CYP46A1-mediated cholesterol turnover induces sex-specific changes in cognition and counteracts memory loss in ovariectomized mice

Abstract

The brain-specific enzyme CYP46A1 controls cholesterol turnover by converting cholesterol into 24S-hydroxycholesterol (24OH). Dysregulation of brain cholesterol turnover and reduced CYP46A1 levels are observed in Alzheimer's disease (AD). In this study, we report that CYP46A1 overexpression in aged female mice leads to enhanced estrogen signaling in the hippocampus and improved cognitive functions. In contrast, age-matched CYP46A1 overexpressing males show anxiety-like behavior, worsened memory, and elevated levels of 5α-dihydrotestosterone in the hippocampus. We report that, in neurons, 24OH contributes to these divergent effects by activating sex hormone signaling, including estrogen receptors. CYP46A1 overexpression in female mice protects from memory impairments induced by ovariectomy while having no effects in gonadectomized males. Last, we measured cerebrospinal fluid levels of 24OH in a clinical cohort of patients with AD and found that 24OH negatively correlates with neurodegeneration markers only in women. We suggest that CYP46A1 activation is a valuable pharmacological target for enhancing estrogen signaling in women at risk of developing neurodegenerative diseases.

Fig. 1.. Behavioral tests in aged Cyp46Tg mice.

(A) Experimental design: Aging female and male Cyp46Tg mice were assessed by elevated plus maze (EPM), Y-maze, and Morris water maze (MWM) tests. After the behavioral tests, the brains were collected for further molecular analysis. The experimental design was created with Biorender.com. (B) Percentage of time spent in the open arms during the EPM test in Cyp46Tg and Tg⁻ females. (C) Percentage of spontaneous alternations during the Y-maze test in female Cyp46Tg and Tg⁻ mice (one-sample t test, P = 0.022). (D) EPM test represented as the percentage of time spent in open arms in Cyp46Tg and Tg⁻ males (unpaired t test, P = 0.0047). (E) Percentage of spontaneous alternations in Y-maze test in male Cyp46Tg and Tg⁻ mice. (F) Escape latency over 4 days acquisition phase in the MWM test in Cyp46Tg and Tg⁻ females [repeated-measures analysis of variance (ANOVA), the effect of days P = 0.0001]. Adapted from data previously shown in (4). (G) Time spent in the platform quadrant during the probe test in Cyp46Tg and Tg⁻ females (unpaired t test, P = 0.09). Adapted from data previously shown in (4). (H) Escape latency over the 4-day acquisition phase in the MWM test in Cyp46Tg and Tg⁻ males (repeated-measures ANOVA, the effect of days P = 0.0094). (I) Time spent in the platform quadrant during the probe test (unpaired t test, P = 0.0130). (J) Representative heatmaps of the MWM probe test in female and male Cyp46Tg and Tg⁻ mice. Red zones display the areas mostly explored by the mice. N = 6 to 10 mice per group per sex. Data are represented as means ± SEM. *P < 0.01 and **P < 0.001.

Fig. 2.. Golgi staining in the hippocampus of Cyp46Tg aged mice.

(A to L) Representative Golgi staining images from stratum radiatum from the CA1 region of 18-month-old female (left) and male (right) Cyp46Tg mice and their corresponding age-matched Tg⁻ controls. Scale bars, 500 μm [(A) to (D)] and 5 μm [(E) to (L)]. (M) Analysis of dendritic spine density measured as spines per micrometer in female Cyp46Tg compared to Tg⁻ mice. (N) Spine area in square micrometer in Cyp46Tg and Tg⁻ females (unpaired t test, P = 0.0416). (O) Spine length in female Cyp46Tg and Tg⁻ mice (unpaired t test, P = 0.0384). (P) Frequency distribution histogram of spine area in female mice (two-sample Kolmogorov-Smirnov test, P < 0.0001). (Q) Frequency distribution histogram of spine length of female Cyp46Tg and Tg⁻ mice (two-sample Kolmogorov-Smirnov test P < 0.001). (R) Dendritic spine density in male Cyp46Tg and Tg⁻ mice (unpaired t test, P = 0.0122). (S) Spine area in male Cyp46Tg compared to male Tg⁻ mice. (T) Spine length in male Cyp46Tg and Tg⁻ mice. (U) Spine area frequency distribution histogram of male Cyp46Tg and Tg⁻ mice (two-sample Kolmogorov-Smirnov test, P = 0.0480). (V) Frequency distribution histogram of spine length from apical collateral dendrites of male Cyp46Tg and Tg⁻ mice (two-sample Kolmogorov-Smirnov test, P = 0.0162 respectively). N = 5 to 6 mice per group per sex. Data are represented as means ± SEM. *P < 0.05 and ***P < 0.001.

Fig. 3.. Sex hormone signaling in the hippocampus of 18month-old Cyp46Tg mice.

(A) Scheme representing the neurosteroid synthesis starting from cholesterol. The genes up-regulated in Cyp46Tg mice are indicated in orange, while the down-regulated ones are indicated in purple. (B) Enzyme-linked immunosorbent assay (ELISA) measurements of E2 levels in the hippocampus from female Cyp46Tg and Tg⁻ mice. (C) DHT levels from male Cyp46Tg and Tg⁻ hippocampi assessed by ELISA (Mann-Whitney test, P = 0.0451). (D and E) Cyp19a1 levels in Cyp46Tg females and males in comparison to their Tg⁻ controls (unpaired t test, P = 0.0053 in females). (F and G) Srd5a1 in Cyp46Tg females and males in comparison to their Tg⁻ controls (unpaired t test, P = 0.00289 and P = 0.0193, respectively). (H and I) Hsd17b10 in Cyp46Tg females and males in comparison to their Tg⁻ controls (unpaired t test, P = 0.0134 and P = 0.0081, respectively). (J and K) Esr2 in Cyp46Tg females and males in comparison to their Tg⁻ controls (unpaired t test, P = 0.0367 in females). (L and M) Esr1 in Cyp46Tg and Tg⁻ females and males. (N and O) Serum levels of 24OH in female and male Cyp46Tg and Tg⁻ mice (unpaired t test, P = 0.0036 and P = 0.0136, respectively). (P and Q) Serum cholesterol levels in aged female and male Cyp46Tg and Tg⁻ mice. (R and S) Brain lanosterol levels in aged female and male Cyp46Tg and Tg⁻ mice (unpaired t test, P = 0.0395 in females). Gene expression is normalized by Gapdh. N = 5 to 11 mice per group per sex for RT-qPCR and N = 4 mice per group per sex for gas chromatography–mass spectrometry analysis. All data are presented as means ± SEM. *P < 0.05 and **P < 0.01

Fig. 4.. 24OH, DHT, and DHT + 24OH treatments in primary neurons.

(A to I) Relative expression levels of different genes normalized by Gapdh in hippocampal neurons treated with 24OH, DHT, DHT + 24OH, or vehicle (CON): (A) Esr2 levels (one-way ANOVA P = 0.0030 followed by Tukey’s multiple comparisons test: CON versus 24OH P = 0.0497, 24OH versus DHT P = 0.0025, 24OH versus DHT + 24OH P = 0.0161). (B) Ers1 levels (C) Arc levels (one-way ANOVA P = 0.0021 followed by Tukey’s multiple comparisons test: CON versus 24OH P = 0.0364, CON versus DHT + 24OH P = 0.0449, 24OH versus DHT P = 0.0040, DHT versus DHT + 24OH P = 0.0048). (D) Bdnf levels. (E) Srebf1 levels (one-way ANOVA P = 0.0078 followed by Tukey’s multiple comparisons test: CON versus 24OH P = 0.0161, 24OH versus DHT P = 0.0220, 24OH versus DHT + 24OH P = 0.0111). (F) Rara levels (one-way ANOVA P = 0.0005 followed by Tukey’s multiple comparisons test: CON versus 24OH P = 0.0029, 24OH versus DHT P = 0.0010, 24OH versus DHT + 24OH P = 0.0008). (G) Cyp26b1 levels (one-way ANOVA P = 0.0304 followed by Tukey’s multiple comparisons test: CON versus 24OH P = 0.0491, 24OH versus DHT P = 0.0476). (H) Cyp19a1 levels (one-way ANOVA P < 0.0001 followed by Tukey’s multiple comparisons test: CON versus 24OH P < 0.0001, CON versus DHT +24OH P < 0.0001, 24OH versus DHT + 24OH P < 0.0001, DHT versus DHT + 24OH P = 0.0001). (I) Hsd17b1 levels (one-way ANOVA P = 0.0449). (J) Proposed mechanism of action for 24OH in aged female (left) and male (right) Cyp46Tg mice: In neurons, 24OH binds to LXR, activating the RARE promoter and RARA. 24OH induces neurosteroidogenesis via LXR/RARA, enhancing estrogen receptor (ER) signaling to promote neuroprotection in transgenic females. In contrast, the presence of high levels of DHT + 24OH in male mice counteracts 24OH up-regulating effects on LXR target genes (Srepf1, Rara, and Cyp19a1) and Ers2. ImageJ was created with Biorender.com. N = 3 to 5 independent experiments, where each experiment was performed in duplicates or triplicates. All data are presented as means ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Fig. 5.. Behavioral tests in gonadectomized mice.

(A) Experimental design: Female and male Cyp46Tg and wild-type mice underwent gonadectomy at 2 to 3 months of age. At 9 months, the mice were assessed for behavioral tests, after their brains were collected for molecular analysis. The experimental design was created with Biorender.com. (B and C) Percentage of time spent in open arms during the EPM test in the female and male groups, respectively. (D) Percentage of spontaneous alternations in the Y-maze test in female mice (one-way ANOVA P = 0.0255 followed by Tukey’s multiple comparisons test: Wt versus GDX-Cyp46Tg P = 0.0202). (E) Percentage of spontaneous alternations in the Y-maze test in male mice. (F) Escape latencies during MWM acquisition phase in the female groups (repeated-measures ANOVA, the effect of days P = 0.0018, effect of genotype P = 0.0030, followed by Tukey’s multiple comparison test: for day 3 GDX Wt versus GDX Cyp46Tg P = 0.0096, for day 5 Wt versus GDX Wt P = 0.0217, GDX Wt versus GDX Cyp46Tg P = 0.0065). (G) Percentage of time spent in the quadrant target during probe test in female mice (one-way ANOVA P = 0.0448, followed by Tukey’s multiple comparisons test: GDX-Wt versus GDX-Cyp46Tg P = 0.0441). (H) MWM acquisition phase expressed as escape latency in the male mice groups (repeated-measures ANOVA, the effect of days P < 0.0001). (I) Percentage of time spent in the quadrant target during the probe test in male mice. (J) Heatmaps from the MWM probe test in female and male mice: The occupancy rate is represented by a color map (in red color the most visited zones). N = 9 to 13 mice per group per sex. Following the exclusion criteria, we removed four males and four females from the analyses of % alternation in the Y-maze and four females and two males from the MWM probe test analyses. All data are presented as means ± SEM. *P < 0.05.

Fig. 6.. Neurosteroid signaling, 24OH, and CYP46A1 levels in gonadectomized mice.

(A) Hippocampal levels of Esr2 in female GDX-Cyp46Tg, GDX-Wt, and Wt mice. (B) Hippocampal levels of Esr1 in female mice (Kruskal-Wallis test, P = 0.0208, followed by Dunn’s multiple comparisons test: Wt versus GDX-Wt P = 0.0242). (C) Hippocampal levels of Cyp19a1 in female mice. (D) Hippocampal levels of Srd5a1 in female mice (Kruskal-Wallis test P = 0.0275, followed by Dunn’s multiple comparisons test: Wt versus GDX-Wt P = 0.0283). (E) Hippocampal levels of Hsd17b10 in female mice (one-way ANOVA P = 0.0032, followed by Tukey’s multiple comparison test: Wt versus GDX-Cyp46Tg P = 0.00393 and Wt-GDX versus GDX-Cyp46Tg P = 0.0027). (F) Hippocampal levels of Esr2 in male mice. (G) Hippocampal levels of Esr1 in male mice. (H) Hippocampal levels of male Cyp19a1 mice. (I) Hippocampal levels of Srd5a1 in male mice. (J) Hippocampal levels of Hsd17b10 in male mice (one-way ANOVA P = 0.0212, followed by Tukey’s multiple comparisons test Wt versus GDX-Wt P = 0.0163). (K) 24OH levels in serum from female GDX-Cyp46Tg, GDX-Wt, Cyp46Tg, and Wt mice. (L) 24OH brain levels from female mice. (M) Cyp46a1 levels in female GDX-Wt and Wt mice and CYP46A1 levels in GDX-Cyp46Tg mice compared to age-matched Cyp46Tg females. (N) Serum 24OH levels from male mice (one-way ANOVA P = 0.0043). (O) Brain 24OH levels in male mice (One-way ANOVA P = 0.0024). (P) Cyp46a1 levels in male GDX-Wt and Wt mice and CYP46A1 levels in male GDX-Cyp46Tg mice compared to age-matched Cyp46Tg mice (unpaired t test P = 0.0212 and P = 0.0005 respectively). N = 4 to 6 mice per group per sex. All data are presented as means ± SEM. *P < 0.05, **P < 0.01, and ****P < 0.0001.