Exercise and estrogen: common pathways in Alzheimer’s disease pathology

Abstract

Alzheimer’s disease (AD) is a neurodegenerative disease that is characterized by progressive declines in cognitive function. Current epidemiological data indicate significant sex-linked disparities, where females have a higher risk of developing AD compared with male counterparts. This disparity necessitates further investigations to uncover the pathological and molecular factors influencing these sex differences. Although the underlying pathways behind this observed disparity remain elusive, recent research points to menopausal estrogen loss as a potential factor. Estrogen holds a significant role in amyloid precursor protein (APP) processing and overall neuronal health through the regulation of brain-derived neurotrophic factor (BDNF), a factor that is also reduced in postmenopausal women. BDNF is a known contributor to neuronal health and its reduced expression is typically linked to AD disorders. Exercise is known to increase BDNF and may provide an accessible activity for postmenopausal women to reduce their risk of AD. This review aims to discuss the relationship between estrogen, exercise, and BDNF in AD pathology.

INTRODUCTION

Alzheimer’s disease (AD) is the most common form of dementia, accounting for 50%–60% of the total cases (1). Classified as a neurodegenerative disease, AD is defined by the deterioration of cognitive functions, including memory, learning, and language (1). In its initial stages, AD typically affects the hippocampus, which is the center for memory, before spreading to other regions of the brain in later stages (1). Furthermore, AD can be classified into two main categories, sporadic (late onset) and familial (early onset). Familial AD accounts for less than 1% of all AD cases (2) and results from autosomal genetic mutations that affect the proteolytic processing of the amyloid precursor protein (APP) along with alterations in presenilin 1 and presenilin 2 (2). Conversely, sporadic AD accounts for the vast majority of AD cases (>70%) and is found to be most prevalent in patients over the age of 65 yr (3). Age is the largest risk factor associated with late onset AD; however, there are several modifiable factors that contribute significantly to AD risk (1). Such additional risk factors include cardiovascular disease, physical inactivity, depression, smoking, obesity, and type 2 diabetes (4). Demographically, women are found to be affected by AD in a disproportionate manner when compared with male counterparts of the same age group (5, 6). Moreover, women make up approximately two-thirds of the total AD cases and have a 1 in 6 lifetime risk of developing AD, compared with a 1 in 11 lifetime risk for men (7). These disparities necessitate further investigations to uncover the pathological and molecular factors influencing these sex differences. One hypothesis points to menopausal estrogen loss as a strong risk factor in the development of AD. Estrogen holds a significant role in APP processing and overall neuronal health through the regulation of brain-derived neurotrophic factor (BDNF), a factor that is also reduced in postmenopausal women (8, 9). Due to estrogen loss being associated with AD, estrogen replacement therapy has been investigated as a means to prevent AD development. However, estrogen replacement therapy is associated with elevating the risks for breast cancer, venous thromboembolism, and cardiovascular disease (10). This prompted researchers to investigate alternative therapeutic factors that can minimize AD progression in postmenopausal women, without introducing additional risks. Currently, exercise is known to increase BDNF and may provide an accessible activity for postmenopausal women to reduce their risk of AD (11). Here, we will review the common pathways between estrogen and BDNF in the brain.

AMYLOIDOGENIC PATHWAY

The accumulation of Aβ peptides into plaques is one of the most significant hallmarks of AD, and increased Aβ production is argued to be associated with the preclinical stages of AD before symptomatic behavioral changes (12). Much research has focused on examining how to reduce the production and accumulation of these peptides (9). Aβ peptides originate from the cleavage of the transmembrane protein, APP (13). Cleavage of APP can be carried out by the β-site APP cleaving enzyme 1 (BACE1), which cleaves APP in its N-terminus region (13). This initial cleavage results in a membrane-bound C99 fragment, along with a soluble APP β fragment that is released into the extracellular region (14). The membrane-bound C99 fragment is then cleaved by γ-secretase, to result in the Aβ peptides (13). Alternatively, in the nonamyloidogenic pathway, APP cleavage is completed by α-secretase (15). The significance of α-secretase’s cleavage comes from the fact that APP is cleaved within the Aβ domain, which avoids Aβ production. Since BACE1 is the rate-limiting enzyme for Aβ production, its increased activity, coupled with reductions in α-secretase, will result in more Aβ peptide production and eventual aggregation.

ESTROGEN AND THE BRAIN

Estrogen is a steroid hormone mainly produced by the ovaries and primarily known for its role in the achievement of sexual maturation and fertility. Apart from its commonly known roles, estrogen also stimulates neuroprotective mechanisms that lead to the improvement of both neuronal health and spatial memory retention (8, 16). These neuroprotective properties of estrogen are typically facilitated by its most biologically active form, estradiol (17, 18). Estradiol is found in circulation and is able to cross the blood-brain barrier (BBB) through lipid-facilitated diffusion (19). After crossing the BBB, estradiol can act on neurons via altering genetic transcription, with the aid of estrogen receptors (ERα/ERβ) (18). In binding to ERα/ERβ, estrogen can elicit its effects through genomic or nongenomic signaling. Estrogen receptors develop homo- and heterodimers (ERα/ERβ) following their activation (19). With genomic signaling, estrogen’s impact is mediated by cytosolic estrogen receptors. The newly formed estrogen receptor homo- and heterodimers translocate to the nucleus, resulting in increases in gene transcription via binding to estrogen response elements (EREs) (17, 20, 21). Conversely, with nongenomic signaling, estrogen binds to its receptor at the cell membrane, resulting in a downstream signaling cascade. Two important downstream effects of genomic and nongenomic estrogen signaling are increased in BDNF content and the regulation of APP processing (Fig. 1).

Figure 1.

Schematic representation of estrogen and brain-derived neurotrophic factor (BDNF) signaling. Estradiol (E2) can act on neurons via genomic and nongenomic mechanisms. With genomic signaling, estrogen’s impact is mediated by cytosolic estrogen receptors (ERs). Estrogen receptors develop homodimers and heterodimers following activation and translocate to the nucleus, resulting in increases in gene transcription via binding to estrogen response elements (ERE). With nongenomic signaling, E2 binds to its receptor at the cell membrane, resulting in a downstream signaling cascade through phosphorylation of cAMP response element-binding protein (CREB). Increased BDNF is an important downstream outcome of both genomic and nongenomic estrogen signaling. BDNF levels can be independently increased through exercise. BDNF signaling through its receptor TrkB, results in CREB phosphorylation and increased BDNF transcription and translation. Further, BDNF signaling has been linked to increased α-secretase activity, while also downregulating β-site amyloid cleaving enzyme 1 (BACE1) activity.

During menopause, however, these physiological processes occur less frequently, as estrogen production declines. Women typically undergo menopause at the age of 50–52 and experience a significant reduction in the synthesis of estrogen as a result (22). The median age of menopause is governed by several factors such as socioeconomic status, body mass index (BMI), physical activity levels, and dietary habits (22). Moreover, menopausal estradiol levels typically decrease at a steep rate, often by 80%–90%, relative to premenopausal levels (17). To contextualize this observed reduction, young, healthy women tend to have estradiol levels of 150 pg/mL, whereas postmenopausal women’s estradiol levels average an approximate 10–15 pg/mL (23).

One of estrogen’s neuroprotective roles includes the regulation of APP processing. It has been hypothesized that, using the ERα receptor, estrogen regulates the activity of BACE1 and thus regulates the production of Aβ peptides (21). To confirm this hypothesis, Nord et al. (24) treated human neuronal cells with estradiol. The investigation demonstrated that estradiol treatment resulted in the downregulation of BACE1 activity, reducing the amyloidogenic processing rate of the APP (24). In another study, Jaffe et al. (25) also found that treating human breast cancer cells with estradiol yielded an accumulation of soluble APPα, suggesting an increase in nonamyloidogenic APP processing in the presence of estrogen. In support of estrogen’s protective role in APP processing in cell culture models, estradiol treatment was found to significantly prevent increases in Aβ accumulation, within ovariectomized (OVX) transgenic AD mice (26). Conversely, the loss of estrogen has been linked to a significant increase in BACE1 activity, as demonstrated in a study by Fukuzaki et al. (27), where hippocampal samples from ovariectomized (OVX) mice showed an increase in BACE1 activity due to the loss of ovarian hormones. Henceforth, based on current literature, it can be inferred that estrogen’s presence is protective against AD-associated amyloidogenic APP processing and the stimulation of nonamyloidogenic cascade.

An important downstream genomic target of estradiol in the context of AD is the expression of BDNF, as it contains an ERE (11, 28). BDNF is known to be responsible for the promotion of neuronal growth, maturation, and maintenance, as well as synaptic plasticity and memory consolidation (29). Previous research points to AD patients expressing significantly lower levels of BDNF (11, 30). Interestingly, OVX-induced estrogen loss in female Sprague–Dawley rats decreased BDNF mRNA in the hippocampus, (11) further demonstrating an important link between estrogen and BDNF. A recent study also highlighted that lower plasma BDNF levels were correlated with a worsened memory performance in postmenopausal women (9). In addition, Sharma et al. (31) showed that long-term estradiol therapy in OVX mice resulted in an increase in synaptophysin in hippocampal neurons, which is indicative of increased synaptic plasticity. To better understand the mechanism behind the observed increase in synaptophysin, the study also demonstrated that estradiol therapy increased the phosphorylated variant of cAMP response element binding (CREB) (31). This is important because once activated via phosphorylation, CREB translocates to the nucleus where it increases BDNF gene transcription (32). Results by Sharma et al. (31) are significant because they explain a mechanism whereby estrogen signaling increases BDNF levels, which in turn induces the neuroprotective pathways associated with BDNF signaling (Fig. 1).

EXERCISE AND ALZHEIMER’S DISEASE

It is now known that physical inactivity is a significant risk factor for the development of sporadic AD (33–35) and that exercise may prevent or alleviate some of the neurodegenerative features of AD (34, 36, 37). Human work by Erickson et al. (38) has demonstrated the link between physical activity, hippocampal volume, and memory. In a randomized controlled trial, 120 older adults were assigned to a moderate-intensity aerobic exercise group or a stretching and toning control group for 2 yr. Hippocampal size was increased by 2% in the exercise group and this was accompanied by improvements in spatial memory and increased serum BDNF content (38). To elucidate the mechanisms underlying exercise-induced improvements in the brain, it is important to examine rodent models. For example, access to a running wheel reduced BACE1 activity in APP transgenic mice fed a saturated fat diet (39) and 5 mo of voluntary exercise reduced Aβ peptides in both the cortex and hippocampus of TgCRND8 mice (40). Work from our laboratory has demonstrated that one exercise bout can reduce BACE1 content and activity in male wild-type mice with obesity (36, 41), providing further support for a direct effect of exercise on the brain. Berchtold et al. (11) found that 5 days of exercise increased BDNF mRNA levels in both intact and OVX 3-mo-old female Sprague–Dawley rats, compared with the sedentary controls, indicating that exercise maintains the ability to increase BDNF in female rodents and importantly in a postmenopausal model. In addition, a study conducted by Brown et al. (42) demonstrated that exercise is beneficial in reducing plasma Aβ levels in adults aged 60–95, provided that they are noncarriers of the apolipoprotein E (APOE) ɛ4 allele. However, the exact mechanisms linking the beneficial effects of exercise in the brain with BDNF are unclear. To further complicate the relationship between exercise and BDNF, work investigating the cardiorespiratory fitness (V̇o_2max) and resting serum BDNF determined that there was an inverse relationship in healthy active adults (43). The authors conclude that these results likely do not translate to the wider population, as the study participants were all engaged in recreational- or sport-based activities. They further speculate that perhaps the peripheral BDNF storage and release system or central nervous system uptake is altered with physical activity in humans (43). More work across research models is therefore necessary to examine a direct effect of exercise on the amyloidogenic pathway and the potential role for BDNF in this process.

Several studies have suggested that the release and increased expression of BDNF plays a role in coordinating the beneficial effects of exercise on the brain (34, 44). A single bout of exercise can increase circulating BDNF levels in male participants and this exercise response appears to be intensity dependent (45); however, it is unknown if a similar relationship exists in females. Although there are several potential sources of this increased circulating BDNF, it has been shown that the brain can be a significant source of exercise-derived BDNF (46). Considering BDNF’s involvement in the development of neurons and its neuroprotective properties, examining its relationship with exercise could uncover the elusive relationship between exercise and the reduction of BACE1 activity.

In relation to APP processing, BDNF has a direct effect on increasing alpha-secretase activity (47). Following 3 wk of voluntary wheel running male transgenic APP/PS1 displayed higher sAPPα and BDNF content and lower Aβ40/42 peptide content in the hippocampus compared with controls (47). To further explore the role of BDNF, the researchers treated SH-SY5Y cells with BDNF, an α-secretase inhibitor (batimastat), or the combination of both and found that BDNF increased α-secretase activity and pushed APP processing toward the nonamyloidogenic pathway. To date, a link between BDNF and BACE1 has yet to be demonstrated; however, given that BDNF content and signaling are increased in the brain after exercise (36, 48) and BACE activity is reduced, it is possible that there is a connection. As discussed previously, women are disproportionately affected by AD due to menopausal estrogen loss and potentially the accompanying reduction in BDNF; however, exercise may be employed as a measure that counteracts the negative impacts amplified by estrogen loss.

In the previous section, research data were presented to establish the linkage between estrogen and BDNF, in which estrogen can increase BDNF levels via CREB (31, 32). Importantly, CREB also acts as a downstream protein within the BDNF signaling pathway. When activated, BDNF interacts with its receptor, tyrosine kinase receptor B (TrkB), which in turn dimerizes, autophosphorylates, and initiates downstream signaling through CREB phosphorylation and activation (37). Consequently, pCREB modulates the expression of BDNF and triggers pathways that enhance synaptic plasticity and neuronal survival (32). Moreover, pCREB interacts with BDNF through a positive feedback mechanism, whereby elevated pCREB levels will lead to the upregulation of BDNF expression (49) (Fig. 1). This relationship between CREB and BDNF is further reinforced in the presence of estrogen, because of estrogen’s ability to increase BDNF, via CREB activation.

Although there exists an established relationship between estrogen and BDNF, there is a lack of data investigating whether exercise can in fact supplement the estrogen-linked BDNF loss in postmenopausal women. Previous studies show that exercise can increase plasma BDNF levels in humans (45) and reduce β-secretase activity levels in male mice (37). A recent preclinical study by Rashidy-Pour et al. (50) on OVX adult female Wistar rats also showed that exercise resulted in increase in hippocampal BDNF levels and significant improvements in the rats’ memory retention during the Morris water-maze task. However, neither of these previous studies investigate the long-term effects of exercise in women, and more particularly, postmenopausal women. Given that women are disproportionately affected by AD pathology, possibly due to estrogen loss, it would be vital to further investigate this relationship in future studies.

CONCLUSIONS

Further research is needed to explain the mechanisms that lead to the observed disparities in AD outcomes between males and females. The data from current research demonstrate that the loss of estrogen due to menopause can contribute to these observed differences. It is possible that the reduction in BDNF that accompanies the loss of estrogen plays a role. Therefore, strategies aimed at increasing or recovering BDNF levels in postmenopausal women may be of benefit. One such strategy is physical activity, which is known to increase BDNF as well as alter the processing of APP. To date most work has focused on the use of male models, and future studies are needed to fully examine the beneficial effects of exercise on brain health in postmenopausal women.

GRANTS

This work was supported by the Scottish Rite Charitable Foundation of Canada (No. 17114) (to R. E. K MacPherson). A. Bagit was the recipient of a Brock Match of Minds award.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

A.B., G.C.H., and R.E.K.M. prepared figures; A.B., G.C.H., and R.E.K.M. drafted manuscript; A.B., G.C.H., and R.E.K.M. edited and revised manuscript; A.B., G.C.H., and R.E.K.M. approved final version of manuscript.