Neuroscientists as cartographers: mapping the crossroads of gonadal hormones, memory and age using animal models

Abstract

Cognitive function is multidimensional and complex, and research in multiple species indicates it is considerably impacted by age and gonadal hormone milieu. One domain of cognitive function particularly susceptible to age-related decrements is spatial memory. Gonadal hormones can alter spatial memory, and they are potent modulators of brain microstructure and function in many of the same brain areas affected by aging. In this paper, we review decades of animal and human literature to support a tertiary model representing interactions between gonadal hormones, spatial cognition and age given that: 1) gonadal hormones change with age, 2) age impacts spatial learning and memory, and 3) gonadal hormones impact spatial learning and memory. While much has been discovered regarding these individual tenets, the compass for future aging research points toward clarifying the interactions that exist between these three points, and understanding mediating variables. Indeed, identifying and aligning the various components of the complex interactions between these tenets, including evaluations using basic science, systems, and clinical perspectives, is the optimal approach to attempt to converge the many findings that may currently appear contradictory. In fact, as discoveries are being made it is becoming clear that the findings across studies that appear contradictory are not contradictory at all. Rather, there are mediating variables that are influencing outcome and affecting the extent, and even the direction, of the effects that gonadal hormones have on cognition during aging. These mediating variables are just starting to be understood. By aligning basic scientific discoveries with clinical interpretations, we can maximize the opportunities for discoveries and subsequent interventions to allow individuals to "optimize their aging" and find their own map to cognitive health as aging ensues.

Figure 1

Tertiary model representing interactions between gonadal hormones, memory and age.

Figure 2

Age alters the ability of animals to handle an increasing working memory load [35]. a) Mean ± SE number of spatial working memory errors committed by young and aged rats averaged into Low, Moderate, and High load blocks. Also depicted is the linear trend for each block, for each group. Both Young and Aged groups exhibited significant linear trends on the Low load block (regression equations: Aged: y = -0.042 x + 0.600, Young: y= -0.058 x+.639), while only the Young group showed a significant linear trend on Moderate and High load blocks (ys = -0.058 x + 1.12 and -0.176 x + 3.76, respectively). Findings suggest that both young and aged animals could learn to handle a low memory load and that while the young animals could also learn to handle moderate and high memory loads, the aged animals could not. b) Mean ± SE number of spatial working memory errors on Low, Moderate, and High load blocks, averaged across the latter testing days, for young and aged rats. As trials increased, the number of elements of information to be remembered increased. Aged animals committed disproportionately more errors on the latter trials, when the spatial working memory load was highest.

Figure 3

Spatial memory performance in several experimentally-induced forms of ovarian hormone loss, VCD-induced follicular depletion (gradual, transitional menopause) and Ovx (surgical menopause), in the middle-aged rat [87]. Also included was a group that received VCD followed by Ovx to model transitional menopause followed by surgical menopause, thereby removing residual hormones released from the follicle deplete ovary. a) After water radial-arm maze initial testing, a 4-hr delay between trials 2 and 3 was imposed. The graph depicts mean ± SE spatial working memory errors for each group for post-delay trials. The VCD-Ovx group made fewer errors relative to all groups, suggesting that transitional ovarian hormone loss before Ovx aids memory retention, and that removal of residual ovarian tissue after transitional ovarian hormone loss improves memory retention relative to retaining residual ovarian tissue, * p < 0.0005. b) Initial baseline performance on the delayed-match-to-sample spatial recent memory task. Trial 1 was the information trial, informing the animal where the platform was located on that test day. Shown are the mean±SE errors for each group for blocks 1 and 2 for the test trials. SHAM and VCD-Ovx groups made fewer errors during initial learning in block 1, relative to the VCD group, * p < 0.05, suggesting that transitional menopause without Ovx impaired performance, while Ovx initiated after the transition enhanced performance. The Ovx group was marginally different from the VCD group, # p = 0.06.

Figure 4

Age alters responsiveness to Ovx and 17β-estradiol treatment, and serum levels of 17β-estradiol correlate with memory scores in young and middle-aged rats [53]. Mean ± SE distance scores (cm) on the reference memory Morris maze for a) young Sham, Ovx and OvxE groups, b) middle-aged Sham, Ovx and OvxE groups, and c) aged Sham, Ovx, and OvxE groups; the inset graphs represent data collapsed across days. Ovx impaired performance on the latter portion of testing for young animals, while there was no effect of Ovx in middle-aged animals. 17β-estradiol treatment enhanced performance in both young and middle-aged groups. For aged Sham, Ovx, and OvxE groups there were no significant main effects or interactions for distance scores due to Ovx or 17β-estradiol replacement. Thus, there was a dissociation between sensitivity to Ovx and responsiveness to estrogen treatment. * p < 0.05, # p = 0.06, *** p < 0.0001. d) For the scatterplot, young animals are represented by filled circles, and middle-aged animals by unfilled squares. Higher circulating 17β-estradiol levels were correlated with better spatial reference memory, as tested on the Morris maze.

Figure 5

Progesterone reverses 17β-estradiol’s effects on spatial memory and neurotrophins [97,184]. a) Mean ± SE distance-to-platform (inches) on the Morris maze task for each treatment group. Tonic and cyclic 17β-estradiol treatments enhanced spatial reference memory performance in middle-aged Ovx rats, while the addition of progesterone reversed these beneficial cognitive effects. *p < 0.05. b) Mean ± SE protein levels of BDNF, NGF, and NT3 in entorhinal cortex for aged Ovx rats receiving no hormone (Aged Ovx), a 17β-estradiol pellet (Aged Ovx E), a 17β-estradiol pellet plus two progesterone pellets (Aged Ovx EP), and a young group also receiving Ovx surgery, with no hormone treatment (Young Ovx). Findings suggest that for all three neurotrophins, 17β-estradiol increased protein levels, while the addition of progesterone reversed these increases.# p < 0.07; * p < 0.05; ** p < 0.01; *** p < 0.0001.

Figure 6

Progesterone reverses the cognitive enhancing effects of Ovx in aged animals, an effect paralleled by the synthetic progestin medroxy progesterone acetate (MPA) [58,91]. a) and b) represent the mean ± SE number of working memory correct and incorrect errors, respectively, committed on the water radial-arm maze for Young, Aged Sham, Aged Ovx, and Aged Ovx P (progesterone) treated animals [for error definitions see 58]. Young Sham and Aged Ovx rats were better able to sustain successful performance as trials and working memory load increased, compared to Aged Sham and Aged Ovx rats treated with progesterone (Aged Ovx P). Thus, on the two orthogonal spatial working memory measures, aged animals showed impaired performance as memory load increased relative to young animals, and Ovx enhanced performance in aged rats. Progesterone abolished the Ovx-induced enhancement. c) represents the working memory load effect whereby Sham, OvxP, and OvxMPA animals exhibited more mean±SE errors than Ovx animals on the last trial, indicating that animals treated with progesterone or MPA were impaired as the demand on working memory increased. This replicated prior findings that Ovx enhanced, and progesterone impaired, performance as memory load increased, and showed that the synthetic progestin, MPA, impaired spatial working memory as memory load increased in aged surgically menopausal rats.

Figure 7

Testosterone supplementation decreased spatial working memory errors in aged male gonadally intact rats [215]. Rats were not castrated in this study in an attempt to conserve the hypothalamic/pituitary/gonadal axis, as this axis is typically conserved in older men. The administered androgen treatments were thus supplemental therapies. The aged male rats were divided into three treatment groups (n = 11 per group): sham (Aged Sham), dihydrotestosterone (Aged DHT), testosterone (Aged T). There was also a young male group which received sham surgery. Aged sham rats made more mean ± SE spatial working memory errors than young sham rats, and aged rats treated with testosterone (Aged T) committed fewer spatial working memory errors than aged rats receiving vehicle treatment (Aged Sham). * p < 0.05, *** p < 0.0005

Figure 8

Circulating LH levels show a curvilinear relationship with memory scores [87]. Regression analysis indicated that in animals without ovaries (Ovx and VCD-Ovx), higher LH was associated with better reference memory for the a) Morris maze and, b) water radial-arm maze. As shown in c), in animals with ovaries (SHAM and VCD), higher LH was associated with worse working memory; indeed, higher LH was associated with more water radial-arm maze working memory errors. Graphically, the significant regression coefficients for these analyses are shown as solid lines; the dashed lines of the other groups are shown for comparison purposes to aid interpretation.

Figure 9

Premarin (CEE) treatment protects against scopolamine-induced amnesia and increases the number of ChAT positive neurons in the VDB of the basal forebrain in middle-aged Ovx rats [165]. Middle-aged Ovx rats received either chronic cyclic treatment with vehicle injection (sesame oil, Ovx-Oil) or one of the following amounts of injectable Premarin (as prescribed to women): 10 (Ovx-Prem-Low), 20 (Ovx-Prem-Medium) or 30 μg (Ovx-Prem-High). a) Mean ± SE total number of errors for each treatment group after scopolamine (0.2 mg/kg, IP) challenge. Premarin dose-dependently protected against scopolamine-induced amnesia on the delayed-match-to-sample working memory plus maze. b) Mean ± SE ChAT immunohistochemically stained cells were quantified in the basal forebrain in behaviorally tested animals. All Premarin-treated groups were collapsed and compared to the Ovx-Oil group. In the VDB, Premarin-treated animals had more ChAT-immunoreactive neurons relative to the Ovx-Oil group, ps< 0.05. Photomicrographs showing ChAT-immunoreactive neurons in sections through the BF from animals in the c) Ovx-Oil, d) Ovx-Prem-Low, e) Ovx-Prem-Medium, f) Ovx-Prem-High.