Background
Estradiol (E
2) modulates a wide set of morphological and physiological endpoints across the lifespan in many vertebrates. While the influence of E
2 on various indices of neuroplasticity has long been established [
1], there is now an emerging role for this steroid in neuroprotection against degeneration and inflammation following insult to the CNS, including traumatic brain injury (TBI; [
2,
3]). Indeed, E
2 is associated with a decreased risk or progression of a variety of clinical diseases, including Alzheimer’s disease, multiple sclerosis, schizophrenia, and some sleep disorders [
4‐
6], and improves outcomes following experimental TBI and stroke [
3,
7‐
9].
In addition to peripheral sources, the brain itself is capable of E
2 synthesis via the expression of aromatase (
estrogen-
synthase) in neurons of the limbic forebrain [
10]. In songbirds and mammals, ischemic or excitotoxic brain injury also induces aromatase expression in reactive astrocytes immediately around the site of damage [
3,
11‐
17]. In songbirds, this induction is particularly rapid, dramatic, and sustained [
11,
15,
18,
19] and results in a robust increase of local E
2 content around the site of injury [
20]. Locally derived E
2 is a potent modulator of cell turnover as it decreases apoptosis and increases cyto- and neurogenesis [
14,
15,
18,
21‐
23]. Glial aromatization also decreases microglial activation following experimental stroke [
9] and other indices of inflammatory signaling following brain injury [
24].
In recent work, we have posited an association between neuroinflammatory signaling pathways and injury-induced increases in neural aromatization in zebra finches (
Taeniopygia guttata). This association reflects a feedback loop of inflammatory and neurosteroidogenic signaling in the injured brain and includes two stages: the induction of aromatase expression by injury-induced inflammatory signals and the subsequent anti-inflammatory effects of injury-induced increases in central E
2. Indeed, inhibition of cyclooxygenase (COX) 1/2-activity with indomethacin mitigates the robust induction of aromatase and E
2 soon after a penetrating brain injury. More specifically, birds who received indomethacin or vehicle delivered via a penetrating needle into contralateral lobes had lower prostaglandin E2 (PGE2), aromatase-expression, and E
2 content in the hemisphere injected with the COX-inhibitor [
25]. These data strongly suggest that aspects of injury-induced inflammatory signaling are, in part, responsible for the induction of aromatase following brain damage. These increases in central E
2 exert powerful inhibitory influence on inflammatory signaling. Specifically, birds injected with the aromatase inhibitor fadrozole alone or with concomitant E
2 demonstrated elevated and decreased COX2 expression and PGE2 content relative to vehicle, respectively [
24]. These data strongly suggest an anti-inflammatory role for injury-induced aromatization via the synthesis of E
2. This unique feedback between neuroimmune and neuroendocrine signaling may serve as a powerful model towards understanding the role of inflammation and steroidogenesis in neuroprotection. Despite these recent findings, the cell-specificity (glial or neuronal) of COX-dependent aromatase expression is unclear, and the mechanisms underlying the interactions of inflammatory and neurosteroidogenic signaling pathways during brain injury are completely unknown. We investigate this in experiment 1 and predict that COX activity increases glial aromatase following brain injury.
PGE2 has a high affinity for four known E-prostanoid (EP) receptors: EP 1-4 [
26]. Binding of PGE2 to these receptors can regulate aromatase and E
2 via modulation of downstream signaling pathways in other systems [
27‐
29]. EP-1 and 2 regulate aromatase in adipose stromal cells [
28], while EP-2 and 4 are necessary for the modulation of aromatase in breast cancer cells [
29]. However, nothing is known about the association of any EP receptors and aromatase during brain trauma. Further, EP-dependent function is surprisingly understudied in songbirds. This impedes progress in our understanding the interactions between neuroimmune and neurosteroidogenic signaling pathways during brain trauma. Correspondingly, there are three known estrogen receptors (ER): ER-α, ER-β, and the g-protein coupled receptor-1 (GPER1). Of these, ER-α is a potent modulator of cell death and infarct size following experimental stroke with correlated effects on NADPH oxidase activation, cytokine release, and microglia activation and following ischemia or lipopolysaccharide administration [
30‐
33]. ER-α along with ER-β may mediate neurogenesis after ischemia [
34], suggesting important roles for ERs in neuroinflammation and cell-turnover. However, which ERs regulate PGE2 and other indices of neuroinflammation during brain injury is unknown, particularly in the songbird. We investigate the role of prostanoid and estrogen receptors in experiment 2 and 3 and predict a receptor(s) will be involved in cox-dependent increases in E
2 and consequent decreases in PGE2.
Here, we document injury-associated, COX-dependent increases in glial aromatase expression and replicate increases in central PGE2 and E2 following penetrating brain injury. Further, we describe injury-dependent changes in PGE2- and ER expression and reveal the necessity of specific EP receptors and ER in the injury-dependent, reciprocal interactions of neuroinflammatory and neurosteroidogenic pathways.
Discussion
Earlier reports suggest that indices of inflammation including prostaglandin signaling may be both necessary and sufficient for the increases in neural aromatase expression and E
2-content following penetrating brain damage in the songbird [
25,
37]. While other reports strongly support an inductive role for brain damage on astrocytic aromatase expression in songbirds [
14,
15,
18,
19,
37,
54] and mammals [
6,
55‐
57], the specific role of COX-activity on aromatase expression in this cell type was unclear. Moreover, the mechanism(s) that supported the inductive role of COX1/2-dependent signaling on aromatization were completely unknown. The present data suggest that glial aromatase is potently affected by local COX1/2 activity following brain damage as evidenced by lower numbers of aromatase-expressing cells of astrocytic morphology around damage associated with indomethacin administration relative to controls. As previous work using double-label immunocytochemistry has established the astrocytic nature of aromatase-expressing cells following mechanical damage [
15] and neuroinflammation [
37], we are confident that aromatization affected by indomethacin is likely expressed in reactive astrocytes around the site of damage.
The majority of the current set of studies focused on the mechanisms responsible for the reciprocal interactions of injury-associated inflammatory signaling and neurosteroidogenesis. Specifically, given the inductive role of COX-activity on neural aromatization [
25] and the anti-inflammatory influence of injury-induced E
2-synthesis [
24], we were interested in how PGE2 may influence neural aromatization following brain trauma and how locally generated E
2 may regulate neuroinflammation.
Role of prostanoid receptors on injury-induced aromatization
In a preliminary study, we first sought to describe injury-induced changes in the four known prostanoid receptors (EP 1-4) in the songbird brain using qPCR for these target gene-products at various times following injury. Surprisingly, EP-1 and EP-2 receptors were not represented in the zebra finch genome [
58] or could not be amplified with two different sets of primers specific to EP-2 under conditions that revealed abundant and specific amplification of other products from the identical cDNA template. The expression of both EP-3 and EP-4 changed in a temporally distinct manner following brain injury. Specifically, at 6 h, females showed increased EP-3 receptor mRNA, and both sexes had higher EP-4 expression. Additionally, females, but not males had elevated EP-3 and EP-4 expression 24 h post-damage. Thus, we chose to antagonize these receptors during injury to test if these receptors are necessary for the induction of E
2 using specific antagonists (L-798, 106, or BCG-20-1531 hydrochloride). Antagonism of prostanoid receptor(s) did prevent injury-induced E
2 but in a sex-specific manner. Antagonism of EP-3 in males prevented E
2 induction at 24 h, and antagonism of EP-4 prevented the induction in females at 6 h post-injury. These time points were chosen based on a previous study in our lab that found that inhibition of cox 1/2 signaling, and therefore PGE2, prevented the induction of E
2 at 6 h for females and 24 h for males, and not vice versa. Given this data, we believe that PGE2 may bind to EP-3 in males and EP-4 in females to achieve the robust induction of E
2 that has been well-documented following penetrating brain injury. However, additional doses of specific antagonists and/or time points are necessary to conclude that this represents a true sex-difference in the mechanism underlying the induction of aromatase by PGE2.
Prostanoid receptors have been shown to regulate aromatase in other systems in the periphery, including adipose stromal cells [
28] and breast cancer cells [
29] via EP-1, 2, or 4. These receptors have been shown to increase cAMP or intracellular calcium concentrations, which may induce aromatase and E
2 content. In some systems, EP-3 decreases cAMP through G
i signaling [
26]. However, it is unknown how these receptors work in the songbird brain. Our current data suggests that EP-3 and EP-4 may stimulate E
2 following brain injury.
Role of ERs in the anti-inflammatory effects of injury-induced E2
In order to determine how estrogen receptors change following brain injury, we measured three known receptors (ER-α, ER-β, and GPER1) using qPCR. Both males and females had elevated ERα and ERβ at multiple time points following injury, but we failed to detect changes in GPER1 at any time point. Thus, we chose to antagonize ER-α and ER-β with specific antagonists (MMP or PHTPP) to test if they are necessary for the reduction of PGE2 following brain injury. Previous work in our lab suggests that E
2 induction acts as a potent anti-inflammatory signal [
24]. Specifically, central E
2 decreases cytokine and cox-2 mRNA, along with PGE2 content 24 h following penetrating brain injury. In the current study, antagonism of ER-α, and not ER-β, results in the prolonged elevation of PGE2 content. Thus, ER-α may be responsible for the anti-inflammatory actions of injury-induced E
2; exploration of additional doses or time points may be necessary to understand this pathway fully.
Previous work has shown that ER-α is a potent anti-inflammatory signal following various types of brain insult [
30‐
33,
59‐
61], but our data is the first to suggest that it may do so by decreasing PGE2 signaling in vivo. Although, similar in vitro work has identified an effect of E
2 on PGE2 production [
62], it seems to be mediated through ER-β [
63]. Identification of the role of estrogen receptors play in regulation of PGE2 may have relevant implications from a therapeutic perspective. Selective estrogen receptor modulators (SERMs) may be appropriate for the treatment of neuroinflammatory disorders [
64,
65]. The overexpression of COX-2 is prevalent in many neurodegenerative diseases or models of trauma, including epilepsy, Alzheimer’s disease, or ischemia [
64]. SERMs can decrease of microglia activation [
65] and have been used to limit inflammatory signaling in experimental models [
65]. Our data suggest that ER-α may be necessary to limit excessive inflammatory signaling following damage and could be a potential therapeutic target.