Causes and Consequences of Age-Related Steroid Hormone Changes: Insights Gained from Nonhuman Primates


  • K. G. Sorwell,

    1. Departments of Neuroscience and Reproductive & Developmental Sciences, Oregon National Primate Research Center, Oregon Health & Science University, Beaverton, OR, USA
    2. Department of Behavioral Neuroscience, Oregon Health & Science University, Portland, OR, USA
    Search for more papers by this author
  • H. F. Urbanski

    Corresponding author
    1. Departments of Neuroscience and Reproductive & Developmental Sciences, Oregon National Primate Research Center, Oregon Health & Science University, Beaverton, OR, USA
    2. Department of Behavioral Neuroscience, Oregon Health & Science University, Portland, OR, USA
    • Correspondence to: Henryk F. Urbanski, Division of Neuroscience, Oregon National Primate Research Center, Oregon Health & Science University, Beaverton, OR 97006, USA (e-mail:

    Search for more papers by this author


Similar to humans, rhesus macaques (Macaca mulatta) are large, long-lived diurnal primates, and show similar age-related changes in the secretion of many steroid hormones, including oestradiol, testosterone, cortisol and dehydroepiandrosterone (DHEA). Consequently, they represent a pragmatic animal model in which to examine the mechanisms by which these steroidal changes contribute to perturbed sleep–wake cycles and cognitive decline in the elderly. Using remote serial blood sampling, we have found the circulating levels of DHEA sulphate, as well as oestradiol and testosterone, decline markedly in old monkeys. Furthermore, using the real-time polymerase chain reaction, we have shown that the genes for the enzymes associated with the conversion of DHEA to oestradiol and testosterone (3β-hydroxysteroid dehydrogenase, 17β-hydroxysteroid dehydrogenase, and aromatase) are highly expressed in brain areas associated with cognition and behaviour, including the hippocampus, prefrontal cortex and amygdala. Taken together, these findings suggest that the administration of supplementary DHEA in the elderly may have therapeutic potential for cognitive and behavioural disorders, although with fewer negative side effects outside of the central nervous system. To test this, we have developed a novel steroid supplementation paradigm for use in old animals; this involves the oral administration of DHEA and testosterone at physiologically relevant times of the day to mimic the circadian hormone patterns observed in young adults. We are currently evaluating the efficacy of this steroid supplementation paradigm with respect to reversing age-associated disorders, including perturbed sleep–wake cycles and cognitive decline, as well as an impaired immune response.

Rhesus macaques (Macaca mulatta) have many attributes that make them suitable for translational neuroendocrine studies. Similar to humans, they are large, diurnal, long-lived primates, with a similar brain morphology and organisation of key neuroendocrine systems. Additionally, during ageing, they show many changes in their physiology, cognition, behaviour and immune function that are similar to those observed in elderly humans [1]. Furthermore, rhesus macaques can be readily maintained under highly-controlled environmental conditions (e.g. photoperiod, ambient temperature, diet and medication), and they can yield high-quality post-mortem tissue from scheduled necropsies. Consequently, they represent a pragmatic animal model in which to investigate the neuroendocrine mechanisms that underlie normal and pathological human ageing. In this review, we highlight some of the key similarities between the neuroendocrine systems of rhesus macaques and humans, and focus on novel insights that have been gained from using this translational nonhuman primate model in ageing research.

Neuroendocrine ageing in humans and nonhuman primates

Communication between different organ systems is essential for normal physiological functions, and this relies on both neuronal and endocrine signalling. From the perspective of primate ageing, many steroid hormones from the gonads (oestradiol, progesterone and testosterone) and adrenal gland [cortisol and dehydroepiandrosterone (DHEA)] are particularly important [2]. Not only do they show marked age-related changes in their circulating levels, but also many of them show attenuation of their 24-h pattern of release [3, 4], which may contribute to the aetiology of perturbed circadian rhythms such as the sleep–wake cycle. In turn, a lack of sleep has been linked to poor cognitive performance and deficits in attention and executive function [5], as well as to an impaired immune response [6-9]. Although the underlying causal mechanism is still poorly understood, human studies have recently shown that insufficient sleep can significantly affect the expression of genes associated with inflammatory, immune and stress responses, amongst other biological processes [10].

The hypothalamic-pituitary-gonadal (HPG) axis and the hypothalamic-pituitary-adrenal (HPA) axis of humans and rhesus macaques are remarkably similar. Within the HPG axis, gonadotrophin-releasing hormone (GnRH) serves as the primary neuroendocrine link between the brain and the anterior pituitary gland, stimulating the secretion of luteinising hormone (LH) and follicle-stimulating hormone (FSH). Interestingly, rhesus macaques and humans are one of the few mammalian species in which two distinct molecular forms of GnRH have been identified (GnRH-I and GnRH-II), suggesting that different subpopulations of GnRH neurones contribute differentially to the regulation of reproductive function in primates [11]. The two pituitary gonadotrophins in turn stimulate gametogenesis and sex-steroid hormone production within the ovary (oestradiol and progesterone) and testis (testosterone). The coordinated release of these reproductive hormones is essential for the onset of puberty and for the subsequent maintenance of fertility in adults, as well as for the development and maintenance of secondary sexual characteristics and other physiological functions. The similarity between the HPG axis of women and adult female rhesus macaques is emphasised by the similar hormonal changes that occur across the menstrual cycle, and the subsequent precipitous loss of sex-steroid output after menopause [12, 13] (Fig. 1), which is considered to be triggered by the loss of ovarian follicles [14, 15], as well as reduced responsiveness of the hypothalamus to GnRH and oestrogen feedback [16-18]. This decline in sex-steroid levels is associated with a decrease in the output of other ovarian hormones, such as anti-Müllerian hormone and inhibin B, as well as a concomitant increase in circulating LH and FSH levels as a result of the loss of negative-feedback to the hypothalamus and pituitary gland [13, 19]. This age-related change within the primate HPG axis differs greatly from that observed in rodents, in which the first sign of reproductive senescence appears to be an attenuation of GnRH signalling [20, 21], resulting in a dampened and delayed preovulatory LH surge [22, 23], further supporting the use of the nonhuman primate as a model for human ageing over that of the rodent.

Figure 1.

Reproductive serum hormone profiles of representative pre-menopausal and post-menopausal rhesus macaques. Plasma oestradiol, progesterone, luteinising hormone (LH) and follicle-stimulating hormone (FSH) values were determined every 2 days for approximately 70 consecutive days. Shaded vertical bars represent days of menstruation. Note the marked attenuation of circulating oestradiol and progesterone levels after menopause, with a compensatory increase in LH and FSH levels. Adapted from Downs and Urbanski [13].

Although male primates show an age-related decline in circulating testosterone levels, this is generally very gradual and less extreme than the precipitous decrease of oestradiol observed in females around the time of menopause [4, 24, 25]. What makes the detection of age-related changes in testosterone output particularly difficult to study, however, is its episodic pattern of release, which is driven by pulsatile secretion of GnRH and LH every few hours. In addition, circulating testosterone levels are characterised by a distinctive 24-h release pattern, which means that single measurements of testosterone in the circulation are unreliable indicators of overall testosterone output. To overcome this problem in rhesus macaques, we have used a remote blood sampling system to serially collect blood samples from young and old males across the 24-h day [26] and have detected a significant age-related decline (Fig. 2a). This was reflected as a significant age-associated decrease in the overall mean, maximum and minimum testosterone levels [4], similar to circadian sampling studies previously reported in humans [27, 28]. It should be emphasised, however, that the attenuated testosterone levels of the old males were still considerably higher than those typically observed before puberty [29], and so the physiological impact of declining testosterone levels during normal ageing is unclear.

Figure 2.

Age-related changes in the 24-h plasma concentrations of (a) testosterone, (b) dehydroepiandrosterone sulphate (DHEAS) and (c) cortisol in male rhesus macaques. Mean hormone profiles are shown from ten adult (~10 years; shown in red) and ten aged (~26 years; shown in black) animals, and the horizontal light and dark bars correspond to the 12 : 12 h light/dark cycle; note that the profiles have been double plotted to facilitate observation of day–night differences. Each hormone showed a distinct 24-h rhythm, with a peak occurring either during the night (testosterone) or early in the morning (cortisol and DHEAS). Both testosterone and DHEAS showed a significant age-related attenuation in the plasma levels, whereas cortisol showed a significant increase. Adapted from Urbanski and Sorwell [9].

Within the HPA axis, corticotrophin-releasing hormone (CRH) from the paraventricular nucleus of the hypothalamus stimulates the release of adrenocorticotrophic hormone (ACTH) from the anterior pituitary gland. In turn, ACTH signals the adrenal cortex to release cortisol from the zona fasciculata and DHEA from the zona reticularis. Under normal circumstances, cortisol exerts negative feedback on both the hypothalamus and pituitary to reduce the secretion of CRH and ACTH, respectively. Although acute increases in cortisol can be adaptive in times of stress, prolonged increases can result in hippocampal excitotoxicity [30-32] and oxidative damage [32, 33]. In addition, high levels of cortisol can decrease hippocampal volume and interfere with the structural changes necessary for learning and memory [34-37]. Because the hippocampus, itself, responds to glucocorticoids by exerting additional negative-feedback on the hypothalamus, these changes result in a disruption of HPA axis activity and further elevations in cortisol [35]. DHEA, meanwhile, can act as a ‘functional antagonist’ of cortisol [38, 39], in part by promoting neuronal and glial survival [40, 41]. An increase in circulating cortisol with advanced age has been observed in both humans [42, 43] and nonhuman primates [44], with a concurrent marked decline in DHEAS (the sulphated form of DHEA) throughout adulthood [44-47] (Fig. 2b, c). This resulting increase in the cortisol : DHEA ratio may have drastic implications for many physiological process, including learning and memory [38, 39], a view that is supported by the finding that higher cortisol : DHEA ratios are associated with greater cognitive impairment [48, 49].

Impact of age-related hormonal changes on cognitive function

Close associations exist between age-related hormonal changes and cognitive decline, even in healthy individuals [50, 51]. Consequently, extensive studies have examined not only the effect of steroid hormone deprivation on learning and memory, but also the therapeutic potential of hormone supplementation. Circulating testosterone levels show an age-related decline in men [28, 52, 53], as well as in male nonhuman primates [4, 24, 25], and aged men with higher levels of endogenous testosterone exhibit greater cognitive performance [54-56]. Additionally, testosterone supplementation in men with low endogenous testosterone levels has been shown to improve certain aspects of cognitive function [57-59]. Although the exact mechanism is unclear, it is possible that the beneficial effects of supplementation are mediated by conversion of testosterone to oestradiol. This rationale is based on the observation that men undergoing androgen deprivation therapy experience cognitive deficits that can be rescued by oestradiol supplementation [50], and that aromatisation of testosterone to oestradiol appears to be necessary for some of the cognitive benefits of testosterone supplementation [60]. In women, oestradiol deprivation is associated with cognitive impairments [61, 62], which can be overcome by oestrogen therapy [62, 63]. As emphasised in Fig. 2, the adrenal steroids cortisol and DHEAS both have significant age-related patterns of secretion, and the endogenous cortisol : DHEA ratio is associated with cognitive performance in aged humans [48, 49]. Interestingly, however, there is little evidence from clinical studies that DHEA supplementation incurs cognitive benefits in the elderly [64-69], despite improvements in episodic memory performance in young men [70]. High endogenous levels of cortisol alone have also been associated with cognitive impairment in both middle age [71] and old age [72-74].

Circadian perturbations associated with age

Rhesus macaques, similar to humans, are diurnal and generally confine their eating and activity to daylight hours. Although the underlying mechanism is still unclear, there is a wealth of evidence showing that the suprachiasmatic nucleus (SCN) of the hypothalamus plays a major role in controlling circadian rhythms in mammals [75, 76]. Furthermore, there is now evidence from several species, including the rhesus macaque, indicating that many peripheral organs, such as the pituitary and adrenal glands, also express circadian clock mechanisms [77-79]. Consequently, the prevailing view is that human circadian physiology is regulated by a hierarchical clock mechanism, involving a master oscillator in the SCN and numerous subordinate peripheral oscillators. It is also clear that the SCN receives photoperiodic information from the retina, which it uses as a primary Zeitgeber to synchronise its circadian rhythm with that of the external environment. The mechanism responsible for synchronising individual peripheral oscillators remains to be elucidated, although it is likely to involve both neural and hormonal cues. Indeed, circadian hormone rhythms have been implicated in the control of many human behaviours, including learning and memory [80-82], and ageing-related changes in the peaks and/or phase relationships may interfere with cognitive processes [5, 80, 83-86]. Given that many steroid hormones, show clear-cut 24-h rhythms (Fig. 2), it is plausible that an age-related attenuation of these rhythms contributes to age-related perturbed sleep–wake cycles, as well as other pathologies [8, 9]; note that, in humans and rhesus macaques, circulating cortisol levels do not decline during ageing but the age-related increase in basal cortisol levels means that there is less circadian information being relayed from the adrenal gland to other peripheral targets, such as the liver and muscle [76].

Intracrinology and the role of precursor steroids

Thus far, we have discussed the impact of gonadal steroids on physiology; however, in primates, the gonads are not the sole source of androgens and oestrogens. Young adult humans and rhesus macaques have characteristically very high circulating levels of the adrenal steroid hormone precursor DHEA, and many tissues are capable of locally converting DHEA into active steroids such as testosterone or oestradiol: a phenomenon termed ‘intracrinology’ [87]. Because of the possible local intracrine conversion of DHEA to testosterone and oestradiol (Fig. 3), circulating levels of these sex steroids may not accurately reflect the amount of hormone acting within individual tissues. This further introduces a mechanism by which individual tissues can titrate the amount of active hormone that they are exposed to. In conditions of low oestradiol, for example, expression of the enzyme aromatase increases within the rhesus macaque hippocampus as compared to conditions of high oestradiol [46], potentially compensating for decreased oestradiol of ovarian origin. This mechanism also provides an additional potential target for hormone therapy (HT) because supplementation of DHEA may be useful to increase local oestrogen levels without impacting circulating oestrogen, thus reducing the risks of negative side effects associated with oestrogen HT.

Figure 3.

Schematic showing the biochemical conversion of dehydroepiandrosterone (DHEA) to oestradiol and dihydroxytestosterone (DHT). Genes coding for the key converting enzymes are depicted in italics. 3BHSD, 3β-hydroxysteroid dehydrogenase; 17BHSD, 17β-hydroxysteroid dehydrogenase, also known as 17-ketosteroid oxidoreductase. Importantly, tissues that express the key converting enzymes have the potential to activate oestrogen and androgen receptors, using DHEA as a sex-steroid hormone precursor.

Intracrine conversion of DHEA to oestradiol appears to be involved in mediating the actions of exogenous DHEA within the rodent hippocampus because the administration of DHEA increases spine synapse density and the effect can be blocked with letrozole, an aromatase inhibitor [88]. We have previously shown that the rhesus macaque hippocampus expresses all of the key enzymes (Fig. 3) that are necessary to convert DHEA to oestradiol [46], suggesting that a similar intracrine mechanism may be operative in primates. To test the hypothesis that DHEA improves cognition though conversion to oestradiol, we recently supplemented reproductively-intact peri-menopausal female rhesus macaques with daily oral DHEA. The animals were tested when they were on and off treatment, allowing for a within-subjects examination of the effect of DHEA. Despite promising evidence from rodent studies that DHEA improves cognitive performance [89-91], we saw no improvement of performance in a delayed match-to-sample test or a spatial delayed response task [92, 93]. Although the lack of an effect is discouraging, it does contribute to the validation of the aged macaque as a model for human ageing because the finding is consistent with the plethora of clinical studies that have failed to observe cognitive benefit of DHEA supplementation in aged men or women [50, 55-69].

Given that DHEA can exert pro-cognitive effects in the rodent brain (via conversion to oestradiol) and cognitive areas of the primate brain express the enzymes required for the DHEA-to-oestradiol conversion, it is puzzling why there is no obvious benefit of DHEA supplementation on cognition in humans and nonhuman primates. One possibility is that there is a significant age-related dampening of the intracrine mechanism within these brain areas. Consequently, a decline in expression or action of any of the enzymes involved in the conversion of DHEA to oestradiol [3β-hydroxysteroid dehydrogenase (3BHSD), 17β-hydroxysteroid dehydrogenase (17BHSD) and aromatase] (Fig. 3) would result in decreased ability to perform such a conversion. Indeed, the discordance between our findings and previous reports of improved cognitive performance in aged female macaques treated with oestradiol [94] suggests that this intracrine conversion may be attenuated during ageing. To examine this possibility, we investigated hippocampal expression of the key enzymes across the life span of rhesus macaques, and observed a significant age-related decrease in expression of 3BHSD, an enzyme responsible for the conversion of DHEA to androstenedione and androstenediol to testosterone (Fig. 3). Importantly, this decrease in 3BHSD was observed in animals that were within the age range used in the DHEA supplementation study [88, 93]. Thus, we have identified three steroidal mechanisms by which ageing may impact cognition: (i) an age-related decline of gonadal oestradiol; (ii) a decline in adrenal DHEA production that serves as the intracrine precursor to oestradiol; and (iii) a decline in the ability of the hippocampus to synthesise oestradiol from circulating DHEA.

Androgen supplementation in the aged rhesus macaque

Although an age-related decline in 3BHSD expression in cognitive brain areas could account for the disappointing findings from clinical DHEA supplementation studies, potential cognitive benefits may still be possible from other steroid hormone supplementations, at least in males. Our rhesus macaque study [88] found that hippocampal aromatase gene expression was maintained even into old age, suggesting that, although DHEA may no longer be efficiently converted to oestradiol, testosterone may still serve as a beneficial intracrine supplement (Fig. 3). Indeed, in human studies, testosterone supplementation has been shown to increase spatial performance [95-97], and the administration of oestradiol can rescue cognitive deficits induced by androgen deprivation in men [50]; this suggests that the effects of endogenous testosterone on at least some aspects of cognition may depend on conversion to oestradiol. Indeed, the benefits of testosterone on verbal memory have been shown to require aromatisation to oestradiol [60], giving further support to the hypothesis that intracrine conversion of steroids within the brain helps to maintain cognitive performance during ageing.


Oestrogen HT has been available for many years and has demonstrated efficacy in alleviating symptoms associate with post-menopausal disorders. On the other hand, increasing concern about possible side-effects of long-term HT has provided the impetus for developing safer HT paradigms, especially ones that do not include oestrogens. The rhesus macaque HPG and HPA neuroendocrine axes are remarkably similar to those of the human, showing the same age-associated changes. This makes the rhesus macaque a pragmatic animal model in which to investigate the mechanisms that underlie normal and pathological human ageing and to develop more effective therapies, such as the administration of safer precursor hormones [87] or nonfeminising oestrogens [98-100]. So far, data from old female rhesus macaques have failed to show significant beneficial effects of DHEA supplementation on cognitive function, which is in general agreement with human studies and in contrast to rodent studies. In part, this is likely to be the result of an ageing-associated dampening of the intracrine mechanism responsible for converting DHEA to sex steroids. It is also possible that existing hormone supplementation paradigms do not adequately mimic the endogenous circadian hormone profiles and so are less effective, or even detrimental. To explore this possibility, we have recently initiated a study involving old male rhesus macaques, in which we are assessing the efficacy of androgen supplementation on a wide range of physiological functions, including sleep–wake cycles, cognition and immune function [92]. What makes this study especially pertinent is that the daily combined testosterone–DHEA supplementation paradigm not only raises mean circulating levels of DHEAS, testosterone, dihydrotestosterone (DHT) and oestradiol to juvenile levels, but also preserves the normal 24-h pattern of these hormones in the circulation (Fig. 2). Thus, testosterone, oestradiol and DHT levels of androgen-supplemented old males continue to show a peak during the night, and DHEAS levels continue to show a peak in the morning (K.G. Sorwell and H.F. Urbanski, unpublished observations). Given that many primate behaviours and physiological functions have strong circadian components [8, 9], we anticipate that more physiological hormone supplementation paradigms may prove to be safer and more efficacious at treating disorders in the elderly.


This research was supported by grants received from the National Institutes of Health: AG-023477, AG-029612, AG-036670, HD-007133 and OD-011092.