Effects of Cholinergic Lesions and Cholinesterase Inhibitors on Aromatase and Estrogen Receptor Expression in Different Regions of the Rat Brain

Abstract

Cholinergic projections have been shown to interact with estrogens in ways that influence synaptic plasticity and cognitive performance. The mechanisms are not well understood. The goal of this study was to investigate whether cholinergic projections influence brain estrogen production by affecting aromatase (ARO), or influence estrogen signaling by affecting estrogen receptor expression. In the first experiment, ovariectomized rats received intraseptal injection of the selective immunotoxin 192IgG-saporin to destroy cholinergic inputs to the hippocampus. In the second experiment ovariectomized rats received daily intraperitoneal injections of the cholinesterase inhibitors donepezil or galantamine for 1 week. ARO activity and relative levels of ARO, ERα, ERß, and GPR30 mRNAs were quantified in the hippocampus, frontal cortex, amygdala and preoptic area. Results show that the cholinergic lesions effectively removed cholinergic inputs to the hippocampus, but had no significant effect on ARO or on relative levels of ER mRNAs. Likewise, injections of the cholinesterase inhibitors had no effect on ARO or ER expression in most regions of the brain. This suggests that effects of cholinergic inputs on synaptic plasticity and neuronal function are not mediated by effects on local estrogen production or ER expression. One exception was the amygdala where treating with galantamine was associated with a significant increase in ARO activity. The amygdala is a key structure involved in registering fear and anxiety. Hence this finding may be clinically relevant to elderly patients who are treated for memory impairment and who also struggle with fear and anxiety disorders.

INTRODUCTION

The goal of this study was to explore potential effects of cholinergic manipulation on aromatase (ARO) activity and estrogen receptor (ERs) expression in different regions of the brain. Estrogens have been shown to have beneficial effects on learning, memory, and attention in multiple species including rats, mice, non-human primates, and in humans (Daniel et al., 1997; Bimonte and Denenberg, 1999; Luine et al., 2003; Gresack and Frick, 2006; Frye et al., 2007; Sherwin and Henry, 2008). Effects often are limited to females and are task specific. In rodents, estrogens (primarily estradiol (E2)) have been shown to enhance performance on a variety of spatial navigation tasks (Daniel et al., 1997; Fader et al., 1998; Gibbs and Johnson, 2008), and to enhance working memory (Bimonte and Denenberg, 1999; Daniel et al., 2006; Bohacek and Daniel, 2007), as well as novel object and object placement recognition (Luine et al., 2003; Frye et al., 2007; Fernandez et al., 2008). In humans beneficial effects have been observed on short-term and long-term verbal memory and logical reasoning (Sherwin, 1988; Krug et al., 2006). Estrogens also have been shown to enhance synapse formation, connectivity, and NMDA receptor expression in the hippocampus, with corresponding effects on synaptic transmission and long-term potentiation (McEwen et al., 2001; Jelks et al., 2007; Mendez et al., 2011). These effects are thought to underlie some of the effects of estrogens on cognitive performance.

We and others also have demonstrated that cholinergic projections from the medial septum (MS) to the hippocampus are significantly affected by estrogens and that these projections can play an essential role in enabling estrogen-mediated effects on cognitive performance. For example, ovariectomy reduces and E2 treatment increases choline acetyltransferase (ChAT) mRNA in the MS and nucleus basalis magnocellularis (NBM), with corresponding effects on ChAT activity (Luine, 1985; Gibbs and Pfaff, 1992; Gibbs et al., 1994; Gibbs, 1996, 1997), high affinity choline uptake and acetylcholine (ACh) release in the frontal cortex and hippocampus (Gibbs et al., 1997; Gibbs, 2000; Gabor et al., 2003). E2 treatment has been shown to mitigate effects of scopolamine on T-maze alternation in rats (Fader et al., 1998), and likewise to mitigate effects of both scopolamine and mecamylamine on cognitive performance in post-menopausal women (Dumas et al., 2006). Notably, selective removal of cholinergic projections to the hippocampus prevents estrogen-mediated enhancement of a delayed-matching-to-position (DMP) spatial navigation task (Johnson et al., 2002; Gibbs and Johnson, 2007). Similar cholinergic lesions also have been shown to block estrogen-mediated increases in synaptic spines on CA1 neurons in the hippocampus (Lam and Leranth, 2003). More recent studies suggest that loss of ERs also contribute to loss of estrogen effects on cognitive function with age (Foster, 2012; Bean et al., 2014; Black et al., 2016), and that increasing ERα expression can enhance cognitive performance (Foster et al., 2008). We have shown that beneficial effects of E2 on DMP acquisition can be restored by treating older rats and rats with partial cholinergic lesions with selective cholinesterase inhibitors (ChEIs) (Gibbs et al., 2009, 2011a, b). Whether levels of ERs also were affected was not explored.

Collectively the findings demonstrate important interactions between basal forebrain cholinergic projections and estrogen effects on performance that impact brain aging and cognition. To date there has been little study of whether cholinergic projections significantly influence estrogen signaling, ER expression, or perhaps even local estrogen production, in the brain.

Estrogens are produced by the aromatization of androgens via the cytochrome P450 enzyme ARO. ARO is encoded by the CYP19A1 gene. In brain there are two isoforms of the gene. One is 430nt in length (ARO_L), and is associated with enzyme activity. The other is a truncated form 300nt in length (ARO_S), the function of which is unknown (Tabatadze et al., 2014). Recent studies demonstrate that local estrogen production in the adult brain can significantly influence brain structure and function (Garcia-Segura, 2008; Roselli et al., 2009; Stocco, 2012; Kato et al., 2013; Fester and Rune, 2015; Bender et al., 2017).

Estrogen effects are mediated by binding with specific estrogen receptors (ERs). Three receptors have been identified. ERα and ERβ are nuclear receptors that act as transcription factors for estrogen-regulated genes (Toran-Allerand, 2004). Studies show that these receptors also are located in specific cytoplasmic compartments where they can activate second messenger signaling pathways such as mitogen-activated protein kinases (MAPK), calcium/calmodulin-dependent protein kinases (CamKII), and cAMP response element-binding proteins (CREB) (McEwen, 2002; Manavathi and Kumar, 2006). A third receptor GPR30 is more recently identified and is a G protein-coupled receptor (Funakoshi et al., 2006; Moriarty et al., 2006; Brailoiu et al., 2007). It is located both intracellularly and on the plasma membrane and promotes rapid estrogen signaling in a variety of cell types. In the rat brain, GPR30 is present in many regions including the basal forebrain, cortex, hippocampus and hypothalamus (Brailoiu et al., 2007; Hazell et al., 2009). Studies from our lab have demonstrated that GPR30 is expressed by the majority of basal forebrain cholinergic neurons (Hammond et al., 2011), and that treatment with a GPR30 agonist increases ACh release in the hippocampus similar to the effects of E2 (Hammond et al., 2011; Gibbs et al., 2014).

Based on these findings, we proceeded to explore whether selective cholinergic lesions, as well as treatment with cholinesterase inhibitors, have effects on ARO expression, ARO activity, and ER expression in different regions of the brain.

EXPERIMENTAL PROCEDURES

Animals

Ninety-four ovariectomized (OVX, 270–350 g, 3 months old) Sprague–Dawley female rats were purchased from Harlan Sprague–Dawley Inc. Rats were individually housed for two weeks in our facility on a 12 h:12 h light/dark schedule with unrestricted access to food and water. All procedures were carried out in accordance with PHS policies and with the approval of the University of Pittsburgh’s Institutional Animal Care and Use Committee.

In the first experiment, forty-six rats were used to test the effect of selective lesions of cholinergic neurons in the medial septum (MS) on aromatase (ARO) mRNA, activity and estrogen receptors (ERs) mRNA in the hippocampus and frontal cortex. Rats received intraseptal injections of 192IgG-Saporin (192IgG-SAP) or vehicle as described below. Two weeks later, rats were anesthetized with an overdose of ketamine (3 mg) and xylazine (0.6 mg). Brains were removed; hippocampal and frontal cortex tissues were collected and stored at −80 °C until use. Of the 46 rats, tissues from 20 rats (10/grp) were analyzed for relative levels of ARO and ER mRNAs using qRT-PCR methods described below. Tissues from the other 26 rats were analyzed for ARO activity (13 rats/grp) in microsomes using a recently validated and highly sensitive UPLC–MS/MS assay (Li et al., 2016). Immunohistochemical detection of ChAT-positive cells in the MS also was performed to confirm the loss of cholinergic neurons.

In the second experiment, a total of 48 rats were treated intraperitoneally with 3 mg/kg donepezil (Sigma–Aldrich, Inc.), 5 mg/kg galantamine (Sigma–Aldrich, Inc.) or saline (as control) injected once daily for 7 days. Donepezil and galantamine are cholinesterase inhibitors (ChEIs) approved for the treatment of memory decline associated with Alzheimer’s disease. Following treatment rats were anesthetized and brain tissues were dissected as above. In addition to collecting hippocampus and frontal cortex, tissues from the amygdala and preoptic area (POA) also were dissected and analyzed. Tissues from 12 rats (4 rats/grp) were analyzed for relative levels of ARO and ER mRNAs using qRT-PCR methods. Tissues from 36 rats (12 rats/grp) were analyzed for ARO activity.

Cholinergic lesions

Rats were anesthetized and placed on a standard stereotaxic apparatus. The skull was exposed and a hole was drilled at midline 0.3 mm rostral to Bregma. A 28-ga stainless steel cannula was lowered −5.6 mm from dura into the medial septum. 2.0 μl of 192IgG-SAP (0.2 mg/ml; Advanced Targeting Systems, Inc.) was injected at a rate of 0.2 μl /min. Previous studies have shown that these injections cause a selective loss of cholinergic cells in the basal forebrain with little non-selective damage to GABAergic neurons and no damage to cholinergic neurons in the NBM (Gibbs, 2002, 2007; Johnson et al., 2002; Rudick et al., 2003; Fitz et al., 2006, 2008; Gibbs et al., 2011; Babalola et al., 2012; Cai et al., 2012). This is also accompanied by decreased activity of choline acetyltransferase (ChAT) and by reduced high-affinity uptake of [³H]choline into cholinergic nerve terminals in the hippocampus, but not the frontal cortex (Rossner et al., 1995b). Controls received intraseptal infusions of saline. The skin was sutured and rats were placed onto a heating pad during recovery. Following surgery, rats received ketofen (Fort Dodge, Inc., 3.0 mg/kg, i.p.) once per day for three days to relieve pain. After 14 days of recovery, rats were dissected and the hippocampus and frontal cortex were collected and analyzed. In addition, tissues containing the MS were fixed by immersion in 4% paraformaldehyde in 50 mM phosphate-buffered saline (PBS, pH7.2) at 4 °C overnight. These tissues were then transferred to 20% sucrose in PBS at 4 °C for several days prior to sectioning and immunostaining.

ChAT assay

ChAT activity in the hippocampus and frontal cortex were measured as previously described (Gibbs and Johnson, 2007). Briefly, 30 mg tissues were sonicated in 300 μl sonication buffer (10 mg tissue/mL) which contains 10 mM EDTA and 0.5% Triton X-100. Samples were run in triplicate. 10 μl of substrate solution, which contains 0.25 mM [³H] acetyl-CoA (50,000–60,000 dpm/tube) was added to each reaction tube. 5 μl aliquots of sample in sonication buffer were added to the tubes and incubated for 30 min at 37 °C. The reaction was terminated with 4 mL sodium phosphate buffer (10 mM) at 4 °C. The production of [³H] acetylcholine was detected by adding scintillation LSC-cocktail (Packard Instruments, Meriden, CT) and counting cpm in the organic phase using an LKB beta-counter. ChAT activity was calculated for each sample as pmol acetylcholine manufactured/h/μg protein.

ChAT immunohistochemistry (IHC)

To further confirm the loss of cholinergic neurons, 40-μm coronal sections through the MS (corresponding to plates 14–21 of Paxinos and Watson (1986) separated from vertical and horizontal limbs of the diagonal band (vDB and hDB) by Mesulam et al. (1983)) were cut and stained for ChAT immunoreactivity (IR) as previously described (Johnson et al., 2002). Briefly, sections were placed in a solution containing primary antibody against ChAT (goat anti-ChAT 1:3500, EMD Millipore AB144P) for three days at 4 °C. Sections were then rinsed with PBS and incubated with a biotinylated secondary antibody (horse anti-goat 1:220, Vector Laboratories, Inc.) for 1 h at room temperature. Sections were then placed in an avidin/biotinyl-peroxidase solution (ABC Elite kit, Vector Laboratories, Inc.) for 1 h and then stained with a solution of 3–3′-diaminobenzidine, H₂O₂, and NiCl₂. Sections were then rinsed with PBS, mounted onto glass slides, dehydrated, cover slipped and examined with a Leitz photomicroscope (Leica, Inc.). ChAT-immunopositive cells were then evaluated in 5–6 sections spanning the anterior septum corresponding to plates 15–17 of the rat brain Atlas (Paxinos and Watson, 1986). MS was distinguished from the vertical limb of the diagonal band of Broca by drawing a line connected the anterior commissures. The lateral ventricles and corpus callosum defined the lateral and dorsal boundaries of the septum.

qRT-PCR

mRNA isolation and reverse transcription.

Detection and relative quantification of ARO and ER mRNAs was conducted as recently described (Li et al., 2016). Tissue samples were sonicated in 1.0 ml Trizol (Invitrogen, Inc.) at 4 °C and extracted with 250 μl chloroform. The organic phase was collected and isopropyl alcohol was added to precipitate the nucleic acids. DNA residues were digested with DNAse. The remaining RNA was re-extracted with phenol–chloroform and precipitated with sodium acetate and ETOH. mRNA was then reverse transcribed to cDNA using the High-Capacity cDNA Reverse Transcription kit (Applied Biosystem, Inc.) as per manufacturer’s recommendations.

qRT-PCR method.

2.0 μl of cDNA, 10 μl of SYBR Green, and 1.2 μl primer pair was pipetted into each well of a 96-well plate (0.1 ml/well). As mentioned above, two isoforms of CYP19A1 mRNA have been discovered in the rat brain. Therefore, in this study, both the total (ARO_T) and long form (ARO_L) CYP19A1 genes were evaluated. The primer pairs used for detection of total CYP19A1 were: (F) 5′-CGTCATGTTGCTTCTCATCG-3′ and (R) 5′-TACCGCAGGCTCTCGTTAAT-3′. The primer pairs used for long-form CYP19A1 were: (F) 5′-CTCCTCCTGATTCGGAATTGT-3′ and (R) 5′-TCTGCCATGGGAAATGAGAG-3′. The primer pairs for ERs were: ER α: (F) 5′-TCCGGCACATGAGTAACAAA-3′ and (F) 5″-TGAAGACGATGAGCATCCAG-3′. ERβ: (F) 5′-AAAGTAGCCGGAAGCTGACA-3′; (R) 5′-ACTGCTGCTGGGAGGAGATA-3′ GPR30 (F) 5′-AGGAGGCCTGCTTCTGCTTT-3′; (R) 5′-ATAGCACAGGCCGATGATGG-3′. GAPDH was used as the control gene: (F) 5′-TGCCACTCAGAAGACTGTGG-3′ and (R) 5′-GGATGCAGGGATGATGTTCT-3′. PCR was conducted using the 7500 Sequence Detection System (ABI) and results were obtained as Ct (threshold cycle number) values analyzed using Sequence Detection System (SDS) software (ABI, Inc.). The sample and reference data points were normalized to the geometric mean of GAPDH (ΔCt). Relative gene expression was normalized to the control group using the 2^−ΔΔCt method, where ΔΔCT = ΔCt (treatment) −ΔCt (control) [37]. For gene expression of total and long form ARO in the ChEI treatments experiment, relative ARO expression was normalized to the amygdala control group to highlight regional differences in the levels of ARO_T and ARO_L.

Microsomal incubation assay

Microsomal extraction and incubation.

The assay to detect ARO activity in the brain tissue has been previously reported by our laboratory (Li et al., 2016). It was necessary to pool tissues in order to have sufficient activity for detection. For this experiment, each hippocampal data point represents tissue from two rats, and each amygdala or preoptic area data point from three rats. Samples were homogenized and centrifuged at 20,000g for 33 min at 4 °C. The supernatant was collected and centrifuged at 140,000g for 1 h to obtain the microsomal pellet. Microsomal pellets were then dissolved in 200 μl Tris buffer. Protein levels were determined by Bio-Rad protein assay (Bio-Rad Laboratories, Inc). To measure ARO activity, 100–200 μg microsomes were added to the microsomal incubation buffer. Testosterone (Sigma–Aldrich, Inc.) was added to each sample to a final concentration of 400 nM. To start the reaction, 50 μl of 0.02 M nicotinamide adenine dinucleotide phosphate (NADPH, Sigma–Aldrich, Inc.) was added to each tube, vortexed for 5 s and then placed at 37 °C. After 30 min, the reaction was stopped by rapid cooling on wet ice. The total volume of the reaction was 1.0 ml.

Estradiol extraction and derivatization.

Samples were spiked with internal standard 25 μl d5–17-beta-estradiol (E2) and extracted with 3 ml n-Butyl chloride. Samples were then centrifuged and the organic layer transferred to silanized culture tubes and dried down under nitrogen. Residues were derivatized with dansyl chloride and transferred to glass vials for UPLC–MS/MS analysis.

Estradiol detection.

E2 was eluted using a Waters Acquity UPLC BEH C18, 1.7 μm, 2.1 × 150 mm reversed-phase column, with an acetonitrile: water (0.1% formic acid) gradient. MS detection and quantification were achieved in the positive mode. Transitions used for analysis were 506 → 171 for E2, and 511 → 171 for the deuterated internal standard. Inter-day and intra-day precision and accuracy of this assay have been described (Li et al., 2016). The limit of detectability for this assay is 2.5 pg/ml.

Aromatase activity calculation.

An E2 standard curve with concentrations ranging from 2.5 pg/ml to 200 pg/ml was prepared in a matrix of 0.2% 2-hydroxypropyl-β-cyclo dextrin (Aldrich, USA), extracted and derivatized at the same time as the microsomal samples. The concentration of estradiol pg/ml in the unknowns was determined by measuring area under the peak, then interpolating from the standard curve and adjusting for differences in volume of the unknowns vs. the standards. ARO activity is reported as pmol estradiol/h/mg microsome. Negative controls included samples that received no microsome or no NADPH.

Statistical analyses.

In experiment 1, differences between the 192IgG-SAP-treated and the control group were analyzed by t-test. In experiment 2, effects of ChEI treatments were analyzed by a one-way ANOVA followed by Tukey’s post-hoc test. All statistical analyses were done using JMP (Pro12), with significance defined as p < 0.05.

RESULTS

Effects of septal cholinergic lesions

Verification of the lesions.

Lesions were evaluated by confirming loss of ChAT-positive cells in the septum, and by loss of ChAT activity in the hippocampus. Immunostaining confirmed that septal infusions of 192IgG-SAP eliminated most of the ChAT-IR cells in the MS (Fig. 1A, B). ChAT activity in the hippocampus also was significantly decreased (>80%) in 192IgG-SAP-treated rats relative to controls (t₉ = 6.15, P < 0.001; Fig. 1C). In contrast, ChAT activity in the frontal cortex was not significantly affected (t₉ = 0.52, P = 0.62). This demonstrates a loss of cholinergic input to the hippocampus, but not the frontal cortex, consistent with previous reports (Gibbs and Johnson, 2007; Gibbs et al., 2011a).

Fig. 1.

Effect of 192IgG- SAP-induced cholinergic lesions on ChAT immunoreactivity by immunohistochemistry (IHC) in the medial septum of rats treated with (A) Saline and (B) 192IgG-SAP as well as on (C) ChAT activity detected in the hippocampus and frontal cortex. Two weeks following intracerebroventricular injection of 0.4 μg 192IgG-SAP treatment, the number of positive ChAT-stained cholinergic neurons was significantly decreased in the medial septum (Panel B) relative to controls (Panel A). Panel C confirms a significant decrease in ChAT activity in the hippocampus, but no significant change in ChAT activity in frontal cortex. In panels A and B, rostral is toward the top and midline extends top to bottom through the middle of the panel. Scale bar=0.1 mm. *Indicates the p < 0.05 compared to control group. N = 23 for each group.

Effects of septal cholinergic lesions on ARO and ER mRNA levels and on ARO activity

Effects of cholinergic lesions on relative levels of ARO mRNA are summarized in Fig. 2A, B. Levels of ARO_T mRNA were slightly higher (27.0%) and levels of ARO_L were slightly lower (18.1%) in the hippocampus of 192IgG-SAP-treated rats vs. controls; however, these differences were not statistically significant. Likewise, there was no significant difference in ARO activity detected in the hippocampus of 192IgG-SAP-treated rats vs. controls (Fig. 2(C)). The frontal cortex, which was included as a negative control, also showed no significant effects of treatment on relative levels of ARO mRNAs (data not shown). As previously reported, levels of ARO mRNA were extremely low in the frontal cortex and ARO activity was not detected in this region (Li et al., 2016).

Fig. 2.

Effect of cholinergic lesion in the medial septum on (A) ARO_T mRNA, (B) ARO_L mRNA expression and (C) ARO activity in the hippocampus. No significant change was found in region-specific ARO mRNA and activity. Bars in (A) and (B) indicate the mean ratio of ARO mRNA relative to OVX controls ± s.e.m., after normalizing to GAPDH. Bars in (C) indicate the estradiol production (pmol/h.mg microsome) ± s.e.m, which represents ARO activity. In the study of ARO mRNA, N = 10 for each group. For ARO activity study, N = 13 for each group.

Relative levels of ER mRNAs detected in the hippocampus are summarized in Fig. 3. No significant effects on the relative levels of ERα, ERß, or GPR30 were detected in 192IgG-SAP-treated rats vs. controls.

Fig. 3.

Effect of cholinergic lesions on ER alpha, ER beta and GPR30 expression in response to saline or 192IgG-SAP infusions. Bars represent ratio of ER mRNA levels that were normalized to control GAPDH mRNA ± s.e.m. N = 10 for each group.

Collectively these findings indicate no significant effects of selective cholinergic denervation on ARO or ER expression in the hippocampus.

Effect of ChEIs on ARO and ER mRNAs and on ARO activity

Effects of ChEIs on ARO mRNA and ARO activity are summarized in Fig. 4. As previously reported, highest expression of both ARO_T and ARO_L was detected in the amygdala, followed by preoptic area, then hippocampus (Fig. 4A, B) (Li et al., 2016). Relative levels of ARO activity mirrored the relative levels of ARO_L mRNA expression in these regions. Levels of ARO_T and ARO_L mRNA in the frontal cortex were extremely low and no ARO activity was detected in this region.

Fig. 4.

Effect of donepezil (3 mg/kg, daily injection for 7 days) and galantamine (5 mg/kg/day for 7 days) on (A) ARO_T mRNA, (B) ARO_L mRNA expression and (C) ARO activity in hippocampus, frontal cortex, amygdala and preoptic regions. Bars in (A) and (B) indicate the mean ratio of total, long form ARO mRNA relative to Amygdala saline treatment controls ± s.e.m. after normalizing to GAPDH. Bars in (C) indicate the mean ARO activity ± s.e.m. *Indicates the p ≤ 0.05 compared to control group. In the study of ARO mRNA, N = 4 for each group. For ARO activity study, N = 12 for each group.

No significant effects of donepezil or galantamine treatments were observed on relative levels of ARO mRNA in any of the brain regions examined. In the amygdala, mRNAs coding for both total and long form ARO showed slight increases in expression in galantamine (~30 ± 13%) treated rats, but this increase was not statistically significant. Slightly larger increases in ARO_T mRNA were observed in the POA following treatment with galantamine (93 ± 17%), and in ARO_L following treatment with donepezil (55 ± 6%). These changes also were not statistically significant although the effect on ARO_L in the POA following donepezil treatment was close to significance (p = 0.07).

With respect to ARO activity, no significant effects of ChEI treatments were detected in the hippocampus, preoptic area or frontal cortex (Fig. 4C). In contrast, ARO activity in the amygdala was 45 ± 0.7% greater in rats treated with galantamine, relative to controls (p < 0.01). Effects of ChEI treatments on levels of ERα, ERß, or GPR30 mRNAs were also detected in the four brain regions (Fig. 5). Although we saw an increase in ERα mRNA (56 ± 9%, p = 0.41) with donepezil treatment and a decrease in GPR30 mRNA (50 ± 14%, p = 0.33) with galantamine treatment in the HPC, neither of these reached statistical significance. Similarly, no significant effects of ChEI treatment on ERα, ERß, or GPR30 mRNAs were detected in any other brain regions.

Fig. 5.

Effect of donepezil and galantamine on ER mRNA expression in (A) hippocampus, (B) frontal cortex, (C) amygdala and (D) preoptic area. Bars indicate the mean ratio of estrogen receptor mRNA relative to saline treatment ± s.e.m. after normalizing to GAPDH. N = 4 for each group.

DISCUSSION

Cholinergic lesions

Our goal was to explore the role of cholinergic projections in regulating ARO mRNA, ARO activity, and the expression of ER mRNAs in specific regions of the rat brain. In the first experiment, we tested the effect of selectively removing cholinergic afferents to the hippocampus. Results show that removal of over 80% of the cholinergic afferents had no significant effect on relative levels of ARO_T, ARO_L mRNAs or ARO activity. This indicates that cholinergic inputs to the hippocampus do not play a significant role in regulating ARO and therefore are not likely to affect local estrogen production in this region. Likewise, loss of cholinergic inputs had no significant effect on relative levels of ERα, ERß, or GPR30 mRNAs in the hippocampus. This indicates that cholinergic inputs also do not play a significant role in regulating ER expression in this brain region. Collectively, these data suggest that cholinergic influences on estrogen effects in the hippocampus (e.g., spine density, NMDAR expression, synaptic plasticity) are not mediated by effects on either local estrogen production or by effects on ER expression.

The frontal cortex, which was used as a negative control, also showed no change in ARO or ER expression. As previously reported (Li et al., 2016), levels of ARO_L in the frontal cortex were very low which is consistent with the inability to detect ARO activity in this region. Our findings show no induction of ARO activity in the frontal cortex as a result of the surgery. In addition, there was not a change in the levels of ARO or ER mRNAs in response to cholinergic denervation of the hippocampus. This shows that the surgical procedures associated with the intraseptal injections of 192IgG-SAP or saline were not sufficient to significantly alter ARO or ER expression in the frontal cortex.

The fact that relative levels of ARO mRNA and activity in the hippocampus were not affected by the cholinergic lesions was unexpected. These lesions are robust, and our analysis of ChAT immunostaining and ChAT activity confirm that >80% of the cholinergic inputs to the hippocampus had been lost. Other studies have shown that ARO expression and activity increases in the brain in response to other types of injury such as trauma and stroke (Carswell et al., 2005; Peterson et al., 2007; Duncan and Saldanha, 2011; Pietranera et al., 2011), primarily due to induction of ARO in reactive astrocytes. 192IgG-SAP lesions have been shown to induce a strong activation of microglia and a moderate astrocytic reaction in the hippocampus (Rossner et al., 1995a,b), but in this case no significant effect on ARO was detected. This suggests that elevated ARO is not a universal response to brain injury, but rather is injury-, and perhaps brain region-specific. It may be that only lesions that produce a strong astrocytic reaction result in upregulation of ARO. It is also possible that the cholinergic lesions resulted in a reduction in ARO in neurons and an increase in ARO in astrocytes, resulting in no net change; however, this has not been observed in response to other lesions. Whether reactive astrocytes that are present in the hippocampus following cholinergic lesions express ARO was not investigated and will need to be explored.

Cholinesterase inhibitors

In the second experiment, we tested the effect of two ChEIs, donepezil and galantamine, on ARO expression and activity in four regions of the brain. Donepezil is a piperidine-based mixed, non-competitive reversible inhibitor of acetylcholinesterase. It has an in vitro IC50 of approximately 6.7 nM and an in vivo ID50 of approximately 2.6 mg (6.8 μMol)/kg brain tissue (Sugimoto et al., 2002). Galantamine is a less potent ChEI than donepezil (Bores et al., 1996), but unlike donepezil has been shown to act as an allosteric enhancer at nicotinic acetylcholine receptors (Samochocki et al., 2003; Schilstrom et al., 2007). Doses in the range of 1.5–5.0 mg/kg in rats have been reported to produce optimal brain concentrations for the allosteric potentiating ligand effect of galantamine (Geerts et al., 2005). Previously we showed that daily injections of these ChEIs at the same doses used in the current study were able to enhance estrogen effects on acquisition of a delay matching-to-position T-maze task in aged rats (Gibbs et al., 2009, 2011a), and in young adult rats with partial cholinergic lesions (Gibbs et al., 2011a).

In the current study, daily injections of donepezil or galantamine had no significant effects on ARO expression or activity in the hippocampus, frontal cortex, or POA of OVX rats. This suggests that the systemic up-regulation of cholinergic activity does not affect local estrogen production in these regions. Likewise, ChEI treatment had no significant effect on relative levels of ER mRNA in any of the regions examined. We conclude, therefore, that any ability of cholinergic inputs to alter estrogen effects in these regions is not mediated by effects on local estrogen production or ER expression.

One exception to these results was the amygdala where ARO activity was significantly increased in rats treated with galantamine, but not donepezil, despite the fact that relative levels of ARO mRNA were not significantly affected. The amygdala is a central structure in limbic circuitry, which plays an important role in regulating emotional expression, emotional experience, emotional memory, and fear (Gallagher and Chiba, 1996; Pessoa, 2010). It receives cholinergic afferents from basal forebrain cholinergic neurons. These inputs play an important role in regulating plasticity and coordinating different memory systems leading to the selection of appropriate behavioral strategies in conditioned-fear behaviors (Nagai et al., 1982; Calandreau et al., 2006; Jiang et al., 2016). Both muscarinic and nicotinic acetylcholine receptors are located in the amygdala (Muller et al., 2013; Pidoplichko et al., 2013). Human and animal studies show that exposure to nicotine can increase the rate of depression and anxiety in adolescent females (Biegon et al., 2012). ARO also is highly expressed in this region and studies have shown that ARO in the amygdala can influence aggression and mood in both males and females (Unger et al., 2015; Li et al., 2016). ARO also has been shown to influence synaptic plasticity in the basolateral amygdala in females, but not in males (Bender et al., 2017). Moreover, studies by Balthazart and co-workers have shown that ARO activity can be regulated by phosphorylation in response to changes in intracellular calcium (Balthazart et al., 2001, 2005). Hence the allosteric potentiation of nicotinic receptors by galantamine might account for the increase in ARO activity detected in the amygdala, in the absence of any effect on the levels of ARO mRNA. Further studies using different doses of galantamine or selective nicotinic receptors agonists and antagonist would help to clarify the effects of galantamine and the role of nicotinic receptors in regulating ARO activity in the amygdala and its role in modulating behaviors. This may be particularly relevant in elderly patients who are treated for memory impairment and who also struggle with fear and anxiety disorders.

Collectively, these results show that increasing cholinergic activity by treating with ChEIs has no significant effect on ARO or ER expression in several brain regions, including the hippocampus where cholinergic inputs have been shown to influence estrogen effects on synaptic plasticity and cognitive performance. One exception was the amygdala where galantamine increased ARO activity, possibly affecting synaptic plasticity in the basolateral amygdala with corresponding effects on fear-related behaviors.

Relevance to cognitive decline associated with aging and neurodegenerative disease

This study is the first to systematically evaluate the effects of hippocampal cholinergic denervation, and daily treatment with cholinesterase inhibitors (ChEIs), on ARO and ERs in different regions of the brain. Though much of the results are negative, these findings are important given the evidence for decreased central cholinergic function with aging and in neurodegenerative diseases like Alzheimer’s disease and Parkinson’s disease (Mufson et al., 2008; Bohnen and Albin, 2011; Ferreira-Vieira et al., 2016), as well as the frequent use of ChEIs in elderly individuals with memory impairment. Studies have shown a loss of ER and ER function with age (Foster, 2012), possibly due to methylation of the receptors and degradation by ubiquitin-proteasomal pathway (Tschugguel et al., 2003; Pinzone et al., 2004). There also is evidence for an association between ARO expression/polymorphism and risk for AD (Hiltunen et al., 2006). Decreased ARO expression has been reported in the hippocampus of AD patients (Ishunina et al., 2007). Iivonen reported that a single SNP in the ARO gene was associated with a 60% increase in AD risk (Iivonen et al., 2004). Baker reported that ApoE4 carriers with SNPs in the ARO gene had a two-fold increased risk of AD (Baker et al., 2006). Given the evidence that ARO and ERs play a role in cognitive decline associated with aging and AD, and given the frequent use of ChEIs in treating elderly patients with memory impairment, it is significant that we report relatively little acute effects of cholinergic manipulations on ARO or ER expression in several brain regions including the hippocampus.

CONCLUSIONS

Collectively, we showed that manipulating cholinergic system by selectively destroying cholinergic projections to the hippocampus, or by treating with cholinesterase inhibitors, had little effect on ARO and ER expression in many regions of the rat brain. Galantamine increased ARO activity in the amygdala, possibly due to allosteric potentiation of nicotinic acetylcholine receptors. This raises the possibility that increasing cholinergic activity may increase local estrogen production in the amygdala and thereby affect amygdala function. To our knowledge, this is the first report of an effect of ChEI treatment on ARO activity in the brain. The clinical significance will require further examination.

Abbreviations:

ACh

acetylcholine

ARO

aromatase

CamKII

calcium/calmodulin-dependent protein kinases

ChAT

choline acetyltransferase

ChEIs

cholinesterase inhibitors

CREB

cAMP response element-binding proteins

DMP

delayed-matching-to-position

ERs

estrogen receptors

IHC

immunohistochemistry

IR

immunoreactivity

MAPK

mitogen-activated protein kinases

MS

medial septum

NBM

nucleus basalis magnocellularis

POA

preoptic area