Young adult [∼2 month-old at the initiation of the experiments (Flurkey and Currer, 2004)] C57BL/6 WT mice (n = 22 females and n = 11males) and transgenic 5xFAD (C57BL/6 background; n = 22 female and n = 11 male) mice were bred and maintained in a colony at Weill Cornell Medicine (WCM). Breeding pairs of hemizygous 5xFAD mice were obtained from the Jackson Laboratory (Bar Harbor, ME; JAX MMRRC stock#034840). Mice were bred in-house and genotyped prior to experimentation. 5xFAD mice express human APP and PSEN1 transgenes with a total of five AD-linked mutations under the control of the Thy1 promoter. This mouse model exhibits early-onset parenchymal Aβ aggregation correlated with cognitive deficits (Xu et al., 2014a; Richard et al., 2015). Amyloid deposition begins in the cerebral parenchyma at 2–3 months of age, with little accumulation in the cerebral vasculature, and amyloid plaques are found throughout the hippocampus and cortex by 6 months (Oakley et al., 2006; Xu et al., 2014b). Astrogliosis and microgliosis begin around 2 months, developing in parallel with plaque deposition (Oakley et al., 2006). Mice were housed in groups of three to four animals per cage and maintained on a 12-h light/dark cycle (lights out 18:00 h) with ad libitum access to water and rodent chow. At euthanasia, mice weighed 23–32 g. All experiments were approved by the WCM Institutional Animal Care and Use Committees and followed the National Institutes of Health guidelines for the Care and Use of Laboratory Animals guidelines.

AOF induction by VCD treatment in mice has been shown to recapitulate the gradual hormonal fluctuations that correspond to peri- and post-menopause in humans (reviewed in (Van Kempen et al., 2011; Marques-Lopes et al., 2018)). The AOF model can be used to separate hormonal effects from aging effects and can be applied to any mouse genotype (Van Kempen et al., 2014; Brooks et al., 2016; Marques-Lopes et al., 2017). Low dose VCD injections selectively deplete ovarian primary follicles without negatively affecting peripheral tissues, kidney, or liver function (Brooks et al., 2016; Haas et al., 2007; Sahambi et al., 2008; Wright et al., 2008). VCD does not directly increase inflammation markers in the brain, including the hippocampus (Van Kempen et al., 2014).

Gonadally intact 53-58-postnatal-day-old female mice received 130 mg/kg VCD (cat. #S453005 Millipore Sigma, St. Louis, MO) in sesame oil (cat. # 8008-74-0 Millipore Sigma) for 5 days per week for 3 weeks (Marques-Lopes et al., 2017; Van Kempen et al., 2011). Control mice received injections of sesame oil only. Prior studies from our lab and others (Van Kempen et al., 2014; Haas et al., 2007; Lohff et al., 2005) established the peri-AOF stage as occurring 58 days after the first VCD injection. At this stage (∼3.5 months old), the mice exhibit irregular, prolonged estrous cycles and elevated plasma follicle stimulating hormone (Lohff et al., 2005; Mayer et al., 2004; Harsh et al., 2007). Behavioral assessments of VCD- and oil-treated females, as well as aged-matched males, were initiated when mice were about 4 months of age (the peri-AOF stage of VCD mice). A timeline of the experimental procedures is shown in Figure 1.

At euthanasia, vaginal smears (Turner and Bagnara, 1971) were collected to determine the terminal estrous cycle stage via cytological examination. Estrous cycle phases were classified as proestrus (high estrogen), estrus (declining estrogen), or diestrus (low estrogen and progesterone). Most females were in estrus or diestrus at euthanasia.

4G8: A mouse monoclonal antibody raised against amino acid residues 17-24 of beta-amyloid (4G8, Biolegend Cat. # 800701) was employed. This antibody recognizes abnormally processed isoforms as well as precursor APP forms (manufacturer’s instructions), and labels both parenchymal and vascular Aβ aggregates (Kövari et al., 2013; Alafuzoff et al., 2009; Marazuela et al., 2022; Davis et al., 2004). GFAP: A rabbit polyclonal antibody (Abcam # ab7260; lot # GR20948-21; RRID:AB_305808) raised against full-length human GFAP was used. On Western blot, this antibody recognized 48 kDa and 55 kDa GFAP bands (manufacturer’s datasheet). Iba1: A rabbit polyclonal antibody raised against a synthetic peptide corresponding to the C-terminus of Iba1 (#SAR6502; 019-19741 FUJIFILM Wako Pure Chemical Corporation) was employed. The antibody reacts with rat, mouse and human Iba1 and recognizes a 17 kDa band protein on Western blot (manufacturer’s datasheet). These antibodies have been used in our prior studies (Milner et al., 2022; Platholi et al., 2023).

Mouse brains were processed for immunocytochemistry using established procedures in our labs (Milner et al., 2011). Briefly, mice were deeply anesthetized with sodium pentobarbital (150 mg/kg, i. p.), and then perfused with saline. The brains were extracted, bisected sagittally, and the right hemisphere fixed in 4% paraformaldehyde in 0.1 M phosphate buffer (PB, pH 7.4) for 24 h on a shaker (70 rpm) at 4°C. The forebrain containing the hippocampus was sectioned (40 µm thick) on a vibratome (VT1000X Leica Microsystems, Buffalo Grove, IL) and stored in cryoprotectant (30% sucrose, 30% ethylene glycol in PB) at −20°C until immunocytochemical processing.

For each experiment, one rostral (−2.00 to −2.70 mm from Bregma (Hof et al., 2000)) or caudal (−2.90 to −3.50 mm from Bregma (Hof et al., 2000)) hippocampal section per animal was selected and then punch coded in the cortex. Tissue sections from each experimental group were pooled into single containers to ensure identical reagent exposure (Milner et al., 2011).

Hippocampal sections from each genotype/sex group were processed for 4G8, Iba1 or GFAP (n = 11/group). Sections were rinsed in 0.1 M Tris-saline (TS; pH 7.6), blocked with 0.5% bovine serum albumin (BSA) in TS for 30 min, and incubated in primary antibodies mouse anti-4G8 (1:4000), rabbit anti-GFAP (1:6000) or rabbit anti-Iba-1 (1:4000) diluted in 0.1% Triton-X and 0.1% BSA in TS for 24-h at room temperature followed by 24-h at +4°C. Next, sections were rinsed in TS and incubated in either biotin-conjugated goat anti-rabbit IgG (for GFAP and Iba1; #111-065-144, Jackson ImmunoResearch Inc., West Grove, PA; RRID:AB_2337965) or goat anti-mouse IgG (for 4G8; # 115-065-166, Jackson ImmunoResearch Inc.; RRID:AB_2338569) in 0.1% BSA and TS. Sections were washed in TS and incubated in Avidin Biotin Complex (ABC; Vectastain Elite kit, Vector Laboratories, Burlingame, CA) at half the manufacturer’s recommended dilution for 30-min. After rinsing in TS, the bound peroxidase was visualized by reaction in 3,3′-diaminobenzidine (Sigma-Aldrich, St. Louis, MO) and 0.003% hydrogen peroxide in TS for 6-min (4G8), 3-min (GFAP), or 8-min (Iba1). All primary and secondary antibody incubations were carried out at 145 rpm, whereas rinses were at 90 rpm on a rotator shaker. Sections were mounted in 0.05 MPB onto gelatin-coated glass slides, dehydrated through an ascending series of alcohol to xylene, and coverslipped with DPX (Sigma-Aldrich).

Quantification for 4G8, Iba1 and GFAP labeling in the hippocampus were performed using previously established densitometric methods (Pierce et al., 2014; Williams and Milner, 2011; Williams et al., 2011). To ensure unbiased data quantification, the analysis was performed by investigators blinded to experimental conditions. Images were acquired using a Nikon Eclipse 80i microscope with a Micropublisher 5.0 digital camera (Q-imaging, BC, Canada) and IP Lab software (Scanalytics IPLab, RRID: SCR_002775). ImageJ64 software (ImageJ, RRID:SCR_003070) was used to measure the pixel density within regions of interest (ROI) in defined hippocampal subregions. ROIs within four subregions of the rostral and caudal hippocampus were selected: 1) CA1: stratum oriens (SO), pyramidal cell layer (PCL), stratum radiatum (SR) and stratum lacunosum-moleculare (SLM); 2) CA2/3a: SO, PCL, near and distal SR; 3) CA3b: SO, PCL, stratum lucidum (SLu) and SR; 4) Dentate gyrus (DG): the supragranular blade (SG), the infragranular blade (IFG) and the central hilus (Cen) and 5) Subiculum (caudal section). Background pixel density from non-labeled regions (e.g., corpus callosum) was subtracted to control for illumination variability and background labeling. Prior studies (Pierce et al., 2014) demonstrated a strong correlation between pixel density and actual transmittance, confirming measurement accuracy.

Mice were tested sequentially over 2 weeks in the Novel Object Recognition, Y maze, and Barnes maze tasks, as described in prior studies (Park et al., 2020; Park et al., 2008). The same investigator conducted all behavioral tests. Testing occurred at the same time each day, with results recorded using ANY-maze (Stoelting Co.). Mice were habituated to the testing room for 2 h daily for 5 days before testing. On testing days, mice were acclimated to the room for 1-h before each session. The order of the Y-maze and Novel Object tests were counterbalanced, with 24-h rest between the tests. Behavioral apparatuses were cleaned with 70% ethanol between trials.

Images were adjusted first for contrast and sharpness in Adobe Photoshop 9.0 (Adobe Photoshop, RRID:SCR_014199). Next, images were imported into Microsoft PowerPoint, where final adjustments to brightness, sharpness and contrast were achieved. Images of different groups in the same figure were similarly adjusted to same brightness and contrast. Adjustments were made to the entire image, none of which significantly altered the appearance of the initial raw image.

Data are presented as means ± SEM. Statistical analyses were conducted using Prism 9 software (Graphpad Prism, RRID:SCR_002798) and significance was set at alpha < 0.05. Group comparisons were performed using analysis of variance (ANOVA; one-, two-, threeway) with Tukey or Sidak’s post-hoc tests. Two-group comparisons used Student’s t-tests. Specific analysis conducted are indicated in figure legends. Graphs were generated in Prism 9 software.

AD dementia is more prevalent in women and may emerge at the onset of menopause. However, there is limited evidence that Aβ levels in the hippocampus, which are a hallmark of AD, are influenced by perimenopause. To evaluate whether Aβ increases in the hippocampus of females at a stage of early ovarian failure, 5xFAD females were treated with VCD, or sesame oil as a control, and Aβ levels quantified at peri-AOF, a stage corresponding to human perimenopause. To obtain a more granular understanding of perimenopause role on AD pathology, analysis was performed across major hippocampal subregions. Given the small size and complex geometric borders of these hippocampal subregions, ELISA and similar methods requiring precise tissue punches were not feasible. Instead, we employed light microscopic immunohistochemistry.

The density of 4G8, a marker of Aβ, was examined in CA1, CA3, DG and subicular subregions in rostral and caudal hippocampal sections of the 5xFAD mice (Figures 2A,B). No immunoreactivity was detected in the WT mice. As described below, 4G8 labeling in the 5xFAD mice varied across hippocampal subregions with sublayer specificity.

CA1 and subiculum: In the rostral CA1, dense 4G8-immunoreactivity (ir) was found in the pyramidal cell layer (PCL), with scattered clusters in the other laminae, particularly in the stratum oriens (SO) (Figure 2C–E). In the rostral CA1 PCL, 5xFAD male mice exhibited lower 4G8 labeling than both 5xFAD-oil treated and 5xFAD VCD-treated female mice following post-hoc analysis (F = 5.080, p = 0.013) (Figure 2F). Similarly, in the stratum radiatum (nSR), 5xFAD males showed lower 4G8 labeling compared to 5xFAD-oil treated female mice (F = 3.783, p = 0.034) (Figure 2F). Analogously, in the deep stratum radiatum (dSR), the density of 4G8 labeling following post-hoc analysis was lower in 5xFAD male mice compared to 5xFAD oil- and 5xFAD VCD-treated female mice (F = 4.178, p = 0.025) (Figure 2F). The 4G8 labeling in the caudal CA1 was similar to that seen rostrally. Moreover, scattered 4G8-positive cells were found in the subiculum though no significant effects were found (Supplementary Figures S1A–D).

CA3: Scattered 4G8-labeled cells were found in the PCL of CA3a and CA3b, with labeled clusters dispersed throughout other laminae (Figures 2G–I). No significant difference in 4G8 density was observed in any sublayers of CA3a (Supplementary Figures S1E–H). However, in the CA3b PCL, 5xFAD VCD females showed greater 4G8 labeling than both 5xFAD oil-treated females and 5xFAD male mice following post hoc analysis (F = 5.049, p = 0.013) (Figure 2J). Additionally, in the stratum lucidum (SLu) of CA3b, 5xFAD VCD-treated females showed higher 4G8 density than 5xFAD males (F = 5.98, p = 0.007; t(20) = 3.374, p = 0.003) (Figure 2J).

Dentate Gyrus (DG): In both the rostral and caudal DG, 4G8-labeling appeared in clusters throughout all laminae, especially in the hilus. No significant difference in 4G8 density was observed in any DG sublayers (Supplementary Figures S1I–P).

Astrocytes facilitate the removal of Aβ from the brain parenchyma by mediating efflux into the cerebral vasculature (Liu et al., 2018). Their function is influenced by sex, partly through the actions of estrogen signaling via its estrogen receptors (Santos-Galindo et al., 2011; Chowen and Garcia-Segura, 2021; Lu et al., 2020). Therefore, astrocytes are expected to play an important role in amyloidosis during ovarian failure. However, evidence is limited regarding changes in astrocyte activity across hippocampal subregions in both intact and reproductively compromised females.

To assess astrocyte activation, the density of the astrocytic marker GFAP was examined in CA1, CA3, DG and subicular subregions within rostral and caudal hippocampus (Figures 3A,B) of WT and 5xFAD oil- and VCD-treated female mice, as well as oil-treated male mice. Consistent with our prior studies (Milner et al., 2022; Platholi et al., 2023), GFAP-labeled cells were found throughout all lamina of CA1, CA3 and DG, with fewer GFAP-positive cells in the pyramidal and granule cell layers (Figures 3,4, ). Representative micrographs of GFAP labeling are shown for rostral CA1 (Figures 3C–H), caudal CA1/subiculum (Figures 3J–O), rostral DG (Figures 4A–F), and caudal DG (Figures 4H–M).

CA1 and subiculum: In the rostral CA1 SO region, there was a significant main effect of genotype (F_genotype = 18.50, p < 0.0001) and treatment (F_treatment = 3.770, p = 0.029). Post-hoc analysis showed that the density of GFAP labeling in SO was greater (p < 0.05) in 5xFAD mice than their WT counterparts (Figure 3I). In the PCL region, a significant main effect of treatment was found (F_treatment = 4.822, p = 0.011),with WT-VCD mice showing greater GFAP labeling than WT male mice (p = 0.048) (Figure 3I). In the SLM region, a significant main effect of genotype was observed (F_genotype = 12.20, p = 0.0009), and post-hoc analysis showed that 5xFAD male mice had significantly greater GFAP labeling than WT male mice (p = 0.006) (Figure 3I).

In the caudal CA1 SO region, there was a significant main effect of genotype (F_genotype = 38.26, p < 0.0001). Post-hoc analysis showed increased GFAP labeling in 5xFAD mice compared to WT controls (p < 0.05) (Figure 3P). In the PLC region, a main effect of genotype was also observed (F_genotype = 7.458, p = 0.008). Post-hoc multiple comparisons showed that WT-VCD mice had greater GFAP labeling than WT-oil mice (p = 0.022) (Figure 3P). Additionally, 5xFAD-oil mice had significant higher GFAP labeling in the PLC than WT-oil mice (p = 0.001) (Figure 3P). In the SR region, of caudal CA1, a significant main effect of genotype was observed (F_genotype = 9.803, p = 0.003), with increased GFAP labeling in 5xFAD-oil compared to WT-oil mice (p = 0.005) (Figure 3P). In the SLM region, there was a significant main effect of genotype (F_genotype = 12.49, p = 0.001), with 5xFAD-oil mice exhibiting greater GFAP labeling than WT-oil mice (p = 0.0071), and 5xFAD-male showing increased GFAP compared to WT-male mice (p = 0.0243) (Figure 3P). Similar to other regions of caudal CA1, in the subiculum, a significant main effect of genotype was found (F_genotype = 61.46, p < 0.0001), with post-hoc comparisons revealing greater GFAP labeling in 5xFAD mice compared to WT controls (p < 0.05) (Figure 3P). Additionally, 5xFAD-oil mice had significantly higher GFAP labeling in the subiculum than 5xFAD-male mice (p = 0.045) (Figure 3P).

CA3: No significant effect of treatment or genotype on GFAP density was observed in any CA3 subregions (Supplementary Figures S2A–N).

DG: In the rostral DG crest, there was a significant effect of genotype (F_genotype = 6.914, p = 0.011) on the density of GFAP labeling. Post-hoc analysis revealed significantly greater GFAP labeling in 5xFAD-VCD compared to WT-VCD mice (p = 0.021) (Figure 4G). In the hilus, a significant main effect of genotype was observed (F_genotype = 12.21, p = 0.001), with post-hoc multiple comparison analysis revealing that 5xFAD-VCD and 5xFAD-male mice had greater GFAP labeling than their WT counterparts (p < 0.05) (Figure 4G).

In the caudal DG crest, there was a significant main effect of genotype (F_genotype = 5.325, p = 0.025), with increased GFAP in 5xFAD-male compared to WT-male mice (p = 0.034) (Figure 4N). In the hilus, a main effect of genotype was also found (F_genotype = 10.18, p = 0.002), with 5xFAD-male mice exhibiting greater GFAP labeling than WT-male mice (p = 0.036) (Figure 4N). In the db region of the DG, a main effect of genotype was observed (F_genotype = 18.47, p < 0.0001), with increased GFAP labeling in 5xFAD-oil compared to WT-oi mice (p = 0.002) (Figure 4N). Similarly to the crest and hilus, 5xFAD-male mice showed significantly more GFAP labeling than WT-male mice (p = 0.031) in the db (Figure 4N).

These patterns of GFAP labeling suggest that peri-AOF is associated with increased astrocyte activity in select hippocampal regions of the rostral and caudal CA1 and DG.

Microglia, the brain’s resident macrophage, regulate neuroinflammation and cognitive function (Cornell et al., 2022). The protein Iba1 is constitutively expressed in microglia and upregulated upon activation (Sasaki et al., 2001; Imai et al., 1996), a common feature of aging and neurodegenerative disorders (Prinz et al., 2021). Significantly, microglia activity is influenced by ovarian hormone changes. Ovariectomy increases Iba1 expression in middle-aged female mice (Sárvári et al., 2017), while estradiol reduces microglia reactivity in the hippocampus of aged ovariectomized animals (Lei et al., 2003). Additionally, ovariectomy elevates macrophage antigen complex-1, another marker of reactive microglia, in the hippocampus of aged mice (Benedusi et al., 2012). In AD mouse models, chronic estrogen deficiency is linked to heightened microglial activation and neurodegeneration (Prat et al., 2011). Given these findings, ovarian failure may alter hippocampal Iba1 expression during amyloidosis, however, direct evidence remains limited.

The density of Iba1 was examined in CA1, CA3, DG and subiculum subregions within the rostral and caudal hippocampus (Figures 5A,B). Consistent with our prior studies (Milner et al., 2022; Platholi et al., 2023), Iba1-labeled cells were scattered throughout all lamina in the CA1, CA3 and DG; however, the pyramidal and granule cell layers contained less Iba1 labeling (Figures 5,6, ). Representative micrographs show the distribution of Iba1 labeling in the rostral CA1 (Figures 5C–H), caudal CA1/subiculum (Figures 5J–O), rostral DG (Figures 6A–F), and caudal DG (Fiures 6H–M).

CA1 and subiculum: In the rostral CA1 SO region, there was a significant main effect of genotype (F_genotype = 10.16, p = 0.0023) on Iba1 labeling. Post-hoc analysis showed that Iba1 density was greater (p < 0.05) in 5xFAD-VCD and 5xFAD-male than their WT counterparts (Figure 5I). In the PCL region, there was a main effect of genotype (F_genotype = 14.12, p = 0.0004) and treatment (F_treatment = 3.449, p = 0.038) on Iba1 labeling. As in the SO region, post-hoc analysis showed increased labeling in 5xFAD-VCD and 5xFAD-male mice (p < 0.05) compared to WT mice (Figure 5I). Additionally, 5xFAD-VCD mice exhibited significantly more Iba1 labeling compared to 5xFAD-oil, mice (p = 0.016) (Figure 5I). In the SLM region of the rostral CA1, a main effect of genotype was observed (F_genotype = 10.89, p = 0.002), with post-hoc analysis showing significantly more Iba1 labeling in 5xFAD-male mice than WT-male mice (p = 0.006) (Figure 5I).

In the caudal CA1 SO region, there was a significant main effect of genotype (F_genotype = 19.61, p < 0.0001). Post-hoc analysis showed that Iba1 density was greater (p < 0.05) in 5xFAD-VCD and 5xFAD-male mice than their WT counterparts (Figure 5P). In the PCL region, a main effect of genotype (F_genotype = 30.99, p < 0.0001) and treatment (F_treatment = 4.272, p = 0.014) was observed. Multiple comparisons revealed significantly more Iba1 labeling in 5xFAD mice than in WT mice (p < 0.05) (Figure 5P). Additionally, 5xFAD-VCD mice exhibited greater Iba1 labeling in the PLC than 5xFAD-male mice, (p = 0.014) (Figure 5P). In the SR region of the caudal CA1, a main effect of genotype (F_genotype = 9.983, p = 0.003) was found, with post-hoc analysis showing significantly more Iba1 in 5xFAD-male mice than in WT-male mice (Figure 5P). In the SLM, there was also a main effect of genotype (F_genotype = 14.19, p = 0.0004), with 5xFAD-male mice exhibiting greater Iba1 labeling than WT-male mice (Figure 5P). Similarly, in the subiculum, a significant main effect of genotype (F_genotype = 67.47, p < 0.0001) and treatment (F_treatment = 3.996, p = 0.024) was found, with post-hoc analysis revealing significantly more Iba1 labeling in 5xFAD mice than WT mice (p < 0.05) (Figure 5P).

CA3: There was no effect of treatment or genotype on the density of Iba1 labeling in any of the sublayers of CA3a or CA3b (Supplementary Figures S2O–AB).

DG: In the rostral DG, there was a significant main effect of genotype (F_genotype = 9.923, p = 0.0025) on Iba1 labeling in the crest. Post-hoc analysis showed greater Iba1 density (p < 0.05) in 5xFAD-VCD and 5xFAD-male mice than in WT mice (Figure 6G). In the hilus, there was a main effect of genotype (F_genotype = 20.20, p < 0.0001) and treatment (F_treatment = 4.349, p = 0.017). Post-hoc analysis showed significantly more Iba1 labeling (p < 0.05) in 5xFAD-VCD and 5xFAD-male mice compared to their WT counterparts (Figure 6G). Additionally, 5xFAD-VCD mice exhibited increased Iba1 density compared to 5xFAD-oil mice (p = 0.036) and 5xFAD-male mice (p = 0.027) (Figure 6G). In the db of the rostral DG, a main effect of genotype (F_genptype = 16.39, p = 0.0002) was observed, with post-hoc analysis showing greater density of Iba1 (p < 0.05) in 5xFAD-VCD and 5xFAD-male mice compared to WT mice (Figure 6G).

In the caudal DG, a main effect of genotype (F_genotype = 10.04, p = 0.002) was found in the crest, with 5xFAD males exhibiting greater Iba1 density than WT male mice (p = 0.003) (Figure 6N). In the caudal hilus, a significant main effect of genotype (F_genotype = 20.42, p < 0.0001) was observed, with post-hoc analysis showing greater Iba1 labeling (p < 0.05) in 5xFAD mice than in their WT counterparts (Figure 6N). Similarly, a significant main effect of genotype (F_genotype = 25.07, p < 0.0001) was determined in the db of the DG, with multiple comparisons revealing significantly more Iba1-labeling (p < 0.01) in 5xFAD mice than in WT mice (Figure 6N).

These results suggest that the 5xFAD genotype at peri-AOF influences microglia activation in multiple subregions of rostral and caudal CA1and DG compared to peri-AOF WT mice. Additionally, peri-AOF further exacerbates microglia activation in 5xFAD mice in rostral CA1 PCL and DG hilus, suggesting sex-dependent vulnerability. These findings, in concert with increased GFAP labeling in select hippocampal regions, suggest enhanced susceptibility to neuroinflammation at the intersection of AD and perimenopause.

Neuroinflammation is intrinsically linked to the progression of cognitive impairment and dementia (Heneka et al., 2015). However, the impact of early ovarian failure on cognitive function in VCD-treated 5xFAD mice, and how this compares to males, remains unknown. To address this, we assessed the behavioral consequences of early ovarian decline in 5xFAD mice compared to non-VCD treated females and males using different cognitive tests sensitive to Aβ pathology, including the Y-maze alternation test, novel object recognition, and spatial navigation in the Barnes maze.

No significant differences were observed in arm alternation behavior (Figure 7A) or in the total number of arm entries in all 3 arms during the testing period (Figure 7B) between 5xFAD and WT mice. No differences were observed in the % novel arm entries respect to total arm entries and in the time spent in the novel arm (data not shown). This may suggest preservation of short-term working memory or that other cognitive domains might be more affected in 5xFAD mice and in VCD-treated females at a young age.

An effect of AOF was seen in locomotor activity as assessed by total distance travelled (F (5, 60) = 2.42, p = 0.05). 5xFAD-male compared to 5xFAD-VCD female mice had a lower total distance (p = 0.06), although not significant (Figure 7C). No significant differences across groups were observed in the total object exploration time or in Novel Object Recognition Index (Figures 7D,E). Interestingly, significant increase in time spent exploring the novel object vs. the familiar object was observed between WT-male (p < 0.0001) and 5xFAD-male mice (p = 0.001) (Figure 7F). For females, significant differences in time spent exploring familiar and novel objects were found between treatment in WT mice (WT-female (p = 0.018); WT-VCD female (p = 0.014)), but did not extend to 5xFAD females (Figure 7F).

There were no significant differences in motor activity, as assessed by distance traveled and mean speed between 5xFAD and WT mice, regardless of genotype or treatment (Figures 8A,B). Although, a significance difference was observed in mean speed between WT females and VCD-treated WT females at day 1 only (p = 0.0086). Also, no differences were observed in learning measures as shown for latency to find the escape and path efficiency (Figures 8C,D).

When cognitive performance was tested 24 h after the final acquisition training session, VCD-treated WT-females spent significantly less time in the target quadrant where the escape hole was located, and more time in non-target quadrants compared to both WT-oil or 5xFAD-VCD females (Figures 8E,F).

Given that the Barnes maze is a measure of spatial learning and memory, the altered performance in VCD-treated WT females suggests a potential adverse effect of early ovarian failure on spatial memory. This effect appears to be mitigated in 5xFAD-VCD females.

Overall, these behavioral data demonstrate that VCD treatment impairs performance in select cognitive tasks. Combined with our anatomical findings on Aβ deposition and glia activation in the hippocampus, these results suggest that early ovarian failure contributes to early-stage AD-like neurobehavioral pathology.