Strain-Related Differences in Susceptibility to Transient Forebrain Ischemia in SV-129 and C57Black/6 Mice
Background and Purpose We explored susceptibility to injury after global ischemia in SV-129 and C57Black/6 mice, two commonly used background strains in genetically engineered mice.
Methods Mice (n=84) were subjected to 15, 30, or 75 minutes of bilateral common carotid artery (BCCA) occlusion followed by reperfusion for 72 hours. BCCA occlusion was performed under halothane or chloral hydrate anesthesia; in one experiment, mean arterial blood pressure and regional cerebral blood flow (laser Doppler flowmetry) were matched by controlled exsanguination. Baseline absolute blood flow measurements were obtained in both strains using a tracer, N-isopropyl-[methyl 1,3-14C]-p-iodoamphetamine, indicator fractionation technique (n=5 per group). Vascular anatomy of the circle of Willis was visualized by intravascular perfusion of carbon black ink (n=10 per group). Cerebrovascular reactivity was assessed by measuring the diameter of pial vessels (intravital microscopy) to acetylcholine (ACh) superfusion (0.1 to 10 mmol/L) in a closed cranial window preparation (n=29).
Results Resting blood flow values did not differ between groups in striatum, cerebellum, and brain-stem regions. SV-129 mice were less susceptible than C57Black/6 mice to ischemic injury (0.0±0.0 versus 1.3±0.3 damage in hippocampal CA1 region after 30 minutes of ischemia in SV-129 and C57Black/6, respectively; P<.01). Cellular damage (grade 1 to 3 injury) comparable to 30-minute BCCA occlusion was achieved only after 75 minutes of ischemia in SV-129 mice (1.1±0.3). Ischemic damage was also significantly less in SV-129 mice after blood pressure and flow were matched during ischemia in halothane-anesthetized SV-129 mice (0.5±0.3 versus 1.4±0.2, P<.05), or after chloral hydrate anesthesia (0.4±0.2 versus 1.5±0.4, P<.05). Hypoplastic posterior communicating arteries were found in all 10 C57Black/6 mice and may explain the greater susceptibility of these mice to injury after BCCA occlusion. More robust vasodilation to ACh in C57Black/6 mice could also indicate genetic differences in responses to vasoactive substances.
Conclusions C57Black/6 mice exhibit enhanced susceptibility to global cerebral ischemic injury, an incompletely formed circle of Willis, and augmented pial vessel dilation to ACh compared with SV-129 mice. Our findings suggest that strain differences may confound results when genetically engineered mice generated from more than a single background strain are used.
Genetically engineered mice provide a unique opportunity to evaluate the role of single gene products in models of cerebral ischemia. SV-129 and C57Black/6 mice are the most commonly used background strains.1 2 3 4 5 SV-129 mice are the source of the embryonic stem cells that, after successful gene targeting, become implanted in blastocysts of C57Black/6 mice (eg, see endothelial, type III, NOS, and neuronal, type I, NOS knockout mice1 2 ). Differences in genetic background can influence the outcome of cerebral ischemia6 7 and therefore may confound the interpretation of results in genetically engineered mice derived from more than a single parent strain.5
We recently characterized a mouse model of global ischemia in which common carotid and basilar arteries were occluded.4 Although this model clearly established a relationship between type I NOS gene deficiency and resistance to stroke damage,4 mortality was sufficiently high and the preparation technically difficult enough that development and characterization of a more simple model seemed justified. Hence, we searched for a technique that was simple, and highly reproducible, in which the neuropathological changes were consistent. Some of these criteria are fulfilled by reversible BCCA occlusion. Using this model, we determined that C57Black/6 mice were significantly more susceptible to global ischemic injury and that the resistance could not be accounted for by differences in blood pressure or choice of anesthesia, but that it was probably related to the presence of a circle of Willis lacking vascular connections between the anterior and posterior circulations and poor collateral blood flow during BCCA. Vasomotor tone and reactivity to vasoactive substances may also contribute to differences in ischemic susceptibility between the two wild-type strains.
Seventy-three SV-129 (18 to 24 g) and 70 C57Black/6 (18 to 24 g) mice were obtained from Taconic Labs (Germantown, NY) and allowed food and water ad libitum.
Thirty Minutes of BCCA Occlusion
SV-129 (n=8) and C57Black/6 (n=12) mice were anesthetized with 2% halothane, maintained with 1% halothane plus 70% N2O and 30% O2 using a Fluotec 3 vaporizer (Colonial Medical), and kept in a supine position. The duration of anesthesia was approximately 5 minutes (4 minutes to occlude BCCA and 1 minute to close the suture). Rectal and scalp temperatures were recorded and maintained at approximately 37°C with a heating pad (Temperature Control; FHC) and heating lamp (Skytron, Daiichi Shomei). After a ventral midline cervical incision, the common carotid arteries were exposed and ligated with 6-0 silk sutures. Immediately after the ligation, anesthesia was turned off and the animals were assessed for level of consciousness and pupillary and corneal reflexes. To reperfuse, animals were briefly reanesthetized with 1% halothane and the ligatures removed. After observing blood flow return, the incision was closed.
Animals were killed at 72 hours after ischemia by injecting euthanasia solution (50 mg/kg IP pentobarbital; Euthanasia-5 solution, Henry Schein). They were then perfused via the ascending aorta with heparinized physiological saline (2 USP heparin Units/mL, Abbott Laboratories) followed by 10% phosphate-buffered formalin (Poly Scientific). The brains were removed and postfixed in 10% phosphate-buffered formalin for 3 days and then embedded in paraffin. Five-μm coronal sections were stained with hematoxylin and eosin.
Hippocampal CA1, striatum, and cerebral cortex were evaluated in two tissue sections by investigators (H.H., M.S.) naive to the treatment group. Striatal damage was assessed at the level of the globus pallidus and triangular nucleus. Hippocampus and cerebral cortex were evaluated at the level of the subthalamic nucleus. The grading scale was as described by Pulsinelli and Brierley8 : 0=normal, 1=a few (<30%) neurons damaged, 2=many neurons (30% to 70%) damaged, and 3=majority of neurons (>70%) damaged. The mean values of both sides were calculated and used for further analysis.
Matching MABP by Controlled Exsanguination
Because of significant hemodynamic differences during ischemia under halothane anesthesia (Table 2A⇓), we next evaluated ischemic damage after withdrawing 200 μL of blood into a heparinized syringe from the femoral artery of SV-129 mice. Both strains (n=8 per group) were then subjected to 30-minute BCCA occlusion. After reperfusion, approximately 180 μL of shed blood was returned carefully over about 1 minute in SV-129 mice. Physiological parameters and histological outcome were compared at 72 hours.
Chloral Hydrate Anesthesia
To determine whether the difference in susceptibility to ischemia was related to anesthetic effects on vascular resistance as reflected by differences in MABP, chloral hydrate (300 mg/kg IP, Sigma Chemical Co) was used instead of halothane anesthesia. SV-129 (n=12) and C57Black/6 mice (n=7) were then subjected to 30 minutes of BCCA occlusion under physiological monitoring. The change in rCBF was matched in the two groups. Hence, 5 of 12 SV-129 mice were rejected because rCBF did not decrease to 25% after BCCA occlusion. Animals were sacrificed after 72 hours.
Effect of Longer Ischemic Duration on Neuronal Damage in SV-129 Mice
Because SV-129 mice were less susceptible to the consequences of 30 minutes of BCCA occlusion than C57Black/6 mice, we next determined whether 75 minutes of BCCA occlusion in SV-129 mice (n=16) would produce injury equivalent to 15 minutes of ischemia in C57Black/6 mice (n=13) under 1% halothane anesthesia. Animals were killed after 72 hours as described above.
MABP, rCBF, arterial blood gases, pH, and rectal and scalp temperatures were measured. MABP was obtained from the left femoral artery. rCBF was measured by laser Doppler flowmetry (Perimed, PF2B) with a flexible 0.8-mm fiberoptic extension. The tip of the probe was affixed to the intact skull over the left cortex (2 mm posterior and 6 mm lateral from bregma). MABP and rCBF were recorded from 10 minutes before ischemia until 30 minutes after reperfusion. Arterial blood gases (Pao2, Paco2) and pH were measured 10 minutes before ischemia and 30 minutes after induction of ischemia in 30-μL samples (Corning 178; Ciba-Corning Diagnostics). Scalp temperature was measured with a needle probe (model BAT-12, Physitemp Instruments) inserted into the right temporal muscle. Rectal and scalp temperatures were maintained at approximately 37°C as described above. Parallel groups of animals were monitored physiologically in experiments described in “30 Minutes of BCCA Occlusion” above, whereas direct measurements were obtained in animals described in experiments “Matching MABP by Controlled Exsanguination” and “Chloral Hydrate Anesthesia,” also described above.
Absolute CBF Measurements
CBF was determined using an indicator fractionation technique described previously9 10 with some modifications. Animals (n=5 per group) were anesthetized with 1% halothane as above. The right femoral artery and jugular vein were cannulated with PE-10 polyethylene tubing. After determining MABP and blood gases, arterial blood was withdrawn continuously from the femoral artery at a rate of 0.3 mL/min (Stoelting). One microcurie of N-isopropyl-[methyl 1,3-14C]- p-iodoamphetamine (American Radiolabeled Chemicals Inc) dissolved in 0.1 mL saline was injected into the jugular vein as a bolus (<1 second). Twenty seconds after injection, the animal was decapitated and the blood withdrawal terminated simultaneously. The brain was removed quickly, frozen in isopentane solution, chilled with dry ice, and then dissected into right and left hemispheres and subdivided into striatum and remaining forebrain, cerebellum, and brain-stem regions. After adding scintigest (Fisher Scientific) and incubating (50°C for 6 hours), scintillation fluid and H2O2 were added. Twelve hours after shaking, radioactivity in brain and blood were measured by liquid scintillation spectrometry (RackBeta 1209, LKB). CBF was calculated according to the method of Van Uitert and Levy9 and Betz and Iannotti.10
Carbon Black Perfusion Study
Animals (n=10 per group) were perfused under pentobarbital anesthesia (50 mg/kg IP) with 10 mL of physiological saline followed by 2 mL of concentrated solution of Pelican carbon black ink (Pelican AG, D-3000) via the left cardiac ventricle until the tissues (eg, tongue, lips, and gums) turned black. After decapitation, the brains were carefully removed into 10% buffered formalin for 24 hours before examination. The vessels of the circle of Willis and their branches were examined with a Leica-Wild M3Z microscope.
Pial Vessel Diameter and Its Response to ACh
Mice (SV-129, n=11; C57Black/6, n=18) were initially anesthetized with 2% halothane and intubated. Mice were artificially ventilated (SAR-830/P) with a mixture of 70% N2O/30% O2. Alpha-chloralose (1%, 80 to 100 mg/kg) was then injected via the femoral vein, after which halothane inhalation was discontinued. End-tidal CO2 was continuously monitored by a microcapnometer (Columbus Instruments).
A stainless steel cranial window ring (8.0 mm in inner diameter, 2.0 mm in height) containing three ports was embedded into a loop of bone wax so as to adhere to the skull. A craniotomy (2.0×1.5 mm) was made in the left parietal bone within the ring of the window. The dura was opened while the brain surface was superfused with aCSF. The volume under the window was approximately 0.1 mL. The composition of aCSF was as follows (in mmol/L): Na+ 156.5, K+ 2.95, Ca2+ 1.25, Mg2+ 0.67, Cl− 138.7, HCO3− 24.6, dextrose 3.7, and urea 0.67. The pH of aCSF was kept at 7.35 to 7.45 by equilibration with 6.5% CO2, 10% O2, and balance N2, and was monitored continuously with a pH meter (Corning Inc). The aCSF was superfused by an infusion pump (0.4 mL/min) via a PE-100 tubing connected to a window port. Intracranial pressure was maintained at 5 to 8 mm Hg by adjusting the outlet tubing to an appropriate height. The temperature of aCSF within the window was maintained at 36.5°C to 37.0°C.
Pial vessels were visualized by an intravital microscope equipped with a video camera. The diameter of a single pial arteriole (20 to 30 μm) was continuously measured (C3161, Hamamatsu Photonics) and recorded using the MacLab data acquisition and analysis system. After obtaining a stable diameter and blood pressure, ACh (0.1, 1, and 10 mmol/L) was sequentially superfused and the change in diameter measured over 3 to 4 minutes.
All data were presented as mean±SEM. For nonparametric data (eg, histological grading) Mann-Whitney U test was utilized, whereas ANOVA followed by Dunnett’s test was used to compare physiological parameters, and Student t test was used to compare maximum percent change in pial vessel diameter during ACh superfusion in the two strains. Values of P<.05 were considered statistically significant.
Values varied between 125 and 141 mL · 100 g–1 · min–1 in the four brain regions of SV-129 and C57Black/6 mice. They did not differ between sides or between groups (Table 1⇓). MABP was 93±3 and 74±4 mm Hg for SV-129 and C57 black/6 mice, respectively.
Thirty Minutes of BCCA Occlusion in SV-129 and C57Black/6 Mice
Histological damage was evident in hippocampus, striatum, and cerebral cortex in C57Black/6 mice but not in any regions of SV-129 mice after 30 minutes of BCCA occlusion in halothane-anesthetized mice (Fig 1A⇓). During ischemia, rCBF decreased in both strains, although it was more reduced in C57Black/6 mice (Table 2A⇓). Other physiological variables (rectal and scalp temperatures, arterial blood gases) did not differ except that MABP was lower in C57Black/6 mice at rest and during ischemia (Table 2A⇓).
Matching MABP by Controlled Exsanguination
After controlled exsanguination, there were no group differences in MABP, rCBF, or other physiological parameters (Table 2B⇑). SV-129 mice remained less susceptible than C57Black/6 mice at 72 hours after ischemia (Fig 1B⇑).
Chloral Hydrate Anesthesia
SV-129 mice were also significantly more resistant to ischemia than C57Black/6 mice under chloral hydrate anesthesia (Fig 1C⇑). However, unlike the results under halothane anesthesia, ischemic injury was detected in both strains, probably as a consequence of the lower MABP and rCBF after chloral hydrate versus halothane anesthesia. There were no differences in MABP, rCBF, or other physiological parameters in these experiments (Table 2C⇑).
Effect of Longer Ischemic Duration on Neuronal Damage in SV-129 Mice
In halothane-anesthetized SV-129 mice, 75 minutes of BCCA occlusion approximated the histological damage after 30 minutes of ischemia in C57Black/6 mice (Fig 1A⇑).
Hippocampal CA1 (particularly the medial aspect) and CA2 neurons were most frequently damaged, whereas neurons in CA3 and CA4 were well preserved. Dentate gyrus and hilus regions also showed ischemic changes in both groups.
Rostal neostriatum was more susceptible than caudal in both groups. Small- to medium-sized striatal neurons appeared abnormal most frequently, whereas large neurons were relatively preserved. Globus pallidus was remarkably resistant to ischemic injury in this model.
In the cerebral cortex, ischemic changes were limited to layers 2 and 3 and less frequently 5 and 6, whereas changes consistent with ischemia were noted in ventral posteromedial, ventral posterolateral, posterior and mediodorsal thalamic nuclei, and zona incerta.
During ischemia, halothane-anesthetized animals continued to breathe spontaneously and showed reversible signs (loss of corneal reflex, righting reflex, mydriasis) that recovered within 1 hour after reperfusion.
No seizure activity was witnessed after 15 minutes of occlusion in C57Black/6. Tonic and clonic seizure activity such as rolling and rapid running was observed in 43% of C57Black/6 mice after 30 minutes of ischemia and in 13% of SV-129 mice after 75 minutes of BCCA occlusion.
There was no mortality in SV-129 or C57Black/6 mice after 30 minutes of ischemia, whereas after 75 minutes, 25% of SV-129 mice died.
Carbon Black Study
The rostral circle of Willis did not differ between strains at the level of the anterior and middle cerebral arteries (Fig 2⇓). However, in all 10 C57Black/6 mice, the posterior communicating artery was hypoplastic on both sides. In contrast, this vessel was well formed in all 10 SV-129 mice studied.
Pial Vessel Diameter and Its Response to ACh
The baseline diameters did not differ between strains (SV-129, n=11, 24.9±1.5 mm versus C57Black/6, n=18, 22.5±0.8 mm). Topical superfusion of ACh induced dose-dependent arteriolar dilation in both strains. More robust dilation was observed at each of three concentrations in C57Black/6 (Fig 3⇓). The values for pHa, Paco2, and Pao2 did not differ between groups (7.26±0.02 versus 7.31±0.02 mm Hg, 35.3±1.9 versus 29.3±0.9 mm Hg, and 104±9 versus 120±8 mm Hg for SV-129 versus C57Black/6, respectively). MABP was higher in SV-129 mice (96±4 versus 78±2 mm Hg, P<.01).
Two murine strains commonly used in genetically engineered mice differ in their susceptibility to brain damage after transient forebrain ischemia. After 30 minutes of BCCA occlusion, ischemic changes were significantly less in SV-129 mice. In fact, 75 minutes of BCCA occlusion in SV-129 mice caused only as much ischemic damage as 30 minutes in C57Black/6 mice. Differences in outcome were observed at both 72 hours and 7 days (SV-129: n=7; C57Black/6: n=8; data not shown). The response to anesthetic may explain some of the strain differences, as rCBF changes after ischemia and MABP were higher in SV-129 mice anesthetized with identical concentrations of halothane. However, C57Black/6 mice were still more susceptible when both MABP and %rCBF decreases were matched by controlled exsanguination. Greater ischemic injury in C57Black/6 mice was probably not due to halothane sensitivity because of the similar results with chloral hydrate anesthesia. Hence, the susceptibility differences between strains were not explained by sensitivity of pial or other vessels to specific anesthetics. The findings are especially interesting because resting CBF values did not differ between strains. Poorly developed vascular connections between the anterior and posterior circle of Willis were the most significant risk factor identified in C57Black/6 mice, which was observed in all 10 animals perfused with carbon black ink. This same anomaly was described in the Mongolian gerbil and makes it particularly useful in ischemia studies.11 Barone et al found a similar anomaly in BALB/C mice.6 An absent posterior communicating artery decreases collateral blood flow between the anterior and posterior circulations. If both carotid arteries and the basilar artery are occluded, possibilities for collateral blood flow between the anterior and posterior circulation are eliminated. In this global ischemia model, ischemic damage does not differ significantly between C57Black/6 and SV-129 mice.4 Hence, the integrity of the circle of Willis appears to be an important risk factor contributing to susceptibility to global ischemia in mice.
Factors other than vascular anatomy may also influence susceptibility to ischemia. We observed more robust dilation to topical ACh in pial vessels of C57Black/6 mice. Higher eNOS activity or the efficiency of ACh receptor coupling may underlie such differences, but a precise explanation was not identified at this time. We previously reported that genetically engineered mice deficient in eNOS exhibit higher vascular resistance (MABP approximately 112 mm Hg) and show greater ischemic damage after MCA occlusion.1 12 Our results showing greater dilation to topical ACh in pial blood arterioles (Fig 3⇑) by nitro-l-arginine–reversible mechanisms (data not shown) could reflect upregulation of eNOS activity within vascular endothelium of C57Black/6 mice, and this speculation could be examined experimentally. C57Black/6 mice appeared more susceptible to hypotension after halothane anesthesia (Table 2A⇑) and were reportedly more susceptible to fatty streak formation when fed an atherogenic diet.13 Factors such as the frequency of spreading depression,14 15 seizures,16 NOS1 2 3 4 and superoxide dismutase activity,17 18 glutamate receptor subtype, and density are theoretical explanations for differences in selective vulnerability. We have shown that the density of glutamate receptor subtypes (NMDA, AMPA, kainate) and muscarinic ACh receptors did not differ between C57Black/6 and SV-129 mice (unpublished data, Waeber et al, 1997). Seizure may have contributed to the greater damage in C57Black/6 mice after 30 minutes of BCCA occlusion; although C57Black/6 mice showed no overt seizure activity after 15 minutes of BCCA occlusion, they were more susceptible to ischemic damage.
Several authors have previously reported that mouse strains differ in their susceptibility to focal and global cerebral ischemias. Hara et al3 observed that C57Black/6 mice developed larger lesions than SV-129 mice during focal ischemia/reperfusion, although such differences were not found after permanent focal ischemia.2 Barone et al, the first group to describe BCCA occlusion as a useful model,6 showed that BALB/C mice were more sensitive to focal ischemia than BDF or CFW mice. In addition, the ddY strain showed higher mortality and memory disturbance than the ICR strain after 2 to 30 minutes of BCCA occlusion.19 20 Approximately 35% of CA1 neurons in BALB/C and C57Black/6 mice were ischemic after 30 minutes of BCCA occlusion, and as noted above, both strains have an incomplete circle of Willis.6 Strain differences are also important in rat models of ischemia.14 21
Because SV-129 and C57Black/6 mice together provide the genomic background of many mutant mice,1 2 3 4 5 our results raise questions as to the choice of control groups when using knockout mice with mixed genetic backgrounds. Our data suggest that if wild-type littermates of heterozygote matings are not available, both parent strains must be tested individually. If both parent strains show the same difference (ie, larger or smaller lesions) from the mutant group, additional controls may not be necessary. However, if only one wild-type strain differs from the mutant, then it becomes essential to develop and test wild-type littermates. Alternatively, mutant mice can be backcrossed with a single parent strain for at least 12 generations to ensure homogeneity of genetic background.5 Although laborious, time-consuming, and expensive,5 by doing so, group comparisons become more valid.
Variations in genetic background can lead to erroneous conclusions about the importance of an identified phenotype to a specifically targeted gene. Hence, the results described herein emphasize the need to use proper controls when studying ischemia and cerebrovascular regulation in genetically engineered mice derived from more than a single parent strain.
Selected Abbreviations and Acronyms
|aCSF||=||artificial cerebrospinal fluid|
|BCCA||=||bilateral common carotid artery|
|CBF||=||cerebral blood flow|
|eNOS||=||endothelial nitric oxide synthase|
|MABP||=||mean arterial blood pressure|
|NOS||=||nitric oxide synthase|
|rCBF||=||regional cerebral blood flow|
This research was supported by Massachusetts General Hospital Interdepartmental Stroke Project Grants (NS10828) and by an unrestricted award in Neuroscience from Bristol-Myers Squibb (M.A.M.). We thank Drs Michael Winking, Christian Waeber, Masao Sasamata, and Cenk Ayata for their helpful assistance.
© 1997 American Heart Association, Inc.
- Received April 7, 1997.
- Revision received May 22, 1997.
- Accepted June 6, 1997.
- Copyright © 1997 by American Heart Association
Huang Z, Huang PL, Panahian N, Dalkara T, Fishman MC, Moskowitz MA. Effects of cerebral ischemia in mice deficient in neuronal nitric oxide synthase. Science. 1994;265:1883-1885.
Pulsinelli WA, Brierley J. A new model of bilateral hemispheric ischemia in the unanesthetized rat. Stroke. 1979;10:267-271.
Van Uitert RL, Levy DE. Regional brain blood flow in the conscious gerbil. Stroke. 1978;9:67-72.
Van Lenten B, Prieve J, Navab M, Hama S, Lusis A, Fogelman A. Lipid-induced changes in intracellular iron homeostasis in vitro and in vivo. J Clin Invest. 1995;95:2104-2110.
Kinouchi H, Epstein CJ, Mizui T, Carlson EJ, Chen SF, Chan PH. Attenuation of focal cerebral ischemic injury in transgenic mice overexpressing CuZn superoxide dismutase. Proc Natl Acad Sci U S A. 1991;88:11158-11162.
Yang G, Chan PH, Chen J, Carlson E, Chen SF, Weinstein P, Epstein CJ, Kamii H. Human copper-zinc superoxide dismutase transgenic mice are highly resistant to reperfusion injury after focal cerebral ischemia. Stroke. 1994;25:165-170.
Muguruma K, Nakamuta H, Tsuji M, Yanagita K, Koida M, Maeda H, Manno K, Narama I, Ogawa Y, Hiramatsu Y. Biochemical and histopathological studies on a mouse brain ischemic model induced by bilateral carotid artery occlusion: comparison with the carbon monoxide inhalation method and other ischemic or anoxic models. Nippon Yakurigaku Zasshi. 1991;98:23-29. Abstract in English.
Payan H, Levine S, Strevel R. Effects of cerebral ischemia in various strains of rats. Proc Soc Exp Biol Med. 1965;120:208-209.