Hormones and Behavior 35, 254–263 (1999) Article ID hbeh.1999.1518, available online at http://www.idealibrary.com on Relationships between Dehydroepiandrosterone Sulfate (DHEAS) and Cortisol (CRT) Plasma Levels and Everyday Memory in Alzheimer’s Disease Patients Compared to Healthy Controls Linda E. Carlson,* Barbara B. Sherwin,*,† and Howard M. Chertkow‡ *Department of Psychology, †Department of Obstetrics and Gynecology, and ‡Department of Clinical Neurosciences, McGill University, Montreal, Canada Received June 25, 1998; revised August 4, 1998; accepted September 23, 1998 Fifty-two age-matched Alzheimer’s disease (AD) pa- and depletion of hippocampal neurons and dendritic tients (26 men, 26 women), mean age 76.2 years, were branching in the CA1 and CA3 layers (Uno, Tarara, assessed with the Rivermead Behavioural Memory Test, Else, Suleman, and Sapolsky, 1989; Sapolsky, Uno, a test of everyday memory, coincident with the measure- Rebert, and Finch, 1990; Uno, Lohmiller, Thieme, ment of plasma cortisol (CRT) and dehydroepiandros- Kemnitz, Engle, Roecker, and Farrel, 1990). In hu- terone sulfate (DHEAS) via radioimmunoassay. The AD patients were compared to a control group of age- and mans, the elevated levels of cortisol (CRT) seen in gender-matched healthy elderly men and women. No individuals with Cushing’s Syndrome were related to differences were found between the AD patients and the a decreased volume of the hippocampus as assessed controls in DHEAS or CRT levels, or in the DHEAS/CRT by MRI (Starkman, Gebarski, Berent, and Shcteingart, ratio. There were no gender differences in DHEAS or 1992). CRT levels, or in the DHEAS/CRT ratio in subjects with Reports of elevated CRT levels in Alzheimer’s dis- AD. However, AD patients with higher levels of DHEAS ease (AD) patients (Leblhuber, Neubauer, Peichl, scored better than those with lower levels on the Reisecker, Steimpartz, Windhager, and Dienstl, 1993; subtests of Remembering a Name associated with a Svec and Lopez-S, 1989; Swaab, Raadsheer, Endert, picture, Digit Span Total and Forward, and the Mini Men- Hofman, Kamphorst, and Ravid, 1994; Maeda, Tani- tal Status Exam. AD patients with higher CRT levels moto, Terada, Shintani, and Kakigi, 1990) led to spec- performed worse on Delayed Route Recall than those with lower levels. These findings suggest that AD pa- ulation that some portion of the degenerative changes tients with higher endogenous levels of DHEAS may associated with this disease may be related to in- perform better on some memory tasks than those with creased exposure to glucocorticoids. In a longitudinal lower levels, while AD patients with lower levels of CRT investigation of 12 AD patients over a 1-year period, may perform better than those with higher CRT. © 1999 CRT levels at baseline correlated positively with cog- Academic Press nitive deterioration over the next 12 months, as as- sessed by the Alzheimer’s Disease Assessment Scale (Weiner, Vobach, Svetlik, and Risser, 1994). However, Increased levels of glucocorticoids, due either to not all studies have found elevated CRT in AD pa- environmental stress or to exogenous administration, tients compared to unaffected controls (Touitou et al., have been associated with a number of adverse phys- 1982; Dodt et al., 1991; Nasman et al., 1996). iological and biochemical actions in the brains of ro- There is also reason to believe that dehydroepi- dents and monkeys, particularly in the hippocampus. androsterone (DHEA), an adrenal androgen, and its These include damage to the hippocampal dendritic sulfate, DHEAS, may influence cognition in people tree (Wooley, Gould, and McEwen, 1990), decreases in with AD. At low concentrations, DHEA and DHEAS hippocampal long-term potentiation (LTP) (Foy, Stan- function as allosteric antagonists of GABAA receptors. ton, Levine, and Thompson, 1987), and degeneration Noncompetitive binding with DHEA and DHEAS in- 0018-506X/99 $30.00 Copyright © 1999 by Academic Press 254 All rights of reproduction in any form reserved. Hormones and Memory in Alzheimer’s Disease 255 hibits GABA-induced neuronal activity, which is as- ratio was found in elderly AD patients compared to sociated with cognitive impairment (Majewska, age-matched controls, most particularly in women Demirgoren, Spivak, and London, 1990). When low (Leblhuber et al., 1993). However, neither the ratio of doses of DHEA and DHEAS were added to 14-day-old DHEAS/CRT nor the levels of CRT or DHEAS inde- embryonic mouse brain cultures, enhanced neuronal pendently were correlated with memory function as and glial survival occurred (Roberts, Bologna, Flood, assessed by the Mini Mental Status Exam (MMSE), or and Smith, 1987). DHEAS administration increased with the duration of AD symptoms. Lower DHEAS the excitability of CA1 hippocampal neurons in hip- levels were associated with increased age, which re- pocampal slices from rat brains, an effect that occurred sulted in a decrease in the DHEAS/CRT ratio in the within minutes and was reversible upon withdrawl of more elderly controls. In the female AD patients, even DHEAS (Meyer and Gruol, 1994). In intact rats, the lower ratios of DHEAS/CRT and higher levels of CRT application of DHEAS to the dentate gyrus resulted in were found compared to the control women. increases in LTP at all doses compared to baseline Finally, the issue of the generalizability of perfor- (Yoo, Harris, and Dubrovsky, 1996), indicating that mance on standard neuropsychological measures of DHEAS can increase the electrical activity of this re- function to everyday life is particularly important for gion. Some (Sunderland et al., 1989; Yanase, Fukahori, AD patients. For this reason, we utilized an ecologi- Taniguchi, Nishi, Sakai, Takayanagi, Haji, and Na- cally valid measurement of everyday memory that wata, 1996) but not all (Leblhuber, Windhager, bridges laboratory-based measures of memory and Reisecker, Steinparz, and Dienstl, 1990; Leblhuber et assessments obtained by questionnaire and direct be- al., 1993; Cuckle, Stone, Smith, Wald, Brammer, Haji- havioral observation (The Rivermead Behavioural mohammedreza, Levy, Chard, and Perry, 1990; Spath- Memory Test (RBMT), Wilson, Cockburn, Baddeley, Schwalbe, Dodt, Dittmann, Schuttler, and Fehm, 1990; and Hiorns, 1989). In this study we investigated the Birkenhager-Gillesse, Derksen, and Lagaay, 1994) relationships between CRT and DHEAS in AD pa- studies have found that plasma DHEAS concentra- tients and in healthy controls in an attempt to further tions were lower in AD patients compared to healthy elucidate the possible roles of these hormones on cog- controls. Sunderland et al. (1989) failed to find corre- nition in AD and in normal aging. lations between levels of DHEAS and dementia sever- ity or baseline cognitive scores in a sample of 10 AD patients. The possible significance of DHEAS levels in METHODS AD thus remains controversial. In animal models, DHEAS has antiglucocorticoid Subjects actions in the liver by blocking the enzymatic effects of glucocorticoids. When administered to mice, DHEAS Subjects were recruited from the pool of patients with blocked dexamethasone (DEX)-induced production of possible or probable Alzheimer’s Disease who were be- tyrosine aminotransferase (Svec and Lopez-S, 1989) ing followed for annual clinical visits at the Jewish Gen- and antagonized the glucocorticoid effect on glucose- eral Hospital/McGill University Memory Clinic, Mon- 6-phosphate dehydrogenase (McIntosh and Berdanier, treal, Canada. Subjects were referred by the clinical staff 1988). Based on these findings, Svec and Lopez-S of the Memory Clinic (neurologists and geriatricians), (1989) suggested that the ratio of CRT to DHEAS and screened for eligibility for the study. Exclusion cri- could serve as an index of glucocorticoid agonist ac- teria included diabetes, recent heart attack, stroke or tivity. If it is the case that AD patients have lower recent head injury, use of any psychotropic medication, DHEAS than age-matched controls (Sunderland et al., or lack of fluency in the English language. The diagnosis 1989), this would result in a ratio of agonist to antag- of possible or probable AD was made by the clinical staff onist much higher than that of the control population, based on the criteria established by the National Institute which could lead to a mild but progressive degener- of Neurological and Communicative Disorders and ative effect of CRT on hippocampal cells. If, in turn, Stroke (NINCDS) of the National Institute of Health and such degeneration caused dysregulation in feedback the Alzheimer’s Disease and Related Disorders Associa- mechanisms to the adrenals, an ever-accelerating, for- tion (ADRDA; McKhann et al., 1984), with the use of CT ward-cascading degree of hippocampal damage, and scans and routine blood work to rule out other possible presumably memory impairment, would follow causes of dementia. Charts were reviewed to determine (Sapolsky, Krey, and McEwen, 1986). the degree of dementia according to the Clinical Demen- Indeed, a lower DHEAS/CRT (antagonist/agonist) tia Rating scale score (CDR; Hughes et al., 1982), and only 256 Carlson, Sherwin, and Chertkow those diagnosed as mildly to moderately demented and Delayed), Route Recall (Immediate and Delayed), (CDR 1 or 2) were included. Only patients assessed as Remembering to Deliver a Message (Immediate and capable of giving their own informed consent were re- Delayed), Orientation, Date, and the Total Score (out cruited. Each eligible subject was approached by his/her of 24), which consists