Brain-derived estrogen exerts anti-inflammatory and neuroprotective actions in the rat hippocampus

Highlights

• Aromatase is highly expressed in neurons in the hippocampus in non-ischemic rats.

• Aromatase and local estradiol increase in astrocytes after global cerebral ischemia.

• In vivo knockdown of aromatase leads to greater global ischemic damage and increased microglial activation.

• Brain-derived estradiol exerts neuroprotection and anti-inflammatory effects in the hippocampus.

Abstract

17β-estradiol (E2) has been implicated to play a critical role in neuroprotection, synaptic plasticity, and cognitive function. Classically, the role of gonadal-derived E2 in these events is well established, but the role of brain-derived E2 is less clear. To address this issue, we investigated the expression, localization, and modulation of aromatase and local E2 levels in the hippocampus following global cerebral ischemia (GCI) in adult ovariectomized rats. Immunohistochemistry (IHC) revealed that the hippocampal regions CA1, CA3 and dentate gyrus (DG) exhibited high levels of immunoreactive aromatase staining, with aromatase being co-localized primarily in neurons in non-ischemic animals. Following GCI, aromatase became highly expressed in GFAP-positive astrocytes in the hippocampal CA1 region at 2–3 days post GCI reperfusion. An ELISA for E2 and IHC for E2 confirmed the GCI-induced elevation of local E2 in the CA1 region and that the increase in local E2 occurred in astrocytes. Furthermore, central administration of aromatase antisense (AS) oligonucleotides, but not missense (MS) oligonucleotides, blocked the increase in aromatase and local E2 in astrocytes after GCI, and resulted in a significant increase in GCI-induced hippocampal CA1 region neuronal cell death and neuroinflammation. As a whole, these results suggest that brain-derived E2 exerts important neuroprotective and anti-inflammatory actions in the hippocampal CA1 region following GCI.

Introduction

17β-Estradiol (E2, estrogen) is a steroid hormone that has been implicated to be neuroprotective against a variety of neurodegenerative disorders, including stroke, Alzheimer’s disease (AD) and Parkinson’s disease, although controversy exists (Brann et al., 2007, Yao and Brinton, 2012, Simpkins et al., 2012, Bourque et al., 2012). With respect to stroke, studies in rats, mice and gerbils found a sex difference in brain injury following cerebral ischemia, with young adult female animals having smaller infarct volume as compared to young adult males (Brann et al., 2007, Alkayed et al., 1998, Park et al., 2006). Similarly, a number of studies have documented sex differences in stroke risk and outcome in humans, with women generally protected against stroke, at least until menopause (Murphy et al., 2004, Di Carlo et al., 2003). Many groups, including our own, have shown that exogenous administration of E2 dramatically reduces infarct volume in cortex and hippocampus following focal or global cerebral ischemia (GCI) in ovariectomized female mice, rats and gerbils, and in male rats and gerbils (Brann et al., 2007, Zhang et al., 2008, Simpkins et al., 1997, Dubal et al., 1998, Shughrue and Merchenthaler, 2003, Zhang et al., 1998).

It has been generally assumed that the neuroprotective effects of E2 are primarily due to ovarian-derived E2. However, work by a number of laboratories has shown that certain areas of the brain exhibit high expression of the E2 generating enzyme, aromatase, which has raised the possibility that brain-derived E2 may have important roles in the CNS. For instance, work within the last decade in rodents, birds, monkeys, and humans has shown that forebrain structures, in particular the hippocampus CA1–CA3 regions, exhibits high expression of aromatase as indicated by in situ hybridization, RT-PCR and immunohistochemical analysis, and can produce significant levels of E2 levels that are equivalent to or even higher than that observed in the circulation (Veiga et al., 2005, Hojo et al., 2004, Azcoitia et al., 2011, Higaki et al., 2012, Yague et al., 2008, Fester et al., 2011, Garcia-Segura, 2008, Mukai et al., 2010, Shen et al., 1994). It should be noted that the cerebral cortex has also been reported to express aromatase (Azcoitia et al., 2011, Stoffel-Wagner et al., 1999, Srivastava et al., 2010), and thus brain-derived E2 may also regulate cortical functions. In support of this possibility, global aromatase knockout mice have been reported to have greater cortical damage following focal cerebral ischemia than wild type ovariectomized mice, suggesting that brain-derived E2 may have neuroprotective actions in the cerebral cortex (McCullough et al., 2003).

With respect to the hippocampus, treatment of cultured mouse hippocampal neurons with an aromatase inhibitor has been reported to result in a significant decrease in axon outgrowth and dendritic spines in the CA1 region (Fester et al., 2011, Mukai et al., 2010, Kretz et al., 2004, Rune and Frotscher, 2005, von Schassen et al., 2006), as well as a significant decrease of long-term potentiation (LTP) amplitude, dendritic spines and synapses in hippocampal slices in vitro (Grassi et al., 2011, Vierk et al., 2012). These results suggest that local E2 in the hippocampus may modulate synaptic function. Interestingly, studies in songbirds have also shown that inhibiting aromatase by intracerebral administration of aromatase inhibitors results in increased damage and apoptosis in the brain after a penetrating injury (Wynne and Saldanha, 2004, Wynne et al., 2008). Aromatase inhibition has also been reported to result in increased hippocampal damage in male rats following excitotoxic injury (Azcoitia et al., 2001).

It is well known that the hippocampal CA1 region is highly vulnerable to GCI, which can occur after cardiac arrest, asphyxiation, and hypotensive shock (Neumann et al., 2013, Harukuni and Bhardwaj, 2006), and can lead to significant neuronal damage, cognitive defect and mortality. It is currently unknown whether brain-derived E2 in the hippocampal CA1 region has a neuroprotective role against GCI, and whether it can modulate neuroinflammation that occurs after GCI. To address these deficits in our knowledge, the goals of the current study were: (1) to access whether aromatase and local E2 levels change in the hippocampus following GCI, (2) to determine the cell types containing aromatase and local E2 expression in ischemic and non-ischemic animals, and (3) to assess whether antisense oligonucleotide knockdown of aromatase and local E2 levels in the hippocampus affects GCI-induced neurodegeneration and inflammation in ovariectomized rats.