Clinically Used Hormone Formulations Differentially Impact Memory, Anxiety-Like, and Depressive-Like Behaviors in a Rat Model of Transitional Menopause

Abstract

A variety of U.S. Food and Drug Administration-approved hormone therapy options are currently used to successfully alleviate unwanted symptoms associated with the changing endogenous hormonal milieu that occurs in midlife with menopause. Depending on the primary indication for treatment, different hormone therapy formulations are utilized, including estrogen-only, progestogen-only, or combined estrogen plus progestogen options. There is little known about how these formulations, or their unique pharmacodynamics, impact neurobiological processes. Seemingly disparate pre-clinical and clinical findings regarding the cognitive effects of hormone therapies, such as the negative effects associated with conjugated equine estrogens and medroxyprogesterone acetate vs. naturally circulating 17β-estradiol (E2) and progesterone, signal a critical need to further investigate the neuro-cognitive impact of hormone therapy formulations. Here, utilizing a rat model of transitional menopause, we administered either E2, progesterone, levonorgestrel, or combinations of E2 with progesterone or with levonorgestrel daily to follicle-depleted, middle-aged rats. A battery of assessments, including spatial memory, anxiety-like behaviors, and depressive-like behaviors, as well as endocrine status and ovarian follicle complement, were evaluated. Results indicate divergent outcomes for memory, anxiety, and depression, as well as unique physiological profiles, that were dependent upon the hormone regimen administered. Overall, the combination hormone treatments had the most consistently favorable profile for the domains evaluated in rats that had undergone experimentally induced transitional menopause and remained ovary-intact. The collective results underscore the importance of investigating variations in hormone therapy formulation as well as the menopause background upon which these formulations are delivered.

Keywords: VCD; anxiety; depression; estrogen; levonorgestrel; memory; menopause; progesterone.

FIGURE 1

Experimental Timeline. Following accelerated follicular depletion, rats received daily hormone treatments and were evaluated on a series of behavior tasks assessing working memory, reference memory, anxiety-like behavior, and depressive-like behavior.

FIGURE 2

Water Radial Arm Maze Error Subtype Learning Curves. (A) Working memory correct errors across days (B) Working memory incorrect errors across days (C) Reference memory errors across days. For all error types, Day 1 was considered Training and was excluded from data analysis. The Early Acquisition Phase was defined as Days 2–5, the Late Acquisition Phase was defined as Days 6–9, and the Asymptotic Phase was defined as Days 10–12. Performance for each error subtype was analyzed separately. The n/group for all WRAM two-group analyses were: VCD-Vehicle n = 10, VCD-E2 n = 10, VCD-PROG n = 9, VCD-LEVO n = 9, VCD-E2 + PROG n = 10, and VCD-E2 + LEVO n = 10.

FIGURE 3

Early Acquisition Phase RM Errors Across All Trials (Two-Group Comparisons). The VCD-E2 + LEVO group showed enhanced reference memory performance compared to the VCD-E2 group (p < 0.05) and compared to the VCD-E2 + PROG group (p < 0.01) during the Early Acquisition Phase. Significance: ^∗ = p < 0.05, ^∗∗ = p < 0.01.

FIGURE 4

WMC Errors During the Asymptotic Phase. (A) While there was no main effect of Treatment collapsed across trials for any two-group comparison, there were Trial × Treatment interactions between the VCD-LEVO and VCD-PROG groups (p < 0.05) and (B) VCD-PROG and VCD-E2 + PROG groups (p < 0.05). (C) When the High Load trials (Trials 3 plus 4) were assessed, VCD-PROG rats made more WMC errors than VCD-E2 + PROG rats (p < 0.05). Significance: ^∗ = p < 0.05.

FIGURE 5

WMI Errors During the Asymptotic Phase. (A) Across all trials, a main effect of Treatment was present between the VCD-PROG group and the VCD-Vehicle group (p < 0.05) as well as compared to the VCD-E2 + PROG group (p < 0.05). (B) VCD-PROG vs. VCD-Vehicle comparison: A Trial × Treatment interaction was present for this comparison (p < 0.05) (C) When High Load trials (Trials 3 + 4) were assessed, VCD-PROG rats made more WMI errors than VCD-Vehicle rats (p < 0.05). (D) VCD-PROG vs. VCD-E2 + PROG comparison: A Trial × Treatment interaction was present for this comparison (p < 0.05). (E) When High Load trials (Trials 3 + 4) were assessed, VCD-PROG rats made more WMC errors than VCD-E2 + PROG rats (p < 0.05). Significance: ^∗ = p < 0.05.

FIGURE 6

RM Errors During the Asymptotic Phase. (A) Across all trials, a main effect of Treatment was present between the VCD-PROG group and the VCD-E2 + PROG group (p < 0.05). (B) VCD-E2 vs. VCD-E2 + PROG comparison: A Trial × Treatment interaction occurred (p < 0.01). Significance: ^∗ = p < 0.05, ^∗∗ = p < 0.01.

FIGURE 7

WRAM Six-Hour Delayed Memory Retention Test. (A) The VCD-Vehicle group exhibited a delay-induced working memory impairment compared to the previous day’s baseline performance (p < 0.01). (B) The VCD-E2 group exhibited a delay-induced working memory impairment compared to the previous day’s baseline performance (p < 0.01). (C) The VCD-PROG group did not display a delay-induced working memory impairment compared to the previous day’s baseline performance. (D) The VCD-LEVO group did not display a delay-induced working memory impairment compared to the previous day’s baseline performance. (E) The VCD-E2 + PROG group exhibited a delay-induced working memory impairment compared to the previous day’s baseline performance (p < 0.05). (F) The VCD-E2 + LEVO group exhibited a delay-induced working memory impairment compared to the previous day’s baseline performance (p < 0.01). Significance: ^∗ = p < 0.05, ^∗∗ = p < 0.01.

FIGURE 8

Morris Water Maze. (A) Swim Distance to Platform Across All Days. (B) VCD-E2 treated rats swam less distance to the platform compared to the VCD-E2 + PROG treated rats (p < 0.05). (C) VCD-E2 treated rats swam less distance to the platform compared to the VCD-E2 + LEVO treated rats (p < 0.05). (D–I) Probe trial. All treatment groups swam a greater proportion of total distance in the previously platformed quadrant vs. the opposite quadrant, indicating that all groups spatially localized to the hidden platform location. Significance: ^∗ = p < 0.05, ^****p < 0.0001.

FIGURE 9

Visible Platform. (A) All subjects decreased latency to platform from the first to last trial. (B) Trial times (means + S.E.M.) for each treatment group are provided. Significance: ^∗ = p < 0.05, ^∗∗ = p < 0.01, ^ = p = 0.06.

FIGURE 10

Open Field Task. (A) Schematic of the OFT arena. Green squares indicate which boxes were defined as Corners, pink squares indicate which boxes were defined as Center, and yellow stripes indicate the Small Center, which was also included in the “Center” measure. (B) The VCD-Vehicle group spent more time in the corners compared to VCD-E2 and VCD-PROG groups. (C) VCD-E2 + LEVO treatment increased time spent in the center compared to VCD-E2 treatment. (D) No significant differences in Small Center Time were detected. (E) The VCD-E2 group had more fecal boli than the VCD-Vehicle and VCD-E2 + PROG group. (F) The VCD-E2 + LEVO group made more entries into the corner compared to the VCD-E2 group as well as the VCD-E2 + PROG group. (G) The VCD-E2 + LEVO group made more entries into the center compared to VCD-E2 treatment alone. (H) The VCD-E2 + LEVO group made more entries into the small center compared to VCD-Vehicle group, VCD-E2 group, and VCD-E2 + PROG group. (I) Total Line Crossing analyses indicate that the VCD-E2 + LEVO group moved more in the OFT compared to VCD-Vehicle, VCD-E2, and VCD-E2+PROG groups. Significance: ^∗ = p < 0.05, ^∗∗ = p < 0.01, ^∗∗∗ = p < 0.001.

FIGURE 11

Forced Swim Test. (A) Both combination hormone therapy groups had a longer latency to immobility when compared to VCD-Vehicle or VCD-E2 groups, suggesting an antidepressant-like effect of combination hormone therapy compared to no treatment or E2 treatment alone. (B) Total immobility was decreased in the combination hormone therapy groups, again suggesting an antidepressant-like effect compared to Vehicle treatment or LEVO-alone treatment. (C) No Treatment differences were indicated in time spent swimming. (D) The VCD-E2 + LEVO group spent more time climbing compared to the VCD-Vehicle group, indicating antidepressant-like effects. (E,F) Total Dive Count and Fecal Boli Counts did not differ among treatment comparisons. Significance: ^∗ = p < 0.05, ^∗∗ = p < 0.01.

FIGURE 12

Serum Hormone Levels. (A) E2 was elevated in VCD-E2, VCD-E2 + PROG, and VCD-E2 + LEVO groups compared to VCD-Vehicle rats. Additionally, combination hormone therapy groups had elevated E2 compared to their respective progestogen-only groups. E2 analysis n/group: VCD-Vehicle n = 9; VCD-E2 n = 10; VCD-PROG n = 9; VCD-LEVO n = 9; VCD-E2 + PROG n = 10; VCD-E2 + LEVO n = 10. (B) Progesterone was elevated in the VCD-PROG group and the VCD-E2 + PROG group compared to the VCD-Vehicle group, VCD-E2 group, and VCD-LEVO group. The combination hormone group had higher progesterone levels compared to the VCD-PROG group alone. Progesterone analysis n/group: VCD-Vehicle n = 10; VCD-E2 n = 10; VCD-PROG n = 9; VCD-LEVO n = 9; VCD-E2 + PROG n = 10; VCD-E2 + LEVO n = 10. (C) All subjects in the VCD-E2 group and VCD-E2 + LEVO group had undetectable levels of androstenedione. Androstenedione was elevated in the VCD-PROG group compared to VCD-Vehicle, VCD-LEVO, and VCD-E2 + PROG groups. Androstenedione analysis n/group: VCD-Vehicle n = 9; VCD-E2 n = 0 [undetectable]; VCD-PROG n = 9; VCD-LEVO n = 9; VCD-E2 + PROG n = 10; VCD-E2 + LEVO n = 0 [undetectable]. Significance: ^∗∗ = p < 0.01, ^∗∗∗ = p < 0.001, ^**** = p < 0.0001.

FIGURE 13

Ovarian Follicle Counts. An independent ovary-intact reference group (n = 10) is included to assess successful follicular depletion following VCD treatment. The letter “a” indicates that this ovary-intact reference group was significantly different from each VCD-treated group. (A) Estimated primordial follicle counts were decreased in the VCD-E2 group compared to the VCD-Vehicle group and the VCD-E2 + LEVO group. (B) Primary follicles were decreased in the VCD-E2 group compared to the VCD-Vehicle group, replicating prior work. (C) Secondary follicle counts were significantly depleted in VCD-treated groups, indicating successful accelerated follicular atresia. (D) Antral follicle counts were significantly depleted in VCD-treated groups, indicating successful accelerated follicular atresia. (E) The VCD-E2 + PROG group had more corpora lutea compared to the VCD-E2 group, suggesting occasional ovulatory cycles in this group during the transition to reproductive senescence. Significance: ^∗ = p < 0.05, ^∗∗ = p < 0.01.

FIGURE 14

Peripheral markers of overall health and uterine stimulation. (A) Body weight changes across the experimental timeline (B) At the end of the experiment, the VCD-E2 + PROG group weighed less than the VCD-Vehicle group and the VCD-PROG group, suggesting combination hormone therapy promotes weight maintenance compared to no hormone therapy treatment or progesterone treatment alone. The VCD-E2 + LEVO group also weighed less than its VCD-LEVO alone counterpart, again suggesting combination hormone therapy promotes weight maintenance. (C) PROG treatment reduced uterine weight compared to VCD-Vehicle, VCD-LEVO, and VCD-E2 + PROG groups. VCD-E2 + PROG uterine weight was attenuated compared to VCD-E2 treatment along, suggesting progesterone blocked uterine proliferation. The VCD-E2 + LEVO group uteri weighed more than those in the VCD-E2 + PROG group, indicating less progestin-induced attenuation of uterine stimulation compared to natural progesterone when in a combined hormone therapy regimen. Significance: ^∗ = p < 0.05, ^∗∗ = p < 0.01, ^**** = p < 0.0001.