in Focal Cerebral Ischemia: An MRI Study at
17.6 Tesla
Mirko Pham1*., Xavier Helluy2., Christoph Kleinschnitz3., Peter Kraft3, Andreas J. Bartsch1, Peter
Jakob2, Bernhard Nieswandt4, Martin Bendszus1", Guido Stoll3"
1Department of Neuroradiology, University of Heidelberg, Heidelberg, Germany, 2Department of Experimental Physics, Section V, University of Wu¨rzburg, Wu¨rzburg,
Germany, 3Department of Neurology, University of Wu¨rzburg, Wu¨rzburg, Germany, 4 Rudolf-Virchow-Center, DFG Research Center for Experimental Biomedicine and
Chair of Experimental Medicine, University of Wu¨rzburg, Wu¨rzburg, Germany
Abstract
Background: Inhibition of early platelet adhesion by blockade of glycoprotein-IB (GPIb) protects mice from ischemic stroke.
To elucidate underlying mechanisms in-vivo, infarct development was followed by ultra-high field MRI at 17.6 Tesla.
Methods: Cerebral infarction was induced by transient-middle-cerebral-artery-occlusion (tMCAO) for 1 hour in C57/BL6
control mice (N = 10) and mice treated with 100 mg Fab-fragments of the GPIb blocking antibody p0p/B 1 h after tMCAO
(N= 10). To control for the effect of reperfusion, additional mice underwent permanent occlusion and received anti-GPIb
treatment (N= 6; pMCAO) or remained without treatment (N = 3; pMCAO). MRI 2 h and 24 h after MCAO measured cerebral-
blood-flow (CBF) by continuous arterial-spin labelling, the apparent-diffusion-coefficient (ADC), quantitative-T2 and T2-
weighted imaging. All images were registered to a standard mouse brain MRI atlas and statistically analysed voxel-wise, and
by cortico-subcortical ROI analysis.
Results: Anti-GPIb treatment led to a relative increase of postischemic CBF vs. controls in the cortical territory of the MCA
(2 h: 44.266.9 ml/100 g/min versus 24 h: 60.568.4; p = 0.0012, F(1,18) = 14.63) after tMCAO. Subcortical CBF 2 h after tMCAO
was higher in anti-GPIb treated animals (45.365.9 vs. controls: 33.664.3; p = 0.04). In both regions, CBF findings were clearly
related to a lower probability of infarction (Cortex/Subcortex of treated group: 35%/65% vs. controls: 95%/100%) and
improved quantitative-T2 and ADC. After pMCAO, anti-GPIb treated mice developed similar infarcts preceded by severe
irreversible hypoperfusion as controls after tMCAO indicating dependency of stroke protection on reperfusion.
Conclusion: Blockade of platelet adhesion by anti-GPIb-Fab-fragments results in substantially improved CBF early during
reperfusion. This finding was in exact spatial correspondence with the prevention of cerebral infarction and indicates in-vivo
an increased patency of the microcirculation. Thus, progression of infarction during early ischemia and reperfusion can be
mitigated by anti-platelet treatment.
Citation: Pham M, Helluy X, Kleinschnitz C, Kraft P, Bartsch AJ, et al. (2011) Sustained Reperfusion after Blockade of Glycoprotein-Receptor-Ib in Focal Cerebral
Ischemia: An MRI Study at 17.6 Tesla. PLoS ONE 6(4): e18386. doi:10.1371/journal.pone.0018386
Editor: Andreas Meisel, Charite´ Universitaetsmedizin Berlin, Germany
Received October 18, 2010; Accepted March 5, 2011; Published April 1, 2011
Copyright: � 2011 Pham et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by a Postdoctoral Fellowship granted to MP from the Medical Faculty of the University of Heidelberg, Germany (www.
medizinische-fakultaet-hd.uni-heidelberg.de) and by the Deutsche Forschungsgemeinschaft (www.dfg.de), SFB 688 (TP B1 granted to BN and GS, TP A13 granted
to CK and Z2 to PJ). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: mirko.pham@med.uni-heidelberg.de
. These authors contributed equally to this work.
" These authors are joint senior authors on this work.
Introduction
Ischemic stroke is a major cause of death and disability [1]. A
significant proportion of strokes are caused by thromboembolic
occlusion of major intracerebral vessels such as the middle cerebral
artery (MCA). The complex cellular and molecular processes
underlying the development of ischemic brain lesions are
incompletely understood [2,3]. This also applies to the situations
in which extended and clinically severe strokes evolve despite
‘‘favorable’’ removal of the vessel occluding clot either spont-
aneously or by thrombolysis giving rise to reperfusion [4,5].
Reperfusion is a prerequisite for replenishing brain areas at risk for
infarction with oxygen and nutritional factors, but, on the other
hand, elicits detrimental processes referred to as reperfusion injury.
We could recently show that interference with critical steps of
platelet tethering to the vessel wall can prevent ischemic stroke in
the mouse model of transient MCA occlusion (tMCAO) [6,7].The
initial tethering of platelets at sites of vascular injury is mediated by
GPIb-V-IX, a structurally unique receptor complex exclusively
expressed in platelets and megakaryocytes [8]. GPIba is
indispensable for platelet adhesion under conditions of high shear
such as in the arterial cerebrovascular system. Inhibition of the von
PLoS ONE | www.plosone.org 1 April 2011 | Volume 6 | Issue 4 | e18386

Willebrand factor (vWF)-binding site of GPIba with Fab fragments
of the antibody p0p/B in wild-type mice abrogated platelet
tethering and adhesion in a model of mechanically induced
arterial thrombosis as well as in ischemic stroke, while unspecific
Fab fragments had no effect [7]. Cerebral infarcts were
significantly smaller when assessed histologically or by 1.5T
MRI. These findings make GPIb an attractive target for clinical
development of an antithrombotic drug in acute stroke.
There is recent evidence that GPIb, so far mainly regarded as
instrumental in haemostasis and thrombus formation, can
profoundly guide inflammation [9]. Thus, the effect of GPIb
blockade in cerebral ischemia could be due to sustained patency of
blood vessels during reperfusion or, alternatively, due to a primary
anti-inflammatory effect [10]. To address this important issue, we
employed multimodal MRI at ultra-high-field strength (UHF-
MRI) to monitor lesion development in relation to cerebral blood
flow in GPIb-Fab-treated mice and naive controls after tMCAO.
As principal finding, we show that cerebral perfusion after
tMCAO is sustained in GPIb-treated mice after removal of the
vessel occluding thread while in normal mice perfusion further
decreases leading to progressive stroke. Thus, anti-GPIb treatment
substantially ameliorates infarct progression during early ischemia
and reperfusion.
Materials and Methods
Experimental design and animal stroke model
All procedures and animal studies were approved by the
Regierung von Unterfranken (Wuerzburg, Germany, approval
number: 55.2-2531.01-23/04 and -55/09) and conducted in
accordance with the recommendations for the performance of
basic experimental stroke studies as previously published [11].
The main experimental group in this study were anti-GP1b
treated mice (adult male C57/BL6 mice weighing 20–25 g
(Charles River, Sulzfeld, Germany) undergoing one hour of
tMCAO (N=10). These mice received 100 mg p0p/B Fab
fragments [12] intravenously 1 hour after tMCAO, that is, after
1 hour of occlusion at the time point at which the thread was
removed. This regimen led to significantly smaller infarcts
compared to control-treated animals in our previous study [7].
In the present study we used naive mice (adult male C57/BL6
mice weighing 20–25 g (Charles River, Sulzfeld, Germany) as
controls for the efficacy to induce full-blown MCA infarcts by
1 hour of tMCAO(N=10) because in our previous study there
was no difference between naive mice and mice treated with an
unspecific Fab fragment [7]. To investigate whether reperfusion is
required for the therapeutic effect of GPIb blockade, additional
anti-GPIb treated mice (N=6) and control mice (N= 3)
underwent permanent MCAO (pMCAO). Furthermore, another
group of control mice underwent sham operation (N=3).
The experimental procedures were performed as described in
detail previously [7,13,14]. Briefly, a standardized suture coated
with silicon rubber (6021PK10; Doccol Company, Redlands, CA,
USA) was introduced into the right common carotid artery and
advanced over the internal carotid artery to the origin of the
MCA. The suture was fixed and left in situ and animals were
allowed to recover. Operation time per animal did not exceed
15 minutes. After 60 min. animals were re-anesthetized and the
suture was withdrawn to allow tissue reperfusion (tMCAO). For
pMCAO, the thread was left within the vessel until the end of the
experiments at day 1. Sham operation included preparation of the
common carotid artery and ligation of its branches without thread
insertion. The operations were performed under inhalation
anesthesia (2.0% isoflurane in a 70%/30% N2O/O2 mixture)
and the body temperature was maintained at 37uC using a servo-
controlled heating pad. All subjects were subsequently followed in-
vivo by serial multimodal UHF-MRI at 2 h and 24 h.
An additional group of anti-GP1b treated mice was investigated
at an even earlier time point after tMCAO, i.e. 1 h after thread
removal, to address the question whether the observed hypoper-
fusion at 2 h is preceded by hyperperfusion. In this scenario, a
deleterious effect of reperfusion on tissue fate (reperfusion injury)
might be functional rather than a beneficial effect of sustained
reperfusion for the prevention of infarct progression under anti-
GP1b treatment. In our local experimental setting of multimodal
UHF-MRI, logistic circumstances restrict the earliest time point
applicable for data acquisition to around 1 h after removal of the
thread.
Multimodal UHF-MRI of experimental cerebral ischemia
in-vivo
A detailed description of the imaging protocol is given in
previous work [15]. Cerebral perfusion was measured using a
modified arterial spin labeling (CASL) method [16,17,18]. To
benefit especially from increased longitudinal magnetization and
the elevation of the T1 relaxation time for detailed anatomical
mapping of CBF and group analysis, all measurements were
performed at ultra-high magnetic field strength (Avance 17.6T,
750 MHz, Bruker BioSpin GmbH, Ettlingen, Germany). Image
maps of cerebral perfusion were calculated on a pixel-by-pixel
basis according to Detre et al. [18]. The degree of the inversion
efficiency was assumed to be alpha= 0.7 [19,20]. In close
approximation to the value recently reported by Leithner et al.
for the mouse brain [21] the brain-blood partition coefficient value
for water was assumed to be lambda= 0.90 mL/g. Slice selective
T1 mapping was measured with a single slice partial saturation
inversion recovery RARE sequence (TI of 0.02 s, 0.5 s, 1.0 s,
1.5 s, 2.0 s, 3.0 s, 5.0 s, 10.0 s). The recovery time after
acquisition of each image was 10 s (echo-train-length = 16,
TEeff = 30 ms). Inversion of magnetization was performed by a
6 mm slice selective adiabatic hypersecant pulse. T1 relaxation
time constants were calculated voxel-wise applying first, a 3
parameter fit to estimate the efficiency of the inversion pulse and
then, a 2 parameter fit with a fixed averaged value for the
inversion pulse efficiency typically between 95% and 97%.
Diffusion weighted imaging (DWI) was performed with a
pulsed-field gradient Setjskal-Tanner-like multislice spin echo
sequence because echo-planar-imaging suffers from extreme
susceptibility artifacts at ultra-high magnetic field strength.
Diffusion sensitization was only performed along the slice direction
to keep the overall acquisition time low [22]. Images with different
b-values, 0 and 800 s/mm2, were acquired to allow for the
calculation of apparent diffusion coefficient (ADC) maps of brain
water. Whole brain coverage was achieved by thirteen coronal
slices acquired with a matrix size of 64664, FOV 1.861.8 cm, in
plane resolution 2816281 mm, slice thickness = 0.5 mm, interslice
distance = 1 mm, TE/TR=22.3/2000 ms. Repeated measure-
ment of the b= 800 s/mm2 DWI experiments (number of
repetition NR=3) led to an overall acquisition time for diffusion
weighted experiments of 8 min. ADC maps were calculated by
applying the common equation ADC=20.001256ln (SIB800/
SIB0). The b value of 800 s/mm
2 was chosen to maintain a high
signal-to-noise ratio for each acquisition, in case motion artifacts of
the spin echo DWI sequence would degrade other acquisitions.
T2 relaxometric mapping was performed for the in-vivo
delineation of infarcted brain tissue at 24 h. Single slice T2-
weighted (T2-w) imaging was performed using a Carr-Purcell-
Meiboom-Gill (CPMG) multi-spin echo sequence collecting thirty
Sustained Reperfusion after Gp1b Blockade
PLoS ONE | www.plosone.org 2 April 2011 | Volume 6 | Issue 4 | e18386

two echoes at TR/TE=4.2/2000 ms. T2 relaxation times
constants were calculated voxel-wise by fitting the intensities of
the 20 first echoes to a monoexponential model. CBF, T1 and T2
relaxometric maps were measured each with the exact same
geometry for a 1.5 mm thick slab centered at the bregma as the
operational definition of the central MCA territory.
For high-resolution structural imaging with whole-brain cover-
age, an additional strongly T2 weighted 2D turbo spin-echo
sequence was acquired (RARE factor 16, TR=8 s, effective
TE=56.44 ms, 2 averages, 13 coronal slices with an image matrix
of 1286128 were acquired, FOV=1.8 cm61.8 cm, slice thick-
ness = 0.5 mm, interslice distance = 1 mm, overall acquisition time
of 2 min).
At the host console measurements and data processing were
performed with the ParaVision software (version 3.02, Bruker
BioSpin GmbH, Ettlingen, Germany). Further image calculation
and fitting procedures were done using MATLABH (The Math-
works Inc., Natick, MA, USA).
During UHF-MRI measurements mice were anesthesized by
2.0% Isoflurane in medical air (21%). The respiratory rate was
monitored using an air-balloon positioned ventrally underneath
the mouse body. The body temperature was constantly measured
on the body surface and actively maintained at 37uC.
Statistical and image analysis
The extraction of brain tissue from the scalp and skull was done
by manual segmentation for each subject and time point. Packages
from the FMRIB Software Library FSL (version 4.1) [23] were used
for motion correction, registration (FLIRT) [24] and statistical
image analysis. Intra-subject linear alignment and registration to a
common standard template [25] was achieved by a step-wise affine
procedure with six degrees of freedom. For voxel-wise statistical
analyses, the global CBF maps were normalized by the overall
average CBF value of the contralateral hemisphere.
CBF values early (at 2 h) and at 24 h after the experimental
procedure were analysed for statistically significant voxelwise
changes (24 h vs. 2 h) within the framework of the General Linear
Model and corrected for multiple comparisons by nonparametric
permutation testing using randomise, part of the FSL software
library [23]. Randomise implements the method of permutation
testing based on randomisation to correct for the multiple
comparisons involved in testing across all image voxels to
adequately protect against false-positive detections as described
in detail by Nichols and Holmes [26]. For quantitative group
comparisons, selected regions-of-interest (ROIs) were delineated in
atlas space: 1) the cerebral cortex in the center of the MCA
territory 2) the subcortex including the ipsilateral caudoputamen
and pyramidal tract. Statistical analysis of ROIs was done by a
262 repeated measures ANOVA with factors of GROUP (tMCAO
anti-GPIb; tMCAO controls) between-subjects, and TIME (2 h;
24 h) within-subjects. Additional groups to control for the
experimental procedure (Sham controls, N= 3) and to control
for recanalization and reperfusion (pMCAO anti-GPIb, N= 6)
were analyzed separately. The risk of cerebral infarction was
determined on within-group probability maps by averaging binary
segmentations results of healthy vs. infarcted brain tissue within-
subject. For each animal binary segmentation of cerebral
infarction was performed in an automated fashion by applying a
threshold of 34 ms T2 relaxation time on the T2 relaxometric
maps at 24 h. Among different segmentation results for stepwise
increasing T2 relaxation times, this cut off showed best agreement
with visual delineation of infarction on T2-w imaging and with
histological 2,3,5-triphenyltetrazolium chloride stain in selected
subjects. Manual input was given only for the removal of
intraventricular CSF. In addition, whole-brain volumetric analysis
of infarcted tissue was retrieved by manual segmentation on T2-w
RARE images.
Results
Cerebral perfusion in naive controls and anti-GPIb treated
mice after transient and permanent MCAO
In naive control mice hypoperfusion extended over cortical and
subcortical ROI’s in the center of the MCA territory at 2 h and
was followed by a significant further decrease in CBF at 24 h after
tMCAO (cortical CBF (ml/100 mg/min): 40.964.4 (2 h) vs.
26.063.2 (24 h), p = 0.022; subcortical CBF: 33.664.3 (2 h) vs.
24.863.2 (24 h), p = 0.009). In contrast, anti-GPIb treated mice
showed significant reperfusion of the cortex (44.266.9 (2 h) vs.
60.568.4 (24 h), p = 0.037). In the subcortex, initial CBF of the
anti-GPIb group was higher than in controls (33.664.3 (controls at
2 h) vs. 45.365.9 (anti-GPIb at 2 h), p = 0.047). Subcortical CBF
remained stable at 24 h in anti-GPIb treated mice (46.967.5) but
further declined in controls (24.863.2). Table 1 gives an overview
of mean CBF values within cortical and subcortical ROI’s in the
center of the MCA territory.
Correspondingly, on voxel-wise analysis, clusters of significant
perfusion activation (reperfusion) and deactivation (deterioration
of hypoperfusion) were found. In anti-GPIb treated mice reperfu-
sion was located in the cortex, mainly in the distribution of the
middle and posterior cerebral artery. In control mice, however,
hypoperfusion deteriorated in the center of the cortical territory of
the middle cerebral artery and in a smaller temporobasal cluster in
the distribution of the posterior cerebral artery. Figure 1 shows the
location of significant clusters of perfusion activation/deactivation
in standard space (blue overlay for the contrast 2 h.24 h; yellow
overlay for the contrast 24 h.2 h).
In contrast, anti-GPIb treated mice with permanent vessel
occlusion (pMCAO) experienced progression of severe hypoper-
fusion (2 h: 42.9611.5 vs. 24 h: 35.465.2) and developed
extended complete MCA infarctions (not shown) similar to naive
controls. This indicates that anti-GPIb treatment is ineffective after
permanent vessel occlusion. Sham operated control mice (N= 3)
did not exhibit any perfusion abnormalities at 2 h or 24 h after the
experimental procedure and did not develop cerebral infarctions
(Figure 1).
In line with the results at 2 h and 24 h, increased cortical CBF
was also observed 1 h after tCMAO (1 h: 28.263.5 ml/100 g/
min vs. 24 h: 110.09610.0; n = 4/group; p = 0.002). This effect
was still robust when evaluating CBF ratios between ipsilateral and
contralateral mirror ROIs: (1 h: 0.1960.01 vs. 24 h: 0.5660.06;
p = 0.005). In addition, quantitative cortical and subcortical T2
values (ms) representing infarct probability in these areas were
similar in comparison with the original group of anti-GP1b treated
mice measured at 2 h and 24 h (cortical ROI 1 h: 30.660.7 vs.
cortical ROI 24 h: 32.262.1; subcortical ROI 1 h: 30.860.3 vs.
subcortical ROI 24 h: 45.762.4).
Probability of cerebral infarction and quantitative ADC
values in naive and anti-GPIb treated mice
Cerebral infarction was determined on T2 relaxometric images
by binary segmentation at a threshold of 34 ms. This cut-off was
previously demonstrated to give accurate estimates of infarct
extension at 17.6 Tesla field strength [15]. Figure 2 shows that the
cut-off value of T2= 34 ms yields best results of infarct extension
when comparing the results of a stepwise segmentation procedure
with increasing quantitative T2 thresholds as compared to high
resolution T2-w RARE imaging.
Sustained Reperfusion after Gp1b Blockade
PLoS ONE | www.plosone.org 3 April 2011 | Volume 6 | Issue 4 | e18386

Probability maps of cerebral infarction were rendered group-
wise using the individual segmentation results of each animal.
They are given along with maps of the mean ADC for each group
and time point in Figure 3. Of note, for the given segmentation
threshold, in control mice after tMCAO cerebral infarction was
already manifest at 2 h in the basal ganglia and covered the
Table 1. All outcome measures per group, time point and location.
Controls Anti-GPIb
CBF (ml/100 mg/min) 2 h vs. 24 h Cortex 40.964.4 vs. 26.063.2 44.266.9 vs. 60.568.4
Subcortex 33.664.3 vs. 24.863.2 45.365.9 vs. 46.967.5
ADC (mm2/s*1024) 2 h vs. 24 h Cortex 6.4860.27 vs. 5.7560.23 7.8860.28 vs. 7.5360.26
Subcortex 6.0860.60 vs. 5.2960.33 7.8660.33 vs. 7.1260.26
qT2 (ms) 2 h vs. 24 h Cortex 37.2461.96 vs. 60.0563.15 28.660.4 vs.29.0 60.97
Subcortex 33.4161.05 vs. 49.8963.15 30.660.3 vs. 37.462.2
Probability of Infarction (%) 2 h vs. 24 h Cortex 60.969.3 vs. 95.162.8 17.462.1 vs. 34.568.1
Subcortex 79.169.9 vs. 10060 21.567.9 vs. 64.8614.5
Values are expressed as group means and corresponding standard errors. As the main finding sustained reperfusion was observed in anti-GPIb treated mice, whereas
controls exhibited significant progression of hypoperfusion from 2 h to 24 h. In the cortex of the MCA territory, reperfusion significantly increased from 2 h to 24 h in
anti-GPIb treated mice presumably related to a larger capacity of collateral blood flow as compared to the subcortex. In the subcortex of anti-GPIb treated mice,
improved reperfusion as compared to controls was reflected by a significantly higher baseline CBF at 2 h. Sustained reperfusion both in the cortical and subcortical
territory of the MCA was associated with a protection from cerebral infarction as evident by a low probability of infarction.
doi:10.1371/journal.pone.0018386.t001
Figure 1. CBF and statistical maps of voxel-wise group comparisons. Sustained reperfusion is demonstrated by significantly elevated cortical
perfusion in anti-GPIb treated mice as compared to persisting severe hypoperfusion in control mice. Color maps of mean CBF are given for each
group and time point (left, CBF). The results of voxel-wise statistical analyses of change in CBF over time are shown on the statistical parameter maps
(right, Statistical maps). The spatial distribution of significant reperfusion in the anti-GPIb treated group is indicated by the yellow overlay (yellow
contrast of 24 h.2 h). The spatial distribution of significant deterioration of hypoperfusion in control mice after tMCAO is indicated by blue overlay
(blue contrast of 2 h.24 h).
doi:10.1371/journal.pone.0018386.g001
Sustained Reperfusion after Gp1b Blockade
PLoS ONE | www.plosone.org 4 April 2011 | Volume 6 | Issue 4 | e18386

complete cortical and deep MCA territory at 24 h. In anti-GPIb
treated mice infarction did not occur with relevant probability at
2 h after tMCAO (17.462.1%) and at 24 h occurred with a
significantly lower probability in the cortex and basal ganglia than
in corresponding ROI’s in controls (cortex: 34.568.1% vs.
95.162.8%, p= 0.0001; subcortex: 64.8614.5% vs. 10060%,
p= 0.01). Quantitative T2 values showed similar group differenc-
es. Cortical and subcortical quantitative ADC values exhibited a
significantly stronger decrease in controls than in anti-GPIb treated
mice (cortex: p = 0.001; subcortex: p= 0.003). Table 1 displays all
values of ADC, quantitative T2 and probabilities of infarction for
each group, ROI and time point.
Whole brain volumetric measurement of cerebral infarction by
manual delineation was performed additionally and showed
similar group differences between control mice and anti-GPIb
mice after tMCAO as observed by automated segmentation in the
center of the MCA territory (Figure 4).
Intracerebral hemorrhage was not observed in any of the anti-
GPIb treated mice which is in accordance with our previous study
[7].
Comprehensive group analyses of CBF response from
cortical and subcortical regions-of-interest
Outcome measures of cerebral perfusion (CBF) and completed
cerebral infarction (qT2) were calculated from cortical and
subcortical ROIs as indicated in atlas space and are plotted in
Figure 5. The cortical ROI was associated with the ipsilateral
cortical MCA territory (yellow overlay in ipsilateral cortex), the
subcortical location comprised the ipsilateral caudoputamen and
deep pyramidal tract (red overlay in ipsilateral subcortex).
Group comparison of CBF from the cortical ROI, similar to
voxel-wise image analysis, demonstrated that the time course of
CBF between both groups went in opposite directions showing
deterioration in controls (2 h: 40.964.4; 24 h: 26.063.2) and
strong recovery of CBF indicating sustained reperfusion in anti-
GPIb treated mice (2 h: 44.266.9; 24 h: 60.568.5). This is
reflected by the significant interaction between the factors GROUP
and TIME in the repeated measures ANOVA (p= 0.0012,
F(1,18) = 14.63).
The significant main effect of improved cortical reperfusion in
anti-GP1b treated mice in ipsilateral ROIs was still robust when
CBF ratios between ipsilateral and contralateral mirror ROIs were
evaluated (2 h: 0.3160.070 vs. 24 h: 4160.07; p = 0.01).
Group comparison of CBF from the subcortical ROI showed
deterioration of severe hypoperfusion in naive controls (2 h:
33.664.4 and 24 h: 24.863.2). In contrast, sustained reperfusion
was observed in anti-GPIb treated mice (2 h: 45.365.9 and 24 h:
46.967.5). The postischemic baseline value of subcortical CBF at
2 h was significantly lower in controls than in the anti-GPIb treated
group (2 h: 33.664.4 vs. 45.365.9; p = 0.04).
Quantitative cortical T2 values were significantly different
between groups (GROUP: F(1,18) = 14.63, p,0.00001), both time
points (TIME: F(1,,18) = 35.83, p,0.00001) and also with a strong
interaction between GROUP and TIME (GROUPxTIME:
F(1,,18) = 35.83, p,0.00001). All T2 measures are given in Table 1.
The additional group of N= 4 anti-GP1b treated mice
investigated very early after tMCAO 1 h after thread removal
also exhibited strong reperfusion, which was most marked in the
cerebral cortex: 28.263.5 ml/100 g/min (cortex at 1 h) vs.
110.1610.0 (cortex at 24 h); p = 0.002. This effect was robust
against evaluating CBF ratios between ipsilateral and contralateral
mirror ROIs: 0.1960.01 (cortex at 1 h) vs. 0.5660.06 (cortex at
24 h); p = 0.005. In this additional series, the observed quantitative
T2 values (ms), and hence the probability of infarction, within the
cortical (1 h: 30.660.7 vs. 24 h: 32.262.1) and subcortical ROI
(1 h: 30.860.3 vs. 24 h: 45.762.4) were similar in comparison
with the original group of anti-GP1b treated mice undergoing
measurements at 2 h and 24 h.
Figure 2. Results of automated stepwise binary segmentation of cerebral infarction. Different segmentation results are displayed for
increasing quantitative T2 values. Accurate estimates of the extent of cerebral infarction as compared to high-resolution T2-w imaging (upper left)
and histological stains with 2,3,5-triphenyltetrazolium chloride were achieved for a segmentation threshold of T2 = 34 ms. Delineation of infarction
was performed on T2 relaxometric images at 24 h after the experimental procedure.
doi:10.1371/journal.pone.0018386.g002
Sustained Reperfusion after Gp1b Blockade
PLoS ONE | www.plosone.org 5 April 2011 | Volume 6 | Issue 4 | e18386