Department of Neurology, Case Western Reserve University School of Medicine, Cleveland, Ohio 44120
Address correspondence and requests for reprints to: Alan J. Lerner, M.D., Alzheimer Center, 12200 Fairhill Road, Cleveland, Ohio 44120- 1013; E-mail:
ALZHEIMERS disease (AD) is the most common form of dementia in Western societies, with major risk factors being advanced age, possession of an apolipoprotein E (Apo E) e4 allele, family history, head trauma, and Downs syndrome (1). Other possible risk factors include environmental exposures such as alcohol consumption, smoking, level of education, estrogen replacement therapy (ERT), and use of anti-inflammatory medications. Gender effects in AD susceptibility may come as a result of differential outside environmental exposures, but also because of differences in internal biochemistry. Over the life span, steroid hormones influence brain development, are neurotrophic, and modulate processes relating to cell vulnerability to stimuli such as excitotoxin damage. This review will focus on issues of AD specific to males, including aspects related to steroid hormone effects, and clinical issues including prospects for new treatments based on these endocrine aspects of AD.
Clinical Aspects of AD and Gender Issues
Epidemiology
While women have greater average longevity, studies of age-specific dementia rates have not always confirmed greater risk for women developing AD (2). Overall dementia rates are similar in both sexes, but men have a greater risk of developing vascular dementia. This may tend to mitigate the likelihood of developing pure AD, which is frequently diagnosed by "exclusion" of other neurological conditions associated with dementia (3, 4). Competing sources of mortality from all other causes may serve to censor cases of AD, introducing another source of bias in AD case finding.
Genetics
Women with at least one Apo E e4 allele are more likely than men to develop AD. In a meta-analysis of Apo E and AD studies, women with at least one e4 allele appear to be at greater risk for AD, with the strongest effects occurring between age 50 and 75 yr (5). Caucasian women with the e3/e4 genotype are 1.5 times more likely to develop AD than males with that genotype. Similar results have been seen with women having an e2/e4 genotype, but they did not reach statistical significance due to sample size issues. Apo E does not appear to be as strong an AD risk factor in African-American or Hispanic samples, but amongst Hispanic women, sex-related trends similar to Caucasian samples are detected. This gender effect was not seen in Japanese women in the meta-analysis study. The reason for this genetic sexual dimorphism is unclear, but it may represent an interaction with post-menopausal estrogen replacement therapy (ERT), or possibly a sex-specific susceptibility amongst e3/e4 heterozygotes (6). Women with a positive family history and an e4 allele may also have an earlier age at onset (7).
Steroids and AD pathogenesis
The cause of AD is unknown, but it can be viewed as a complex genetic disease with both genetic and environmental inputs affecting all aspects of the disease (e.g. occurrence, age of onset, and course). Evidence from clinical cases indicates that there may be a very long preclinical phase, perhaps lasting decades (8).
Steroid hormones affect neural development from the prenatal period and have been implicated in multiple aspects of brain sexual dimorphism. Insofar as we conceive of the adult post-mitotic brain as a stable unit, this forms the biological substrate in which AD develops. Evidence for steroid effects in AD includes the role of glucocorticoids in mediating damage to hippocampal neurons and the multiple roles implicated for estrogens, from maintaining synapses to enhancing cerebral blood flow. Androgens may have anti-glucorticoid effects and may be aromatized to estrogens in the brain. There are no clinical methods for determining human levels of estrogens at the cellular level, but the aromatase effect may spare males from relative estrogen deficiency with aging (9). Aromatase distribution is more concentrated in temporal lobe than frontal lobe, the former being a brain region important for encoding new memories (10). Astrocytes also express aromatase mRNA, suggesting another way in which they may support neuronal viability across the life-span (11).
Structure-activity relationships and neuroprotection
The structure and function steroids may play a role in their
ability to act as neuroprotective agents. In a model of neuroprotection
involving in vitro cell viability after serum
deprivation, only steroids with a phenolic A ring were found to be
protective. These compounds include 17ß-estradiol,
diethylstilbesterol, and compounds such as 17-estradiol that binds
only weakly to estrogen receptors. Testosterone, dihydrotesterone,
progesterone, prednisolone, aldosterone, and cholesterol, lacking a
phenolic A ring, had no protective effects in that assay (12).
Steroidal structure also affects its antioxidant properties. Estriol and 17ß-estradiol have relatively strong antioxidant properties, while cortisone and corticosterone are mildly antioxidant in an assay involving generation of peroxy radicals. However, testosterone, progesterone, androstenedione, DHEA, and estrone had no antioxidant effect (13, 14, 15). RU 486 has also been reported to have strong anti-oxidant properties, preventing peroxide accumulation caused by ß-amyloid, hydrogen peroxide, and glutamate in mouse and rat hippocampal neurons (16).
Using a different model, Bastianetto et al. (17) showed that DHEA could protect hippocampal neurons from oxidative stress-induced damage. DHEA at concentrations of 10100 µM protected hippocampal neurons in vitro from toxicity induced by hydrogen peroxide and sodium nitroprusside.
Using an informatics approach linking estrogens to AD, Smalheiser and
Swanson (18) identified that estrogenic antioxidant activity needed
further study. They identified processes such as estrogen regulation of
calbindin D28k, induction of cathepsin D and other proteases,
inhibition of Apo E levels, and enhanced neuronal response to glutamate
as possible mechanisms related to estrogen effect at the molecular
level, to which the link to AD is largely unexplored. Antioxidants such
as -tocopherol and selegiline may slow progression of AD to defined
end-points such as loss of activities of daily living (19), but it is
unknown whether treatment with any steroidal compounds, including ERT,
would meaningfully affect AD course.
Androgens, aging, and AD
Androgen production falls with aging, with falls in the
concentrations of dehydroepiandrosterone (DHEA) levels and
its sulfated metabolite (DS) well-documented (20, 21). There is
a substantial literature regarding the beneficial effects of
DHEA administration in elderly patients (22, 23, 24). Reported
effects of DHEA include increased energy, increased muscle
mass, and a psychological sense of well-being, without demonstrable
virilzing effects. Insulin-like growth factor-I is increased, and
insulin growth factor binding protein-I levels fall with
DHEA administration. Effects of DHEA on
immune function include increased monocyte levels and enhanced B cell
mitogenic response. Total T cells and T-cell subsets are unchanged, but
T-cell mitogenic responses are also increased by DHEA.
Functional activation of T cells, as measured by percentage of cells
bearing IL-2 receptors was also reported. DS, but not
DHEA, also activates peroxisome proliferator-activated
receptor (PPAR-
) (21). DS has been reported to decrease Il-2
mediated activity of natural killer cells in aged and AD individuals
(24, 25, 26). Furthermore, DS protected hippocampal neurons against
glutamate-induced excitotoxin damage, an effect not produced by
DHEA and which could be blocked with suppression of a
B-dependent transcription factor with decoy oligonucleotides (27).
These latter findings are intriguing in light of the immune hypothesis
of AD pathogenesis, for which there is evidence of an acute phase
reaction limited to brain occurring in AD. Large numbers of proteases
are found in association with ß-amyloid, microglial activation occurs
in AD, and anti-inflammatory medications are reported in some, but not
all, studies to be relatively protective against AD (28, 29, 30).
In a study of cortisol and DHEA levels and cognition in AD, Miller et al. (31) found an inverse correlation between basal morning cortisol levels and cognitive impairment, but lower DHEA levels were associated with better cognitive performance. DHEA did not correlate with any noncognitive (i.e. behavioral) measures. In another reported study, DHEA replacement did not improve memory in elderly men given 50 mg/day for 2 weeks, but subtle improvements did occur in the P3 component of an event-related potential, indicating improved attention processing (32).
From these preliminary studies it would appear that androgen replacement in the elderly is well tolerated and is associated with positive effects on endocrine, musculoskeletal, immune, and psychological variables. There is less data on whether DHEA in particular significantly enhances neuropsychological functioning (32, 33) and no data on the efficacy of DHEA replacement in AD. Larger cross-sectional studies and clinical trials in an AD population may be able to elucidate some of these issues better.
Estrogens and antiandrogens in men and AD
Although not confirmed in all studies (34, 35), ERT has been found to be a protective risk factor for AD in a number of studies, including community surveys, population studies, and in AD case control studies (36, 37, 38). The development of selective estrogen receptor modulators (SERM) such as raloxifene and its congeners have held out hope that similar findings in AD prevention may occur with these compounds. However, there are no controlled clinical trials of SERMs in AD to date, and its long-term safety in males is unknown.
A number of case reports have reported use of estrogenic compounds, medroxyprogesterone or leuprolide in demented males. Men are more likely to show physical aggression as well as apathy in the course of their AD (39). In two men treated briefly with diethylstilbesterol, agitation was reported to occur less often. (40). Medroxyprogesterone use has been reported in a male with vascular dementia and unwanted sexual behavior. Although positive cognitive effects were not observed, the patient responded successfully after other medications had failed (41). Three demented patients with aggression, agitation, and pacing restlessness improved greatly with medroxyprogesterone and leuprolide (42). A single case of dementia and features of the Kluver-Bucy syndrome with verbal, physical, and sexual aggression responded to leuprolide (43). Four additional patients with disruptive sexual behavior responded favorably to medroxyprogesterone and were treated for one year without side effects (44).
Neurosteroids
Neurosteroids are synthesized in the central nervous system, and
their supply is generally independent of the endocrine system (45).
Many of them are progesterone metabolites, with particular actions as
antagonists of the -amino butyric acid (GABA) type A receptors. The
GABAergic system has multiple effects on the central nervous system,
including modulation of multiple other neuroendocrine systems.
Modulation of GABA-A receptors has also been implicated in memory
formation, and animal studies have shown that GABA-A antagonists can
improve memory in a number of paradigms (46, 47, 48). DHEA,
DS, pregnenolone, and pregnenolone sulfate, but not progesterone,
acting at the
-1 receptor attenuated memory deficits in mice
injected with fragments of ß-amyloid. (49). DHEA has
also been implicated in a cellular model of amyloid precursor protein
processing (50). These findings are of potential significance insofar
as ß-amyloid deposition is a neuropathological hallmark of AD, and as
ß-amyloid is widely felt to be the major neurotoxic moiety in AD.
Neurosteroids are under clinical investigation for use as anxiolytics and anticonvulsants. Anxiety is an important noncognitive symptom target in AD, frequently treated with benzodiazepines, which can lead to sedation and paradoxical agitation increases. The development of agents with predictable psychotropic properties and better side-effect profiles would be of tremendous importance in treating AD patients.
Conclusion
AD is a major public health problem in the modern world, and demographic trends associated with increased longevity have raised the specter of a society whose resources are drained by a large population of dependent elderly individuals. The changes in multiple body systems, including endocrine changes with aging, may play a role in AD pathogenesis. Although AD is vastly more complex than describing it as a form of accelerated aging, we must recognize that aging is the most consistent risk factor, and thereby likely plays a major role in pathogenesis. Perhaps the clues contained in the changing relationships of the immune, endocrine, and neural systems with aging will give us important insights into AD pathophysiology and will contribute to improved quality of life.
Footnotes
1 This study was supported in part by a grant from the National
Institutes of Health, P50-AG08012.
Received May 26, 1999.
Accepted June 16, 1999.
References