The site of embolization related to infarct size, oedema and clinical outcome in a rat stroke model - further translational stroke research.

Abstract

Reliable models are essential for translational stroke research to study the pathophysiology of ischaemic stroke in an effort to find therapies that may ultimately reduce oedema, infarction and mortality in the clinic. The purpose of this study was to investigate the relation between the site of arterial embolization and the subsequent oedema, infarction and clinical outcome in a rat embolic stroke model.
Thirty-six male Sprague-Dawley rats were thromboembolized into the internal carotid artery. The site of occlusion was demonstrated by arteriography. Following histological preparation and evaluation, the size of the hemispheres and the infarcts were measured by quantitative histology and planimetry. Another parallel stroke model study was subsequently examined to investigate if the conclusions from the first study could be applied to the second study.
The median size of the infarct was 40% of the ipsilateral hemisphere in both the 19 animals with occlusion localised to the intracranial part of the internal carotid artery and in the 11 animals where the main trunk of the middle cerebral artery was occluded. In 5 animals, occlusion of the extracranial part of the internal carotid artery resulted in significantly smaller infarcts compared to other groups (p < 0.01). Another independent study re-confirmed these results. Furthermore, significant correlations (R > 0.76, p < 0.0001) were found between 1) cortical, subcortical, and total infarct volumes, 2) oedema in percent of the left hemisphere, 3) clinical score before termination and 4) postoperative weight loss.
Distal occlusions of the intracranial part of the internal carotid or middle cerebral arteries resulted in comparable large sized infarctions and oedema. This indicates that investigators do not need a similar number of such occlusions in each experimental group. Contrary to observations in the clinic, distal internal carotid artery occlusions did not result in worse outcome than middle cerebral stem occlusions, but this finding may be explained by the controlled emboli size in this experimental stroke model.

Full text

Page 1

RESEARCHOpen Access
The site of embolization related to infarct size,
oedema and clinical outcome in a rat stroke
model - further translational stroke research
Karsten Overgaard1*, Rune S Rasmussen1, Flemming F Johansen2
Abstract
Background and purpose: Reliable models are essential for translational stroke research to study the
pathophysiology of ischaemic stroke in an effort to find therapies that may ultimately reduce oedema, infarction
and mortality in the clinic. The purpose of this study was to investigate the relation between the site of arterial
embolization and the subsequent oedema, infarction and clinical outcome in a rat embolic stroke model.
Methods: Thirty-six male Sprague-Dawley rats were thromboembolized into the internal carotid artery. The site of
occlusion was demonstrated by arteriography. Following histological preparation and evaluation, the size of the
hemispheres and the infarcts were measured by quantitative histology and planimetry. Another parallel stroke
model study was subsequently examined to investigate if the conclusions from the first study could be applied to
the second study.
Results: The median size of the infarct was 40% of the ipsilateral hemisphere in both the 19 animals with
occlusion localised to the intracranial part of the internal carotid artery and in the 11 animals where the main trunk
of the middle cerebral artery was occluded. In 5 animals, occlusion of the extracranial part of the internal carotid
artery resulted in significantly smaller infarcts compared to other groups (p < 0.01). Another independent study re-
confirmed these results. Furthermore, significant correlations (R > 0.76, p < 0.0001) were found between 1) cortical,
subcortical, and total infarct volumes, 2) oedema in percent of the left hemisphere, 3) clinical score before
termination and 4) postoperative weight loss.
Conclusions: Distal occlusions of the intracranial part of the internal carotid or middle cerebral arteries resulted in
comparable large sized infarctions and oedema. This indicates that investigators do not need a similar number of
such occlusions in each experimental group. Contrary to observations in the clinic, distal internal carotid artery
occlusions did not result in worse outcome than middle cerebral stem occlusions, but this finding may be
explained by the controlled emboli size in this experimental stroke model.
Introduction
Commonly human stroke patients suffer from direct
occlusion of the middle cerebral artery (MCAO), or
from occlusions of the internal carotid artery (ICAO),
which reduce or prevent sufficient blood flow through
the middle cerebral artery (MCA).
Recently, the importance of establishing reliable ani-
mal models is clear, as hundreds of effective neuropro-
tective therapies in animals failed to show beneficial
effect in the treatment of human stroke [1]. It has been
argued, that the quality of many animal models provided
unreliable results due to insufficient study design and
type of model, randomisation, blinding, inclusion and
exclusion criteria, measurement of physiological para-
meters etc. [2]. The animal model used in the present
investigation is different from several other MCAO ani-
mal models by using obstructing autologous blood clots
that compromise cerebral arterial circulation in focal
brain regions. The advantage of this model in transla-
tional stroke research is its similarity to human ischae-
mic stroke, and this advantage has been exemplified by
the fact that this model provided results of thrombolytic
* Correspondence: karove01@geh.regionh.dk
1Stroke Unit, University Hospital of Copenhagen, Gentofte Hospital, Hellerup,
Denmark
Full list of author information is available at the end of the article
Overgaard et al. Experimental & Translational Stroke Medicine 2010, 2:17
http://www.etsmjournal.com/content/2/1/17
© 2010 Overgaard et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative
Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and
reproduction in any medium, provided the original work is properly cited.

Page 2

efficacy in rats before benefit of human thrombolytic
treatment was documented [3-5]. The primary disadvan-
tage of this embolic stroke model is a large variation of
infarct volumes, but although the variation in infarct
volume makes it necessary to use relatively large animal
group sizes, the model may provide a sound basis for
translational stroke research.
The main purpose of this study was to investigate the
relationship between the site of arterial occlusion and
the subsequent size of the resulting infarct in a rat
embolic stroke model. If the consequences of emboliza-
tion in this animal model are shown to be similar to
consequences of embolization in humans, the model
will be of high translational significance due to its prog-
nostic abilities.
Measurement of infarct volumes generally has been
the main endpoint of investigations in animal stroke
models, but many factors may interfere with measure-
ments of infarcted tissue such as brain swelling by
oedema or brain shrinkage by resorption. In this study
we investigated several aspects associated to the embolic
location, namely infarct volume, oedema size, mortality
and effects of re-embolization. An endpoint to the
investigation was to examine whether or not evenly dis-
tributed locations of the emboli and the consequently
degrees of ischemia among treatment and control
groups were necessary to avoid unwanted selection bias,
especially when groups are relatively small in numbers.
Thus, in the present study, the consequences of emboli
localisation especially upon infarct volume were investi-
gated for the first time and further related to both the
formation of oedema and the clinical outcome.
Materials and methods
Animals in this study have previously been investigated
for other purposes [6]. Re-using animals to provide
extended results and observations is in favour of the
methodology encompassed in the concepts of refine-
ment, reduction and replacement within animal labora-
tory science [7,8]. This ethical methodology corresponds
to the aim of medical researchers to use as few animals
and as responsibly as possible.
All experiments were approved and conducted accord-
ing to the guidelines by the Animal Research Committee
of University of Copenhagen. Thirty-six male Sprague-
Dawley rats weighing from 240 to 420 grams were used.
The rats were housed in a 12-hour light-dark cycle at a
room temperature of 22 ± 1°C, with unrestricted access
to water and chow before and after surgery. Anaesthesia
was induced with subcutaneous administration of Hyp-
norm® (fluanison 4 mg/kg and fentanyl 0.12 mg/kg, Jans-
sen Pharmaceutica, Beerse, Belgium) combined with
atropine 0.05 mg/kg and intraperitoneal administration
of diazepam 2 mg/kg (Apozepam 5 mg/ml; Apothekernes
Laboratorium, Lier, Norway). The body temperature was
kept at 37 ± 0.5°C by rectal temperature monitoring and
a thermostat controlled heating lamp.
Surgical procedure and preparation of the emboli
Small white ‘arterial-like’ elongated autologous micro-
clots rich in fibrin were produced in vitro as reported
earlier [4,5,9]. Then by modification and amendment of
this method, the suspension of microclots was injected
into a 120-cm long PP10 (PP10; ID 0.28 mm, also corre-
sponding to clot diameter) air filled tube [10]. The
microclots in the suspension accumulated to several sin-
gle macroclots of larger and variable length by this
method [10]. A macroclot with a length of about 8 mm
(0.44 mm3) was selected for embolization by cutting a
piece of the tube with the macroclot enclosed. Both the
right femoral vein and the external carotid artery (ECA)
were catheterised, while the pterygopalatine artery was
ligated as described earlier [9]. Clotting associated with
the inserted catheters was avoided by continuous flow
of heparinised (5 IU/ml) saline through the line at the
speed of 0.5 ml/h until the catheter was removed at the
end of surgery. The mean arterial blood pressure
(MABP), blood glucose, haemoglobin, PaO2, O2-satura-
tion, PaCO2and pHawere measured (Radiometer ABL
2, Copenhagen).
Embolization produced arterial occlusion documented by
arteriography
The right common carotid artery (CCA) was clamped
temporarily using a ligature, and the embolus was gently
injected through the saline-filled ECA catheter by the
use of 0.3 ml saline. This procedure caused embolization
of the right hemisphere by occlusion of the internal car-
otid artery (ICA) or MCA. After this procedure carotid
arteriography was performed immediately by bolus
injection of 0.20 ml heparinized (5 IU/ml) iohexol
(Omnipaque, 300 mg I/ml; Nycomed, Roskilde, Den-
mark) through the ECA catheter. Angiograms were pro-
cessed onhigh-resolution
Fotochemische Werke, Berlin, Germany). Since the CCA
was unclamped during the angiographic procedure, this
procedure did not affect the embolization.
Arteriograms were evaluated without knowledge of the
fixation procedure in accordance with a modification of
a previously reported topographically rank ordered scale
[9]. All arteriograms were evaluated individually and
assigned one of the following scores: 0 = patent arteries,
1 = middle cerebral artery occlusion (MCAO in one or
more branches), 2 = MCAO at the stem, 3 = distal and
intracranial ICA occlusion, 4 = extracranial and proxi-
mal ICA occlusion. The angiographic scores are illu-
strated on Figure 1. During surgery using an unblinded
evaluation of the arteriograms, fourteen (39%) of the
films (RetinaXOE;
Overgaard et al. Experimental & Translational Stroke Medicine 2010, 2:17
http://www.etsmjournal.com/content/2/1/17
Page 2 of 10

Page 3

animals achieved a score of 0, 1 or 1.5 with an occlusion
in the distal half of the MCA. These 14 animals were
considered poorly embolized and were embolized again
by repeating all embolization procedures, and the final
grade of occlusion was determined by the repeated
arteriography. In two animals embolization and arterio-
graphy according to the above procedures were per-
formed 3 times, while the other 12 animals were
successfully embolized on the second attempt. In order
not to expose animals to an excessive amount of dislo-
cated embolus material, no more than three emboliza-
tion attempts were allowed in this study. After the final
arteriography and then ligations of vessels, femoral and
neck wounds were closed. The effect of anaesthesia was
reverted with naloxone 0.3 mg/kg s.c. (Narcanti®, Du
Pont Pharmaceuticals, U.S.A.).
Clinical Evaluation
About two hours after recovering from anaesthesia, the
clinical neurological status was evaluated using a rank
ordered scale for rats developed by Bederson et al. [11];
a score of 0 reflected normal neurology and no observa-
ble deficit (grade 0), a score of 1 corresponded to mod-
erate clinical damage observed by forelimb flexion
(grade 1), a score of 2 reflected severe clinical damage
and decreased resistance to lateral push (and forelimb
flexion) without circling (grade 2), while a score of 3
corresponded to maximum clinical damage with the
same behaviour as grade 2 with circling (grade 3). A
Bederson score has been found to correlate to the
infarct sizes [11]. Twenty-four hours after anaesthesia all
animals were weighted, re-evaluated clinically as above
and terminated.
Preparation of brain tissue and measurement of infarct
size
Twelve animals were perfusion fixed with formalin. In
this group and in a second group of 13 other animals
the brains were carefully removed and immersed in for-
malin. A third group of 11 animals was left dead for
three hours at room temperature before the brains were
dissected out and immersed in formalin. The 3 different
fixation procedures were evenly distributed without sig-
nificant differences between the 3 groups with different
grades of vascular occlusion. As shown previously, fixa-
tion procedures had no influence on the results of this
current investigation, because infarct and oedema
volumes calculated as percentages of the corresponding
hemisphere were independent of fixation procedures [6].
Standard histological procedures were used and coro-
nal serial sections of the brain with a distance of 400
μm between each section (about 20-25 sections per
brain) were stained with hematoxylin-eosin [12]. This
distance between sections was used, because a tripling
of the distance between brain sections can decrease the
accuracy of the measurements by up to 4% [13-15].
Volume calculations using image analysis and planimetry
All brain sections were examined carefully by conven-
tional light microscopy by investigators blinded to the
experimental protocol. Borderlines of the cortical and
subcortical infarcts were delineated and separated from
healthy brain regions using a pen, and the infarcted tis-
sue could be recognised easily (Figure 2). All brain sec-
tions were scanned into a personal computer by the use
of an image analysis program (Sidney Data, Copenhagen,
Denmark),providing volumemeasurements of
Figure 1 The ventral aspect of carotid-vertebral arteries in the
rat. All embolized animals should obtain occlusion scores of 1.5 to
4. The following arteries are identified: (clamped) common carotid
artery [CCA], (ligated) pterygopalatine artery [PTY], (ligated) external
carotid artery [ECA], internal carotid artery [ICA] and middle cerebral
artery [MCA].
Overgaard et al. Experimental & Translational Stroke Medicine 2010, 2:17
http://www.etsmjournal.com/content/2/1/17
Page 3 of 10

Page 4

hemispheres and infarctions. The accuracy of the plani-
metry was secured by performing repeated control-scans
of a rectangle sized 115 mm2placed in a horizontal as
well as in a diagonal position.
Calculations and analysis of weight, oedema and statistics
Animals were weighted preoperative and immediately
before the animal was terminated at 24 hours. Post-
operative weight loss in percent of preoperative weight
was calculated as: 100 × (preoperative weight - pre-ter-
mination weight)/preoperative weight. Under the
assumption that differences between right and left hemi-
sphere volumes constituted oedema sizes, the volumes
of oedema in percent of the left hemisphere was calcu-
lated as: 100 × (RH - LH)/LH, with RH and LH repre-
senting volumes of the right and left hemispheres,
respectively. Oedema volumes in percent of infarct
volumes were calculated as: 104× (RH - LH)/(INF ×
RH), where INF represented infarct sizes in percent of
RH. Non-parametric statistical analyses of data were
performed, because all data were either categorical, or
could be placed on rank ordinal or ratio-interval scales
with no apparent normal distribution. Mann-Whitney
U-test for unpaired observations, Wilcoxon matched-
pairs test for paired observations and the Spearman test
for correlation of ranked pairs were used. P-levels below
0.05 were considered significant.
Results
On the final blinded evaluation of the arteriograms the
surviving animals were divided into 3 different sub-
groups: one subgroup of 11 animals had a grade 2 (stem
of the MCA) occlusion, a second subgroup of 19 ani-
mals had a grade 3 (intracranial ICA) occlusion and the
remaining subgroup of 5 animals had a grade 4 (extra-
cranial ICA) occlusion. One animal found dead at
20 hours was excluded from the results, and all other
animals survived 24 hours. Figures 3A and 3B illustrate
grade 2 and 3 embolizations. The 3 different fixation
methods were evenly distributed without significant dif-
ferences between the 3 groups on vascular occlusions.
Physiological parameters
No significant differences were observed between any of
the 3 subgroups in physiological parameters. Results
(median and interquartile range) in all 35 animals were:
MABP: 76 mmHg (70-85), haemoglobin: 10.4 mmol/l
(10.1-10.7), glucose: 11.3 mmol/l (11-13), pHa: 7.395
Figure 2 (a + b + c). Hematoxylin-eosin stained paraffin-embedded coronal section of an immersion fixed brain from an embolized rat with a
proximal middle cerebral artery occlusion (A). Arrows indicate the border of the infarct, constituting 47% of the volume of the right hemisphere.
The borderline of infarction is not easily identified by differences of structure or colour intensity. Medium power magnification (B). The border
between healthy and dead brain is more obvious (arrows). High power magnification (C) of the narrow penumbral zone. The border between
healthy and live neurons now distinctively exposed (arrows). A rim of ‘dark’ neurons adjacent to the normal neurons is observed and deeper in
the infarct the cell bodies have shrunken from the surrounding neuropil [33,34].
Overgaard et al. Experimental & Translational Stroke Medicine 2010, 2:17
http://www.etsmjournal.com/content/2/1/17
Page 4 of 10

Page 5

(7.38-7.41), PaCO2: 39 mmHg (37-42), PaO2: 87 mmHg
(82-92), O2-saturation: 96% (96-97), preoperative weight
of animals: 350 g (313-397), size of both hemispheres
(excluding cerebellum, brainstem and olfactory bulbs):
795 mm3(740-892), respectively (comparing all sub-
groups, all p > 0.05, ns). Similar to earlier reports by
other investigators [14], we found no influence of rat
age or preoperative weight on infarct volume.
Infarct size was dependent on the efficacy of
embolization documented by grade of arteriographic
occlusion
As shown in table 1 the 11 animals with a MCAO had
median cortical, subcortical and total infarct volumes of:
17.2%, 20.5% and 40% respectively of the right hemisphere
volume. The 19 animals with intracranial ICAO had
almost identical median infarct volumes of 16.6%, 19.9%
and 40% respectively (all p > 0.45). Thus animals with a
grade 2 or 3 occlusion obtained similar infarct volumes.
Two of the 5 animals with an extracranial ICAO had
in addition an unintended insufficient ligation of the
pterygopalatine artery documented by the arteriography.
One of the animals with an open pterygopalatine artery
had a total infarct of only 1.7% of right hemisphere
volume even embolized three times with the first two
arteriograms showing patent arteries, the third showing
extracranial ICA occlusion. These 5 animals all had sig-
nificantly lower size of infarcts (table 1); total infarct
was in all 5 animals below 7% (median 3.5%) of the
right hemisphere volume.
Figure 3 (a + b). Arteriograms showing proximal occlusion of the main-stem of the middle cerebral artery (A) and a distal occlusion of the
internal carotid artery (B), arteriographic score values of 2 and 3 respectively. The following arteries are identified: internal carotid artery [ICA],
anterior cerebral artery [ACA], middle cerebral artery [MCA], posterior cerebral artery [PCA], common carotid artery [CCA] and catheter [CAT].
Table 1 Main results
Type of
occlusion
N Percent post-operative
weight loss
Oedema in percent of the volume of the
left hemisphere
Volume of infarct in percent of the right
hemisphere
SubcortexCortex Total infarct
MCA 115.5 (5-6) 5.6 (2-11) 20.5 (9-29) 17.2 (9-18)40 (23-47)
Intracranial
ICA
195.9 (5-8) 6.6 (2-12)19.9 (12-34)16.6 (13-23) 40 (22-56)
Extracranial
ICA
5 4.8* (1-5)0.8* (0-1)0.7*** (0-1)2.0** (1-3)3.5*** (2-5)
Main results: postoperative weight loss, volumes of oedema and infarction. Median values and values in brackets are 25 and 75 percentiles. The 5 animals with
an extracranial ICAO significantly differed from all the other 30 animals with an intracranial site of occlusion. * = p < 0.05, ** = p < 0.01, *** = p < 0.001.
Overgaard et al. Experimental & Translational Stroke Medicine 2010, 2:17
http://www.etsmjournal.com/content/2/1/17
Page 5 of 10