Address all correspondence and reprint requests to Christian J. Pike, PhD, University of Southern California, Andrus Gerontology Center, 3715 McClintock Avenue, Los Angeles, CA 90089–0191, USA. E-mail: email@example.com
As a normal consequence of aging, men experience a significant decline in androgen levels. Although the neural consequences of age-related androgen depletion remain unclear, recent evidence suggests a link between low androgen levels and the development of Alzheimer's disease (AD). Here, we test the hypothesis that androgens act as endogenous modulators of β-amyloid protein (Aβ) levels. To investigate this possibility, brain and plasma levels of Aβ were measured in male rats with varying hormonal conditions. Depletion of endogenous sex steroid hormones via gonadectomy (GDX) resulted in increased brain levels of Aβ in comparison to gonadally intact male rats. This GDX-induced increase in Aβ levels was reversed by DHT supplementation, demonstrating a functional role for androgens in modulating brain levels of Aβ. These findings suggest that age-related androgen depletion may result in accumulation of Aβ in the male brain and thereby act as a risk factor for the development of AD.
In contrast to abundant evidence that estrogen depletion is associated with increased risk of developing Alzheimer's disease (AD) (Hogervorst et al. 2000), little is known about the relationship between androgen depletion in men and AD. It is well established that aging men exhibit significant declines in bioavailable levels of testosterone and its active metabolite 5α-dihydrotestosterone (DHT) but maintain relatively normal levels of estrogen (Kaiser and Morley 1995). This normal, age-related reduction in androgen levels often results in functional impairments in androgen responsive tissues leading to a clinical syndrome termed ‘androgen deficiency in aging men’ (Morley and Perry 2000). It is unknown as to how age-related androgen depletion affects the brain, an established target of androgen action. Interestingly, recent reports demonstrate lower circulating testosterone levels in AD vs. non-demented men (Hogervorst et al. 2001), suggesting that androgen depletion in males may promote AD pathogenesis.
One mechanism by which sex steroid hormones may affect the development of AD is regulation of β-amyloid protein (Aβ) levels. Estrogen has been shown to decrease Aβ production in cultured cells (Jaffe et al. 1994; Xu et al. 1998) and in both wild type rodents and transgenic mouse models of AD (Petanceska et al. 2000; Levin-Allerhand et al. 2002; Zheng et al. 2002). Unclear is whether androgens regulate Aβ levels. This issue is complicated by the fact that the primary male sex steroid testosterone is converted to both DHT (via 5α-reductase) and estradiol (via aromatase) and thus activates androgen and estrogen signaling pathways. Although recent evidence shows that testosterone can increase non-amyloidogenic processing of amyloid precursor protein (APP) and thereby decrease Aβ (Gouras et al. 2000), testosterone's modulation of APP processing is dependent upon its aromatization into estradiol (Goodenough et al. 2000).
Here we directly investigate the contributions of estrogen and androgen pathways to regulation of Aβ levels. To accomplish this, we compared Aβ levels in brain and plasma of normal adult male rats with rats that were androgen depleted then treated with either placebo, the non-aromatizable androgen DHT, or the estrogen 17β-estradiol.
Materials and methods
Adult 90 day-old Sprague Dawley male rats were purchased (Harlan Laboratories) in gonadectomized (GDX) and sham GDX conditions. Animals were individually housed with ad libitum food and water in a vivarium with a 12-h light/dark cycle. Two weeks after surgery, all animals received four weeks of hormone replacement consisting of sequential administration of two 21-day slow release subcutaneous pellets (Innovative Research of America) containing either 12.5 mg DHT, 0.72 mg 17β-estradiol, or vehicle. Rats were subsequently killed and (i) brains quickly removed, halved midsagitally, and hemi-brains frozen at − 80°C (for Aβ measurement) or immersion fixed for 48 h in fresh 4% paraformaldehyde/0.1 m PBS (for immunohistochemistry) and (ii) trunk blood collected and separated into serum (for hormone assays) and plasma (for Aβ measurement).
Testosterone, DHT and estradiol were quantified from serum samples by radioimmunoassay following purification by celite column partition chromatography (Slater et al. 2001).
Aβ was measured in brain extracts (diethylamine (DEA) homogenates of soluble protein) and blood plasma using a well-characterized sandwich capture ELISA (Suzuki et al. 1994; Bimonte-Nelson et al. 2003). Aβ was captured using antibody BNT77 coated plates, followed by detection of Aβx-40 and Aβx-42 peptides with the horseradish peroxidase-conjugated antibodies BA27 or BC05, respectively. Plates were developed using TMB as substrate (reaction stopped by addition of 6% phosphoric acid) and optical density read at 450 nm. Raw data were converted to pmol Aβ by comparison to a standard curve of synthetic Aβ.
Brain lysates were immunoblotted using standard methodology. Briefly, soluble APPα and full-length APP were analyzed using DEA and sodium dodecyl sulfate (SDS) brain extracts, respectively. Protein corrected DEA and SDS extracts were diluted into reducing sample buffer, electrophoresed on 7.5% polyacrylamide gels, then transferred onto 0.45 µm polyvinylidene difluoride membranes (Millipore; Medford, MA). Blots were processed with the primary Aβ antibody 6E10 (Senetek), visualized using enhanced chemiluminescence reagents (Amersham) and exposure onto Hyperfilm (Amersham), then quantified by band densitometry of scanned films using NIH Image 1.61 software. Raw data were normalized to an internal standard (control brain lysate) and are expressed as a percentage of intact control values.
Fixed hemi-brains were sectioned (40 µm) in the horizontal plane using a vibratome (Ted Pella). Free-floating sections were immunostained for androgen receptor (AR) with PG-21 antibody (0.25 mg/mL) (Upstate Biochemicals) using previously described techniques (Ramsden et al. 2003).
Raw data from the hormone assays, ELISA and western blot experiments were statistically analyzed by anova followed by pair-wise comparisons using the Fisher LSD test.
Results and discussion
To evaluate the potential role of androgens in Aβ regulation, we investigated the effects of sex steroid depletion caused by GDX in comparison to sham GDX, and GDX rats with either DHT or estrogen replacement. To confirm that the experimental manipulations yielded the anticipated changes in hormonal status, we (i) measured serum levels of testosterone, DHT and estradiol (Table 1) and (ii) evaluated levels of AR in the bed nucleus of the stria terminalis (BNST), an androgen responsive brain region in which AR expression is positively regulated by androgens (Lu et al. 1998). Robust AR immunoreactivity was observed in BNST of sham GDX rats but was markedly decreased in GDX males (Figs 1a and b). Hormone replacement with DHT but not estradiol fully restored AR immunoreactivity, confirming functional neural efficacy and specificity of androgen treatment (Figs 1c and d).
Table 1. Effects of GDX and hormone replacement on serum hormone levels
To begin evaluating the regulatory effects of androgens on Aβ, we assessed how hormone manipulation affected soluble Aβ levels in brain. Depletion of endogenous sex steroid hormones via GDX caused an ∼25% increase in brain levels of Aβ40 and Aβ42 (Figs 2a and b). DHT replacement completely reversed the GDX-induced increase in Aβ, significantly lowering brain levels of both Aβ species (Figs 2a and b). The possibility that androgens may regulate Aβ levels has been suggested (Gouras et al. 2000; Gandy et al. 2001) but until now has remained unclear. In gonadally intact male rats with normal androgen levels, DHT supplementation resulted in a non-significant trend of reduced brain Aβ levels (Bimonte-Nelson et al. 2003), a finding consistent with our observation that DHT treatment in GDX rats reduced Aβ to levels below those found in sham GDX rats (Figs 2a and b). In neuronal cultures, Gouras et al. (2000) demonstrated that chronic exposure to supra-physiological levels of testosterone induced a reduction in soluble Aβ. However, since neuron and glia cultures express aromatase activity (Poletti and Martini 1999), these data are consistent with an indirect activation of established estrogen action as demonstrated by Goodenough et al. (2000). In our study, the use of non-aromatizable DHT and comparison to 17β-estradiol allowed us to unambiguously demonstrate that androgens but not estrogens modulate Aβ levels in male brain.
In contrast to the effects of DHT replacement, we observed that estradiol treatment of GDX male rats did not significantly affect Aβ levels in brain (Figs 2a and b) although it has clearly been shown to reduce Aβ levels in female rodent brain (Petanceska et al. 2000; Levin-Allerhand et al. 2002; Zheng et al. 2002). This lack of an estrogen effect in male brain may be due to gender specific neural actions of estrogen and androgens. For example, hippocampal spine density in male rats is regulated by DHT but not estradiol (Leranth et al. 2003), although estradiol is an established regulator of hippocampal spine density in female rats (Woolley and McEwen 1992). This sexual dimorphism mirrors our results in that a known estrogen function in female brain appears to be assigned to androgens in male brain. Similarly, we have recently observed that, whereas estradiol is neuroprotective in female rat hippocampus, DHT but not estradiol is neuroprotective in male hippocampus (unpublished observation, M. Ramsden and C.J. Pike).
Next, we investigated whether hormonal status also affects plasma levels of Aβ. In contrast to the increase in Aβ levels induced by GDX in brain, we observed no effect of GDX on plasma Aβ (Figs 2c and d). Interestingly, estradiol supplementation in GDX rats significantly reduced plasma levels of Aβ40 and Aβ42 in comparison to sham GDX and GDX conditions. Why hormone depletion via GDX did not affect circulating levels of Aβ is not clear and is in contrast to the effects of chemical castration observed in men undergoing prostate cancer therapy (Gandy et al. 2001). In the study by Gandy et al. (2001), hormone suppressive therapy resulted in a precipitous reduction in testosterone and estradiol and a parallel two-fold increase in plasma Aβ. Effects on plasma levels of Aβ resulting from GDX and hormone replacement in female rodents were not reported in prior studies (Petanceska et al. 2000; Levin-Allerhand et al. 2002; Zheng et al. 2002). The difference in hormonal regulation of Aβ between brain and plasma may suggest tissue specificity of estrogen and androgen actions. To what extent plasma effects of sex steroid hormones may be sexually dimorphic remains to be determined.
How hormones modulate Aβ levels in brain is not known. In cell culture paradigms, estrogen increases secretion of soluble APPα (Jaffe et al. 1994; Xu et al. 1998), suggesting that estrogen reduces Aβ levels by promoting non-amyloidogenic processing and or trafficking of APP (Greenfield et al. 2002). To begin investigating whether androgens act by a similar mechanism, we assessed brain levels of soluble APPα and full-length APP across the different treatment groups. Analysis of western blots revealed that hormonal status in male rats was not associated with significant differences in the levels of either soluble APPα or full-length APP (Fig. 3). Whether this result dismisses the possibility of androgen regulation of APP processing is not clear. For example, while phorbol esters both reduce Aβ levels and increase soluble APPα levels in vitro, their application in vivo decreases brain Aβ levels but does not affect soluble APPα (Savage et al. 1998). Further, studies showing reduction of Aβ levels in brain following estrogen treatment can be associated with (Levin-Allerhand et al. 2002) or without (Petanceska et al. 2000; Zheng et al. 2002) evidence of corresponding changes in soluble APPα levels. Alternatively, as recently demonstrated for the hormone insulin-like growth factor-1 (Carro et al. 2002), androgens and or estrogen may affect clearance rather than generation of Aβ in brain, possibilities that remain to be explored.
In summary, our data reveal that androgens beneficially regulate levels of Aβ in male rat brain. These experimental in vivo data provide insight into the potential consequences of natural androgen reduction in aging men. Approximately 50% of men over the age of 50 years have testosterone concentrations that are significantly lower than those in young adult men (Korenman et al. 1990), a situation that can precipitate adverse clinical consequences in androgen-responsive tissues (Morley and Perry 2000). Given that androgens have neuroprotective effects against AD-related insults (Pike 2001; Papasozomenos and Shanavas 2002) and beneficially regulate Aβ levels, we propose that men exhibiting significant androgen depletion may be at increased risk for the development of AD. Consistent with this possibility, recent studies show that men with AD have lower testosterone levels than non-demented age-matched control subjects (Hogervorst et al. 2001). Importantly, our data demonstrate that DHT treatment reverses the increase in brain Aβ induced by androgen depletion, suggesting potential therapeutic role for androgens in combating brain aging.
The authors would like to thank Ms. T.V. Nguyen for technical assistance. This study was supported by grants from the NIA (AG14751 to CJP) and NINDS (NS39072 to TEG). MPM is supported as a Robert and Clarice Smith Fellow in Neurodegenerative Disease.