Differential effects of aging and sex on stroke induced inflammation across the lifespan

Abstract

Aging and biological sex are critical determinants of stroke outcome. Post-ischemic inflammatory response strongly contributes to the extent of ischemic brain injury, but how this response changes with age and sex is unknown. We subjected young (5–6 months), middle aged (14–15 months) and aged (20–22 months), C57BL/6 male and female mice to transient middle cerebral artery occlusion (MCAO) and found that a significant age by sex interaction influenced histological stroke outcomes. Acute functional outcomes were worse with aging. Neutrophils, inflammatory macrophages, macrophages, dendritic cells (DCs) and microglia significantly increased in the brain post MCAO. Cycling females had higher Gr1⁻ non-inflammatory macrophages and lower T cells in the brain after stroke and these correlated with serum estradiol levels. Estrogen loss in acyclic aged female mice exacerbated stroke induced splenic contraction. Advanced age increased T cells, DCs and microglia at the site of injury, which may be responsible for the exacerbated behavioral deficits in the aged. We conclude that aging and sex have differential effects on the post stroke inflammatory milieu. Putative immunomodulatory therapies need to account for this heterogeneity.

Introduction

Stroke is the fourth leading cause of death and the leading cause of adult disability in the US. Sex differences in stroke risk have been well documented. These have been predominantly attributed to the protective effects of estrogen in pre-menopausal women (Lisabeth and Bushnell, 2012; Reeves et al., 2008). Yamori et al. replicated this sexual dimorphism in the laboratory using young stroke prone spontaneously hypertensive rats which showed a significant survival advantage in ovary-intact females (Yamori et al., 1976). Importantly, the Framing-ham Heart Study reported that although women have lower risk of stroke than men, this epidemiology reverses with age leading to a disproportionately higher risk in elderly women compared to age matched men (Petrea et al., 2009). Disparities not only exist in stroke risk with aging but are also seen in stroke outcomes. Elderly females have more severe strokes, poorer recovery and more disability after stroke (Appelros et al., 2009; Fukuda et al., 2009; Lai et al., 2005; Niewada et al., 2005), in part due to the older age at which women experience their first stroke. These sex disparities in aged populations have been confirmed in preclinical models of ischemic stroke (F. Liu et al., 2009). Aging is the most important non-modifiable risk factor for stroke, and the majority of stroke patients are over 65 years of age (Go et al., 2013; Rothwell et al., 2004). Despite this, most pre-clinical stroke studies exclusively examine the response to stroke in young male animals (Feigin et al., 2003). This is in part due to the difficulty of performing surgeries in aging animals, the high animal costs and poor survival rates in aged animals (Liu and McCullough, 2011). Most researchers also exclude female animals due to the variability that can occur with cyclical hormone changes with estrus, which is well known to affect histological damage after injury (Carswell et al., 2000; Zucker and Beery, 2010). Overall, there has been a paucity of stroke studies in females and aging animals, leading to limitations in the translational relevance of potential therapeutics identified in experimental models.

Cerebral ischemia has profound effects on both the central and peripheral immune response. Stroke activates resident microglia and also leads to the recruitment of blood derived leukocytes into the brain (Huang et al., 2006; Iadecola and Anrather, 2012). This recruitment process is modulated by several cell adhesion molecules and cytokines which when induced, act upon the vascular endothelium to increase the expression of ICAM-1, P-selectin, and E-selectin, leading to further local accumulation and adhesion of leukocytes (Danton and Dietrich, 2003; Elkind, 2010; Han and Yenari, 2003; Ishikawa et al., 2004). After gaining entry into the CNS (through the damaged blood brain barrier), infiltrating leukocytes release cytokines and chemokines amplifying the innate (microglial) inflammatory response in the brain and causing further injury (Huang et al., 2006; Jin et al., 2010; Wang et al., 2007). This initial inflammatory cell influx in the brain is followed by a systemic immunodepression (Dirnagl et al., 2007). Secondary lymphatic organs including the spleen and thymus contract after focal cerebral ischemia, potentially increasing the risk of post-stroke infections (Dirnagl et al., 2007; Elkind, 2010; Emsley and Hopkins, 2008; Offner et al., 2006a), which are linked to increased morbidity and mortality. The temporal dynamics of the specific leukocyte populations infiltrating the brain parenchyma after an ischemic event has been studied (Gelderblom et al., 2009; Stevens et al., 2002) but these studies were all performed in young male animals. Aging leads to a chronic low grade inflammatory state and an altered innate immune environment in the brain, which contributes to the development of age-related disorders including stroke (Dorshkind et al., 2009; Gruver et al., 2007; Krabbe et al., 2004; Lucin and Wyss-Coray, 2009). Enhanced astroglial response and proinflammatory cytokine production has been seen in aged animals after stroke (Dinapoli et al., 2010; Popa-Wagner et al., 2007). The interaction between immunosenescence and loss of gonadal hormone secretion with aging is complex and has not been well studied. The primary gonadal female sex hormone, estrogen, appears to be anti-inflammatory in young and pro-inflammatory in the aged brain (Johnson and Sohrabji, 2005; Liu et al., 2012b; Nordell et al., 2003; Wise et al., 2001). Thus, we hypothesized that both sex and aging would shape the post-stroke inflammatory response, leading to differences in histological and behavioral outcomes between the sexes at different time points in the lifespan. We used 5–6 month old female mice as young cycling females, 14–15 month old females as a translational model of females with irregular cycles transitioning to anovulation (perimenopausal in humans), and 20–22 month old females as a model of menopause as mice are anovulatory and persistently acyclic at this age (Nelson et al., 1982; Parkening et al., 1980). The primary objective of this study was to characterize the differential effects of sex and aging on the acute peripheral and central inflammatory response to ischemic stroke.

Materials and methods

Experimental animals

Young (5–6 months), middle aged (14–15 months) and aged (20–22 months) C57BL/6 male and female mice were kept separately in cages of two animals per cage on sawdust bedding (light cycle 12/12 h light/dark). The average weight of males was 37.7 ± 1.1 g (young males), 44.4 ± 1.56 g (middle age males) and 43.6 ± 1.89 g (aged males). Female mice weighed 32.4 ± 2.72 g (young females), 34.4 ± 0.86 g (middle age females) and 29.4 ± 1.58 g (aged females). All mice had access to chow and water ad libitum.

This study was conducted in accordance with the National Institute of Health guidelines for the care and use of animals in research and under protocols approved by the Center for Lab Animal Care at the University of Connecticut Health Center.

Ischemia model

Animals were randomly and blindly assigned to MCAO or control groups using simple random number randomization. Only one surgeon performed the surgeries and was blinded to the age of the animals. Focal transient cerebral ischemia was induced by 60 min of middle cerebral artery occlusion (MCAO) under isoflurane anesthesia, followed by re-perfusion, as described previously (McCullough et al., 2005). In mice weighing greater than 35 g, a larger diameter (0.23 mm) silicon coated suture was utilized to achieve occlusion (F. Liu et al., 2009). Rectal temperature was maintained between 36.5 and 37.5 °C during surgery through an automated temperature control feedback system. Cerebral blood flow (CBF) was measured by Laser Doppler flowmetry (LDF, Moor Instruments Ltd, England) in all animals as previously described (McCullough et al., 2005).

Neurological scores

Neurological scores were recorded at the time of reperfusion (60 min) and 24 h after MCAO surgery. The scoring used was 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 as described previously (McCullough et al., 2003).

Terminal histopathology for cresyl violet staining and immunohistochemistry

Two cohorts of control and MCAO animals were sacrificed at 24 hour endpoints with pentobarbital overdose (i.p.). There was no mortality in any of the cohorts 24 h after surgery. Mice with subarachnoid hemorrhages and/or tumors in the brain were excluded from the study (n = 2 (1 in MA and 1 in aged) and 2 (both in aged) respectively). One cohort was used for cresyl violet staining and immunofluorescence studies and the second for flow cytometry studies. In the cresyl violet cohort, n = 6–8/group, blood was collected from the right ventricle of the heart using heparinized syringes at sacrifice. Transcardial perfusion was performed with cold PBS followed by 4% paraformaldehyde. Brain, spleen and uteri were collected. Splenic and uterine weights were recorded and quantified as mg/g of body weight (pre-operative weight for mice subjected to MCAO). All the weights were taken by a single observer blinded to age and MCAO. The brains were fixed for 24 h and placed in cyroprotectant (30% sucrose), frozen and then sliced into 30-μm free-floating sections on a freezing microtome; every eighth slice was stained by cresyl violet stain to evaluate infarct volumes. The images were digitalized and infarct volumes were measured using computer software (Sigma scan Pro5) as previously described (F. Liu et al., 2009). The infarct volumes were calculated as percentage of contralateral hemispheric structure and corrected for edema using Swanson’s method (Swanson et al., 1990).

Immunohistochemistry (IHC)

Immunohistochemical studies were performed on 30-μm brain sections as described previously (Liu et al., 2010; Manwani and McCullough, 2011; Manwani et al., 2011). Briefly, brain slices were mounted onto gelatin-coated slides, allowed to air dry and then blocked in 0.1 M phosphate buffer (PB) with 0.3% TritonX-100 (sigma) and 10% goat serum (PBTGS) for an hour. Primary antibody (Iba1, Wako, USA) was added overnight, washed with PBTGS followed by incubation with secondary antibody (1:1000) and 4′,6-diamidino-2-phenylindole, dihydrochloride (DAPI, 1:1000, Invitrogen, Carlsbad, CA). Secondary antibody (1:1000, goat anti-rabbit conjugated to Alexa-488) was removed with three consecutive washes in PBTGS, 0.1 M PB, and 0.05 M PB respectively. Images were acquired with immunofluorescence confocal microscopy using Zeiss image acquisition software (Zeiss Axiovert 200 M). Brain slices were taken at the same distance from bregma (0.5 mm anterior to bregma) and three 20× fields/animal (n = 3–4 animals/group) were analyzed in the penumbral area of the infarct. Iba1 positive cells were counted using MacBiophotonics ImageJ software with DAPI (nuclear stain). The average of the total number of cells/field of view was used for statistical analysis as described previously (Liu et al., 2010; Manwani et al., 2011).

Estradiol ELISA

Blood samples collected at the time of sacrifice were centrifuged at 6000 rpm for 10 min at 4 °C to yield serum for hormone detection. Serum was stored at −80 °C until use. Enzyme-linked Immunoassay (ELISA) for 17β-estradiol (BQ, San Diego, CA) was utilized following the manufacturer’s protocol.

Brain leukocyte preparation

The second cohort of control (n = 3–5/group) and MCAO (n = 6–8/group) animals was sacrificed using i.p. pentobarbital overdose for flow cytometry studies. The animals were perfused with cold 1× PBS until the liver turned pale. Brains and spleens were harvested from these animals. Brain leukocyte preparation was done as previously described (Wilson et al., 2005). The brains were dissected to remove olfactory bulbs and cerebellum and only the ipsilateral brain hemisphere (right hemisphere) was used. The dissected brains and spleens were placed in RPMI 1640 complete medium (10% fetal calf serum, 1% sodium pyruvate, 1% non-essential amino acid, 0.1% β Mercaptoethanol, 100 U/ml of penicillin and 100 μg/ml of streptomycin) in separate tubes on ice. The brains were mechanically dissociated and incubated with 100 μl of collagenase/dispase (1 mg/ml, Roche Diagnostics) and 300 μl DNAse I (10 mg/ml, Roche) for 45 min at 37 °C. After incubation, the brain homogenate was passed through 1 ml pipette tip several times and harvested in 20 ml complete RPMI. The cells were pelleted at 1200 G, 4 °C for 10 min, resuspended in 40 ml complete RPMI, passed over a 70 μm cell strainer and pelleted again. The filtered cells were resuspended in a 60%/30% Percoll gradient (Sigma) and spun at 2000 rpm for 25 min at room temperature with no brake. Myelin was removed from the top and cells collected from the interface into 40 ml complete RPMI. The cells were washed and re-suspended in 1 ml RPMI for staining and counting.

Preparation of splenocytes

Spleens collected for flow cytometry were pulverized through 70 um cell strainers into fresh RPMI complete medium. The cells were pelleted at 1600 G for 5 min and then red blood cells lysed in 0.85% ammonium chloride. The cells were then washed once with RPMI complete medium and re-suspended for staining and counting.

FACS analysis

Cell staining and analysis were achieved as described previously (Sansing et al., 2011; Wilson et al., 2005). In short, leukocytes from brain and spleen were blocked with Fc block [895 μl FACS buffer, 100 μl anti-CD16/CD32 (10 μg/ml, BD Biosciences), 5 μl normal rat IgG (Invitrogen)] for 15 min. The cell suspension was then stained with CD45.2-FITC (BD Pharminogen), CD3-PE (ebioscience), APC efluor 780-CD11b (ebioscience), PECy7-CD11c (ebioscience), PerCp-Cy5-Gr1 (ebioscience) for 20 min in the dark. The cells were counted with a LSR II cytometer (BD Biosciences) and data analyzed using FlowJo software (Tree Star, Ashland, OR).

Statistical analysis

All normally distributed data are presented as Mean ± SEM and the neurological scores (non-parametric data) as Median (IQR). Cerebral blood flow (flux) is also expressed as % of baseline (preischemic). Leukocyte populations of interest were quantified as percentage of total leukocytes for brain and spleen samples. Pearson correlation coefficient was used to compare leukocyte subset numbers with estradiol levels. Three way ANOVA with factors of stroke, sex and age was used for comparing spleen weights and microglia counts on IHC. Two way ANOVA with factors of sex and age was used for analyzing the infarct volumes and flow cytometry data from brain and spleen. A post-hoc Bonferroni correction was used, when appropriate, to compare differences in groups. Independent t test with unequal variance was used for the comparison of the means of absolute cell numbers in the brain. The ordinal data (neurological score) was analyzed using Kruskal Wallis analysis of variance with post hoc for factors of aging and sex respectively. p < 0.05 was considered statistically significant. IBM SPSS v.21 was used for statistical analysis.

Results

Stroke injury is dependent on age and sex

Female mice become sexually mature at 6 weeks and cycle until approximately 14 months of age (Nelson et al., 1982; Parkening et al., 1980). Consistent with this, we found that serum 17β estradiol levels and uterine weights in the young and middle-aged female mice were significantly higher than in the aged females (anovulatory/non-cycling) (Figs. 1A and B). 60 min of MCAO was induced in young, middle aged and aged male and female mice. Cerebral blood flow was measured using LDF (Supplementary Fig. 1). There was a cerebral blood flow reduction of 87.9 ± 1% in young males; 87.4 ± 0.9% in young females, 88.4 ± 1.1% in middle aged males, 88.3 ± 0.9% in middle aged males; 86.8 ± 0.8% in aged males and 87 ± 0.9% in aged females. There was no significant difference in the reduction of CBF (% baseline) in any of the cohorts, p > 0.05. In order to assess the influence of age and sex on acute stroke outcomes, we measured infarct volumes at 24 h in these groups (n = 6–8/group) (Figs. 1C and D). The mean infarct volume in young males was significantly higher than that in young females (young male 42.6 ± 5.7% vs. young female 26.5 ± 3.9%, p < 0.05). This phenotype reversed in middle aged mice. Middle aged females had significantly higher infarct volumes as compared to middle aged males (middle-aged females 43.9 ± 4.5% vs. middle aged males 23.05 ± 5.7%, p < 0.05). There was no significant difference in the infarct volumes of aged males (31 ± 7.1%) versus aged females (25.32 ± 3.2%). A significant age × sex interaction was seen [F (2, 36) = 4.22; p < 0.023] suggesting that a complex interaction of aging and sex influences the extent of histological damage post stroke.

Fig. 1.

A and B. Uterine weight (mg/g of body weight) and 17β estradiol levels (pg/ml) in females. Young and middle aged females had significantly higher uterine weights corresponding to higher estradiol levels versus aged females, p < 0.05. C. Representative cresyl violet stained sections of the brains (60 min MCAO, 24 h) in the young, middle aged (aging) and aged males and females. D. Total hemispheric infarct percentage at 24 hour endpoint after 60 min of MCAO. A significant age × sex interaction was seen, p < 0.02; young males had significantly higher infarct volumes vs. young females, p < 0.05; while middle aged females had significantly higher infarct volumes than middle aged males, p < 0.05. Aged males and females had no significant difference in their infarct volumes (n = 6–8/group).

Neurological scores were recorded at 24 h after MCAO to assess acute functional outcome (Table 1). Young males had significantly worse neurological scores as compared with females, p < 0.03. Overall, increasing age was associated with worse neurological deficit scores, χ²(2) = 12.171, p < 0.002.

Table 1.

Neurological scores at 24 hour endpoint after 60 min of MCAO.

Neurological Scores
	YOUNG	MIDDLE AGED	AGED
Male	2.5 (1.75)	3.5 (1)	3.5(1)#
Female	1 (0)*	3(1)	4(1)#

^#Age had a significant main effect on neurological scores (functional outcome), p < 0.05. Middle aged and aged mice had worse neurological scores as compared to the young mice.

^*Young female mice had significantly better neurological scores as compared to young males, p < 0.05.

Since age and sex have a robust effect on stroke outcome, we investigated the differential effects of sex and aging on the central and peripheral immune response to ischemic stroke.

Splenic contraction after stroke is higher in males and acyclic females

Ischemic stroke is a systemic disease. Peripheral lymphoid organs including the spleen respond to brain injury by releasing inflammatory cells into the circulation, reflected by a decrease in spleen size (Offner et al., 2006b). Supplemental Fig. 2 shows spleen weights normalized to body weights of all cohorts (n = 3–4/group in control; n = 4–6/group for stroke). There was a significant main effect of stroke [F (1, 37) = 34.06; p < 0.001], age [F (2, 37) = 3.7; p < 0.03] and sex [F (1, 37) = 109.3; p < 0.001] on splenic weight. Post hoc comparisons revealed a statistically significant decrease in spleen weight after MCAO in male mice of all age groups and aged females. Young and middle aged females (cycling females) did not show a significant reduction in spleen weight after stroke. Interestingly, aged females (with low estrogen levels) had a significant reduction in splenic weights after stroke as compared to control females. In control mice, a significant sex effect [F (1, 18) = 25.4; p < 0.001] was seen. Female mice had significantly heavier spleens as compared to males. Significant age associated splenic volume loss was seen only in males (a 34% decrease in young male 2.43 ± 0.1 mg/g vs. aged male 1.5 ± 0.1 mg/g, p < 0.001).

Together, this shows that stroke leads to splenic contraction in males. Females are protected from the stroke induced splenic volume loss until they are post-menopausal/anovulatory. Moreover, at baseline, females have a higher absolute splenic weight than males and unlike males; this does not decrease with aging (data not shown). Interestingly, the pattern seen in splenic contraction did not mirror the changes seen in infarct volumes, indicating that the splenic response is not the only factor underlying the influences of age and sex on stroke injury.

Leukocyte recruitment to the brain after stroke

Brain leukocytes from control and stroke mice of all the age groups (young, middle aged and aged) and both sexes (males and females) were quantified using flow cytometry. The strategy for identification of leukocyte populations has been detailed in Fig. 2 and Table 2. Blood-derived leukocytes were distinguished from microglia based on levels of CD45 expression with peripheral leukocytes identified as “high” CD45 expression and microglia as CD45 intermediate expression and CD11b⁺ (Sedgwick et al., 1991).

Fig. 2.

Representative plots showing gating strategy in the ipsilateral (lesion side- right) hemisphere of brain. A. Plot showing CD45 expression. The plot is gated on CD45 high (right) and CD45 intermediate (left) cell populations. B. CD3 gate is drawn on CD45 high population, T cells are the CD3⁺ cells (right). The gate on the left is drawn on CD45high/CD3⁻ cell populations. C. Plot on CD45high/CD3⁻/CD11b⁺ cell population gated on CD11c. DCs are the CD11c⁺ cells on the right. D. Plot on CD45high/CD3⁻/CD11b⁺/CD11c⁻ cells, gated on Gr1⁺ cells (red arrow, right) which are neutrophils and inflammatory monocytes/macrophages. Plot on CD45 high/CD3⁻/CD11b⁺/CD11c⁻ cells, gated on Gr1⁻ cells which are monocytes/macrophages. E. Plot on CD45 intermediate cells (green arrow down from 1a) gating on CD11b⁺ cells which are the microglia.

Table 2.

Absolute leukocyte counts in the ipsilateral brain hemisphere 24 h after MCAO (60 min) using flow cytometry. There was a significant increase in the number of CD45high cells after MCAO. DCs, neutrophils, monocytes and inflammatory monocytes also significantly increased in the brain post MCAO.

Gating (cell subsets)	Controls	Stroke	p value
CD45 high (leukocytes)	1998.3 ± 493.1	6253.6 ± 1555.1*	<0.01
CD45hi/CD3+ (T cells)	844.4 ± 221.2	709.2 ± 128.9	<0.60
CD45hi/CD3−/CDllb+/CDllc+ (dendritic cells)	253.8 ± 33.3	1040.7 ± 265.9*	<0.005
CD45hi/CD3−/CDllb+/CDllc/Grl+ (neutrophils and inflammatory macrophages)	5.5 ± 0.87	1421.95 ± 703.1*	<0.05
CD45hi/CD3−/CDllb+/CDllc−/Grl− (macrophages)	156.5 ± 20.6	1448.7 ± 528.5*	<0.02
CD45intermediate/CDllb+ (microglia)	17837.3 ± 1471.4	31034.7 ± 2584.27*	<0.001

^*p < 0.05

A significant influx of leukocytes was seen in the ipsilateral brain hemisphere after stroke in the pooled data from all age groups and both sexes. There was a tripling of the number of leukocytes in the ipsilateral hemisphere by 24 h after stroke (Table 2). The numbers of DCs, neutrophils and inflammatory macrophages, macrophages and microglia were all significantly increased in the brains of mice subjected to stroke compared to control mice (Table 2). T cell counts did not significantly increase in the brain at this time point. Together, these results demonstrate that stroke induces a local immune response, with near doubling in numbers of microglia, in addition to a profound systemic immune response leading to rapid infiltration of blood-derived leukocytes (macrophages, neutrophils and DCs) into the ischemic brain. We next analyzed the effect of age on sex on the numbers of these cells in both the brain and spleen.

T cell numbers higher in the brains of males and acyclic females post stroke

A comparison of the effects of sex and age on CD45high/CD3⁺ T cells in the brain of control mice showed no significant differences (Supplementary Fig. 3). This suggests that neither aging nor sex alone significantly changes the number of T cells that reside in the brain under resting conditions. T cells in the brain did not increase significantly in control vs. stroke mice (in pooled data across ages and sex) at 24 hour endpoint (Table 2). However, even at this early 24 hour endpoint, the dynamics of the changes in T cell percentages with age and sex were easily discernable. After stroke, a significant main effect of age [F (2, 33) = 10.4; p < 0.001], sex [F (1, 33) = 4.04; p < 0.05], and a significant age × sex interaction [F (2, 33) = 3.25; p < 0.05] was seen in the T cell percentages. In females, there was a negative correlation between serum estradiol levels and T cell counts (r = −0.54, p < 0.014). Females had a lower percentage of T cells in the brain after stroke compared to males until they were aged or anovulatory, when T cell infiltration increased and became equivalent to that seen in males. Notably, aged animals had significantly higher numbers of T cells than young mice after stroke.

The splenic T cell response to injury as a function of age and sex was then examined. In control mice (Fig. 3B), there was a significant main effect of age [F (2, 18) = 6.15; p < 0.001] on the percentage of T cells. The number of splenic T cells decreased by 14.6% in aged males compared to young males and by 15.3% in aged females vs. young females. This suggests that aging itself leads to a reduction in T cells in peripheral lymphoid organs such as the spleen. After stroke, a significant effect of sex [F (1, 27) = 16.59; p < 0.001] and a significant age × sex interaction [F (2, 27) = 9.24; p < 0.001] was seen (Fig. 3C). Young females had higher splenic T cell percentages than young males. The differences in the spleen and brain suggest that there is selective recruitment of T cells in a sex and age-dependent pattern to the brain after stroke.

Fig. 3.

A. T cells in the brains of stroke mice. A significant main effect of age, p < 0.001; sex, p < 0.05 and a significant age × sex interaction, p < 0.001 was seen. A significant negative correlation between estrogen in females and T cells, r = −0.54, higher T cell numbers in aged acyclic females than young and middle aged females. B. T cells in the spleen of control mice. A significant effect of age (p < 0.001) on the frequency of T cells in the spleen in control mice. Young vs. aged, p < 0.001*; T cells decrease significantly with age. C. T cells in the spleen of stroke mice. A significant effect of sex (p < 0.001) on the frequency of T cells in the spleen in MCAO mice. T cells are significantly lower in young males vs. females, p < 0.001*.

Stroke induces robust DC response in females and aged males

DCs are the professional antigen-presenting cells and activate T cells after injury and infection (Huang et al., 2012; Itano et al., 2003). DCs increased dramatically in the brain 24 h after stroke (Table 2). After stroke, females had a clear enhancement in the recruitment of DCs into the brain compared to young and middle-aged males [F (1, 33) = 7.79; p < 0.001] (Fig. 4A). In control mice, DCs did not differ in the brains or spleens by sex or age (Supplementary Figs. 3 and 4). In contrast, a significant effect of both age [F (2, 27) = 6.98; p < 0.001] and sex [F (1, 27) = 5.09; p < 0.03] was seen in the splenic DC response after stroke, with middle-aged females having the highest proportion of splenic DCs (Supplementary Fig. 4). There was a positive correlation of DCs with serum estradiol levels, which was not significant, r = 0.404, p < 0.08. These results suggest that DC recruitment to the brain after stroke is most robust in females.

Fig. 4.

A. DCs in the brains of stroke mice. A significant effect of sex (p < 0.001) on the % of DCs in the brain in MCAO mice. DCs are significantly lower in young males vs. female* and middle aged males vs. females**, respectively. B. Neutrophils and Gr1⁺ macrophages in the brains of stroke mice. No significant effect of age or sex on the percentage of neutrophils in the brain, p > 0.05. C. Neutrophils and Gr1⁺ macrophages in the spleen of stroke mice. A significant effect of age*, p < 0.04 on the frequency of neutrophils and Gr1⁺ macrophages in the spleen. D. Gr1⁻ macrophages in the brain of stroke mice. A significant effect of sex (p < 0.001)*, and age x sex interaction (p < 0.04) was seen. Aged females had significantly lower Gr1⁻ macrophages as compared to cycling females** (positive correlation with estrogen, r = 0.58, p < 0.009).

No age or sex effect on stroke induced neutrophil and inflammatory macrophage recruitment in the brain

Control brains had <1% Gr1⁺ cells (data not shown), consistent with the exclusion of these inflammatory cells from the brain under resting conditions. Gr1⁺ cells increased significantly in the brain after stroke (Table 2). However, there was no significant effect of age or sex on neutrophil and inflammatory macrophage populations in either control or stroke brains (Fig. 4B), p > 0.05, suggesting that the recruitment of these cell populations is age and sex-independent.

Similarly, in the spleen, the percentage of neutrophil and inflammatory macrophages was equivalent in all control cohorts (<1% Gr1⁺ cells, data not shown). The percentage of these cells increased in the spleens after stroke, p < 0.03. This was primarily driven by a significant increase of these cells in middle aged and aged mice after stroke (main effect of age [F (2, 33) = 3.3; p < 0.04]) (Fig. 4C).

Cycling females selectively recruit more Gr1⁻ macrophages to the brain after stroke

No differences were seen in the Gr1⁻ macrophage population in the brains of control mice (Supplemental Fig. 7) among the different ages or sexes. Macrophages in the brains of mice subjected to stroke (Fig. 4D) did show a significant effect of sex [F (1, 33) = 4.02; p < 0.001] and age × sex interaction [F (2, 33) = 3.38; p < 0.04]. The Gr1⁻ macrophage percentages were higher in young and middle aged females and decreased in reproductively senescent females. Serum estradiol levels positively correlated with the percent of brain macrophages, r = 0.58, p < 0.009. In the older cohorts there was no difference in macrophages between males and females. In the periphery (spleen), no significant effect of age or sex was seen on the macrophage percentage in either control (data not shown) or stroke mice (Supplemental Fig. 8). Similar to the findings in splenic contraction, the data shows that the Gr1⁻ macrophage population is recruited selectively to the brain after stroke in the cycling females and is closely associated with serum estrogen levels.

Microglia numbers increase rapidly after stroke

Microglia, the resident immune cells of the CNS, have long been known to respond to ischemic insult (Jin et al., 2010). However, there have been limited investigations into the age and sex-dependent response of microglia after stroke. No significant difference was seen in microglia numbers with age or sex in control mice using flow cytometry (Fig. 5A). Consistent with previous studies, we found a robust increase in microglia after stroke by flow cytometry (Table 1, Fig. 5B). However, the use of flow cytometry to quantify microglia, which examines the entire hemisphere after stroke is problematic, as large numbers of microglia may be dead in the core of the infarction. Therefore, microglia (Iba1 positive cells) were also counted in the ischemic penumbra in stroke brains and the corresponding area in the control brains using immuno-histochemistry (Fig. 5D). We found a significant effect of age in control [F (2, 7) = 6.9; p < 0.02] and stroke brains [F (2, 10) = 3.9; p < 0.05]. The increased numbers of microglia with aging was limited to males, as females did not show such significant increase with age (Figs. 5E, F).

Fig. 5.

A. Microglia in the brain of control mice. No significant effect of age or sex was seen. B. Microglia in the brain of stroke mice. There is no significant effect of age or sex. C. Iba1+ (microglia) cells in stroke and control mice quantified as average number of cells per field of view through IHC. An effect of stroke and age was seen, p < 0.05. Significantly higher number of microglia were seen with increasing age (*) and also with stroke (#). D. Cresyl violet stained coronal section of the brain showing the penumbral region where Iba1+ cells were counted using IHC. E. Iba1+ microglia in representative sections from young and aged male and female control mice showing increased number of Iba1+ cells in the aged (20×). F. Iba1+ microglia in representative sections from young and aged male and female MCAO mice showing increased number of Iba1+ cells in the aged (20×). Iba1 – green, DAPI – blue. Images were processed with a 1.6 gamma adjustment. Scale bars indicate 50 μm.

Discussion

This is the first study to assess acute stroke outcomes and the peripheral and central inflammatory response to ischemic stroke across the lifespan in both sexes. We used young (equivalent to mature adult human ~25–30 years old), middle aged (equivalent to middle aged human ~47–52 years old) and aged (elderly human ~62–65 years old) male and female mice in this study (Flurkey et al., 2007). Such lifespan studies in stroke are particularly useful in females where senescence is associated with a dramatic decline in estrogen levels, a hormone that has been implicated to be neuroprotective in stroke. Consistent with other studies (Felicio et al., 1984), young and middle aged female mice had significantly higher circulating estrogen levels and larger uteri than aged females. This study design enabled us to have a broader understanding of the effects of senescence and the complex interplay of sex, hormones, inflammation and aging.

We first assessed histological and functional stroke outcomes in these animals. We found a complex age and sex interaction that influenced the degree of injury after ischemia. Young adult females have smaller infarct volumes and better neurological function compared to young males (Alkayed et al., 1998; Zhang et al., 1998). This is consistent with the known neuroprotective effects of estrogen in young females (Alkayed et al., 1998). Infarct volumes increased in middle aged perimenopausal females, but this phenotype reversed in aged acyclic female mice. The higher infarct volumes seen in middle aged female mice have been reported previously (F. Liu et al., 2009) and may be an effect of the fluctuations in estrogen levels and irregular cycles at this age, although the average total circulating estradiol remains similar to young female mice. Aged mice were subjected to equivalently severe ischemic insults (as also measured by qualitative cerebral blood flow analysis) and ongoing work in our lab has shown no difference in collateral formation in aged animals (Manwani et al., unpublished data). We have previously found that aging leads to a reduction in infarct volumes in males (F. Liu et al., 2009; Manwani et al., 2011) and similar mechanisms might be responsible for this effect in aged females. Paradoxically, despite these smaller infarcts, middle aged and aged mice have worse functional outcomes, regardless of the sex, as has been seen in other studies (DiNapoli et al., 2008; Manwani et al., 2011; Moore et al., 2011; Popa-Wagner et al., 2007).

One limitation of our study is that we focused exclusively at the 24 h endpoint. The emphasis of this initial study is on the early inflammatory response to stroke between the sexes over the life span. Previous work in our lab has shown that the peak of inflammatory infiltrate is at 24 h in young animals, and this is the point where therapeutic intervention would be considered. We also picked this endpoint as aged animals are extremely frail and have a high mortality after an induced ischemic stroke (Liu and McCullough, 2011). Follow up at a later time point may have introduced a survival bias in the results. Previous work in our laboratory has also shown that stroke size is relatively stable at this point in both young and aged males (F. Liu et al., 2009; Manwani et al., 2011), but less is known regarding chronic infarct changes in females, or the later inflammatory responses in the aged brain. As behavioral deficits are more severe and long-lasting in the aged brain, later inflammatory events and their relationship to functional recovery need to be examined.

We next examined both the central and peripheral inflammatory response to ischemic injury to better understand the underlying mechanism leading to these differences in functional and histological stroke outcomes. Cerebral ischemia leads to a peripheral immune deficiency syndrome which increases the susceptibility to fatal infections in stroke patients (Meisel et al., 2005). Peripheral lymphoid organs, such as the spleen contract after stroke in young mice (Offner et al., 2006b; Prass et al., 2003) or in sex- mixed young mice (Prass et al., 2003). We saw a significant decrease in splenic weights even at an acute endpoint of 24 h in both males of all ages and aged females animals. This profound splenic contraction may be the hallmark of peripheral immunosupression and in aged animals may explain their high mortality (F. Liu et al., 2009) due to susceptibility to infections after stroke. Importantly, young and middle aged females were protected from stroke induced splenic atrophy. This was associated with circulating estrogen levels, as this protection was lost in aged acyclic females. Earlier studies have demonstrated that estrogen improves stroke induced peripheral immunosuppression (Zhang et al., 2010). The enhanced splenic contraction in aged females provides evidence of this in a model of natural gonadal senescence.

The post stroke inflammatory cascade in the brain has been well characterized in the literature (Gelderblom et al., 2009; Stevens et al., 2002). However, few studies have examined the chronology of recruitment of peripheral immune cells in aged animals, especially in aged females, a population that now bears the brunt of stroke related mortality and disability (Go et al., 2013). Consistent with other studies, we found an increase in CD45high leukocytes specifically, neutrophils, monocytes/macrophages, and DCs in the brain 24 h after stroke (Felger et al., 2010; Jin et al., 2010). Likewise, we also found a robust post-stroke increase in microglia, the intrinsic immune cells of the brain (Gelderblom et al., 2009; Jin et al., 2010; Kamel and Iadecola, 2012). Thus, innate immune cells rapidly increase in the brain after stroke. We did not find an overall increase in T cells in the brain at 24 h after stroke, consistent with other studies that found that they infiltrate the brain, but peak relatively late (3–4 days post ischemic injury) (Campanella et al., 2002; Gelderblom et al., 2009; Jander et al., 1995; Jin et al., 2010; Stevens et al., 2002). However, even at the early time point examined here, a differential pattern of T cell recruitment after ischemia in the brain was evident.

T lymphocytes are integral to the adaptive immune response. SCID (severe combined immunodeficient) (Hurn et al., 2007) and Rag1^−/−(recombinase activating gene-deficient) mice (Kleinschnitz et al., 2010; Yilmaz et al., 2006), lacking both T and B cells, have better outcomes after experimental stroke. This effect was specific to T cells (not B cells), as reconstitution of B cells did not change the protective phenotype seen in Rag1^−/− mice (Kleinschnitz et al., 2010; Yilmaz et al., 2006). In our study, T cell numbers negatively correlated with circulating estradiol levels, suggesting that estrogen may be normally suppressing the accumulation of T cells in the brain after injury, consistent with previous reports (Dang et al., 2011). This immunosuppressive effect of estrogen was lost with the age-related loss of estrogen in females, and their T cell numbers became equivalent to that of aged males. Higher number of T cells in males versus females in the brain post-stroke has also been noted by others (Brait et al., 2010; Yan et al., 2011), but age-related recruitment of T cells in the brains of aged mice is a novel finding. The specifics of which T cell subsets are recruited to the brains of aged mice after stroke and their function will be determined in future studies.

DCs, the antigen presenting cells, are a bridge between innate and adaptive immune response. DCs have been known to accumulate in the ischemic brain as early as 1 h after injury (Felger et al., 2010; Gelderblom et al., 2009; Kostulas et al., 2002; Yilmaz et al., 2009). Consistent with these studies, we saw a dramatic increase in DC numbers after stroke. Interestingly, this accumulation was more robust in females and aged animals of both sexes. These DCs may be either the resident brain DCs, as suggested by previous bone marrow chimera studies (Felger et al., 2010), or blood-derived DCs (Yilmaz et al., 2006) as DC progenitors have been known to decrease in the blood post stroke. Gr1⁺ monocytes are preferentially recruited to sites of injury, and these cells then differentiate into DCs de novo (Kovats, 2012; Serbina et al., 2008; Shi and Pamer, 2011). Recently, estradiol has been shown to mediate the differentiation of DCs in ex vivo culture models (Kovats, 2012). Thus, it is possible that the preferential accumulation of DCs in ischemic brain in young and middle aged females vs. males is due to estrogen driven differentiation of Gr1⁺ monocytes into DCs. The mechanisms causing DC accumulation may be entirely different in aged males and females, as advanced age alone leads to the formation of inappropriately matured DCs with increased basal levels of activation (Agrawal et al., 2012). We found both an increase in the antigen presenting cells (DCs) and the cells which execute the antigen dependent immune response (T cells) in the aged brain post stroke. This may lead to increased proinflammatory cytokine production (Dinapoli et al., 2010) and be responsible for the poorer functional outcomes in aged animals.

Innate immune cells, neutrophils and monocytes infiltrate the brain after an ischemic event (Amantea et al., 2009; Jin et al., 2010; Kriz, 2006; Price et al., 2004). Gr1 is a classic marker for neutrophils (Egan et al., 2008), however, monocytes/macrophages also express Gr1 (Getts et al., 2008; Howe et al., 2012). There are two separate populations of circulating monocytes, Gr1⁺ “inflammatory” monocytes, and Gr1-monocytes (Auffray et al., 2009). Gr1⁺ inflammatory monocytes are thought to cause the M1 type macrophage response, inflammatory cytokine production and proteolysis, while the Gr1⁻ non inflammatory monocytes patrol the blood vessels and differentiate into M2 type “healing” macrophages at later time points after injury (Hammond et al., 2012; Martinez et al., 2006; Nahrendorf et al., 2007). We saw a significant increase in both the Gr1⁺ and Gr1⁻ cell populations in the brain after stroke. Unlike Gr1⁺ macrophages, the Gr1⁻ macrophages displayed an age and sex specific response post stroke. Interestingly, there were significantly more Gr1⁻ macrophages in young and middle aged females as compared to males at this early time point. Since the numbers of non-inflammatory Gr1⁻ macrophages decreased in aged acyclic females, this is the first suggestion of estradiol dependent recruitment of Gr1⁻ non inflammatory macrophages in the brain after stroke.

We also examined the age and sex effects of stroke on microglia, the resident innate immune cells (Iadecola and Anrather, 2012). Microglia are the first responders to ischemic injury in the brain (Jin et al., 2010) and their activation in stroke may be both beneficial (Denes et al., 2007) and detrimental (del Zoppo et al., 2007; Jin et al., 2010; Pun et al., 2009; Yenari et al., 2006). It is difficult to histologically distinguish microglia from blood-derived monocytes/macrophages in the brain as both express similar cell surface markers (Jin et al., 2010). We used differential expression of CD45 to separate microglia from blood-derived monocytes/macrophages. Similar to previous reports (Gelderblom et al., 2009; Schroeter et al., 1999), we saw a large and rapid increase in microglia numbers after stroke. As others have reported an increase in microglia numbers and activation state in aging (Ogura et al., 1994; Sheffield and Berman, 1998), it was somewhat surprising that we found the microglial response to be age independent using flow cytometry. Since, stroke causes significant cell death in the infarct core; we also performed immunohistochemistry to examine the morphology and numbers of microglia in the penumbra. We found an increase in microglia numbers with aging and stroke. This enhanced baseline microglial response in the aged injured brain (Sandhir et al., 2008) may be acting as a powerful stimulus for further leukocyte recruitment, leading to further T cell activation, creating the overall enhanced inflammatory response after stroke in the aged brain.

In summary, we found an effect of both sex and aging on the inflammatory response to ischemic stroke (Fig. 6). The sex effects found closely mirrored estradiol levels. Estrogen is robustly neuroprotective in most injury models and has pleiotropic mechanisms of action (Jia et al., 2009; M. Liu et al., 2009; McCullough and Hurn, 2003), including actions as an anti-inflammatory agent on a number of cell types involved in the immune response, which has been well characterized in young subjects. The estrogen receptor, more specifically ERα is expressed on all immune cells of the body, T cells, B cells, DCs, monocytes and macrophages (Kovats, 2012). Previous studies have found estrogen mediated suppression of activation/entry of circulating immune cells in the brain after stroke (Johnson and Sohrabji, 2005). Our data suggests that estrogen may mediate its protective effects by suppressing T cell accumulation and enhancing Gr1⁻ macrophage recruitment in the brain. Estrogen may also be protecting cycling females from stroke-induced splenic contraction. Future studies are needed to confirm this effect using ovariectomized and estrogen supplemented groups. This anti-inflammatory phenotype was reversed to a pro-inflammatory response in aged females. Aging leads to a basal chronic inflammatory state that may be due to increased reactivity to self-antigens (Agrawal et al., 2012). Consistent with this hypothesis, we found increased T cells, DCs and microglia in the brains of aged mice after stroke. Moreover, stroke induced peripheral splenic contraction was also more profound in the aged animals. We speculate that this enhanced inflammatory response in the brain combined with peripheral immunosuppression may be responsible for the poor functional outcomes seen in aged animals after an ischemic event. It appears that therapies such as hypothermia that target the inflammatory response may be efficacious in improving stroke outcomes in the aged. Interestingly although hypothermia has been shown to be neuroprotective in some studies using aged animals (Florian et al., 2008; Joseph et al., 2012), other studies have shown it to be less efficacious (F. Liu et al., 2012) and this therapy has not yet translated into an effective therapy for patients with focal stroke. Whether this is due to age-related differences in the recruitment of some inflammatory cells still needs to be determined.

Fig. 6.

Bar graphs showing the differential leukocyte percentage (expressed as a percentage of CD45 high) in the brains of males (6A), and females (6B), across the lifespan after MCAO.

We conclude that the inflammatory response in the ischemic brain is differentially regulated by both age and biological sex. Immunomodulatory therapies will likely need to be specifically tailored for age and sex to be effective.