|Home | About | Journals | Submit | Contact Us | Français|
We previously demonstrated that dorsal hippocampal extracellular signal-regulated kinase (ERK) activation is necessary for 17β-estradiol (E2) to enhance novel object recognition in young ovariectomized mice (Fernandez et al., 2008). Here, we asked whether E2 has similar memory-enhancing effects in middle-aged and aged ovariectomized mice, and whether these effects depend on ERK and phosphatidylinositol 3-kinase (PI3K)/Akt activation. We first demonstrated that intracerebroventricular (ICV) E2 or intrahippocampal (IH) E2 infusion immediately after object recognition training enhanced memory consolidation in middle-aged, but not aged, females. The E2-induced enhancement in middle-aged females was blocked by IH inhibition of ERK or PI3K activation. IH or ICV E2 infusion in middle-aged females increased phosphorylation of p42 ERK in the dorsal hippocampus 15, but not 5, min after infusion, an effect that was blocked by IH inhibition of ERK or PI3K activation. Dorsal hippocampal PI3K and Akt phosphorylation was increased 5 min after IH or ICV E2 infusion in middle-aged, but not aged, females. ICV E2 infusion also increased PI3K phosphorylation after 15 min, and this effect was blocked by IH PI3K, but not ERK, inhibition. These data demonstrate for the first time that activation of dorsal hippocampal PI3K/Akt and ERK signaling pathways is necessary for E2 to enhance object recognition memory in middle-aged females. They also reveal that similar dorsal hippocampal signaling pathways mediate E2-induced object recognition memory enhancement in young and middle-aged females, and that the inability of E2 to activate these pathways may underlie its failure to enhance object recognition in aged females.
Although recent clinical trials report that hormone therapy increases risks of global cognitive decline and dementia in post-menopausal women (Rapp et al., 2003; Shumaker et al., 2004), other studies demonstrate that estrogen therapy can reduce risks of cognitive decline and dementia in younger, recently menopausal, women (Henderson et al., 2005). These findings support the “critical period hypothesis” which posits that estrogen therapy preserves cognition only when initiated near menopause onset (Sherwin and Henry, 2008). Although compelling, the neurobiological mechanisms underlying this critical period have not been clearly elucidated in women or animal models. Because current estrogen therapies are associated with negative health outcomes (Rossouw et al., 2002), understanding the neural mechanisms through which estrogens protect memory in menopausal females could be instrumental to developing safer and more effective estrogen treatments for reducing age-related cognitive decline.
To this end, we recently demonstrated that activation of the extracellular signal-regulated kinase/mitogen-activated protein kinase (ERK/MAPK) signaling cascade in the dorsal hippocampus is necessary for the potent estrogen, 17β-estradiol (E2), to enhance novel object recognition in young ovariectomized mice (Fernandez et al., 2008). In this study, a single intraperitoneal injection of E2 given immediately after training significantly enhanced object recognition tested 48 hrs later and increased phosphorylation of the p42 isoform of ERK in the dorsal hippocampus 1 hr after a subsequent injection (Fernandez et al., 2008). These effects were blocked by inhibiting MAPK kinase (MEK), the exclusive upstream activator of ERK (Fernandez et al., 2008). Post-training intraperitoneal E2 injection also enhances 48-hr object recognition in middle-aged, but not aged, ovariectomized mice (Gresack et al., 2007a, b), raising the possibility that age-related alterations in E2's ability to activate cell signaling underlie its differential effects on object memory in aging females. However, it is unknown if E2 modulates object memory via similar signaling mechanisms in the aging female brain, given the myriad of age-related alterations that affect the hippocampus. Indeed, E2 protects against the ovariectomy-induced decrease in phosphorylated p42 ERK levels in young, but not aged, female rats (Bi et al., 2003). Nevertheless, whether E2 increases hippocampal ERK activation in middle-aged rodents has not been examined, so it is unclear whether ERK activation mediates the beneficial effects of E2 on object recognition at this age.
Further, ERK can be activated by other signaling cascades that are critical for hippocampal-dependent memory formation, including phosphatidylinositol 3-kinase/Akt (PI3K/Akt) (Chen et al., 2005). PI3K is crucial for hippocampal plasticity and object recognition (Kelly and Lynch, 2000; Lin et al., 2001; Horwood et al., 2006), is rapidly activated by E2 (Yokomaku et al., 2003; Manella and Brinton, 2006), and is vital for neuroprotection (Singh, 2001), which could be critically important for the aging brain. As such, hippocampal PI3K/Akt signaling may play a pivotal role in E2-induced object recognition enhancement in aging females, perhaps by activating p42 ERK.
The present study determined whether dorsal hippocampal ERK and PI3K activation are necessary for the beneficial effects of E2 on object recognition in aging females. The data provide the first evidence that dorsal hippocampal PI3K and ERK activation is necessary for E2 to enhance object memory consolidation in middle-aged female mice, and that the inability of E2 to activate these enzymes in aged females may underlie its failure to enhance object recognition in advanced aging.
Young (3 months old), middle-aged (17 months old), and aged (21 months old) female C57BL/6 mice were obtained from the National Institutes on Aging colony at Charles River Laboratories. Mice were housed 4-5 per cage until surgery and were handled briefly prior to use. Mice were maintained on a 12 h light/dark cycle and had ad libitum access to food and water. All procedures followed the National Institutes of Health Guide for the Care and use of Laboratory Animals and were approved by Yale University Animal Care and Use Committee.
Mice were ovariectomized and implanted with intracranial guide cannulae in the same surgical session as described previously (Fernandez et al., 2008; Lewis et al., 2008). All mice were implanted with stainless-steel guide cannulae (Plastics One, Roanoke, VA) aimed at the dorsal hippocampus (bilaterally), dorsal third ventricle, or both brain regions. Mice were anesthetized with isoflurane gas (5% for induction, 2% for maintenance). Using a stereotaxic apparatus (Kopf Instruments, Tujunga, CA), guide cannulae (C232GC, 26 gauge, Plastics One) with inserted dummy cannulae (C232DC) were directed toward the dorsal hippocampus (-1.7 mm posterior to Bregma, ±1.5 mm lateral to midline, -2.3 mm (injection site) ventral to skull surface), dorsal third ventricle (-0.5 mm posterior to Bregma, ± 0.0 lateral to the midline, -3.0 (injection site) ventral to the skull surface), or both the hippocampus and dorsal third ventricle (triple guide; same coordinates as above for both regions) (Paxinos and Franklin, 2003). Each cannula was fixed to the skull with dental cement that also served to close the wound. Mice recovered 5-7 days before testing or drug treatment.
Cyclodextrin-encapsulated E2 (Sigma-Aldrich, St. Louis, MO) was dissolved in physiological saline to a dose of 5.0 μg/0.5 μl and infused at 0.5 μl/min for 1 min per side of the dorsal hippocampus. This dose in young ovariectomized mice infused into the dorsal hippocampus facilitates object memory consolidation (Fernandez et al., 2008). Intracerebroventricular (ICV) infusions were conducted at the same speed for 2 min. The vehicle, 2-hydroxypropyl-β-cyclodextrin (HBC) (Sigma-Aldrich), was dissolved in saline to the same concentration of cyclodextrin as in the cyclodextrin-E2 solution.
To demonstrate that E2-enhanced object recognition consolidation was dependent on dorsal hippocampal ERK activation, the MEK inhibitor 1,4-diamino-2,3-dicyano-1,4-bis (o-aminophenylmercapto) butadiene (U0126) (Promega, Madison, WI), at a concentration of 0.5 μg/0.5 μl in 50% dimethyl sulfoxide (DMSO) in saline, was infused at a rate of 0.5 μl/min for 1 min per side of dorsal hippocampus, alone or concurrently with ICV E2 infusion. In young ovariectomized mice, intrahippopcampal (IH) infusion of this dose of U0126 does not block object recognition tested 24 hrs after training (Fernandez et al., 2008). The vehicle control for U0126 was 50% DMSO in saline.
To demonstrate that E2-enhanced object recognition consolidation was dependent on dorsal hippocampal PI3K activation, the PI3K inhibitor LY294002, which inhibits phosphorylation of the p85 regulatory subunit of PI3K (Geltz and Augustine, 1998). LY204002 (Cell Signaling Technology, Danvers, MA) was dissolved in 100% DMSO to 3 μg/μl as a stock solution and then serially diluted in physiological saline for infusion of various doses. We first wanted to determine a dose of LY294002 that did not impair object recognition tested 24 hrs after training. Previous studies in adult rats have shown that post-training infusions of LY294002 at very low doses (equal to or less than 0.75 ng/side of hippocampus) do not cause retrograde amnesia in a step-down inhibitory avoidance task (Barros et al., 2001). Therefore, we infused very low doses in our 24-hr pilot study; all doses (0.5, 0.005, and 0.0005 μg/side of the hippocampus) of LY294002 were infused at a rate of 0.5 μl/min and a volume of 0.5 μl/side. Only a dose that did not impair 24-hr object recognition was used in subsequent studies with E2. The vehicle control for LY294002 was 50% DMSO in saline.
During drug infusions, mice were gently restrained. Dummy cannulae were replaced with injection cannulae (C232I; IH: 26 gauge, extending 0.8 mm beyond the 1.5 mm guide, ICV: 28 gauge, extending 1.0 mm beyond the 2.0 mm guide) attached to polyethylene tubing (PE50) connected to a 10 μl Hamilton syringe. Infusions were controlled by a microinfusion pump (KDS 100, KD Scientific; New Hope, PA). Infusion cannulae remained in place for 1 min after infusion to prevent diffusion back up the cannula track. For behavioral testing, all mice were infused immediately after object recognition training.
The object recognition task was conducted as described previously (Frick and Gresack, 2003; Fernandez and Frick, 2004), and was used to assess non-spatial hippocampal-dependent memory (Clark et al., 2000; Baker and Kim, 2002). The task consisted of habituation, training, and testing phases conducted on separate days. During habituation, mice were allowed to freely explore an empty white box (60 cm W × 60 cm L × 47 cm H) for 5 min. No data were recorded. Twenty-four hrs later, mice were rehabituated in the same box for 1 min, and then placed in a holding cage while two identical objects were placed in the left and right corners (~5 cm from the walls) of the box. Mice were then immediately placed back into the box and allowed to freely investigate until they accumulated a total of 30 s exploring the objects. Exploration was recorded when the front paws or nose contacted either object. Mice were then immediately infused and returned to their home cage (or in some cases, killed 15 min later). After 24 or 48 hrs, object recognition was tested, using the same procedure as in training except that a novel object was substituted for one of the familiar training objects. Novel object location was counterbalanced across mice. Time spent with each object was recorded. Mice inherently prefer to explore novel objects; thus, a preference for the novel object (more time than chance (15 s) with the novel object) indicates intact memory for the familiar object. The use of 30 s total exploration time rather than a fixed trial duration minimizes confounding influences of group differences in activity. Young ovariectomized mice remember the familiar object after 24, but not 48, hrs (Gresack et al., 2007a), allowing impairing effects of drugs to be observed using the 24-hr delay and enhancing effects of drugs to be observed using the 48-hr delay. Elapsed time to accumulate 30 s of exploration was recorded to control for group differences in activity levels, but none were observed in any experiment (Table 1).
Mice were decapitated 5 or 15 min after intracranial infusions, or 15 min after object recognition training, and the dorsal hippocampus was immediately dissected bilaterally on ice and frozen at -80°C until homogenized. Western blotting was conducted as described previously (Fernandez et al., 2008; Lewis et al., 2008). Briefly, all tissue samples were resuspended 1:50 w/v in lysis buffer and homogenized with a probe sonicator (Branson Sonifier 250, Danbury, CT). Homogenates were electrophoresed on 10% Tris-HCl gels and blotted to Immobilon-P polyvinylidene fluoride (PVDF) membranes (Millipore, Bedford, MA). Western blots were blocked and incubated with either anti-phospho-p44/42 MAPK (1:2000), anti-phospho-PI3K (1:1000), anti-phospho-Akt (1:1000), or total p44/42 MAPK (1:2000), total PI3K (1:1000), total Akt (1:1000) primary antibodies (Cell Signaling Technology, Beverly, MA) overnight. Blots were then incubated with the appropriate secondary antibody conjugated to horseradish peroxidase (Cell Signaling) and developed using West Dura chemiluminescent substrate (Pierce Laboratories, Rockford, IL). Blots were then stripped and re-probed with antibodies for total p44/42 MAPK, total PI3K, or total Akt antibodies for phospho-protein normalization or with anti-ß-actin antibody (1:5000; Sigma, St. Louis, MO) for total-protein normalization. Densitometry was conducted using Kodak 1D 3.6 software (Kodak Scientific Imaging Systems, New Haven, CT) on the Kodak Image Station 440 CF (Eastman Kodak, Rochester, NY).
In middle-aged mice, one side of dorsal hippocampus was infused with vehicle and the other side was infused with E2. Fifteen minutes later, the mice were anesthetized with sodium pentobarbital (50 mg/kg, i.p.) and perfused intracardially with 0.9% saline, followed by 4% paraformaldehyde in 0.1 M phosphate buffer (PB). Brains were post-fixed and cryoprotected, frozen, and stored at -80°C. Before sectioning, a notch was placed along the left cortex to maintain laterality throughout the analyses. Forty micrometer frozen, serial, coronal sections were cut throughout the brain, and stored in cryoprotectant for immunocytochemistry on free-floating sections. Every sixth section was collected and immunostained with phospho-p44/42 MAPK antibody as in (Tsuda et al., 2008) using Vectastain ABC Kit (Vector Labs, Burlingame, CA). Briefly, sections were blocked in 3% normal goat serum and incubated in rabbit anti-phospho-p44/42 MAPK antibody (Cell Signaling; 1:1,000) overnight at room temperature. For control sections the primary antibody was omitted. Sections were then incubated in biotinylated goat anti-rabbit IgG (1:200) for one hour at room temperature, washed, and incubated in avidin-biotin solution (Vectastain ABC Kit, Vector Labs; 1:500) in tris-buffered saline (TBS) for one hour at room temperature. Sections were reacted with 3′, 3′-diaminobenzidine (DAB) solution (Sigma), washed, mounted on gelatin-coated slides, air-dried, dehydrated, cleared and cover-slipped.
For object recognition data, separate one-sample t tests were performed for each group to determine whether the time spent with the novel object differed from 15 s. This analysis was used because time spent with the objects is not independent; time spent with one object reduces time spent with the other object (Gresack and Frick, 2004, 2006). For Western blotting experiments comprising multiple groups, differences between vehicle and treatment groups were evaluated using one-way ANOVA followed by Fisher's least significant difference (LSD) post hoc tests. For Western blotting experiments with only two groups, separate two-tailed unpaired Student's t-tests were performed between the vehicle and treatment group. For experiments examining the effects of object recognition training on phosphorylated and total protein levels, two-way ANOVA was first used to examine the effects of Age and Training, followed by Fisher's LSD post hoc tests. Next, two-tailed unpaired Student's t-tests were performed between the training groups within each age to examine effects of training on protein levels within each age. Significance was determined at p < 0.05.
We previously demonstrated that a single post-training intraperitoneal injection of E2 enhances object recognition in middle-aged, but not aged, ovariectomized mice (Gresack et al., 2007a, b). One possible reason for the lack of effect in aged females is that too little of the systemically injected hormone reached the dorsal hippocampus to be behaviorally effective. Thus, we first sought to determine if direct intrahippocampal (IH) or intracerebroventricular (ICV) infusions of the same dose of post-training E2 that enhances object recognition in young ovariectomized mice (Fernandez et al., 2008) could also enhance object recognition in middle-aged (n = 8-16/group) and aged (n = 9/group) ovariectomized mice. Immediately post-training, mice were infused with vehicle or E2 into the dorsal hippocampus (5 μg/side) or dorsal third ventricle (10 μg total), and retention was tested 48 hrs later. In middle-aged females (Fig. 1), both IH and ICV E2 infusions significantly increased the time spent with the novel object relative to chance (15 s) (t(7) = 2.72, p < 0.05 and t(15) = 3.26, p < 0.01, respectively), suggesting that both IH and ICV E2 enhanced novel object recognition in middle-aged female mice. Vehicle-treated mice did not exhibit a preference for the novel object (t(9) = 0.13, p > 0.05). In contrast to the beneficial effects of E2 in middle-aged females, aged females were unaffected by E2. Neither the vehicle nor the E2-treated groups exhibited a preference for novel object 48 hrs after training (t(8) = 1.10-1.93, p > 0.05; Fig. 1), suggesting that doses of intracranially infused E2 that enhance object recognition in young and middle-aged females do not affect object memory consolidation in aged females.
To determine if the beneficial effects of E2 on object recognition in middle-aged females were dependent on dorsal hippocampal ERK activation, the MEK inhibitor U0126 was bilaterally infused into dorsal hippocampus concurrently with ICV E2 infusion. To first demonstrate that effects of MEK inhibition on E2-induced enhancement of 48-hr object recognition did not result from general prevention of memory formation in this task, we identified a dose of U0126 that did not interfere with object recognition at a shorter delay. Previously, we showed in young ovariectomized mice that immediate post-training infusion of U0126 into the dorsal hippocampus, at a dose of 0.5 μg/side, did not impair object recognition tested 24 hrs after training (Fernandez et al., 2008). Thus, we tested if this dose of U0126 interfered with 24-hr object recognition in middle-aged females. Mice were infused bilaterally into the dorsal hippocampus with vehicle (n = 10) or 0.5 μg U0126 (n = 8) immediately after training. Twenty-four hours later, both the vehicle and U0126 groups spent significantly more time with the novel object than chance (19.45 ± 1.27 sec for vehicle and 20.52 ± 0.6 sec for U0126; t(9) = 3.51, p < 0.01; t(7) = 9.23, p < 0.001, respectively). These data suggest that 0.5 μg U0126 does not prevent general memory formation in this task in middle-aged females. Therefore, this dose was used in all subsequent studies.
We then examined if IH infusion of 0.5 μg U0126 affected the E2-induced enhancement of object recognition by infusing U0126 bilaterally into the dorsal hippocampus concurrently with ICV E2 infusion. As in our previous work with a membrane-impermeable form of E2 (Fernandez et al., 2008), E2 was infused ICV in this study to provide E2 to the brain while preventing tissue damage from repeated infusions into the dorsal hippocampus. As shown in Fig. 2a, the ICV E2-induced enhancement in 48-hr object recognition was completely blocked by IH U0126 (t(9) = -0.27, p > 0.05 relative to vehicle), suggesting that E2-induced facilitation of object recognition in middle-aged female mice is dependent on dorsal hippocampal ERK activation.
The fact that U0126 blocked the effects of E2 on object recognition suggests that E2 increases ERK activation in the dorsal hippocampus of middle-aged females. Therefore, we next sought to determine whether and when that activation takes place. Previous work showed that p42/p44 ERK activation in the dorsal hippocampus is increased 5 min after ICV E2 infusion in young male rats (Kuroki et al., 2000). Further, we recently demonstrated that p42 ERK activation in the dorsal hippocampus is increased 5 min after dorsal hippocampal infusion of a membrane-impermeable form of E2 in young ovariectomized mice (Fernandez et al., 2008). As such, we examined whether IH or ICV E2 could increase ERK activation in middle-aged females 5 min after infusion. Middle-aged females received IH or ICV infusions of vehicle (n = 8) or E2 (IH, n = 7; ICV, n = 6), and the dorsal hippocampus was collected bilaterally 5 min later for measurement of phosphorylated and total p42 and p44 ERK levels using Western blotting. Neither IH nor ICV E2 infusion affected phosphorylation of p42 ERK (F(2,18) = 0.39, p > 0.05) or p44 ERK (F(2,18) = 0.77, p > 0.05; Fig. 2b). This result led us to hypothesize that the temporal dynamics of ERK activation may differ in the young and aging female brain. Evidence for altered ERK activation as a result of aging comes from studies in which p42/p44 ERK activation in the hippocampus and basal forebrain of aged male rats is diminished relative to young males in response to growth factor treatment (Gooney et al., 2004; Williams et al., 2006; Williams et al., 2007). Therefore, we next examined ERK activation at a later time point, 15 min after infusion. To visualize where in the dorsal hippocampus this activation might occur, we infused mice (n = 4) on one side of the dorsal hippocampus with vehicle and the other side with 5 μg E2, perfused them 15 min later, and stained coronal sections for phospho-p42/p44 ERK immunoreactivity. Immunoreactivity was increased primarily in the hilus of the dentate gyrus (DG) and in CA3. In the hilus (Fig. 3c and 3d), E2 treatment increased phospho-ERK immunostaining in cells and processes relative to vehicle; estrogen receptors α and β (ERα and ERβ) have been reported in interneurons, glia, dendrites, axons, and axon terminals of the hilus (Milner et al., 2001; Milner et al., 2005). In CA3 (Fig. 3e and 3f), increased immunolabeling was evident only in processes in the E2-treated side compared with vehicle; ERα and/or ERβ have been localized to dendrites, dendritic spines, axons, and axon terminals of CA3 (Milner et al., 2001; Milner et al., 2005). Given the observed increase in ERK phosphorylation evident at 15 min immunocytochemically, we then quantified this increase using Western blotting. In addition to mice infused with vehicle and E2 alone, a fourth group was added that received ICV E2 + IH U0126. The main effect of treatment was significant for p42 ERK phosphorylation (F(3,22) = 6.33, p < 0.01), but not p44 ERK phosphorylation (F(3,22) = 0.39; p > 0.05; Fig. 2c). Both IH E2 (n = 6) and ICV E2 (n = 6) infusions significantly increased dorsal hippocampal phosphorylated p42 ERK levels relative to vehicle (n = 7; p < 0.01 and p < 0.05, respectively; Fig. 2c). The ICV E2-induced increase in p42 ERK phosphorylation was completely blocked by IH infusion of U0126 (n = 7; p > 0.05 relative to vehicle; Fig. 2c).
The PI3K/Akt pathway, which can activate ERK (Chen et al., 2005), is also critical for hippocampal synaptic plasticity and memory formation, (Kelly and Lynch, 2000; Lin et al., 2001; Horwood et al., 2006), and can be activated in vitro within minutes of E2 treatment (Yokomaku et al., 2003; Manella and Brinton, 2006). As such, we thought that PI3K and Akt might also play a role in the E2-induced enhancement of object recognition. To examine this issue, we first measured the effects of IH and ICV infusions of E2 on dorsal hippocampal PI3K and Akt phosphorylation using Western blotting. Middle-aged females received infusion of vehicle (IH or ICV; n = 7), IH E2 (5 μg/side; n = 7), or ICV E2 (10 μg; n = 6), and the dorsal hippocampus was dissected bilaterally 5 min later. Levels of phosphorylated PI3K protein differed significantly among the groups (F(2,17) = 5.29, p < 0.05), such that both IH and ICV E2 significantly increased PI3K phosphorylation relative to vehicle (p < 0.05; Fig. 4a). Phosphorylated Akt protein levels also differed significantly among the groups (F(2,16) = 15.68, p < 0.001), and again, both IH and ICV E2 infusions significantly increased Akt phosphorylation relative to vehicle (p < 0.001 and p<0.05, respectively; Fig. 4b). To determine if the failure of E2 to enhance object recognition in aged females might be due to an inability of E2 to activate cell signaling pathways, aged females were infused with vehicle (n = 6) or E2 (n = 5) into the dorsal hippocampus, and phosphorylation of PI3K and Akt in the dorsal hippocampus was measured 5 min later. Unlike in middle-aged females, E2 in aged females did not increase phosphorylated PI3K (Fig. 4e) or Akt (Fig. 4f) levels relative to vehicle (t(9)=0.35, p > 0.05 and t(9)=0.27, p > 0.05).
We next wanted to determine if dorsal hippocampal PI3K activation was necessary for the beneficial effects of E2 on object recognition in middle-aged females, and therefore, we conducted several studies with LY294002, an inhibitor of PI3K activation. We first demonstrated that the effects of PI3K inhibition on E2-induced enhancement object recognition did not result from general prevention of object memory formation, by identifying doses of LY294002 that did not interfere with object recognition after a 24 hr delay. Immediately after training, middle-aged females received IH infusions of vehicle (n = 10) or one of 3 doses of LY294002: 0.5 μg/side (n = 9), 0.005 μg/side (n = 6), or 0.0005 μg/side (n = 8). Twenty-four hrs later, mice receiving vehicle or 0.0005 μg LY294002 exhibited a significant preference for the novel object relative to chance (t(9) = 3.51, p < 0.01 and t(7) = 2.38, p < 0.05, respectively), whereas mice receiving 0.5 μg LY294002 or 0.005 μg LY294002 did not (t(8) = 0.07, p > 0.05 and t(5) = -0.65, p > 0.05, respectively, Fig. 4c). Because these data suggest that 0.0005 μg LY294002 did not prevent general object memory formation, the 0.0005 μg dose was used for all subsequent studies.
We then infused mice IH with LY294002 (0.0005 μg/side) and ICV with E2 (10 μg; n = 10) immediately after training, and tested memory retention 48 hrs later. As shown in Fig. 4d, the ICV E2-induced memory enhancement was completely blocked by IH LY294002 (t(9) = 1.49, p > 0.05 relative to chance), suggesting that the beneficial effects of E2 on object recognition are dependent on dorsal hippocampal PI3K activation in middle-aged females.
Our data indicate that both ERK and PI3K activities are required for E2 to enhance object recognition in middle-aged females, and previous studies have shown that PI3K can activate ERK either directly or indirectly (Lin et al., 2001; Perkinton et al., 2002; Chang et al., 2003; Chen et al., 2005). Therefore, we hypothesized that E2 might first activate the PI3K/Akt pathway, which could then activate ERK and lead to memory facilitation. To this end, we examined the effect of PI3K inhibition on ERK activation in the dorsal hippocampus 15 min after ICV E2 infusion. Middle-aged mice received IH or ICV vehicle (n = 7), ICV 10 μg E2 (n = 6) or ICV 10 μg E2 plus IH 0.0005 μg LY294002 (n = 8), and the dorsal hippocampus was dissected bilaterally 15 min later. Phosphorylation of p42 ERK differed among the groups (F(2,18) = 5.07; p < 0.05), but p44 ERK phosphorylation (F(2,18) = 0.21; p > 0.05) did not (Fig. 5a). The ICV E2-induced increase in p42 ERK phosphorylation (p < 0.01 relative to vehicle) was completely blocked by IH infusions of LY294002, suggesting that PI3K activation is upstream from ERK.
To further demonstrate that E2-induced increases in PI3K or Akt are upstream from ERK activation, we next examined PI3K and Akt phosphorylation 15 min after infusion with vehicle (IH or ICV, n = 6), ICV 10 μg E2 (n = 6), ICV 10 μg E2 plus IH 0.5 μg/side U0126 (n = 7) or ICV 10 μg E2 plus IH 0.0005 μg/side LY294002 (n = 8). Phosphorylated PI3K levels differed significantly among the groups (F(3,26) = 3.00, p < 0.05; Fig. 5b); ICV E2 treatment significantly increased PI3K phosphorylation relative to vehicle (p < 0.05), and this effect was completely blocked by IH infusion of LY294002. However, the increase produced by ICV E2 was not blocked by U0126 (p > 0.05 relative to vehicle; Fig. 5b), suggesting that E2-induced PI3K activation is not dependent on ERK activation. Phosphorylated Akt levels did not significantly differ among the groups 15 min after infusion (data not shown), suggesting the E2-induced increase in Akt phosphorylation seen 5 min after infusion is only transient.
Age differences in the ability of E2 to enhance object recognition and activate the ERK/MAPK and PI3K/Akt signaling pathways may be due to age-related decline in the function of these signaling pathways or of the ability of object recognition training to stimulate these pathways, irrespective of E2 treatment. To demonstrate whether object recognition training differentially affects ERK, PI3K, and Akt levels across ages, we examined dorsal hippocampal phospho- and total-ERK, -PI3K, and -Akt levels 15 min after object recognition training in young, middle-aged, and aged ovariectomized mice (n = 7 or 8 per group). Mice were habituated to the object recognition apparatus and then 24 hrs later, were placed in the apparatus with no objects or with two identical objects. Mice with objects were allowed to accumulate 30 sec with the objects as in our standard training protocol. Control mice without objects were yoked to spend the same amount of time in the apparatus as mice that explored objects. Dorsal hippocampi were collected bilaterally 15 min after training based on data from Figs. 2c and and5b5b showing that E2 significantly increased ERK and PI3K phosphorylation in middle-aged females at this time point. The main effect of Age was significant for phosphorylated p42 ERK levels (F(2,40) = 3.978; p < 0.05; Fig. 6a), and posthoc tests indicated that levels were significantly reduced in aged females relative to young and middle-aged females (p < 0.05). However, the main effect of Training and Age × Training interaction were not significant, suggesting no effect at any age of object training on p42 ERK phosphorylation measured after 15 min. Neither the main effects of Age and Training, nor the Age × Training interactions, were significant for phosphorylated p44 ERK, PI3K, or Akt (Fig. 6b-6d), suggesting no effect of age or training on activation of these proteins 15 min after training. To determine if total protein levels were affected by age or training, we also examined total ERK, PI3K, and Akt protein levels normalized to ß-actin in the same mice. Two-way ANOVA showed no significant main effects of Age and Training, or interactions, for total-p42, -p44, -PI3K, or -Akt levels (Fig. 7a-7d). Additional two-tailed unpaired Student's t-tests performed between the training groups within each age (data normalized to the age-matched control group mean) also showed no significant differences in levels of all phosphorylated or total proteins, suggesting no effects within each age of training on protein levels measured after 15 min.
This report provides the first demonstration that intracranial E2 infusion enhances object memory consolidation in middle-aged female mice, and that this effect is dependent on both PI3K and ERK activation, with an early activation of PI3K/Akt observed 5 min after infusion and a later activation of ERK 15 min after infusion. Further, studies with PI3K and ERK inhibitors suggest that E2 first activates PI3K, which then activates ERK, illustrating a potential directionality of cell signaling that leads to enhanced object recognition. The involvement of PI3K signaling in estrogenic modulation of object recognition memory has not previously been demonstrated at any age, so these data also provide novel insights into the molecular mechanisms underlying these effects beyond those previously demonstrated in young females. Of importance to aging in particular, the data suggest that E2 modulates object recognition via similar signaling mechanisms in young and middle-aged females, and that these mechanisms remain functional in middle-aged, but not aged, females. Although we would expect similar cell signaling mechanisms to be involved in E2-induced enhancement of other types of hippocampal memory, e.g., spatial memory, where post-training E2 also enhances memory consolidation in females (Packard and Teather, 1997b, a; Gresack and Frick, 2006), this generalization has yet to be tested systematically at any age.
The fact that intrahippocampal and intracerebroventricular E2 infusion enhanced object recognition in middle-aged females is consistent with effects of systemic post-training E2 injections on object recognition at this age (Gresack et al., 2007a). These data are also consistent with findings from young female mice (Fernandez et al., 2008) that dorsal hippocampal ERK activation is necessary for E2 to enhance object recognition, as demonstrated by the fact that intrahippocampal U0126 blocked the beneficial effects of intracerebroventricular E2 on object recognition. Interestingly, however, the time at which E2 increased phosphorylated p42 ERK levels differed in the two ages, such that levels were increased 5 min after infusion in young females (Kuroki et al., 2000; Fernandez et al., 2008) and 15 min after infusion in middle-aged females. This delay in the temporal dynamics of E2-induced ERK activation in the dorsal hippocampus of middle-aged females may result from reduced estrogen receptor (ER) expression in the hippocampus. Hippocampal p42 ERK is activated in vitro by ERα and ERβ agonists (Zhao and Brinton, 2007) and in vivo by a form of E2 that binds to membrane-bound ERs (Fernandez et al., 2008). Hippocampal ERβ expression is decreased by 12 months of age in female rats (Yamaguchi-Shima and Yuri, 2007), and age-related alterations in the expression or function of any ER could lengthen the time necessary for E2 to increase p42 ERK phosphorylation. The delayed ERK activation in middle-aged females also could be due to age-related alterations in N-methyl-D-aspartate (NMDA) receptors, which, in young female mice, activate ERK and are involved in the E2-inducd enhancement of object recognition (Lewis et al., 2008). In aged female rats, E2 affects the synaptic distribution of NR2B subunits (Adams et al., 2004), which could influence downstream signaling activation. Interestingly, the magnitude of ERK activation induced by intracranial E2 in middle-aged females (about 50% over vehicle) is similar to that seen in young females (unpublished observations), suggesting that although the ERK response is delayed in middle-aged mice, E2 can produce a similar degree of p42 ERK phosphorylation in young and middle-aged females.
This study provides the first evidence that PI3K activation is necessary for E2 to enhance memory at any age. This conclusion is supported by the fact that the beneficial effects of E2 on object recognition were blocked by intrahippocampal infusion of the PI3K inhibitor LY294002, and is consistent with literature demonstrating that hippocampal PI3K is required for several types of memory (Barros et al., 2001; Chen et al., 2005). The PI3K/Akt pathway is rapidly activated by E2 in cultured hippocampal and neocortical neurons (Singh, 2001; Yokomaku et al., 2003; Manella and Brinton, 2006), however, this report is the first to our knowledge showing that E2 can increase phosphorylation of PI3K itself. Phosphorylation of Akt is typically PI3K-dependent and plays an important role in E2-induced neuroprotection (Wilson et al., 2002). E2 increases hippocampal Akt phosphorylation in vitro (Akama and McEwen, 2003; Yokomaku et al., 2003), and phosphorylated Akt is increased during the high-estradiol phase of the rodent estrous cycle (Znamensky et al., 2003; Spencer et al., 2008). Interestingly, Akt in the present study was only activated 5 min after E2 infusion, whereas PI3K was activated 5 and 15 min after infusion, suggesting that cell survival may not be the primary outcome of E2 treatment in our mouse model, compared to ischemic models where Akt phosphorylation can last 2 hrs after injury (Wilson et al., 2002). It should be noted that systemically injected progesterone can rapidly increase PI3K expression and phosphorylation of ERK and Akt in the hippocampus (Nilsen and Brinton, 2002, 2003; Guerra-Ariaza et al., 2009), and so examining the effects of intrahippocampal infusions of E2 plus progesterone on ERK, PI3K, and Akt activation should be addressed in future studies.
The fact that E2 activated PI3K and Akt earlier (5 min) than ERK (15 min), suggests that E2 activates PI3K and Akt upstream from ERK in middle-aged females. This conclusion is supported by the fact that intrahippocampal infusion of LY294002 blocked the E2-induced activation of both PI3K and ERK, whereas intrahippocampal infusion of U0126 blocked only ERK activation. Additional support comes from studies in embryonic neocortical cell culture, where an interaction between ERα and the p85 regulatory subunit of PI3K leads to activation of Akt and ERK within the same neurons (Manella and Brinton, 2006). Other studies in young adult rodents report that PI3K activates ERK directly or indirectly in the hippocampus, amygdala, and striatum (Lin et al., 2001; Perkinton et al., 2002; Chen et al., 2005), so a similar effect to that seen here would be expected in the hippocampus of young female mice.
Collectively, the present data suggest a model of the cell signaling mechanisms in the dorsal hippocampus underlying the beneficial effects of E2 on object recognition memory. E2, binding to membrane or nuclear ERs, activates PI3K, which then activates ERK directly or indirectly through Akt. ERK activation then leads to altered gene transcription and protein translation, consistent with previous work showing that ERK activation promotes both processes (Adams and Sweatt, 2002; Kelleher et al., 2004). Altered gene transcription and protein translation then lead to enhanced memory, possibly through multiple routes such as spinogenesis, neurogenesis, and increased growth factor levels (Woolley and McEwen, 1992; Miranda et al., 1999; Tanapat et al., 1999; Bimonte-Nelson et al., 2004; Fernandez and Frick, 2004). In support of this notion, both PI3K/Akt and ERK signaling have been shown to regulate hippocampal dendritic morphology through the mammalian target of rapamycin (mTOR) translational pathway (Kumar et al., 2005). Although other signaling pathways may also be involved, this model provides a framework for future investigations of E2-regulated memory modulation.
Another important finding from this study is that intracranial E2 infusion enhanced object recognition memory in middle-aged, but not aged, female mice, which is consistent with findings from other studies that have compared effects of E2 on hippocampal memory in middle-aged and aged female rodents (Savonenko and Markowska, 2003; Gresack et al., 2007a; Talboom et al., 2008). The present and previous studies support the “critical period hypothesis”, which maintains that E2 benefits cognition only in recently menopausal women (Maki, 2006; Sherwin and Henry, 2008). The decreased responsiveness of the aged female hippocampus to E2 may result from age-related reductions in hippocampal ER immunoreactivity and mRNA levels (Adams et al., 2002; Mehra et al., 2005; Yamaguchi-Shima and Yuri, 2007), and/or, as demonstrated in this report, by impairments in cell signaling. We show here that dorsal hippocampal ERK and PI3K activation are necessary for E2 to enhance object recognition in middle-aged females, and thus, the failure of a single post-training E2 injection to enhance object recognition in aged females may result from the inability of E2 to activate PI3K (Fig. 4e) or of age-related reductions in phospho-p42 ERK protein (Fig. 6a). However, it should be noted that aged female mice are not completely unresponsive to E2, as E2 can enhance various forms of hippocampal memory in aged rodents (e.g., (Miller et al., 1999; Frick et al., 2002; Heikkinen et al., 2002; Vaucher et al., 2002; Foster et al., 2003)), although the response of aged females to E2 is influenced by numerous factors, including timing of treatment relative to testing (e.g., pre- vs. post-training), priming (Savonenko and Markowska, 2003), duration of treatment, and dose (see (Frick, 2009) for review). In particular, timing and treatment duration may be important factors for object recognition, as chronic pre-training E2 treatment improves object recognition in aged female mice (Vaucher et al., 2002). Dose may also be an issue here, as the doses used in this study may have been too low to enhance memory in aged females. Although this explanation is less likely given that both systemic (Gresack et al., 2007a, b) and intracranial (present study) E2 treatments fail to enhance object recognition in aged female mice, testing higher intracranial doses should help resolve this issue.
Finally, the data also indicate that in the absence of E2, object recognition training does not activate the ERK/MAPK and PI3K/Akt pathways at any age after 15 min. This finding might suggest that object training does not activate these pathways, and that the E2-induced enhancement of object recognition is mediated by a different set of cellular mechanisms than those that normally take place during object memory consolidation. Although this possibility cannot be conclusively discounted at present, it is not supported by previous studies in young mice showing that ERK phosphorylation in whole hippocampus is increased 10 min (Wang et al., 2004) or 30 min (Goeldner et al., 2008) after object recognition training. Moreover, our own data from young (Fernandez et al., 2008) and middle-aged (present study) vehicle-treated females shows that preventing ERK or PI3K activation in the absence of E2 blocks 24-hr object recognition. Together, these studies indicate that object recognition training activates ERK, and that preventing this activation blocks learning. However, ERK and PI3K may be activated more slowly in middle-aged females, and thus, the 15 min time point may have been too soon to observe E2-induced increases in these enzymes. Alternatively, ERK and PI3K may be activated more slowly in the absence of E2, irrespective of age, with the more rapid E2-induced activation leading to longer lasting memory (i.e., 48 hrs rather than 24 hrs). Measuring protein levels at times longer than 15 min would help demonstrate the time-course of training-induced activation of these signaling pathways.
In conclusion, the present study demonstrates that the E2-induced enhancement of object recognition in middle-aged female mice is dependent on PI3K and ERK activation in the dorsal hippocampus. These data provide the first evidence that hippocampal cell signaling molecules are critically necessary for estrogenic enhancement of object memory consolidation in middle-aged females. Interestingly, middle-aged females are responsive to the same doses of E2 as young females, and the signaling pathways that mediate effects of E2 on object recognition are similar in young and middle-aged females, suggesting that the mechanisms through which E2 modulates object recognition in females remain largely intact during middle age. The failure of E2 to enhance object recognition in aged females may be due to its inability to activate dorsal hippocampal PI3K and ERK. Together, the present data provide important new insights into the molecular mechanisms underlying estrogenic modulation of memory in aging females that could be pivotal to the design of future hormone therapies.
This project was supported by Yale University and a grant from the National Institute on Aging (AG022525) to K.M.F.