Using qRT-PCR-based Taqman low-density gene arrays and IPA-based bioinformatics, we investigated the bioenergetics, metabolism and inflammation-related genomic and associated functional impact of ovarian hormonal depletion induced by OVX and exposure to E2/CoP4/CyP4 in the adult female rat hippocampus. Compared to Sham-OVX, OVX induced a number of significant gene expression changes, which were to a large extent reversed by co-administration of E2 and cyclic P4 (E2+CyP4). Co-administration of E2 and continuous P4 (E2+CoP4) induced a markedly disparate profile from that induced by E2+CyP4; the E2+CoP4 gene expression profile exhibited some degree of resemblance to the OVX profile.
OVX-induced changes differentiating E2+CoP4 and E2+CyP4 included genes involved in mitochondrial energy and redox metabolism, with major genes including Atp5a1 and Prdx5, respectively. Atp5a1 encodes the α subunit of mitochondrial ATP synthase catalyzing ADP to ATP production 
. ATP synthase, the complex V which is the last complex of the mitochondrial respiratory chain, plays a central role in synthesis of the majority of ATP demanded in aerobic organs, particularly in metabolically active organs such as brain 
. Regulation of ATP synthase is involved in physiological processes underlying normal nervous system functions, as well as neurodegenerative processes/diseases such as AD 
. In a recent study, APP and Aβ bound to ATP synthase α subunit and regulated its activity at the surface of neural cells, suggesting a novel mechanism in Aβ-mediated AD pathology 
. The ATP synthase α subunit also is particularly vulnerable to oxidative damage compared to other components of the mitochondrial respiratory chain at earlier AD stages, suggesting that the oxidation of ATP synthase and alterations of its biological functions may be an early and essential step in the pathogenesis of AD 
Peroxiredoxins (Prxs) are a family of thiol-dependent antioxidant enzymes that catalyze the reduction of peroxides 
. Different from conventional antioxidants, Prxs exhibit unique properties including high enzymatic activity through utilization of cysteine at their active site and high cellular abundance and involvement in both antioxidant defense and redox signaling 
. Six Prx isozymes have been identified in mammalian cells sharing a common peroxide-scavenging activity despite differences in their cellular distribution and subcellular localization. Prx1 and Prx6 are expressed in astrocytes and glial cells; Prx2, Prx3, Prx4, and Prx5 (encoded by Prdx5) are expressed in neurons 
. Subcellularly, Prx1, Prx2 and Prx6 are mainly localized in the cytoplasm, Prx4 in the endoplasmic reticulum, Prx3 in the mitochondria; Prx5 is expressed in various compartments including peroxisomes and mitochondria 
. Since mitochondria are a major site of generation of and target for hydrogen peroxide, the mitochondrial Prxs, including Prx3 and Prx5, play an important role in regulation of mitochondrial redox homoeostasis 
. Prx5 in particular may present broader activity than other antioxidant mitochondrial enzymes 
. Overexpression of Prx5 has a positive impact on neuronal survival and extension of lifespan; Prx5 deficiency produces signs of oxidative stress, apoptosis, and shortened lifespan 
Our findings of deficits in the expression of genes involved in mitochondrial energy and redox metabolism in the hippocampus of OVX rat brain are consistent with our previous observations indicating vulnerability of CNS bioenergetic processes to female endocrine aging. For instance, in female mice, we discovered a close link between reproductive senescence and neuronal bioenergetic deficit demonstrated by a significant decline in complex IV cytochrome c oxidase activity and mitochondrial respiration 
. The vulnerability of neuronal bioenergetics appeared to be more prominent in AD brain, indicated by our finding in 3xTg-AD mice that mitochondrial bioenergetic deficits in the brain preceded the development of AD pathology 
from the present study emphasize compromised mitochondrial energy and redox metabolism with ovarian disruption and are consistent with earlier reports of age and AD related deficits in bioenergetics 
. A regimen of E2 combined with a cyclic, not continuous, P4, could be an effective strategy for rescuing brain from hormonal deficiency-related brain hypometabolism and increased oxidative stress.
Another major change induced by OVX and differentially regulated by E2+CoP4 and E2+CyP4 was in expression of Igf1. On bioinformatics analysis, Igf1 appeared to act as a hub in the most significant network associated with E2+CyP4. Igf1 encodes the insulin-like growth factor I (Igf-I) that serves as a master regulator of a wide spectrum of biological functions and homeostasis in multiple tissues 
. Igf-I plays an essential role in the modulation of neural activities from early development to adult neurogenesis, to cognition, and to protection against neurodegenerative diseases 
. Igf-I modulates brain levels of Aβ via regulation of proteins involved in major pathways leading to catabolic removal of Aβ 
. Igf-I is also regulated by a number of other factors, including the sex steroids E2 and P4. Substantial evidence indicates crosstalk between E2 and Igf-I signaling pathways to regulate multiple neural responses including cognition 
. Relative to E2, the impact of P4 on the Igf-I system has been less explored. A recent report suggests that P4 may positively regulate the Igf-I system in glial cells, which could be directly linked to a P4-mediated protection against demyelination 
. Data from the present study supports the close interactions between Igf-I and E2/P4; Igf-I deficit associated with hormonal deficiency could be effectively reversed by chronic exposure to E2+CyP4.
Additional changes induced by OVX and differentially regulated by E2+CoP4 and E2+CyP4 included genes involved in cholesterol homeostasis and APP intramembranous processing, major players including Nr1h3 and Bace 1 respectively. Nr1h3 encodes the α subtype of the liver X receptor (Lxrα), which, along with Lxrβ (Nr1h2), are oxysterol receptors that function as master transcription factors mediating cholesterol homeostasis in the periphery 
. In the brain, emerging evidence indicates that, in addition to the regulation of cholesterol transport and metabolism, activation of Lxrs also attenuates inflammatory responses and Aβ production, supporting their therapeutic potential in AD management and other neurological disorders involving dysregulation of cholesterol homeostasis 
. Bace1 encodes β-secretase, the β-site APP-cleaving enzyme I. 
. Since Bace1 cleavage of APP is a pre-requisite for Aβ generation, and the level and activity of Bace1 are elevated in AD, Bace1 has been suggested as a biological candidate marker for early detection of AD 
. Data from the present study suggest that loss of ovarian hormones may increase risk for the development of AD through deficits in cholesterol and APP metabolism, which may be prevented by chronic exposure to E2+CyP4. Compared to pathways discussed above that were regulated in different direction by E2+CoP4 versus E2+CyP4, the APP processing pathway appeared to be a less dramatic differentiator since the key genes (Bace1, Psenen) involved in the pathway were repressed by both E2+CoP4 and E2+CyP4.
The disparity between E2+CoP4 and E2+CyP4 was also reflected on their differential regulation of P4 receptors, Pgr versus Pgrmc1. Pgr is the classical nuclear PR and has been associated with many neural responses to P4 
. In contrast to most reports utilizing acute HT, we found that chronic HT treatment paradigms used in the present study including E2, CoP4 and E2+CoP4 decreased Pgr. In chick embryos, acute P4 exposure increased Pgr in hypothalamus and posterior pituitary 
. In adult OVX rats, E2 and combined E2 and P4 increased Pgr in several hypothalamic nuclei, while P4 alone was not different from OVX 
. Although some studies document that P4 can antagonize E2-mediated upregulation of Pgr 
, others did not find P4 antagonism of E2 induction of Pgr 
. Recent data from our group indicate regional differences in Pgr regulation by E2 and P4 within the adult rat hippocampus 
. Moreover, in the hypothalamus, OVX decreased Pgr by 80% in the medial preoptic nucleus but by only 20% in the ventromedial nucleus 
. We know of only one other report of Pgr response to chronic hormone exposure 
. Continuous E2 treatment for 8 weeks in OVX rats did not increase Pgr in the hypothalamus and hippocampus 
. Together, these data suggests that Pgr regulation by E2 and P4 could be highly dependent upon the duration of hormone treatment, as well as brain region.
In contrast to Pgr, Pgrmc1 is a membrane-associated PR and a member of a multi-protein complex regulating a spectrum of hormone-signaling pathways 
. Pgrmc1 is expressed in multiple brain regions, mediates P4-induced responses in multiple neural cell types, and is suggested to play a role in the promotion of neonatal dendritic growth, spino- and synapto-genesis 
. Further, our previous analysis revealed that Pgrmc1 mediated P4-induced increase in cell cycle gene expression and proliferation in rat neural progenitor cells 
. The present study demonstrated that CyP4 alone or with E2 induced an up-regulation of Pgrmc1. E2 and CyP4 interacted synergistically since the response to E2+CyP4 was greater than the response induced by CyP4 alone. In contrast, an antagonistic interaction occurred between E2 and CoP4 where Pgrmc1 expression decreased. Together, these results lead to the hypothesis that the disparities in gene expression patterns between E2+CoP4 and E2+CyP4 could be closely associated with their differential regulation of Pgrmc1.
It should be pointed out that, technically and statistically, this study presents both strengths and limitations. One issue arisen was the consideration of adjustment for multiple hypothesis testing: an approach such as consideration of the false positive discovery rate, in particular, is usually required in microarray studies where thousands of genes are analyzed simultaneously. Because of the relatively large chance for false positive discoveries involved in such high density microarray studies, findings usually need to be further validated by real-time RT-PCR analysis. In our study, we directly used the real-time RT-PCR technique following a carefully developed and validated protocol, and as part of this protocol, our low-density array consisting of a pathway-focused set of target genes was custom designed based on prior literature and hypothesized associations with neurological aging dysfunctions as well as hormone therapy interventions. Therefore, in consideration of three factors: 1) the well-designed and quality-controlled technical strengths, 2) the intent to reduce false negatives in this first-ever analysis, 3) the by and large internal agreement of our findings with published functional data, we did not control for multiple hypothesis testing, and instead, chose to present unadjusted p-values in this study. Given these reasons, we however recognize the possibility of false positive associations and the need for independent validation in larger studies.
Collectively, our findings from the present study demonstrate that, consistent with many preclinical and clinical observations, surgically-induced ovarian hormone loss exerts a negative impact on multiple biological processes in the brain, including mitochondrial energy production, antioxidant defense/redox homeostasis, cholesterol trafficking, insulin signaling and amyloid metabolism, all of which, as discussed above, play a critical role in neurological health and disease. Further, data contained herein indicate that systems of gene expression in brain are differentially impacted by exposure to cyclic versus continuous hormone interventions and treatment paradigms. A regimen of E2 combined with a cyclic exposure to P4, which more closely mimics the natural female hormone pattern, induced gene expression profiles consistent with the ovary intact brain. In stark contrast, a regimen of E2 combined with a continuous combined exposure to P4, which more closely mimics the clinical hormone therapy regimen, induced gene expression profiles consistent with the ovarian hormone deficient brain. These preclinical data require confirmation in a model of natural menopause but suggest that a hormone therapy of E2 and cyclic exposure to P4 would be a promising therapeutic strategy for preventing or reducing hormonal deficiency-induced compromise on neurological health and function.