Changes in experimental stroke outcome across the lifespan

Abstract

Acute ischemic stroke is a leading cause of mortality and disability in the elderly. Age is the most important non-modifiable risk factor for stroke, yet many preclinical models continue to examine only young male animals. In addition, the majority of stroke deaths now occur in older women, who also shoulder the burden of stroke-related disability, yet no preclinical studies have examined the effect of aging on stroke outcome in female mice. It remains unclear how experimental stroke outcomes change with aging and with biological sex. If sex differences are present, it is not known if these reflect an intrinsic differing sensitivity to stroke or are secondary to the loss of estrogen with aging. To address these questions, we subjected both young and aging C57BL/6 mice of both sexes to middle cerebral artery occlusion (MCAO). Young female mice had smaller strokes compared to age-matched males, an effect that was reversed by ovariectomy. Interestingly, stroke damage increased with aging in female mice whereas male mice had decreased damage after MCAO. Blood Brain Barrier (BBB) permeability changes correlated with infarct size. However edema formation exhibited a striking age-dependent effect that was independent of sex and histological damage, with a significant decrease in edema formation in the aging brains of both sexes. Differences in the cellular response to stroke occur across the lifespan in both male and female mice. These differences need to be considered when developing relevant therapies for stroke patients, the vast majority of whom are elderly.

Introduction

Acute ischemic stroke (AIS) is a disease which primarily affects the elderly, and the strongest independent risk factor for stroke is age (Rojas et al. 2007). With increasing life expectancy, the world is facing a rapid expansion in its elderly population, and old subjects will soon constitute the majority of stroke victims (Rojas et al. 2007). The prevalence of ischemic stroke increases from 11% to 43% as patients age increases from 55 to 85 (Howard et al. 1998). Due to its high mortality and morbidity in elderly patients, stroke is a major healthcare problem worldwide, exerting an enormous financial toll and utilizing vast medical resources. Despite the clear evidence that stroke is a disease of the elderly, the vast majority of preclinical studies continue to utilize young male animals. As the most promising neuroprotective agents have failed when moved to clinical trial, stroke researchers need to address possible flaws in our experimental models if we hope to develop efficacious therapies. Clearly age is an important factor in how the brain responds to stroke. Despite high mortality in the elderly, death is often from co-morbid illnesses rather than from the stroke itself. Edema is more robust in the young brain, suggesting that the cellular response to ischemia may differ across the lifespan (Desbordes and Cohadon 1987; Wagner and Lutsep 2005).

Several previous studies have shown that histological stroke damage may be surprisingly lower in aging male rodents, despite higher mortality and poorer functional outcomes compared with young males (Shapira et al. 2002; Wang et al. 2003). Although the majority of stroke now occurs in elderly women (Rosamond et al. 2008), there is a paucity of information on the response to stroke in aging female rodents, and no previous studies have examined stroke outcomes in aging female mice. It is becoming increasingly recognized that stroke is a sexually dimorphic disease both experimentally and clinically, although these sex differences are still poorly understood. Most international databases consistently demonstrate that women enjoy a lower incidence of stroke relative to men until advanced age (Broderick et al. 1998), yet functional outcomes are worse in women than in men after stroke (Glader et al. 2003), with higher associated disability and mortality presumably due in part to the advanced age at which strokes occur in women. Sex differences in stroke epidemiology have been attributed to the vascular and neuroprotective effects of premenopausal estrogen (17β-estrodial or E₂) exposure. However recent clinical trials examining the effect of hormone replacement have shown a surprising increase in vascular risk in hormone treated women (Prentice et al. 2006). A major criticism of these trials was that E₂ treatment was begun years after the menopause. In an attempt to model and explain these findings researchers have moved back to the bench. Recent animal studies have demonstrated that a sustained period of hypoestrogenicity prior to the initiation of hormone replacement ameliorates the protective effects of E₂ in female mice, however, only young animals subjected to surgical menopause (via ovariectomy) were examined (5-6 months) (Suzuki et al. 2007). Whether E₂ supplementation can reduce infarction in aging female mice is not known.

The overall goal of the present study was to investigate how age and sex affect ischemic stroke outcome in male and female mice. As clinical data suggest that older women have worse outcomes after stroke and that loss of E₂ may contribute to these differences, we hypothesized that aging female mice would have larger strokes than younger female and that this increase in injury could still be ameliorated with peri-menopausal E₂ replacement. Effects on histological and functional outcomes, blood-brain barrier permeability, and edema formation were assessed in mice of both sexes.

Materials and methods

Animals

C57BL/6 mice were purchased from Charles River Laboratories. All experiments were performed according to National Institutes of Health guidelines for the care and use of animals in research and under protocols approved by the University of Connecticut Animal Care and Use Committee. Both young mice (9-12 weeks; 21-25gms) and aging mice (16 months, males 35-50gms, females 26-38gms) of both sexes were utilized.

Ischemic Model

Focal transient cerebral ischemia was induced by MCAO (0.21mm silicone coated suture) for 90 minutes followed by reperfusion as described previously (McCullough et al. 2005). In aging mice a larger 0.23mm silicone coated suture was utilized to achieve occlusion. Sham animals were subjected to sutures of the same size but the suture was not advanced into the middle cerebral artery. Cerebral blood flow (CBF) was measured by laser Doppler flowmetry (LDF, Moor Instruments Ltd, England) during the surgery as previously described (McCullough et al. 2005). Only the mice in which CBF in MCA area showed a sharp drop of over 85% of control immediately after MCA occlusion were included.

Neurological deficit was confirmed and scored as follows: 0, no deficit; 1, forelimb weakness and torso turning to the ipsilateral side when held by tail; 2, circling to affected side; 3, unable to bear weight on affected side; and 4, no spontaneous locomotor activity or barrel rolling. Monitoring of physiological variables was performed in companion cohorts for all groups prior to MCAO and 60 minutes after reperfusion as described previously (McCullough et al. 2005).

In ovariectomized (Ovx) females the ovaries were surgically removed 10 days prior to MCAO as described previously (McCullough et al. 2005). In E₂ treated mice 17β-estradiol was delivered by subcutaneous SILASTIC capsule (0.062 inch inner diameter; 0.125 inch outer diameter) filled with 0.035 ml of 17β-estradiol (180μg/ml; Sigma) (McCullough et al. 2005) in sesame oil implanted at the time of ovariectomy. Serum 17β-estradiol and levels of the inflammatory marker IL-6 was measured in each group by ELISA (E₂: IBL HAMBURG, Hamburg, Germany; IL-6: eBioscience, San Diego, CA). Uteruses of female mice were also weighed at sacrifice to confirm end-organ estrogen effects and ELISA values.

Terminal Histopathology

24 hours after stroke mice were euthanized and the brains were removed and cut into 5 2-mm slices. The slices were stained with 1.5% 2,3,5-Triphenyltetrazolium Chloride (TTC) solution at 37°C for 10 minutes. The stained slices were fixed with 4% formalin, images were digitalized, and the infarct volumes (% cotralateral hemisphere structure, corrected for edema and the different brain sizes in male and female mice) were analyzed using computer software (Sigmascan Pro5) as previously described (McCullough et al. 2005).

Evans Blue (EB) Extravasation

The integrity of the blood brain barrier (BBB) was investigated using EB extravasation, according to Bahcekapili et al (Bahcekapili et al. 2007). One hour before mice were sacrificed, 4ml/kg of 2% Evans Blue (Sigma, St. Louise, MO) in normal saline was injected intravenously. For quantitative measurements, brain hemispheres were homogenized in 1 ml of 50% Trichloroacetic Acid (Sigma, St. Louise, MO), and centrifuged. EB extravasation was quantified in the supernatants by spectrophotometry.

Measurement of Edema

24 hours after stroke, brain edema was measured by comparing wet to dry tissue weight ratios as described previously (Liu et al. 2008). Briefly, the brain was quickly removed after the animal was sacrificed. Then the brain was blotted to remove residual absorbent moisture, and dissected through the interhemispheric fissure into right and left hemispheres. Wet weight was determined with a resolution of 0.1 mg. Dry weight of whole ipsilateral and contralateral hemispheres was determined after heating the tissue for 3 days at 100°C in a drying oven. Tissue water content was then calculated as % H₂O = (1 – dry wt/wet wt) × 100%.

Western Blotting

After 24h of stroke, mouse brain samples were obtained by rapid removal of the brain from the skull, resection of the cerebellum, followed by immediate dissection into the right (R; ischemic) and left (L; non-ischemic) hemispheres. Brain samples were further dissected into core and penumbral regions as follows: The hemisphere was cut 1mm from the midline, and 2mm from either olfactory/frontal or cerebellum/occipital pole; then a wedge of tissue was removed from a coronal section as infarct core along the boundary between infarct and non-infarcted tissue, and the remaining section was assessed as penumbra (Figure 4A). Samples were flash-frozen and stored in -80°C until use. Samples were homogenized in 750 μl of ice-cold RIPA buffer containing protease inhibitor tablet (Roche Diagnostics) and 1mM PMSF using a dounce homogenizer on ice and briefly sonicated on ice. Extracts were immediately centrifuged at 14K, for 20min at 4°C. The resulting supernatant was removed and protein concentrations were determined using a BCA kit (BioRad). The homogenates were mixed with equal volume of 2× sample buffer and boiled for 5 min before being loaded onto a 4-15% gradient SDS-PAGE gel. Equal amounts of protein (40 μg) were loaded onto each lane and then transferred to polyvinylidene difluoride membranes (BioRad). Membranes were blocked and then incubated with Aquaporins (AQP1&4) (1:1000, Chemicon), Claudin-5 (1:1000, Chemicon), Connexin 43 (Cx43) (1:1000, Cell Signaling) or β-actin (1:5000, Sigma). Actin was used as a loading control. All blots were incubated overnight in primary antibody at 4°C in TBS buffer containing 5% bovine serum albumin and 0.1% Tween20. The secondary antibodies (1:2000) were either goat anti-rabbit or goat anti-mouse depending on the primary antibody (Santa Cruz Biotechnology). The signal was visualized with SuperSignalWest Pico Chemiluminescent Substrate (ThermoScientific). The densitometry of Western Blotting images was performed with computer software (Scion Image).

Figure 4.

Cx43 expression in infarct core (C), penumbra (P), and sham (S). (A) The area of infarct core and penumbra were determined as diagram shows; (B) Western Blot analysis was performed on lysates derived from infarct core and penumbra in ipsilateral hemispheres and sham brains. Cx43 protein was blotted as doublets (41 & 39 kDa); β-actin served as loading control. (C) The optical density of samples was expressed as the ratio of Cx43 bands to control band. n=6, P*<0.05 versus sham.

Statistics

All values are expressed as mean±SEM and analyzed with a t-test for two groups and one-way analysis of variance (ANOVA) with Turkey post-hoc correction, when appropriate, for multiple groups. All assessment was performed by a blinded investigator except behavioral scoring, which was blinded by sex. Due to the higher body weights and severe behavioral deficits in aging mice, experimenters were aware of the treatment group (aging vs. young). The criterion for statistical significance was P<0.05.

Results

Physiological Parameters Are Normal and Equivalent in All Study Groups

To assess whether young and aging mice have equivalent physiological parameters, mean arterial pressure (MAP), pH, pO₂, pCO₂, and blood glucose were measured. There were no differences in pre-ischemic or intra-ischemic arterial blood pressure or blood gas measurements between young and aging mice (Table 1). The intra-ischemic laser LDF signal showed that CBF in the area blood-supplied by MCA was reduced equivalently in all groups (Figure 1A). Three aging mice and two young mice were excluded from the experiment because their CBFs did not drop below 15% of baseline after MCAO.

Table 1. Physiological measurements in young and aging mice.

Group		pH	pO2	pCO 2	Glucose	MAP
Young female	pre-ischemia	7.39 ± 0.042	136 ± 12.4	39.6 ± 8.7	140 ± 11.1	76 ± 6
	ischemia	7.30 ± 0.085	128 ± 11.6	40.0 ± 3.6	146 ± 22.6	72 ± 4
Aging female	pre-ischemia	7.36 ± 0.082	136 ± 17.1	35.8 ± 6.3	133 ± 16.6	76 ± 5
	ischemia	7.25 ± 0.102	121 ± 18.9	43.8 ± 6.7	148 ± 19.5	71 ± 5
Ovx	pre-ischemia	7.36 ± 0.074	132 ± 15.0	37.8 ± 7.6	160 ± 3.9	80 ± 6
	ischemia	7.31 ± 0.060	121 ± 13.2	40.0 ± 5.9	161 ± 10.4	75 ± 4
Young male	pre-ischemia	7.35 ± 0.095	137 ± 16.1	38.0 ± 7.8	144 ± 19.1	75 ± 5
	ischemia	7.29 ± 0.079	131 ± 11.4	39.1 ± 7.4	144 ± 22.2	73 ± 5
Aging male	pre-ischemia	7.40 ± 0.052	140 ± 15.0	41.1 ± 7.2	151 ± 8.4	80 ± 6
	ischemia	7.25 ± 0.047	116 ± 16.0	39.8 ± 7.7	168 ± 4.2	74 ± 5

No differences were seen in physiologic variables between young and aging mice of either gender prior to MCAO and 60 min after reperfusion. Hg, Hemoglobin; MAP, mean arterial pressure. n=4/gp.

Figure 1.

CBF, E₂ concentration, and uterine weights measurement. (A) CBF was measured continuously after mice were anaesthetized by LDF; (B) Heart stick was performed and the blood was drawn from the right ventricle; E₂ concentration was detected in the serum; (C) Uteruses of both young and aging female mice were separated, cut, and weighed after brains were removed and all the uteruses were handled by the same operator. n=6/gp, P*<0.05 versus any other group. Y, young; A, aging; F, female; M, male.

E₂ Concentration and Uterine Weights Were Higher in Intact Young Female Mice

To evaluate the contribution of estrogen to stroke outcome across the lifespan, we measured the levels of E₂ and uterine weights in both young and aging mice. Young intact female mice had the significantly higher levels of E₂ (15.6±3.7 pg/ml, n=6) than any other group (P<0.05). There were no significant differences in the levels of E₂ between any other groups (ie., ovx vs. aging P>0.05) (Figure 1B), demonstrating that our aging females had equivalent estrogen levels to aging males and were functionally post-menopausal. In order to confirm that the loss of E₂ led to uterine atrophy, we also examined uterine weights in all female mice after stroke, as this parameter may more reliably reflect the physiological response to estrogen exposure. As expected, young intact female mice had significantly heavier uteri (82.8±12.7 mg, n=6) than either young Ovx (50.5±13.4 mg, n=6) or aging female mice (56.5±19.6 mg, n=6) (P<0.05) (Figure 1C).

Age and Sex Differences Were Found in Infarct Volumes Induced by MCAO

To examine histological outcomes after stroke, we used standard staining with 1.5% TTC (Figure 2A). Quantitative analysis (Figure 2B) demonstrated that aging female mice had significantly larger infarct volumes than young females (Total infarct: aging vs. young: 36.86 ± 5.49 vs. 20.49 ± 2.20%; P<0.05, n=8), whereas aging males had significantly smaller infarct volumes than young males (Total infarct: aging vs. young: 31.23 ± 3.26 vs. 49.11 ± 5.99%; P<0.05, n=7). Interestingly, aging had its most dramatic effect in aging females, as older female mice had significantly larger strokes than age matched males (P<0.05), a complete reversal of the pattern seen in younger animals, where females are protected compared to age-matched males.

Figure 2.

Stroke outcomes in young and aging mice after 24 hours of MCAO. (A) Coronal sections of brain stained by TTC. White color indicates infarction; red color indicates normal tissue. (B) Quantitative analysis of infarct volumes based on TTC staining was presented as % contralateral hemisphere structure. (C) Neurological deficits were assessed and scored at 3h & 24h of MCAO onset. P* < 0.05 (B: *versus aging female or young+Ovx or young male group; C: *versus 3h score). Y, young; A, aging; F, female; M, male; Ovx, ovariectomy.

Estrogen Was Neuroprotective in Aging Animals of Both Sexes

To determine whether the increase in infarct damage in aging females was due to the loss of estrogen, a separate cohort of young females was ovariectomized prior to MCAO (surgical menopause). As expected, young ovariectomized females had significantly larger strokes (Total: 47.11 ± 5.24%, n=6) than young gonadally intact females (P<0.05). We then evaluated the effect of exogenous estrogen in aging females to determine if E₂ maintains its neuroprotective effect in a model of “natural” age-related hormone loss. E₂ pellets were implanted into both aging male and female mice 10 days prior to stroke. Both male and female mice treated with physiological levels of E₂ had smaller infarcts (aging female: 22.63 ± 4.53%, n=6; aging male: 14.07 ± 6.72%, n=6) than their sex and age-matched cohorts without E₂ replacement (P<0.05), indicating that E₂ is neuroprotective in aging mice (Figure 2B).

Neurological Deficits Are Severe in Aging Mice

Behavioral deficits were scored at either 3h or 24h after stroke. Both young and aging mice showed clear neurological deficits at 3h of stroke, however young mice had a significant improvement in their deficits as rapidly as 24 hours post-stroke (female: 3h 2.1±0.2 vs. 24h 1.1±0.1, n=7, P<0.05; male: 3h 2.3±0.2 vs. 24h 1.1±0.1, n=6, P<0.05) whereas aging mice continued to show significant deficits (female: 3h 3.5±0.3 vs. 24h 3.3±0.3, n=8, P>0.05; male: 3h 3.4±0.2 vs. 24h 3.4±0.2, n=7, P>0.05) (Figure 2C). Given that aging males had smaller infarcts than young males yet both male and female aging mice had worse behavioral deficits than young mice, we can infer that the functional damage was exacerbated in aging mice independently of histological damage. Aging mice not only had significantly worse functional outcomes compared to young mice but also had significantly higher stroke-induced mortality (29%, n=6) than young mice (4%, n=1) in the 24 hours following stroke. Many of the aging mice that did not survive for 24h after stroke likely died of pneumonia or pulmonary complications as they exhibited evident clinical signs of heavy moist rales and dyspnea in the reperfusion period. There was no gross evidence of cerebral tumors or subarachnoid hemorrhage on brain removal.

EB Extravasation and Expression of Cx43 Levels Reflected the Degree of Infarction

As it has been demonstrated that reproductive hormones regulate the permeability of the BBB (Wilson et al. 2008), we hypothesized that the loss of estrogen with aging would lead to increased permeability of the BBB after stroke and EB extravasation through BBB was evaluated. The EB concentration in both the ipsilateral and contralateral hemispheres of mice after MCAO was measured to determine an EB index (EBI: ipsilateral/contralateral hemisphere EB concentrations). As expected, the EBI reflected the degree of ischemic damage, with large infarcts leading to a higher amount of dye extravasation. The EBI in both aging and Ovx female mice was significantly higher than that of young intact female mice (P<0.05, n=6), while aging males had decreased EBI compared to both young males and aging females (P<0.05, n=6) (Figure 3 A&B).

Figure 3.

Evans Blue extravasation in brains after 24 hours of stroke. (A) Blue colors indicate extravasated EB. (B) Evans Blue Index (EBI) was calculated by putting EB concentration in ipsilateral over that in contralateral hemisphere. n=6/gp, P*<0.05 versus aging female or Ovx or young male group.

To further assess the integrity of BBB after stroke, we also examined the expression of Cx43, a major gap junction protein that is hormonally regulated (Wilson et al. 2008), involved in the response to stroke. Cx43 expression was very low in the core of the infarct in all groups of mice suggesting complete loss of BBB integrity. However analysis of penumbral tissue showed striking differences between groups. Aging female mice had significantly decreased expression of Cx43 in the penumbra compared to sham, as did young male mice, whereas Cx43 levels were relatively preserved in both young females and aging males. This was consistent with both infarction volumes and EB leakage (Figure 4 B&C). As tight junction proteins may also play a role in the response to stroke, we also examined Claudin-5 levels and found no differences with aging (data not shown)

Aging Mice Had Less Edema Formation than Young Mice

Edematous reactions surrounding brain lesions are less extensive in old patients (Desbordes and Cohadon 1987). Young patients who experience extensive middle cerebral artery strokes are more likely to develop fatal brain edema than older patients (Wagner and Lutsep 2005). Edema interacts with stroke and is believed to be a contributor to propagation of the infarct. In order to determine if there were differences in edema formation induced by MCAO, we measured the water content in a separate cohort of animals after stroke. Young animals had dramatically more edema even upon gross visual inspection of the brain compared to sex-matched aging mice. Water content in both the ipsilateral and contralateral hemispheres was measured and an edema index (EI) was calculated (EI=water content in the ipsilateral over the contralateral hemisphere). EIs were significantly higher in young mice of both sexes compared to aging mice (P<0.05, n=6/gp), indicating an increase in edema formation (Figure 5A). We examined levels of AQP1 and 4, two important water channels in brain, to determine if levels differed in aging mice. We found no differences in brain AQP1 and 4 levels with aging as well as no differences between the sexes, consistent with recent work by others (Liu et al. 2008) (Figure 4B, data not shown for AQP1).

Figure 5.

Edema in brains and serum IL-6 levels in young and aging mice after stroke. (A) Edema index (EI) was calculated by putting water content in ipsilateral over that in contralateral hemisphere. (B) Blood was collected by heart stick at sacrifice and serum IL-6 level was analyzed by ELISA; control mice were non-operated. n=6, P*<0.05 (A: *versus young mice; B: *versus aging female stroke).

Serum Levels of the Pro-inflammatory Cytokine IL-6 Were Lower in Aging Males

Inflammation plays an important role in the response to ischemic stroke (Huang et al. 2006). In order to assess the inflammatory response induced by stroke, and to determine if levels were related to the striking reduction in edema volume in aging mice, we measured the serum levels of IL-6, a well-known pro-inflammatory cytokine, in mice of all groups 24h after stroke. As expected, serum concentrations of IL-6 increased in all stroke groups compared to sham (Figure 5B). Aging males had the smallest increase IL-6. Although IL-6 levels tended to be lower in young females compared to aging females, this differences was not significant. However there was a significant difference in IL-6 levels in aging females vs. aging males (Aging female vs. aging male: 292.6 ± 97.6 vs. 159.9 ± 41.2 pg/ml, P<0.05, n=6/gp), suggesting that aging females suffered a more robust inflammatory reaction after stroke compared to age-matched males. These differences were consistent with IL-6 levels in the brain, which were significantly higher in stroke mice than in sham. Similar to peripheral IL-6 levels, aging males showed the most modest increase after stroke (5.1 pg/ml in aging males vs. 7.8 pg/ml in aging females).

Discussion

The present study represents the initial steps in the investigation of sex and age differences in the response to focal stroke in mice, and demonstrates several important findings. Firstly, age related differences exist in behavioral and histological outcomes after MCAO in mice. Aging female mice had a significant exacerbation in ischemic damage compared to young female mice. This appears to be more related to estrogen loss rather than aging per se, as infarct volumes in young surgically menopausal females were equivalent to those of aging females, and E₂ replacement in aging females reduced infarct size back to levels seen in young gonadally-intact females. Secondly, aging males had decreased infarct volumes compared to young males. Although this initially was surprising due to the significant increase in functional deficits and mortality seen in aging males, this is consistent with findings from other laboratories (Shapira et al. 2002). Both female and male aging mice have higher mortality and more severe functional impairment at 24 hours of stroke than young mice. This reflects an effect of aging rather than infarct size, as high mortality rates (25%) continued to be seen in aging females supplemented with exogenous estrogen whose infarct volumes were reduced to levels seen in young intact females. Additionally, mortality remained low (5%) in Ovx females despite a dramatic increase in ischemic damage. Thirdly, EB extravasation and tight junction protein Cx43 expression in the penumbra suggests that aging female mice have higher BBB permeability than young females. In contrast, aging males have less EB extravasation and CX43 loss than young males after stroke, consistent with, and likely reflective of, the degree of ischemic damage. Finally, edema formation after stroke shows dramatic age-related differences. Aging mice of both sexes had significantly less stroke-induced edema compared to younger animals. Interestingly, edema formation was dissociated from the amount of EB extravasation, Cx43 expression, and infarct size, suggesting that aging animals have less intrinsic edema formation compared to young mice.

Estrogen has numerous effects on neurons, astrocytes, endothelial cells, and microglia, and has potent anti-inflammatory and anti-apoptotic actions in the injured brain (Brann et al. 2007). Stroke outcomes in female mice show a clear estrogen-dependent response in our studies, i.e. both aging and Ovx females had increased infarct volumes compared to intact females. Plasma estrogen levels in both aging or Ovx females were significantly lower and the uteruses were lighter than those of intact females. Our results are consistent with previous work that demonstrated that E₂ treatment at physiologically relevant concentrations reduced infarction after stroke in ovariectomized or aged reproductively senescent female animals, as well as in males, even when given after stroke (see review (Alonso de Lecinana and Egido 2006)). Despite the robust pre-clinical evidence for E₂, recent clinical trials have questioned the ability to translate this effect into human populations. The Woman's Health Initiative (WHI), the largest clinical trial of E₂ replacement for stroke prevention, showed a surprising increase in stroke incidence in E₂-treated women (Wassertheil-Smoller et al. 2003). The explanation for these findings has been debated extensively in the literature, but the timing of replacement may be a major factor. For example, women involved in the WHI trial were considerably older (average age of 63) and well past menopause (which occurs at an average age of 51) prior to randomization and initiation of therapy.

A recent animal study (Suzuki et al. 2007) designed in an attempt to explain these results, found that E₂ exerted profound neuroprotective effects when administered immediately upon ovariectomy, but not when administered after 10 weeks of hypoestrogenicity. Robust stroke-induced inflammatory changes were seen only in the animals that had sustained a long period of hypoestrongeneicity prior to E₂ replacement. A second recent study found that “mature” (7-8 month) females that had normal but lengthened estrous cycles retained the neuroprotective response to E₂ after ovariectomy and E₂ replacement whereas slightly older rats that were acyclic and “reproductively senescent” (9-11 months) had an exacerbation of injury after ovariectomy and E₂ replacement (Selvamani and Sohrabji 2008). These studies support the hypothesis that the neuroprotective efficacy of E₂ depends on the timing of initiation of therapy and that estrogen replacement can be deleterious in the aging brain.

However, our studies suggest that E₂ plays an important protective role in aging mice, and replicate earlier work that demonstrated E₂ retained its neuroprotective efficacy even in aging animals (Toung et al. 2004). There are several explanations for these findings. In the negative studies discussed above, much younger animals were used (i.e., 5-11 months) and all were subjected to surgical ovariectomy prior to injury. We examined older mice (15 months) that had progressed through “natural menopause” as confirmed by low serum E₂ levels and uterine atrophy. It is important to note that these animals are reproductively senescent and aging but not “elderly”. Rodents undergo a transition in midlife (at 12-14 months of age) referred to as the “estropause” which is characterized by irregular, usually prolonged, estrous cycles and eventual acyclicity (Chakraborty and Gore 2004). Therefore our studies reflect the effect of relatively early E₂ replacement initiated just beyond the middle age. This design more accurately reflects the ongoing clinical trial, the Kronos Early Estrogen Prevention Study (KEEPS; Available at: http://www.keepsstudy.org/) in which younger women are given estrogen replacement therapy within months after acyclicity, the results of which will be available in 2012. The effect of late E₂ replacement in aged (20 months) mice that have undergone a prolonged period of natural hypoestrogenicity is not yet known. In the age cohort used here, it is clear that estrogen maintains its neuroprotective efficacy in the aging brain, an effect not limited to females, as E₂ was also able to protect aging males. Whether this has any relevance to the clinic is unclear, as clinical studies have focused on utilizing chronic replacement therapy for the primary or secondary prevention of ischemic events (i.e. reductions in stroke incidence) rather than the acute protection that has been extensively studied experimentally. Our data suggest that estrogen is an efficacious neuroprotective agent when used in this setting, even in the aging brain.

The difference in infarct volumes between aging and young male animals is in agreement with previous work by others that demonstrated increased histological damage in young (3 months) compared to old (24-26 months) male rats (Shapira et al. 2002). Despite the smaller infarcts, our aging mice had a significantly higher mortality (29%) and more severe behavioral deficits than young mice, again consistent with several other studies (Badan et al. 2003; Popa-Wagner et al. 2007), which showed that aging is associated with rapid infarct development and a delayed, suboptimal functional recovery in the post-stroke period. The brains of aged rats demonstrate an increased astrocytic and microglial reaction to injury and an accelerated development of glial scarring leading to a possible subsequent stagnation of recovery in old animals (Badan et al. 2003). Frailty and weak resistance to infections may also partly account for the higher mortality and poorer functional recovery in aging mice.

The mechanism of the decreased histological damage in aging male mice remains unclear. A former study has demonstrated a widespread decline in NMDA-receptor binding in aged, as compared to young male rats (Wenk et al. 1991). There is also a significant age-dependent decrease in NMDA-receptor mediated responses, without changes in GABA-ergic and cholinergic activity, indicating a reduced excitotoxic potential in male aging rats (Gonzales et al. 1991). Whether a similar reduction in excitotoxic potential occurs in the aging female brain is unknown, and the effects may be dwarfed by the dramatic pro-inflammatory changes that occur with menopause.

It is now widely accepted that inflammation plays a critical role in stroke, and post-ischemic inflammatory responses strongly contribute to the extent of ischemic brain injury (Huang et al. 2006). In the elderly, immunodeficiency has been suggested due to low detected quantities of T cells which play critical roles in the development of ischemic injury by producing proinflammatory cytokines such as IL-6, IL-1β, and TNF-α, etc (Cakman et al. 1996). However, there is considerable controversy in the literature regarding the inflammatory responses in aging, possibly because many studies were done only in male animals or gender-mixed subjects (Cakman et al. 1996; Gabriel et al. 2002), or the sex of the animal used was not emphasized. Young female animals are known to have a decreased inflammatory response to injury compared to males and a considerable body of evidence strongly implicates E₂ as a major regulator of inflammatory pathways (Maggi et al. 2004). Both serum and brain levels of the pro-inflammatory marker IL-6 were significantly lower in aging males compared to aging females. Obviously, many cell adhesion molecules, cytokines, and chemokines contribute to the inflammatory response to stroke and IL-6 alone is simply one marker. Studies are ongoing examining the effect of aging on the inflammatory response to stroke in both sexes.

Ischemic stroke induces impairment of the BBB, and increased BBB permeability has been shown to play an important role in pathophysiology of post-ischemic reperfusion injury (Belayev et al. 1996). Ischemia leads to loss of microvascular integrity and enhanced vascular leakage. Our results derived from EB extravasation suggest that BBB disruption may be more reflective of the degree of injury or hormone loss rather than an enhanced intrinsic BBB dysfunction with aging. Several other studies have shown that E₂ loss enhanced brain injury and BBB permeability after ischemic insults (DiNapoli et al. 2008; Wilson et al. 2008). Surprisingly, there is a paucity of data in the literature examining BBB disruption in aging animals after focal stroke. Importantly, the most recent and complete study performed demonstrated an increase in BBB breakdown, stroke damage, and enhanced behavioral deficits after MCAO utilized only aging female rats (DiNapoli et al. 2008). No mention was made of hormonal status, and the enhanced damage is likely reflective of the loss of estrogen and is consistent with our findings.

The integrity of BBB is maintained by both tight junction proteins such as zona occludens 1 (ZO-1) and ZO-2, occludins and claudin-5, as well as gap junction proteins such as Cx43, Cx45, Cx32 and Cx26 between endothelial cells (Nagasawa et al. 2006). We examined Cx43 as it is a major astrocytic gap-junction protein that is upregulated after injury (Daleau et al. 2001),and is hormonally responsive (Wilson et al. 2008). Previous work by others has shown that Cx43 protein is decreased in the post-ischemic cerebral cortex suggesting an impairment of gap junction intercellular communication (Haupt et al. 2007). In the present study Cx43 protein expression in the penumbra was consistent with both our EB and infarction data. Aging female mice had the lowest expression of penumbral Cx43, suggesting the greatest disruption of the BBB after MCAO. Interestingly we didn't find any difference in Claudin-5 expression between young and aging mice of both sexes, which may suggest that not all junction proteins in BBB are affected by sex and age. In this study we only assessed some junction protein levels in brain homogenates by Western, therefore immunohistochemical studies that examine localization of BBB proteins are needed in future studies.

Extravasation of Evans blue is related to leakage of proteins after acute destruction of the endothelium. In contrast edema formation is more complex and differentiation of cytotoxic vs. vasogenic brain edema can be difficult (Xiao 2002). To our surprise there was no sex difference in edema formation in our study despite the striking sex differences in infarction and BBB disruption as measured by EB and Cx43. We examined the expression of AQP4, the most abundant water channel in brain. However, we saw no reduction in AQP4 (or AQP1, data not shown) expression in aging brain despite the significant reduction in edema that was evident in both male and female mice (Figure 5A). This is consistent with several other studies showing a lack of sex differences in cerebral edema formation and AQP 4 expression after brain injury (Carswell et al. 2000; Liu et al. 2008). Others have shown an effect of estrogen on edema formation after brain (O'Donnell et al. 2006). One explanation for these differences is that we specifically examined effects in the aging brain. Acute hormone loss with ovariectomy likely has very different effects than the gradual hormonal decline with aging, as is suggested by the high EI seen in our young ovariectomized females. This, and the similar paucity of edema in aging males, suggests that this is an intrinsic change that occurs with aging. Interestingly our data show that the EI was not correlated with the degree of injury, as aging females had much larger strokes than young females, yet younger mice continued to have significantly more cerebral edema. This suggests that BBB breakdown and infarct are dissociated from the mechanism that leads to cerebral edema, and the most important factor for edema development is age. This suggests that targeting younger patients for edema-reducing therapies may be of more benefit than targeting our elderly. As clinicians it is frequently noted that the amount of clinically relevant cerebral edema is lower in older patients compared to young patients, a finding often attributed to brain volume loss with aging. In a recent study it was found that among patients with massive cerebral infarcts, older patients (>70) had a surprisingly lower in hospital mortality (Chen et al. 2007) and died later of systemic illnesses. In contrast patients under 70 tended to die within the first 2 weeks, coinciding with the peak of cerebral edema. This concept of more robust edema formation in the young brain has been confirmed in post-mortem studies (Jaramillo et al. 2006) and is part of the rationale for proposing an upper age limit of 60 for hemicraniectomy.

The present study has limitations. We only chose the endpoint of 24h after stroke to assess outcomes though former studies (Dereski et al. 1993; Lu and Sun 2003) suggested that infarction become complete by 24 hour of transient MCAO in young animals, and the effect of E₂ or aging on the evolution of infarct needs to be further investigated. Of note, our former study (Liu et al. 2009) has shown that at 48 hours of stroke, both male and female aging mice exhibited similar histological changes of the brain as in the present study. We only measured absolute water content in the brain as an indicator of edema formation, which cannot differentiate cytotoxic vs. vasogenic edema that may predominate at different time phases after ischemic insult. Lastly, we campared E₂ implanted aging mice with intact ones rather than vehicle implanted mice to assess the effect of E₂ due to the limited source of aging animals. However, no studies have reported that the vehicle (sesame oil) has any effect on stroke.

In conclusion, dramatic sex and age related differences occur after experimental stroke that are reflected in the outcomes of injury, BBB disruption, and edema formation. Female mice go through a stroke resistant-followed by a stroke vulnerable pattern of response to MCAO across their lifespan, and this pattern appears to be estrogen dependent. Surprisingly males have the opposite pattern which may be related to a reduction in the inflammatory response to stroke with aging. E₂ retains its neuroprotective effect in aging animals of both sexes. EB extravasation and Cx43 expression in penumbra are consistent with the degree of ischemic damage in both sexes, while edema formation shows age-related differences. These results strongly suggest that therapeutic interventions for stroke need to be tailored to the sex and age of the patient.