APOE4 genotype exacerbates the impact of menopause on cognition and synaptic plasticity in APOE‐TR mice

The impact of sex and menopausal status in Alzheimer’s disease remains understudied despite increasing evidence of greater female risk, particularly in APOE4 carriers. Utilizing female APOE‐TR mice maintained on a high‐fat diet background we induced ovarian failure through repeated VCD injections, to mimic human menopause. At 12 months of age, recognition memory and spatial memory were assessed using object recognition, Y‐maze spontaneous alternation, and Barnes maze. A VCD*genotype interaction reduced the recognition memory (P < .05), with APOE4 VCD‐treated animals unable to distinguish between novel and familiar objects. APOE4 mice displayed an additional 37% and 12% reduction in Barnes (P < .01) and Y‐maze (P < .01) performance, indicative of genotype‐specific spatial memory impairment. Molecular analysis indicated both VCD and genotype‐related deficits in synaptic plasticity with BDNF, Akt, mTOR, and ERK signaling compromised. Subsequent reductions in the transcription factors Creb1 and Atf4 were also evident. Furthermore, the VCD*genotype interaction specifically diminished Ephb2 expression, while Fos, and Cnr1 expression reduced as a consequence of APOE4 genotype. Brain DHA levels were 13% lower in VCD‐treated animals independent of genotype. Consistent with this, we detected alterations in the expression of the DHA transporters Acsl6 and Fatp4. Our results indicate that the combination of ovarian failure and APOE4 leads to an exacerbation of cognitive and neurological deficits.


| INTRODUCTION
The development of sporadic late-onset Alzheimer's disease (AD) is complex and multifactorial with more than 20 susceptibility loci including Apolipoprotein E (APOE). 1 APOE4 is the strongest prevalent genetic risk factor with an allele frequency of 14% in the general population and is concentrated in AD (allele frequency ~40% in Caucasians). 2 The APOE4 gene increases AD risk in a gene dose-dependent manner, rising from 2-to 3-fold for heterozygous APOE4 carriers and 12-fold in homozygotes, 3,4 with age of onset considerably reduced. However, a large proportion of APOE4 carriers do not develop AD indicating that the genotype is predictive rather than prognostic and may be modulated by other environmental and biological factors. 5 There is accumulating evidence that the penetrance of APOE4 genotype is influenced by sex, 4,6 which may underlie the finding that almost two thirds of AD cases are females. 7 In a large meta-analysis Neu and colleagues reported increased MCI and AD risk in females compared to male APOE4 carriers between 55-70 years and 65-75 years, respectively. 8 Despite this observation, little is known of the etiology underlying this susceptibility. However, given the age at which females are at increased risk 8 and the fact that earlier onset of menopause correlates with poorer cognition in later life, 9 it is plausible that menopause acts as a mediating factor.
Within the central nervous system APOE is the main lipid transporter. 10 In comparison to other organs, the brain is highly enriched with the n-3 polyunsaturated fatty acid (PUFA) docosahexaenoic acid (DHA) accounting for ~15% of total brain lipids. 11 We have recently reported lower DHA and specialized pro-resolving mediators in old APOE4 female mice, indicating how sex, APOE4, and age contribute to the development of cognitive decline and AD pathology. 12 DHA is specifically concentrated within synaptosomal membranes, 13 influencing membrane dynamics, membrane protein function, secondary messenger systems, and neurotransmitter concentrations. 14 Thus it is unsurprising that reduced DHA status is consistently linked to poorer cognitive outcome and increased AD risk. 15 Synaptic loss and dysfunction which has been characterized in both menopause and APOE4 individuals, is directly associated with cognitive decline and occurs in the initial stages of AD. 16 With an etiological link accounting for the greater female APOE4 susceptibility distinctly lacking, 17 we posit menopause as a contributing factor, given the neuroprotective properties of estrogen, 18 and greater cognitive decline associated with early menopause. 9 Here, in a transgenic mouse model we assessed the combined impact of APOE genotype and ovarian failure on cognitive performance hypothesizing that APOE4 animals will be more susceptible to estrogen loss. Furthermore, we determine if changes in cognitive performance relate to alterations in brain fatty acid profiles and synaptic plasticity. Our model system combines the humanized-targeted replacement (APOE-TR) mouse model with 4-vinylcyclohexene diepoxide (VCD) treatment credited for its ability to establish an intermediary human-like "perimenopause" phase, while maintaining ovarian tissue integrity. 19

| Study approval
All experimental procedures and protocols used in this study were reviewed and approved by the Animal Welfare and Ethical Review Body (AWERB) and were conducted within the provisions of the Home Office Animals (Scientific Procedures) Act 1986.

| Animal model and experimental design
Forty eight C57BL/6 background female humanized APOE3 (B6.129P2-Apoe tm2(APOE*3)Mae N8) and APOE4 (B6.129P2-Apoe tm2(APOE*4)Mae N8)-targeted replacement mice homozygous for the human APOE3 or APOE4 gene (Taconic, Germantown, NY, USA) were used in these experiments. 12,20,21 Mice were maintained in a controlled environment (21 ± 2°C; 12-h light-dark cycle; light from 07:00 hours) and fed ad libitum on a standard chow diet (RM3-P, Special Diet Services, Essex, UK) until the age of 4 months, ensuring normal development. Following this run-in period, mice were switched to a semi-purified high-fat diet (45 kCal% fat) (D17080301, Research diets, New Brunswick, NJ, USA) for the remaining duration of the experiment (See Table S1 for full dietary composition). A high-fat diet was utilized to mimic a human "western-style" high-fat diet and exacerbate AD age-related cognitive decline. 22 At 8 months of age and to assess the impact of menopause, mice from each genotype received intra peritoneal (i.p.) injections of either VCD (160 mg/kg body weight) diluted in sesame oil, or sesame oil vehicle (sham) for a total of 14 injections over 3 weeks. The VCD model system is a wellestablished method for inducing a "menopause" like state in rodents 23 and is well-tolerated leading to a targeted degradation of ovarian follicles. 24 Eight months was selected as it is roughly midlife for the animals (when human menopause occurs) and before natural ovarian failure is known to occur in C57BL/6 mice. 25,26 Following completion of the final behavioral test, 12-month aged animals were sedated with isoflurane (1.5%) in a mixture of nitrous oxide (70%), and oxygen (30%) and transcardially perfused with an ice-cold PBS-containing protease (SIGMAFAST Protease inhibitor, Sigma-Aldrich, Devon, UK) and phosphatase (1 mM sodium pyrophosphate and 50 mM sodium fluoride, Sigma-Aldrich, Devon, UK) inhibitors. Sera were isolated via centrifugation at 2,000 ×g for 10 minutes. Brains were rapidly removed, halved, snap frozen, and stored at −80°C until biochemical analysis. Ovaries were rapidly removed and processed for histology (see Section 2.4). Graphical representation of the experimental design is given in Figure 1.

| Behavioral assessment
All behavioral tests were performed when mice reached 12 months of age and immediately prior to sacrifice. A visual placing test was performed on each animal on the first day of testing, to ensure animals were not visually impaired. 27 Spatial learning and memory were evaluated with the Barnes Maze as previously described. 28 Briefly, the maze consisted of a brightly illuminated (800 lux lighting) circular platform (92 cm diameter), with 20 evenly distributed holes located around the circumference and visual cues (4 simple shapes) placed at the periphery. The experiment was conducted over a 5-day period, with each mouse tested/trained on ability to locate the escape box four times per day during days 1-4. On day 5, a probe test was conducted, the maze was rotated 90°, the escape box was removed, and mice were placed in the center of the maze in which they were free to navigate for 1 minute. Percentage time in the correct quadrant was determined using the Smart 3.0 tracking software (Panlab, Kent, UK).
The novel object recognition (NOR), a measure of recognition memory, was performed as described previously, 29,30 with slight modifications. Briefly, on day 1 mice were habituated in gray 50 × 50 × 50 cm apparatus illuminated with low lux 100 lux lighting, mice were placed into the empty maze, and allowed to move freely for 10 minutes. On day 2, mice were conditioned to a single object for a 10-minute period. On day 3, mice were placed into the same experimental area in the presence of two identical objects for 15 minutes, after which they were returned to their respective cages and an inter-trial interval of 1 hour was observed. One familiar object was replaced with a novel object. Mice were placed back within the testing area for a final 10 minutes. Videos were analyzed for a 5-minute period, after which if total object exploration time failed to reach an accumulative 10 seconds, analysis continued until the 10 seconds was met. Animals not achieving 10 seconds were excluded from the analysis. 31 Similarly, animals not achieving a cumulative 10 seconds with the familiar object were excluded. Discrimination index was calculated as follows: DI = (TN − TF)/(TN + TF), where TN is the time spent exploring the novel object and TF is the time spent exploring the familiar object.
Y-maze spontaneous alternation test, a measure of spatial working memory was performed on the final day of behavioral testing as previously described. 32 Briefly, the Y-maze apparatus comprised of white Plexiglas in the following F I G U R E 1 Experimental overview. At 4 months of age APOE4 and APOE3 TR mice were provided with a high-fat diet to model a western style diet. At 8 months of age mice received i.p. injections with 160 mg/kg VCD for 14 days in order to deplete ovarian follicles and mimic menopause by 12 months of age. At 12 months of age, mice were assessed cognitively through a battery of behavioral tests including Barnes maze, Y-maze, and object recognition. Upon completing the behavioral tests animals were immediately sacrificed. Created with BioRender.com dimensions (38.5 × 8 × 13 cm, spaced 120° apart) was illuminated with low lux (100 lux) lighting. Mice were placed into the maze and allowed to explore freely for 7 minutes while tracking software recorded zone transitioning and locomotor activity (Smart 3.0 tracking software, Panlab, Kent, UK). Spontaneous alternation was calculated using the following formula: Spontaneous Alternation = (Number of alternations/Total Arm entries − 2) × 100.

| Histological analysis
Ovaries were trimmed of fat and fixed in 10% formalin for 24 hours before being paraffin-embedded, and processed for hematoxylin and eosin (H&E) staining as described previously. 33

| Fatty acid profile in the brain
Total lipids were extracted from sub-cortical brain tissues (n = 5/6 per group) using the Folch extraction method. 34 Fatty acid methyl esters (FAME) were obtained using acidcatalyzed transesterification. 35 FAME were evaporated under oxygen-free nitrogen, and re-suspended in 500 µL of isohexane, SPE silica cartridges (Clean-up 203 Cusil 156, UCT) were used for purification purposes. The purified FAMEs were separated by gas chromatography using a Trace 1300 series Gas Chromatograph (Thermo Fisher Scientific, Stafford House, Boundary Way, Hemel Hempstead. HP2 7GE, U.K.) equipped with a PTV injector modified for on-column injection, and a flame ionization detector. A fused silica capillary column (TraceGOLD TG-5MS Guard GC Column with SafeGuard column, 30 m × 0.32 mm × 0.25 µm; Thermo Fisher Scientific, Hemel, Hempstead, UK) was employed, and hydrogen was used as carrier gas. The temperature gradient was from 50 to 150°C at 40°C/min and then, to 200°C at 2°C/min followed by 214°C at 1°C/min and finally to 220°C at 40°C/min, where it was held for 5 min. Individual methyl esters were identified by comparison to known standards (Marine oil FAME mix RESTEK #35066). Data were collected and processed using the Chromeleon software package (version 7.2).
Follicle-stimulating hormone (FSH) concentrations were determined by ELISA (Abnova, Taipei, Taiwan; ref KA2330) in sera samples as per the manufacturer's instructions and are displayed as ng/mL. Brain-derived neurotrophic factor (BDNF) concentrations were determined using DuoSet ELISA kit (R&D systems Minneapolis, MN, USA; ref DY248) from brain homogenates, as per the manufacturer's instructions, and the concentration (pg/mL) was normalized to total protein content (pg/mg total protein).

| RNA isolation and qRT-PCR
RNA isolation, cDNA synthesis, and qRT-PCR were carried out as previously described. 37 Briefly, total RNA was isolated from the brain samples using the Qiazol reagent (Qiagen, UK). One μg of total RNA was treated with DNase I (Invitrogen, UK) and used for cDNA synthesis using Invitrogen Oligo (dT) primers and M-MMLV reverse transcriptase. Quantitative real-time PCR (qRT-PCR) reactions were performed using SYBR green detection technology on the Roche light cycler 480 (Roche Life Science, UK). Results are expressed as relative quantity scaled to the average across all samples per target gene and normalized to the reference gene glyceraldehyde 3-phosphate dehydrogenase (Gapdh), which was identified as the optimal housekeeping selection using the software RefFinder. 38 Primer sequences are given in Supplementary Table S2.

| Statistical analysis
All data are presented as mean ± S.E.M. Data analysis was performed in GraphPad Prism version 8 (GraphPad Software, CA, USA). After identifying outliers using the Grubbs method, data were checked for normality/equal variances, performing log or box-cox transformation if necessary. Comparisons among groups were performed on normally distributed data | 5 of 16 PONTIFEX ET al.
using two-way ANOVA, followed by post hoc Tukey's test when two-way ANOVA resulted in a significant interaction effect. Pearson's correlation was used in behavioral analysis to assess the association of travel distance and movement speed on corresponding behavioral test performance. P values of less than .05 were considered statistically significant.

| Repeated injections of VCD result in ovarian failure independently of APOE genotype
Repeated injections of VCD in APOE3-TR and APOE4-TR mice resulted in a considerable loss of ovarian follicles (Figures 2A and S1) and a fourfold increase in serum FSH levels (VCD effect: F(1, 21) = 27.64 P < .0001 Figure 2B), irrespective of APOE genotype. Neither VCD treatment nor genotype had any impact on body weight gain (P > .05 Figure  2C) nor food intake (P > .05; Supplementary Figure S2).
3.2 | While APOE4 impairs working spatial memory, VCD causes additional deficits to recognition memory Y-maze, Barnes maze, and NOR were employed to establish the impact of genotype and VCD treatment on cognitive performance ( Figure 3). VCD had no effect on spatial learning and memory tasks, as assessed by the Y-maze and the Barnes maze ( Figure 3A,B). However, a genotype-dependent effect was observed, with APOE4 animals displaying ~12% lower cognitive performance than their APOE3 counterparts as assessed by the Y-maze (genotype effect: F(1, 38) = 7.75 P < 0.01 Figure 3A). This was consistent with the Barnes maze, where APOE4 animals spent ~40% less time in the correct quadrant, (genotype effect: F(1, 37) = 8.09 P < .01 Figure 3B). A genotype effect was also apparent during the learning phase (Supplementary Figure S3). Pearson's correlation revealed that these effects were not influenced by movement speed nor travel distance ruling out these as potential confounding factors ( Figure 3C-F) Representative trajectory maps for Barnes probe test are shown in Figure 3G.
F I G U R E 2 VCD treatment leads to ovarian failure independently of APOE genotype in APOE3-TR and APOE4-TR mice. A, Representative images of APOE3-TR and APOE4-TR female mice ovaries stained with hematoxylin and eosin show important loss of ovarian follicles following repeated i.p. injections of VCD (arrows indicate follicles); B, VCD-injected groups display elevated serum follicle-stimulating hormone (FSH) levels (n ≥ 5); C, Body weight was unaffected by VCD treatment or genotype (n ≥ 8); Data are presented as mean ± S.E.M *P < .05, **P < .01, ***P < .001, ****P < .0001 Unlike spatial learning and memory, recognition memory as assessed by the NOR test was influenced by the menopausal status, and in a genotype-dependent manner, with VCD-injected APOE4 animals losing the ability to distinguish between novel and familiar objects (VCD effect: F(1, 31) = 5.40 P < .05, Interaction effect: F(1, 31) = 5.52 P < .05 Figure 3H). Again, these effects were not influenced by movement speed nor travel distance ( Figure 3I,J).

| VCD injections reduce brain DHA levels independently of APOE genotype
A significant 13% genotype independent lower brain DHA concentration was observed in VCD-injected animals independent of their APOE genotype (VCD effect: F(1, 18) = 17.94 P < .001 Figure 4A). Higher total mono-unsaturated fatty acids (MUFAs) were observed in APOE3 carriers F I G U R E 4 HF VCD treatment reduces brain DHA levels. A, GC-FID analysis revealed lower brain DHA levels in VCD-treated groups (n = 5/6); B-F, Expression of the key transporter Aclsl6, Fatp4, Fatp1, Mfsd2a, and Fabp5 involved in DHA transport and uptake at the blood-brain barrier level. Data are presented as the mean ± S.E.M *P < .05, **P < .01, ***P < .001  Although no interaction was detected, DHA: AA ratio was 8% lower in APOE4 animals but remained constant in APOE3 regardless of treatment.

Fabp5
To identify whether the observed lower DHA related to deficits in DHA transport, gene expression profiles of key DHA transporters in the brain were assessed. Interestingly, a member of the long-chain acyl-CoA synthetase family Acsl6 known to be highly abundant in the brain 39 and present in brain microvessels, 40 was reduced by 10% in VCD-treated animals (VCD effect: F(1, 20) = 4.76 P < .05 Figure 4B). Expression of the predominant fatty acid transport proteins in the BBB endothelial cells, Fatp4 and Fatp1, 41 was also determined. Fatp4 expression was 14% lower as a result of APOE4 genotype, with the reduction particularly evident in APOE4 VCD-treated animals although the interaction did not reach significance (genotype effect: F(1, 20) = 9.29 P < .01; Interaction effect: F = 3.35 P = .08 Figure 4C). A similar genotype trend was observed for Fatp1 (P = .069 Figure 3D), while no effect was observed for major facilitator superfamily domain-containing 2a Mfsd2a or fatty acid-binding protein 5 Fabp5 (Figure 4E,F).

| DISCUSSION
Several studies report that female carriers of the APOE4 genotype are at a higher risk of AD, 8 however little is known as to how this predisposition manifests. 17 Here, we sought to determine whether the cumulative APOE4*menopause effect contributes toward this greater risk. With VCD treatment emerging as a robust model to assess the implications of menopause in neurological disease, 23 we investigated for the first time the impact of VCD-mediated ovarian failure in the APOE-TR mouse model, focusing on cognition, brain fatty acid profiles, and synaptic plasticity-related signaling. Deficits in spatial learning and memory were related to the APOE4 genotype, with an additive effect of menopause and APOE4 carrier status on recognition memory. Furthermore, the VCD insult lowered brain DHA status while both factors influenced synaptic plasticity-related signaling genes and proteins. Unlike previously published data by us and others reporting a strong impact of APOE4 genotype on body weight gain in male mice, 42,43 no effect of APOE genotype nor VCD injections was observed in our female mice. This observation is in agreement with previous reports indicating limited impact of menopause 44 and APOE4 45 on body weight gain in female mice.
Y-maze and Barnes maze revealed APOE4-dependent deficits in spatial learning and memory. Deficits in spatial memory performance mediated via APOE4 genotype have been reported in several studies. [46][47][48] Interestingly, the extent of the deficit appears to be greater in APOE4 female mice or absent in APOE4 male mice. 46 No exacerbation of spatial deficits were detected as a result of menopause in agreement with reports indicating that the impact on spatial memory is subtle, 49 and dependent upon age, which represents a key factor in the extent of decline. 50 Conversely, recognition memory was significantly influenced by the menopause mimic VCD with the effect restricted to APOE4 animals. Impairment of recognition memory has been previously reported in chronic ovariectomized C57BL/6 mice, 51 and aged rhesus monkeys, 52 suggesting that the medial temporal lobe and surrounding cortical areas are particularly sensitive to ovarian failure. Interestingly, we report that this effect was exacerbated by the APOE4 genotype, highlighting this specific brain area as a focus for future endeavors. Together, the behavioral test results indicate both genotype-and genotype*VCD-mediated cognitive deficits and show how the combination of factors result in broader cognitive impairment.
Reduced DHA intake and status are associated with neuropathology, cognitive decline, and higher AD risk. 15,53,54 Both APOE4 and ovarian function impact n-3 PUFA and DHA homeostasis. [55][56][57][58] To the best of our knowledge, the combined impact of both ovarian failure and APOE4 genotype on brain DHA levels has not been previously explored. Here, brain DHA levels were equally diminished in both genotypes following VCD treatment. Interestingly, we observed lower brain ApoE levels in both APOE4 groups, independent of VCD insult, suggesting DHA concentration was uncoupled from ApoE protein levels. This contrasts a previous report indicative of ApoE involvement in brain DHA levels, 55 however as previously mentioned, this may be sex-dependent and more prominent in male mice. Although this is the first time brain fatty acid levels have been analyzed in an APOE-TR VCD model, results from OVX models demonstrate similar reductions in DHA, 59 indicating a clear relationship between reproductive hormones and brain DHA maintenance. Surprisingly, the reduction in DHA did not translate to an observable behavioral deficit in APOE3-treated animals suggesting that protection/tolerance is conferred by APOE3, perhaps via a compensatory mechanism. For example, the DHA:AA ratio in APOE4 VCDtreated animals was more extensively reduced, offering one possible explanation for the greater resilience shown by their APOE3 counterparts. Indeed, the importance of DHA:AA as a determinant of neuro-inflammatory status, given that AA and DHA are precursors of potent pro-inflammatory eicosanoids 60 and specialized pro-resolving mediators, respectively, 12,61 offers a plausible mechanistic basis. Furthermore, in APOE3 the VCD-mediated fluctuations of higher total MUFA and lower total SFA (absent in APOE4) may have influenced cognition to some extent. Higher MUFA intake has been previously shown to improve the brain function while lower MUFA brain levels are associated with AD, aging, and depression. 62 A number of pathways have been proposed to explain the benefits of MUFAs including the maintenance of membrane flexibility 63 and the actions of MUFA's as anti-inflammatory and antioxidant derivatives 64 which have been shown to modulate the neuroinflammation in ApoE KO mice. 65 Although the mechanism by which DHA enters the brain (passive/active) remains to be fully determined, current evidence points out it is likely to be in part mediated by a small number of specific binding/transporter proteins. MFSD2A, ACSL6, FABP5, and the FATPs 1 and 4 are expressed at | 11 of 16 PONTIFEX ET al.
the BBB 40 and have been previously associated with DHA transport into the brain. 17 Of the DHA transporters measured, Acsl6 described from knockout studies as being a key mediator of neuroprotective DHA within the brain 39 and believed to be critical in maintaining brain DHA levels, 66 was downregulated in response to VCD treatment. Acls6 may, therefore, account for the reduction in DHA observed and should be a focus for future studies evaluating the impact of menopause on brain DHA levels. Fatp1, but particularly Fatp4 displayed genotype deficits; these fatty acid transporters have been shown to bind DHA and facilitate its transport across the endothelial cell membrane. 67 This lower expression may in part explain the genotype reductions in DHA observed by other groups particularly if exacerbated by age. Previous work has indicated that blood-brain barrier transporter protein cell localization and membrane shedding are influenced by APOE and neuropathology. This could also influence the capacity to uptake DHA into the brain. 68,69 Moreover, as previously mentioned the change may be independent of transport and instead relate to metabolic disturbances which lead to greater β-oxidation of DHA. 70 This is conceivable given the role estrogens play in bioenergetic systems within the brain. 71,72 Given the role of ApoE protein in neurite outgrowth 73 and neuronal repair processes, 74 APOE4-specific reductions in spatial memory might be expected in light of the diminished protein levels observed in this experimentation. Further investigation may be warranted to establish if these lower APOE4 levels are constant across sexes given that spatial memory deficits appear to be APOE4 female specific. 46 As with human menopause, VCD treatment is associated with increased FSH levels. Research evaluating the impact of elevated FSH on cognition is limited, however in OVX mice a reduction of all gonadotropins, appears equally as effective at preserving cognition as providing estrogen. 75 Additionally, in humans FSH:estradiol has been used as a predictor of MCI. 76 Moreover, estrogen is a well-established regulator of synaptic plasticity in key areas of the brain including the hippocampus. 77 In this study, we observed that both the menopause mimic and APOE4 genotype influenced synaptic signaling in the in the brain. First, the brain neurotrophic factor BDNF, described to be multifaceted and considered a flexible hub for synaptic plasticity and cognitive functioning 78 was reduced as a result of APOE4 genotype and VCD treatment. Such observation is in agreement with previous studies focusing on APOE4 79 and VCD treatment separately. 80 Interestingly BDNF is known to influence both mTOR and ERK activation. 81 mTOR signaling was impaired by the APOE4*VCD combination. To the best of our knowledge, such an effect has not been previously reported in APOE4 mice, but the effect of menopause has been observed in OVX models. 82-84 mTOR activation, likely triggered via the PI3K-AKT pathway, is critical for neuronal cell survival, and is intriguingly associated with maintenance of glucose homeostasis, with similar deficits established in the diabetic rat brain, 85 thus highlighting potential metabolic/bioenergetic disturbances in the APOE4 VCDtreated brain. Furthermore, AKT and mTOR activation in AD mouse models ameliorates deficits in synaptic plasticity improving associated learning and memory. 86,87 It is, therefore, possible that the impaired object recognition performance observed in this study stems from impairment of PI3K-AKT-mTOR signaling, which is already known to be estrogen and ER sensitive. 77 ERK signaling represents another important cascade which is similarly intertwined with BDNF and synaptic plasticity. 88 ERK activation was found to be reduced as a result of APOE4 genotype and the interactive impact of VCD and genotype. This is in line with Yong and colleagues who reported diminishing ERK phosphorylation in aging female APOE4-TR mice. 89 Interestingly both Yong and our group show a reduction in brain APOE protein levels in the APOE4 female mice, highlighting how this may be coupled to ERK phosphorylation. Additionally, ERK activation has been reported in OVX studies. 90,91 Given DHA's proposed influence upon both PI3K-AKT-mTOR 92 and ERK signaling pathways, 93 the altered DHA/PUFA status associated with VCD treatment, and the subsequent reduction in BDNF may offer a plausible mechanistic basis for their disruption, although further confirmatory experiments are required. As discussed mTOR and ERK are central components of synaptic plasticity-related signaling and are involved in various down-stream pathways explaining the diminished transcription of the CREB family transcription factors (Creb1 and Atf4), both of which have integral roles in synaptic plasticity. 94,95 Further down-stream effects reflected previous observations with Arc expression nominally reduced in APOE4, particularly VCD-treated animals. Referred to as the master organizer of long-term synaptic plasticity, 96 Arc has been reportedly regulated by both APOE4 and estrogen with both MAPK and PI3K pathways implicated. 97,98 We also noted diminished Ephb2 expression as a result of the APOE4*VCD interaction. Interestingly, Ephb2 plays a fundamental role in learning and memory, with implications ranging from synapse maintenance and synaptogenesis to AMPA and NMDA receptor expression, localization, and function. [99][100][101][102][103] Indeed, increased Ephb2 expression has been found to compensate for AD-related NMDA receptor impairment. 104 Ephb2, therefore, offers additional mechanistic basis for the cognitive impairment associated with APOE4*VCD. Further to this, APOE4 genotype led to a downregulation of Fos an indicator for neuronal activation, 105 and may be linked to the altered ERK signaling profile to which it is known to be influenced. 106 Finally, with implications in hippocampal synaptic plasticity, adult neurogenesis, and subsequent memory consolidation we assessed endocannabinoid, specifically Cnr1 regulation. 107 An APOE4-dependent reduction in Cnr1 expression was observed, CNR1 receptor antagonism and subsequent endocannabinoid dysregulation has been associated with deficits in learning and memory. 108 Recently CNR1 activation has been reported to be neuroprotective conferring specific improvement in spatial memory, 109 therefore, diminished CNR1 may also contribute to the spatial memory deficits associated with APOE4. Research connecting APOE and endocannabinoid signaling, particularly within the brain is surprisingly limited, and warrants further investigation in light of these results.
As with all animal models some limitations associated with the model system should be mentioned. First, although well-established, 23 particularly in cognitive and cardiology research 24,110,111 and believed to be non-toxic/ovarian follicle specific, 24,112 the VCD menopause induction may introduce as yet unidentified off-target impacts. The timing of the VCD and cognitive assessment (ie, middle-aged animals) could have resulted in some sham animals being spontaneously acyclic. However, given that VCD was introduced before natural ovarian failure is known to occur in C57BL/6 mice, 25,113 and that we observed primordial follicles in sham animals ( Figure S1), the risk of this is low, consistent with the increase in FSH levels in only VCDtreated animals. 114,115 Furthermore, even if some animals had become partially senescent over the later stage of the assessment period, the VCD-treated animals would have undergone the process much earlier and thus would have been exposed to the detrimental effects of ovarian failure for a longer duration of time.

| CONCLUSION
Despite its well-established impact on late-onset AD risk, the etiological basis of the APOE4 genotype-associated cognitive deficits and neuropathology remains elusive. Sex, and menopausal status remain overlooked factors that likely influence the progression of neurological diseases such as AD. Here we provide evidence of menopause-related risk and suggest a greater sensitivity in APOE4 carriers, with APOE4 carriers displaying greater cognitive impairment and more extensive deficits in synaptic plasticity-related signaling.

ACKNOWLEDGMENTS
Pontifex's PhD is funded by an Alzheimer's Society UK PhD studentship (AS-PhD-2015-023). Minihane's research in the area of n-3 fatty acids, APOE genotype, and cognition is funded by the Biotechnology and Biological Sciences Research Council (BBSRC), UK, Alzheimer's Research UK (ARUK) and the Centre for Nutrition Learning and Memory, University of Illinois, US. Vauzour's research in the area of n-3 fatty acids, APOE genotype, and cognition is funded by the BBSRC, UK, and the Centre for Nutrition Learning and Memory, University of Illinois, US. The authors thank the staff of the Disease Modeling Unit at the University of East Anglia for expertize and help with the conduct of the rodent studies. The authors also thank James Dick and the University of Stirling Institute of Aquaculture, Stirling, UK, for their expertize and help with the conduct of the lipid analysis. Finally, the authors also thank Dr Julie Deguil, Faculté de Médecine, Université de Lille for her contribution to the establishment of the animal behavioral testing in our research facility.

CONFLICT OF INTEREST
There are no actual or potential conflict of interest.