Neuroprotection of Sex Steroids

Abstract

Sex steroids are essential for reproduction and development in animals and humans, and sex steroids also play an important role in neuroprotection following brain injury. New data indicate that sex-specific responses to brain injury occur at the cellular and molecular levels. This review summarizes the current understanding of neuroprotection by sex steroids, particularly estrogen, androgen, and progesterone, based on both in vitro and in vivo studies. Better understanding of the role of sex steroids under physiological and pathological conditions will help us to develop novel effective therapeutic strategies for brain injury.

1. Introduction

Sex steroids are involved in functions that extend beyond reproduction. The effects of sex steroids are implicated in cognition, synaptic plasticity, memory, neurogenesis, and neuroprotection. In this article, we review the neuroprotective effects of estrogen, progesterone, and androgens following brain injury, particularly ischemic brain injury, and molecular mechanisms that are implicated in this neuroprotection.

Stroke is a major cause of death and disability worldwide. Ischemic stroke affects more than 700,000 individuals every year in the United States. Epidemiological studies have shown that stroke incidence and mortality rates are higher in men relative to age-matched women worldwide,¹ suggesting that stroke is a sexually dimorphic disease. However, stroke event rates increase with age, and more women are affected by stroke after menopause.¹ Long-term exposure to ovarian estrogens may protect against ischemic stroke, an effect that seems to cease with menopause. Thus, this loss of protection has been one reason for using postmenopausal hormone replacement therapy.

Several different animal models have been used for ischemic brain injury studies, including transient or permanent middle cerebral artery occlusion (MCAO), transient forebrain ischemia, and transient global ischemia. MCAO serves as a valuable in vivo model for human stroke. In addition, oxygen-glucose deprivation (OGD) has been used as an in vitro ischemic model. In experimental stroke, the interpretation of cellular and molecular data in animals is relatively straightforward and consistent with neuroprotection by estrogen. Despite these studies, recent clinical randomized controlled trials have failed to show the beneficial effects of hormonal replacement therapy in estrogen-treated groups. A possible explanation for the apparent discrepancy between clinical trials and animal studies has been ascribed to differences in timing of exposure, dose, duration, and route of administration. However, a careful review of the extensive literature suggests that data from clinical trials and other recent studies should be interpreted within the narrow context of the study design.² For example, in some studies, beneficial effects of estrogen against stroke may not be observed because estrogen was introduced in most of the subjects after an extended period of hypoestrogenicity. Therefore, it is important to use appropriate models to perform studies on the effect of sex steroids on brain injury. This will not only advance our understanding of the relationship of sex steroids and neuroprotection but will also improve the clinical management of stroke thereby reducing brain damage and disability.

2. Estrogen and neuroprotection

Over the past several decades, estrogen has been extensively studied as an endogenous neuroprotective agent in cardiovascular and neurological diseases including ischemic brain injury. An accumulating body of evidence from in vivo (Table 1) and in vitro (Table 2) studies demonstrates that estrogen provides powerful protection against ischemic brain injury through multiple molecular mechanisms.

Table 1.

Neuroprotection of Estrogen in vivo Studies.

Model	Protection	Mechanisms/Results	Sex	Species	References
Transit MCAO	Yes	Increases Aromatase	Female, OVX	Mouse	49
Transit MCAO	Yes	Increases EDHF	Male	Rat	33
Gonadectomy	Yes	Decreases ROS production	Female, male, OVX, ORX	Rat	35, 36
Transit MCAO	Yes	Increases BcI-2	Female, Male, OVX	Rat, mouse	3, 8
Permanent MCAO	Yes	Reduces inflammation, iNOS	Female and OVX	Mouse	24
Permanent MCAO	Yes	Increases neurogenesis	Female, OVX	Mouse	37
Gonadectomy	Yes	Increases angiogenesis	Female, OVX	Rat	38, 39
Transit MCAO	Yes	Increases P-STAT3	Female, OVX	Rat	45
Transit MCAO	Yes	Increases CART	Female, OVX	Rat	43
Transit MCAO	Yes	Increases blood flow, eNOS	Female, OVX	Rat	3,23,40
Transit MCAO	Yes	Decreases sEH	Female, OVX	Mouse	40
Transit MCAO	Yes	nNOS	Female, Male, OVX	Mouse	28
Permanent MCAO	Yes	Proestrus	Female	Rat	5
Permanent MCAO	Yes	Reduces edema	Female, OVX	Rat	41
Permanent MCAO	NO	N.D.	Female	Rat	9

Abbreviations: OVX: ovariectomized; ORX: Orchiectomized; MCAO: middle cerebral artery occlusion; EDHF: endothelium-derived hyperpolarization factor; nNOS: neuronal nitric oxide synthase; N.D. not determined.

Table 2.

Neuroprotection of Estrogen in vitro Studies.

Cell Type	Conditions	Protection	Mechanisms/Results	Sex	Species	References
Astrocytes	OGD	Yes	Increases aromatase	Female/Male cells	Rat and Mouse	15, 16
Astrocytes	17β-estradiol	Yes	Reduces inflammation	Mixed cell	Rat	26
Astrocytes	17β-estradiol	Yes	Increases TGF-β	Mixed cell	Rat	42
Neurons	17β-estradiol	Yes	Reduces oxidative stress	Mixed cell	Rat	35
Neurons	Glutamate	Yes	Reduces apoptosis	Mixed cell	Rat, Mouse	10, 11
	or Aβ exposure
Neurons	Glutamate	Yes	Increases BcI-2	Mixed cell	Rat	23
NT2 Neurons	Glutamate	Yes	Increases BcI-2	Mixed cell	Rat, Human	11, 22
	H2O2 exposure
Endothelial cell	OGD	Yes	Estrogen reduces cell death	Mixed cell	Mouse	17
Endothelial cell	17β-estradiol	Yes	Increases eNOS	Mixed cell	Human and Bovine	27
Oligodentricytes	Peroxynitrite	Yes	Decreases cytotoxicity	Mixed cell	Rat	13
Oligodentricytes	OGD	Yes	Estrogen reduces cell death	Mixed cell	Rat	14
Oligodentricytes	Cystine deprivation	Yes	Reduces oxidative stress	Mixed cell	Rat	14
Microglial	LPS	Yes	Suppresses iNOS	Mixed cell	Rat	25

Abbreviations: OGD: oxygen glucose deprivation; Aβ: amyloid beta; eNOS: endothelial nitric oxide synthase; iNOS: inducible nitric oxide synthase; LPS: Lipopolysaccharides; H₂O₂: Hydrogen peroxide; TGF-β: Transforming growth factor β.

2.1. Estrogen and ischemic brain injury – in vivo studies

The neuroprotective effects of estrogen have been widely documented in various animal models of experimental stroke in different species and different genetic strains. Overall, adult female rats sustain smaller infarcts after experimental stroke induced by 2 hours middle cerebral artery occlusion (MCAO) compared to age-matched males.³ We have repeated this study in different animal strains under pathological conditions that are known to be high risk factors for stroke in humans, such as hypertension and diabetes. We found that this sex difference in ischemic brain injury persists in genetic models of hypertension³ and diabetes,⁴ suggesting that fundamental mechanisms and pathways of ischemic brain injury may be different between males and females. Moreover, in female animals, the estrous cycle also influences the outcome of ischemic brain injury, eg, the infarct size is smaller following ischemic brain injury induced by permanent MCAO during the proestrus (high endogenous estrogen levels) stage of the estrous cycle compared to the metestrus (low endogenous estrogen levels) stage.⁵ In addition, sex differences in infarct size disappear after MCAO in reproductively senescent rats.⁶ Taken together, these differences in outcome of cerebral ischemia are in part due to the protective effect of female sex hormones, especially estrogen,⁷ since ovariectomy increases ischemic brain damage in female rats³ and mice;⁸ and estrogen replacement is protective against cerebral ischemia in ovariectomized rats³ and mice⁸ and in reproductively senescent male and female animals.⁶ However, findings from some animal studies show that estrogen is not neuroprotective and can be deleterious under certain circumstances following stroke induced by permanent MCAO.⁹ Discrepancies in results among different studies have been attributed to differences in dose, formulation, route of administration, or length of treatment. As a whole, the above studies provide strong evidence that estrogen contributes to neuroprotection against cerebral ischemic damage in both male and female animals.

2.2. Estrogen and neuroprotection – in vitro studies

In vitro, estrogen protects neurons against insults induced by glutamate, glucose deprivation, and beta-amyloid peptide.^10,11 A number of studies have suggested that estrogen may suppress microglial activation, an effect that could help mediate estrogen neuroprotection.¹² Other studies demonstrate that 17β-estradiol protects against cell death after OGD in primary oligodendrocyte cultures and that protection by estrogen is dose-dependent.^13,14 In both male and female astrocytes, physiological levels of 17β-estradiol added to the culture medium prevents cell death following OGD.^15,16 Other evidence indicates that both long-term and short-term pretreatments with 10 nM 17β-estradiol protect cerebral endothelial cells after OGD.¹⁷

2.3. Mechanisms of neuroprotection by estrogen

Genomic and non-genomic actions mediated by estrogen receptors

Estrogen actions are mediated by two estrogen receptor (ER) isoforms – ERα and ERβ. Both of these subtypes are expressed in cortical astrocytes, neurons, and endothelial cells and thus could mediate neuroprotection by estrogen. Administration of the potent ER antagonist ICI 182,780 dramatically increases infarct size in intact female mice following MCAO,¹⁸ suggesting that the neuroprotective effect of estrogen is mediated via ER. Infarct size is reduced after permanent MCAO in estrogen-treated wild-type (WT) and ERβ knockout mice, but not in ERα knockout mice, suggesting that ERα mediates estrogen protection against ischemic brain injury.¹⁹ Furthermore, the protective effects of estrogen against endothelial cell death following OGD is mediated by ERα, but not ERβ.¹⁷ Upregulation of estrogen receptors influences the expression of genes involved in cellular proliferation and other physiological changes. There is now general acceptance that estrogen, like other steroid hormones, can act in a classical way as a transcription factor by binding to nuclear receptors, translocating to the nucleus, and interacting with estrogen response elements located in the promoter of target genes.

Recent work has led to the discovery of a third estrogen receptor, the membrane-associated G-protein-coupled receptor 30 (GPR30). GPR30 binds estradiol with high affinity and is expressed in breast cancer cells and various other tissues in the body including brain.²⁰ GPR30 has been recognized as a putative membrane receptor for estrogen that mediates a series of non-genomic signals from estrogen.²⁰ Non-genomic actions of estrogen have a rapid onset and include modification of protein phosphorylation and levels of intracellular second messengers such as cyclic adenosine monophosphate (cAMP) and calcium.²¹ However, the role of GPR30 in neuroprotection by estrogen against ischemic brain injury is unknown.

Increases Bcl-2

Bcl-2 and cell survival in cerebral ischemia and apoptosis has been extensively studied. Estrogen exhibits neuroprotection by increasing Bcl-2 expression in NT2 neurons²² and in primary cortical neurons¹¹ when these cells are exposed to hydrogen peroxide or glutamate. In an ischemic stroke animal model, Bcl-2 mRNA and protein expression increases accompanied by smaller infarct size in female or 17β-estradiol-treated rats compared with WT males and estrogen-deficient animals.³ Furthermore, ovariectomy increases infarction after MCAO in WT females, but not in ovariectomized female mice overexpressing Bcl-2.⁸ These results suggest that estrogen prevents ischemic injury, at least in part, by mediating Bcl-2 upregulation.

Regulates MAPK /PI3K/AKT signaling pathways

In the brain, the mitogen-activated protein kinases (MAPK) and the phosphatidylinositol-3 kinase (PI3K/AKT) pathways are associated with the beneficial effect of estrogen.²³ Several studies report that estrogen enhances activation of PI3K/AKT, which is an important mediator of cell survival signaling pathways,²¹ i.e., in cultured neurons, 17β-estradiol reduces cell death via the PI3K/AKT signaling pathway following various injuries, such as glutamate toxicity.

Reduces inflammatory response

Ovariectomy with estrogen treatment reduces infarct size following permanent MCAO in WT female mice, but not in inducible nitric oxide synthase (iNOS) knockout female mice²⁴, suggesting that estrogen-mediated neuroprotection may involve suppressing inflammatory reactions. Physiological levels of estradiol suppress both microglial activation and the iNOS-mediated immune response as well as the production of several pro-inflammatory mediators including metalloproteinase, prostaglandin E2 (PGE2), and cyclooxygenase 2 (COX-2).²⁵ In addition to microglial cells, astrocytes have been shown to mediate inflammatory action following injury. One study shows that 17β-estradiol pretreatment reduces the protein expression of interleukin 1-beta (IL-1β) and tumor necrosis factor-α (TNFα) and suppresses matrix metallopeptidase 9 (MMP-9) activity in astrocyte media following a 24-hour treatment with lipopolysaccharide (LPS).²⁶ Together, these studies demonstrate that post-ischemic inflammation strongly contributes to the extent of brain injury, and estradiol exhibits anti-inflammatory effects that protect against ischemic injury.²⁵

Modulates nitric oxide synthase

Endothelial dysfunction is an important component of initiating and contributing to the pathogenesis of ischemic damage. In ischemic brain injury, estrogen effects have been linked to nitric oxide (NO). NO is produced from L-arginine by nitric oxide synthase (NOS). Endothelial injury may increase the inflammatory response through loss of normal NO production due to inhibition of NOS. One mechanism by which estrogen protects the ischemic brain may involve increasing cerebral blood flow by enhancing the activity of endothelial NOS (eNOS). This effect has been demonstrated both in vivo and in vitro.^23,27 Indeed, treatment of ovariectomized rats with estrogen activates eNOS in cerebral blood vessels, enhancing NO production²³; and eNOS protein expression increases in endothelial cell cultures after incubation with physiological concentrations of estrogen.²⁷

Another study shows that histological brain injury is significantly increased after MCAO in female neuronal NOS knockout (nNOS^−/−) mice compared to WT females. In contrast, ovariectomy with estradiol treatment has no effect on ischemic damage compared to vehicle-treated ovariectomized female nNOS^−/− mice. These results suggest that the neuroprotective effect of estradiol following ischemic brain injury is mediated via an nNOS pathway in the female.²⁸

Regulates endothelium-derived hyperpolarization factor

Estrogen plays an important role in the regulation of endothelial-dependent vasodilation. Endothelium-dependent relaxation is an important endothelial function and is known to be mediated by three different endothelium-dependent relaxing factors: prostacyclin (PGI2), nitric oxide (NO), and endothelium-derived hyperpolarizing factor (EDHF). Although NO is recognized as the primary relaxing factor at the level of conduit arteries, a growing body of evidence suggests that EDHF also plays an important role in the regulation of regional vascular resistance and blood flow, especially prominent in resistance arteries.²⁹ The EDHF-mediated response involves hyperpolarization with subsequent relaxation maintained even when NO and PGI₂ synthesis is inhibited.

EDHF-mediated vasodilations have been studied in both peripheral and cerebral circulations. Estrogen deficiency has no effect on NO-mediated endothelial relaxation in rat mesenteric arteries. However, EDHF-mediated relaxations and hyperpolarization induced by acetylcholine are significantly increased in arteries from intact females and overiectomized rats treated with estrogen compared to male and untreated ovariectomized rats.^30,31 Similarly, the EDHF response is significantly reduced in mesenteric arteries during the diestrus stage, a good experimental model of short-term estrogen deficiency, when compared with estrus controls.³⁰ These results suggest that both long-term and short-term estrogen deficiency can alter EDHF-mediated relaxation and hyperpolarization.³⁰

Although the actual identity of EDHF remains unknown, several putative candidate factors have been proposed including potassium ions, c-type natriuretic peptide (CNP), hydrogen peroxide (H₂O₂), and metabolites of arachidonic acid such as epoxyeicosatrienoic acids (EETs). Alternatively, EDHF may be an electrical coupling through myoendothelial junctions.²⁹ These gap junction structures are composed of several different members of the connexin protein family including connexin-37, connexin-40, and connexin-43. In mesenteric arteries, ovariectomy significantly reduces connexin-43 protein expression; and treatment with estradiol prevents the reduced expression of connexin-43 protein, indicating that estrogen deficiency may impair the EDHF-mediated response by changing the expression of connexin-43.³² This data suggests that estrogen alters EDHF-mediated relaxation in part by enhancing gap junctional communication.³²

Others have shown upregulation of EDHF-dependent vasodilation in rat middle cerebral arteries after 2 hours MCAO and 24 hours reperfusion compared with control and that potentiated EDHF-mediated dilations after ischemia/reperfusion are due to altered endothelial Ca²⁺ regulation.³³ These results suggest that EDHF plays an important role in cerebral blood flow after cerebral ischemia and may be neuroprotective. Elucidating the impact of estrogen on the cerebral vasculature after ischemic stroke is very important. More investigation is needed to better understand how estrogen affects cerebrovascular function under different physiological and pathological conditions.

Reduces oxidative stress

Cerebral ischemia and reperfusion are responsible for oxidative stress due to the generation of free radicals, which cause deleterious effects during pathogenesis. Mitochondria are a major source of reactive oxygen species (ROS) and oxidative stress and a key player in apoptosis.³⁴ Neurons are particularly sensitive to oxidative stress due to their high energy demand. Damage to mitochondria causes disruptions in ATP production and an increase in ROS that compromise antioxidant defense systems of the cell and lead to both necrosis and apoptosis. Studies have shown that estrogen suppresses brain mitochondrial ROS production in both male and female rats, and similar effects of estrogen on neural-like PC-12 cells confirmed that estrogen inhibits mitochondrial superoxide production.³⁵ Estrogen increases mitochondrial efficiency and reduces oxidative stress by stabilizing ATP production, promoting cell survival.³⁴ Furthermore, a physiological concentration of estrogen is neuroprotective as it modulates antioxidant enzyme activity, including superoxide dismutase, catalase, and glutathione.³⁶

Increases neurogenesis

The investigation of neuroprotective effects of estrogen in neurogenesis is new and interesting. Neurogenesis is restricted to two major brain regions – the dentate gyrus of the hippocampus and the subventricular zone (SVZ) lining the lateral ventricle.³⁷ Adult brain generates new neurons under normal and neurodegenerative conditions. Estrogen influences the process of adult neurogenesis.³⁷ Estrogen promotes the migration of newly generated neurons toward the damaged brain region, eg, from the SVZ to ischemic regions, thus facilitating brain remodeling and repair after ischemic injury. Suzuki et al showed that the number of newborn neurons in the SVZ significantly increased in estrogen-treated ovariectomized mice compared to vehicle-treated ovariectomized mice at 96 hours after permanent MCAO,³⁷ indicating that estrogen enhances neurogenesis following ischemic stroke. Since these newborn neurons of the SVZ can migrate to ischemic regions to potentially replace damaged neurons, this finding may have clinical implications. More research is needed to determine whether these newborn neurons have any effect on functional recovery following ischemic brain injury.

Increases angiogenesis

Angiogenesis, the growth of new blood vessels, is essential for organ development and plays a critical role in the long-term outcome following ischemic injury. Estrogen modulates angiogenesis under physiological and pathological conditions. In the brain, cerebral capillary density in the frontal cortex is significantly increased in estrogen-treated ovariectomized females compared to vehicle-treated ovariectomized female rats.³⁸ In addition, the levels of mRNA and protein expression of vascular endothelial growth factor (VEGF) in cerebral vessels are significantly decreased in ovariectomized female rats, and estrogen replacement increases VEGF expression to levels similar to intact females.³⁸ Others have shown that ovariectomy with estrogen treatment increases angiopoietin-1 mRNA and protein expression compared to vehicle-treated ovariectomized animals.³⁹ Both VEGF and angiopoietin-1 are endothelial cell-specific growth factors and can promote angiogenesis. Thus, these studies demonstrate that estrogen enhances pro-angiogenic molecular expression and promotes cerebral angiogenesis. However, whether angiogenesis will promote functional recovery following ischemic brain injury is unclear.

Other mechanisms

Other potential mechanisms by which estrogen mediates neuroprotection have been studied. (1) Estrogen influences blood flow during ischemic stress and reperfusion. In both normotensive wistar and stroke-prone spontaneously hypertensive rats, females subjected to ischemia have a smaller infarct size and higher cerebral blood flow during ischemia compared with males and ovariectomized females, suggesting that neuroprotection by estrogen is mediated by cerebral blood flow enhancement.³ (2) Estrogen suppresses soluble epoxide hydrolase (sEH) expression, enhancing the levels of epoxyeicosatrienoic acid (EET), which leads to increased blood flow during MCAO in female mice compared to male and ovariectomized female mice. Sex differences in blood flow disappear in sEH knockout mice (sEHKO) using in vivo quantitative optical microangiography,⁴⁰ suggesting that EET is involved in the neuroprotective action of estrogen. (3) Estrogen also modulates blood-brain-barrier permeability and attenuates stimulation of BBB cotransporter activity, reducing edema formation during stroke.⁴¹ (4) In vitro, 24 hours pretreatment with 10 nM of 17β-estradiol significantly reduces cortical neuronal cell death induced by toxicity from glutamate,¹¹ suggesting that estrogen protects primary cortical neurons from glutamate toxicity. (5) Another mechanism involved in the neuroprotective actions of estrogen is the release of growth factors such as glial-derived transforming growth factor-β (TGF-β), which promote neuronal survival.⁴²

Studies have shown that estrogen increases cocaine- and amphetamine-regulated transcript (CART) mRNA and protein expression in cortical neuronal cultures after OGD; and estrogen also increases CART expression in ischemic cerebral cortex and reduces cortical infarct size after MCAO,⁴³ suggesting that CART plays a role in neuroprotection. More recently, Mao et al reported that CART can directly interact with subunit B of the mitochondrial enzyme succinate dehydrogenase (SDHB) based on yeast two-hybrid system screening and in vitro pull-down assay.⁴⁴ Interestingly, treatment with nanomolar (nM) concentrations of CART significantly increases SDH function, complex II activity, and ATP production in purified mitochondria and protects primary cultured cortical neurons at baseline and after OGD.⁴⁴ Collectively, these results suggest that CART may have a mitochondrial protective function through directly interacting with SDH after cerebral ischemic injury.

Ovariectomized rats treated with 17β-estradiol have increased phosphorylation of the signal transducer and activator of transcription-3 (P-STAT3) and reduced infarct size after transient focal cerebral ischemia compared to vehicle-treated ovariectomized female rats. This protective effect of estradiol on infarct size is abolished in the presence of P-STAT3 inhibitor, suggesting that estrogen mediates protection against cerebral ischemic injury via P-STAT3. Furthermore, immunohistochemistry shows that P-STAT3 is expressed in cells staining positive for microtubule-associated protein 2 (MAP2) and Bcl-2, demonstrating that P-STAT3 is expressed in surviving neurons.⁴⁵

3. Aromatase and neuroprotection

In addition to circulating estrogen, growing evidence indicates that local estrogen synthesized by aromatase also protects against cerebral ischemic injury. Aromatase, cytochrome P450 19 (P450 aromatase), is encoded by the cyp19 gene and has been implicated as beneficial in injured brain because it can synthesize protective estrogen from androgen precursors.^46,47 P450 aromatase expression and activity have been detected in multiple brain regions of animals including the cerebral cortex. P450 aromatase expression has also been detected in the human cerebral cortex.⁴⁸

Previously, the role of brain aromatase was believed to be restricted to regulating neuroendocrine events and behavior linked to reproduction. Recent findings have revealed a novel neuroprotective role for brain aromatase. Brain injury induces P450 aromatase expression in astrocytes, which may be neuroprotective.⁴⁷ More recently, Liu et al first reported that female astrocytes sustain less cell death from OGD than do male astrocytes. This endogenous protection in female astrocytes is associated with greater basal P450 aromatase activity and higher P450 aromatase mRNA and protein expression relative to male astrocytes. Furthermore, the sex-dependent response to OGD disappears in male and female cells treated with a selective P450 aromatase inhibitor Arimidex.¹⁵ Interestingly, astrocyte-conditioned media from female cultures protects against OGD-induced cell death in male cells, whereas astrocyte-conditioned media from male cultures has no effect on OGD-induced cell death in female cells. However, the protection of male astrocytes is lost in media from female astrocytes treated with P450 aromatase inhibitor.¹⁵ These results strongly suggest that endogenous P450 aromatase and local estrogen synthesis protect astocytes following OGD.

These new and exciting observations of cytoprotection by aromatase are also confirmed in astrocyte cultures from WT mice and mice with targeted deletion of the P450 aromatase gene (ArKO). ArKO mice of both sexes have abnormal reproductive phenotypes with plasma estradiol levels below detection. Consistent with data generated in sex-specific rat astrocytes, cell death is significantly reduced in WT female mice astrocytes following OGD compared to WT male cells.¹⁶ It is important to note that this sex-dependent response to OGD is observed across species. We also demonstrated that cell death significantly increases in ArKO female astrocytes compared to WT female astrocytes, but not in ArKO male astrocytes vs. WT male cells; and sex differences in astrocyte cell death are not observed in ArKO cells.¹⁶ These findings demonstrate that P450 aromatase is expressed in astrocytes and confirm that aromatase is instrumental in mediating astrocyte survival following OGD. Therefore, P450 aromatase plays a critical role in endogenous neuroprotection in females. These data are consistent with our in vivo findings: brain injury is significantly increased in female ArKO mice vs. WT females after experimental stroke and in female WT mice chronically treated with fadrozole, a well known aromatase inhibitor, compared to female WT mice treated with vehicle. Furthermore, female ArKO mice sustain more brain damage than ovariectomized WT animals, demonstrating that extragonadal synthesis of estradiol is involved in neuroprotection.⁴⁹ Taken together, these studies demonstrate that brain P450 aromatase is an important factor for regulating brain function under physiological as well as pathological conditions.

In summary, both in vivo and in vitro studies demonstrate that estrogen exerts powerful neuroprotective actions following ischemic brain injury. The role of other sex steroids in neuroprotection has also been studied. Below, we will review the evidence for neuroprotection by progesterone and androgens following ischemic brain injury.

4. Progesterone and neuroprotection

The relative “female protection” observed following experimental stroke has been largely attributed to protective effects of estrogen (see above). For this reason, progesterone has been understudied in the context of cerebral ischemia. Nonetheless, the past decade has provided evidence that progesterone is a strong neuroprotectant in various models of brain injury and disease.

4.1. Progesterone and neuroprotection – in vivo studies

To date, over 20 published studies have used animal models to examine the neuroprotective effects of progesterone against cerebral ischemic injury. Only a few of these studies report that progesterone does not significantly decrease damage (see table 3 for references). Interestingly, most studies have focused on male animals; however, progesterone has been shown to protect both male and female brain from damage following MCAO. The duration of the insult is typically 1 or 2 hours. Histological improvement (infarct size) was observed as late as 7 days post ischemia, and cognitive improvement was observed up to 21 days after MCAO. The level of protection varied depending on the duration of ischemia and the dose of progesterone, but typically ranged between 20% and 50% reduction in infarct size. Significant neuroprotection by progesterone has also been observed following global cerebral ischemia (see Table 5). Using a 4-vessel occlusion model (4-VO) in male rats, 8 mg/kg progesterone significantly reduced hippocampal damage 21 days after global ischemia.⁵⁰ Similarly, progesterone completely prevented neuronal death in the caudate nucleus⁵¹ and significantly reduced neuronal death in the hippocampus⁵² following 15 minutes cardiac arrest in male cats. Our group observed that 8 mg/kg provided only modest protection to cerebellar Purkinje cells in male mice following cardiac arrest (unpublished observation), yet the progesterone metabolite allopregnanolone (8 mg/kg) significantly decreased Purkinje cell damage.⁵³

Table 3.

Neuroprotection of Progesterone in vivo Studies.

Model	Protection	Mechanisms/Results	Sex	Species	References
Transit MCAO	Yes	Reduces inflammation	Male	Rat	54-57,62,64,65
		Reduces edema		Mouse
		Reduces NO synthase
Transit MCAO	Yes	N.D.	Female, OVX	Rat	6, 61
Transit MCAO	NO	N.D.	Female, OVX	Rat	63
Permanent MCAO	Yes	Reduces edema	Male	Rat	58, 65-70
		Reduces inflammation, Reduces iNOS		Mouse
Global ischemia	Yes	N.D.	Male, Female, OVX	Cat, Rat	50-52, 59
Partial	Yes	Reduces inflammation,	Male	Mouse	60
Global Ischemia		Increases antioxidants
(BCAO)

Abbreviations: MCAO: middle cerebral artery occlusion; OVX: ovariectomized; NO: nitric oxide; iNOS: inducible nitric oxide synthase; N.D. not determined; BCAO: bilateral common carotid artery occlusion.

Table 5.

Neuroprotection of Androgen in vivo Studies.

Model	Protection	Mechanisms/Results	Sex	Species	References
Transit MCAO	Yes (low doses)	Low doses of T and DHT protect	Male	Mouse	87
T and DHT	NO (high doses)	High doses exacerbate ischemic damage.
pre-injury treatment
Transit MCAO	Yes	T removal decreases tissue damage	Male,	Rat	83
Remove of T before injury		as compared to animals with endogenous T
Transit MCAO	NO	Positive correlation	Male	Rat	84
		Plasma T and tissue damage	(castration with or without T repletion)
Transit MCAO	NO	NO change in tissue damage,	Male	Rat	86
		but improved behavioral	(castration with or without T repletion)
Transit MCAO	NO	DHT increase damage	Male	Rat	89
			(castration with or without T repletion)
Transit MCAO	Yes	T and DHT reduces tissue damage	Male	Rat, Mouse	88
			(Aged animals received T or DHT)

Abbreviations: AMPA: α-amino-3-hydroxyl-5-methyl-4-isoxazole-propionate; Aβ: amyloid beta; DHT: dihydrotestosterone; NMDA: N-methyl-D-aspartic acid; T: testosterone.

Jiang et al first reported that a dose of 8 mg/kg progesterone, by intraperitoneal injection, significantly reduces infarct size; most subsequent studies have used that dose and route of administration.⁵⁴ Significant neuroprotection was not observed with 4 mg/kg.^54-56 However, Cai et al reported significant neuroprotection with 4 mg/kg progesterone following 1 hour MCAO,⁵⁷ and Betz and Coester found that 2 mg/kg progesterone reduces brain edema following 4 hours MCAO.⁵⁸ Neuroprotection has also been observed using higher doses of progesterone: 10 mg/kg with a 4-VO⁵⁹ model and 15 mg/kg with a bilateral common carotid artery occlusion (BCAO) model.⁶⁰ Interestingly, 20, 30, and 60 mg/kg doses of progesterone do not protect against MCAO and may even exacerbate infarct levels.⁶¹

Pre- and post-treatment regimens have been used to test the ability of progesterone to protect against ischemic damage. With the exception of a few studies that administered progesterone intravenously,^50,55 subcutaneously,^51,52 or with subcutaneous pellets,⁶ most studies used intraperitoneal injections. Pre-treatments have consisted of administering progesterone 7 days, ^{6,51,52,59,62} 48 hours,⁵⁷ 1 hour, ^57,58 or 30 minutes^54,60,61,63 prior to the ischemic insult. Post-treatments have been initiated at reperfusion^{54,56,62,64,65} or at occlusion⁶⁶ and 1 hour post-occlusion^65,67-70 in permanent MCAO models. One study administered progesterone 20 minutes after reperfusion in a 4-VO model and achieved significant neuroprotection.⁵⁰ The neuroprotective potential of progesterone in experimental ischemia is clear; however, further research is needed to optimize dose, timing, and route of administration. Regardless, these findings suggest great therapeutic potential for treating ischemic stroke with progesterone.

4.2 Progesterone and neuroprotection – in vitro studies

The protective efficacy of progesterone has been tested in a variety of in vitro models that mimic various damaging insults experienced following cerebral ischemia (Table 4). Specifically, progesterone protects primary cultured cortical neurons,^41,71,72 hippocampal neurons,⁷³ and cerebellar Purkinje cells⁷⁴ from glutamate toxicity and in vitro ischemia (OGD).⁷⁴ Progesterone can also protect primary astrocytes,⁶⁶ the macrophage cell line RAW264.7,⁶⁶ and the microglial cell line BV-2⁷⁵ from inflammatory models of ischemia in vitro.

Table 4.

Neuroprotection of Progesterone in vitro Studies.

Cell Type	Conditions	Protection	Mechanisms/Results	Sex	Species	References
Neurons	OGD	Yes	Increases GABAA	Female/Male cells	Rat	74
			Receptor activity	Mixed cell
Neurons	Glutamate/	Yes	Increases BDNF,	Mixed cell	Rat	71-73
	NMDA, Toxicity		Increases BcI-2		Mouse
Astrocytes	IFNγ, IL-1β	Yes	Reduces iNOS	Mixed cell	Mouse	66
Macrophages	IFNγ, LPS	Yes	Reduces iNOS	Mixed cell	Mouse	66
Microglia	LPS	Yes	Reduces iNOS	Mixed cell	Mouse	75

Abbreviations: OGD: oxygen glucose deprivation; GABAA receptor: γ-Aminobutyric acid A receptor; iNOS: inducible nitric oxide synthase; LPS: Lipopolysaccharides; BDNF: Brain-derived neurotrophic factor; NMDA: N-methyl-D-aspartic acid; IFNγ: Interferon-gamma; IL-1β: Interleukin-1.

4.3. Mechanisms of neuroprotection by progesterone

Decreases inflammation

An anti-inflammatory role for progesterone has been reported. In particular, progesterone has been observed to suppress the expression of various pro-inflammatory cytokines following cerebral ischemia, minimizing the ischemia-induced increase in IL-1β,⁶⁵ TNF-α,⁶⁰ and TGF-β.⁶⁵ Additionally, progesterone can decrease the expression of inducible nitric oxide synthase, iNOS or NOS-2, in vivo ^65,66 and in vitro. ^66,75

Reduces oxidative stress

Progesterone can reduce the level of malondialdehyde, a marker of oxidative stress, and prevent ischemia-induced decreases in glutathione following 10 minutes 4-VO.⁵⁹ Similarly, increased expression of the antioxidant enzymes superoxide dismutase, glutathione peroxidase, and catalase have been reported following progesterone administration and bilateral common carotid artery occlusion (BCAO).⁶⁰ As a result, progesterone attenuation of oxidative stress may contribute to decreased inflammation.

Decreases edema

Progesterone can decrease cerebral edema following MCAO^58,65,68,70 and traumatic brain injury.⁷⁶ Suppression of injury-induced inflammation by progesterone likely contributes to reduced cerebral edema formation. However, there may be other progesterone- mediated mechanisms of osmoregulation. For example, following traumatic brain injury, progesterone can regulate expression of a water-permeable channel, aquaporin-4, in a time- and region-specific manner.⁷⁷

Other mechanisms

Various molecular targets and pathways have been proposed to underlie progesterone-mediated protection. For example, Kaur et al found that progesterone protection from glutamate toxicity is associated with increased levels of BDNF transcript and protein in cortical neurons using an organotypic slice model.⁷² Activation of mitogen-activated protein kinase (MAPK)^71,72 and phosphoinositide-3 kinase (PI3K)⁷² are also associated with, and required for, progesterone-mediated neuroprotection against glutamate toxicity. In addition, Cai et al reported that the timing of progesterone administration alters the mechanism of protection.⁵⁷ Specifically, progesterone receptor-mediated neuroprotection is the major contributor when progesterone is administered 48 hours prior to MCAO. In contrast, administration of progesterone 1 hour prior to insult results in progesterone receptor-independent protection, mediated via antagonizing the σ receptor, which inhibits the NMDA receptor thereby reducing the injury-induced rise in intracellular calcium.

Another mechanism for neuroprotection by progesterone is attributed to its metabolism to allopregnanolone, a strong neuroprotectant⁷⁸ and potent modulator of GABA_A receptors that can increase the inhibitory drive to offset excitotoxicty. In cerebellar Purkinje neurons, progesterone can protect against ischemic injury in a GABA_A receptor- dependent manner; and importantly, protection is abolished by finasteride, an inhibitor of 5α-reductase that metabolizes progesterone to allopregnanolone.⁷⁴ Ischemia can also cause a downregulation of GABA_A receptors early in reperfusion that is prevented by allopregnanolone.⁵³ Aside from acting to stimulate and stabilize GABA_A receptors, there is evidence that allopregnanolone acts as a neuroprotectant by activating various effector pathways to decrease apoptotic and necrotic cell death.⁷⁸

5. Androgens and neuroprotection

Male sex steroids have been much less studied than those of the female. This statement is surprising because male sex is a well-recognized stroke risk factor in humans. Stroke sensitivity, e.g., brain damage that occurs in response to a known ischemic insult, is also greater in male animals and male cultured cells. Whether these male characteristics are under the control of androgen availability remains highly unclear. For example, low circulating testosterone levels are associated with risk for stroke and worse outcome in stroke survivors.^79-82 However, androgen levels are known to be stress-sensitive, so this evidence may reflect a generalized response to acute insult.

5.1 Androgens and neuroprotection – in vivo studies

In bench studies that control steroids, androgens have been reported to either protect or exacerbate ischemic damage (Table 5). In overview, animal data suggest that reduction of androgens via stress or castration improves histological damage after cerebral ischemia.^83-85 On the other hand, exogenous androgens administered after experimental stroke accelerate functional recovery,⁸⁶ an apparent inconsistency that we currently do not understand. We have recently shown that maintaining testosterone or dihydrotestosterone (DHT) plasma levels within the low physiological range during stroke confers protection to both adult castrates and gonadally intact aged rodents with naturally declining androgens. The protection is presumably mediated via androgen receptor (AR) mechanisms because improved tissue outcomes are reversed by administration of the AR antagonist flutamide.^87,88 Using a similar hormone implantation technique, we also observed that high but physiologically relevant androgen levels exacerbate ischemic damage in castrated males,⁸⁹ emphasizing that androgens have complex dose-dependent effects in vivo.

5.2. Androgens and neuroprotection – in vitro studies

As in the animal work, both beneficial and deleterious effects of androgens have been reported in neuronal and glial cultures exposed to injury conditions such as oxidative stress, excitotoxicity, serum deprivation, and amyloid β (Aβ) exposure (Table 6).^90-92 In primary neurons, testosterone either protects against or exacerbates damage depending on concentration. For example, supra-physiological concentrations (10 μM) amplify glutamate-induced excitotoxic neuronal death, while protection is observed at 10 nM because testosterone is aromatized to 17β-estradiol.⁹³ In other studies, testosterone and non-aromatizable DHT salvage injured neurons, suggesting direct protection rather than actions through aromatization.^94-96 One interpretation of the apparent paradox of beneficial vs. detrimental effects of androgens is that two potentially competing signaling mechanisms are engaged in vitro. Such mechanisms remain to be elucidated.

Table 6.

Neuroprotection of Androgen in vitro Studies.

Cell Type	Conditions	Protection	Mechanisms/Results	Sex	Species	References
HT22 Neurons	Glutamate exposure	NO	Exacerbates cell death	Mixed cell	Mouse	83
	T pre-injury treatment
Oligodendrocytes	AMPA or kainite receptor	NO	Enhances excitotoxicity;	Mixed cell	Rat	90
	T pre-injury treatment		Reverse by flutamide;
			But not by aromatase inhibiton
Neurons	NMDA exposure/	Yes	Enhances excitotoxicity and	Mixed cell	Mouse	93
Astrocytes	T pre-injury treatment		protection, depending on T concentration.
Cerebellar	H2O2 exposure/	Yes	Protects against oxidant injury;	Mixed cell	Rat	91
granule cells	T pre-injury treatment		reverse by flutamide
Neurons	Aβ exposure/T or DHT	Yes	Protect across 1-1000 nM range	Mixed cell	Rat	95, 96
	Pre-injury treatment
Neurons	Serum deprivation/	Yes	Reduces apoptosis after	Mixed cell	Human	92
	T co-treatment		serum deprivation; reveres by flutamide.

Abbreviations: AMPA: α-amino-3-hydroxyl-5-methyl-4-isoxazole-propionate; Aβ: amyloid beta; DHT: dihydrotestosterone; NMDA: N-methyl-D-aspartic acid; T: testosterone; H₂O₂: Hydrogen peroxide.

5.3. Mechanisms of neuroprotection by androgens

ARs are expressed in neurons throughout the brain, including cortical and striatal regions impacted by our focal ischemia models.⁹⁷ Therefore, it is possible that AR-regulated transcription is an important mechanism underlying androgen actions. We used microarray and real time PCR to identify gene candidates induced by non-aromatizable DHT and observed that high physiological doses enhance pro-inflammatory gene expression after ischemic injury. (For full list of genes, see reference ⁸⁹.) Non-genomic, rapid signaling pathways also have been implicated in androgen neuroprotection in cerebral ischemia via MAPK/MEK and PI3K/AKT signaling, and CREB activation.^96,98-100 Continued investigation of transcriptional and non-genomic signaling may provide important insights into how androgens impact evolving brain damage.

6. Conclusions

Significant sex differences in brain injury have been observed in animal models as well as in clinical and epidemiological studies. Since sex steroids play a critical role in brain function under physiological and pathological conditions, they represent the most salient factor related to sex differences in brain injury. Physiological levels of the female sex steroids 17β-estradiol and progesterone are neuroprotive both in vivo and in vitro. The role of male steroids in neuroprotection is less clear. More recently, sex differences in the response to ischemic injury have also been found in male and female cells. Therefore, further studies on the effects of sex steroids on brain injury will help us to better understand the cellular and molecular basis of sex differences in susceptibility to stroke injury. Elucidating sex-specific mechanisms of brain injury will advance the development of more effective neuroprotective therapies.

Figure 1.

Mechanisms of neuroprotection by estrogen in the brain.