Abstract
Background: Increased hypothalamic–pituitary–adrenal (HPA) axis activity manifested by elevated cortisol levels is observed in AD and may contribute to AD by lowering the threshold for neuronal degeneration. Presence of the APOE-ε4 allele increases risk for AD. Increased cortisol concentrations in apoE-deficient mice suggest that APOE genotype may influence cortisol concentrations in AD.
Methods: The authors measured cortisol levels in CSF and determined APOE genotypes for 64 subjects with AD and 34 nondemented older control subjects.
Results: CSF cortisol was significantly higher in AD than in control subjects. CSF cortisol concentrations differed with respect to APOE genotype in both subjects with AD (ε4/ε4 > ε3/4ε > ε3/ε3) and normal older control subjects (ε3/ε4 > ε3/ε3 > ε2/ε3). Comparison of CSF cortisol concentrations within the ε3/ε4 and ε3/ε3 genotypes revealed no differences between AD and control subject groups.
Conclusions: Higher CSF cortisol concentrations were associated with increased frequency of the APOE-ε4 allele and decreased frequency of the APOE-ε2 allele in AD subjects relative to control subjects. This effect of APOE genotype on HPA axis activity may be related to the increased risk for AD in persons carrying the APOE-ε4 allele and decreased risk for AD in persons carrying the APOE-ε2 allele.
Hypothalamic–pituitary–adrenal (HPA) axis activity is increased in AD, as manifested by increased concentrations of cortisol in plasma and urine and an increased plasma cortisol response to stress.1-8⇓⇓⇓⇓⇓⇓⇓ Because glucocorticoid administration or stress-induced glucocorticoid elevations in rats and tree shrews produce hippocampal dendritic atrophy and neuronal loss,9-11⇓⇓ increased cortisol level in AD may lower the threshold for neuronal degeneration in this disorder.12 The mechanism responsible for increased HPA axis activity in AD is unknown. Transgenic mice lacking a functional APOE gene (APOE-−/−) have aging-related increases in both basal plasma corticosterone concentrations and an enhanced plasma corticosterone response to stress.13 In humans, APOE genotype influences risk for developing AD: The ε4 allele increases risk,14 and the ε2 allele decreases risk.15 Because apoE influences corticosteroid levels in the mouse, the increased frequency of the APOE-ε4 genotype in AD could contribute to increased glucocorticoid concentrations similar to those found in APOE knockout mice.
Substantially increased cortisol concentrations in AD have been demonstrated in CSF samples obtained post mortem.16 Effects of AD on CSF cortisol concentrations obtained ante mortem have not been reported. Cortisol in CSF is a measure more relevant to CNS effects of cortisol than cortisol in plasma17-19⇓⇓ and provides an approximation of the cortisol concentration to which brain neurons are exposed. Here we asked whether CSF cortisol concentrations are elevated in AD ante mortem and, if so, whether APOE genotype influences CSF cortisol concentrations.
Methods.
Subjects.
This study was approved by the Human Subjects Committee of the University of Washington. Subjects included 64 persons with AD (44 men and 20 women; age = 67 ± 1 years [mean ± SEM; range 35 to 85 years]) and 34 cognitively normal healthy older community volunteers (22 men and 12 women; age = 71 ± 1 years [range 61 to 86 years]). Subjects were nonsmokers in good general health and had been free of medications (except occasional nonprescription analgesics or laxatives) for at least 1 month before lumbar puncture. All were normotensive at screening examination (<150/90 mm Hg) and within 25% of ideal body weight (Metropolitan Life Insurance tables, 1983). All subjects were free of past or present major psychiatric or neurologic disorders (other than AD), renal or hepatic disease, diabetes mellitus, and symptomatic cardiac disease. All subjects with AD met clinical diagnostic criteria for probable AD of the National Institute of Neurological and Communicative Disorders and Stroke20 and Diagnostic and Statistical Manual of Mental Disorders (3rd edition, revised) criteria for dementia of the Alzheimer type. The mean Mini-Mental State Examination (MMSE)21 score of the patients with AD was 16 ± 1. The mean Clinical Dementia Rating (CDR)22 scale score of the patients with AD was 2.8 ± 0.1. All subjects with AD were cooperative with the experimental procedure and free of disruptive agitation on the morning of the study. Normal older subjects had MMSE scores between 26 and 30, no history or evidence of cognitive decline, and CDR scores of 0.
CSF collection.
Lumbar punctures were performed at the VA Puget Sound Health Care System. Subjects were studied between 0900 and 1100 hours after fasting since midnight. Patients were placed at bedrest for 90 minutes. Lumbar puncture was performed atraumatically with a 25G spinal needle while the patient was maintained in the lateral decubitus position. CSF cortisol was measured in the 9th to 13th mL of CSF removed. CSF samples were frozen immediately on dry ice and stored at −70 °C until assay.
Cortisol radioimmunoassay.
CSF cortisol was assayed by radioimmunoassay with 125I kits from Pantex (Santa Monica, CA) using a modification of the commercial protocol to increase sensitivity. The detection limit of this assay is 1 ng/mL (100 pg using a 0.1-mL sample). Intra- and interassay coefficients of variation are 3.3 and 7.2%, respectively.
DNA allele typing for APOE.
APOE genotypes were determined using previously described PCR conditions23 and the HhaI restriction digest method.24
Statistical analysis.
Variables are expressed as means ± SEM. Differences in age and CSF cortisol concentration between the AD and normal control groups, both overall and within genotype, were compared by Student’s t-test (unpaired, two tailed). Differences in CSF cortisol concentrations among APOE genotypes both within AD and control groups and the combined AD plus control group were evaluated by one-way analysis of co-variance (ANCOVA) using age as the co-variate. Post hoc Fisher’s protected least significant difference (PLSD) tests were used to evaluate differences between specific genotypes after a significant ANCOVA.
Results.
Age and genotype relationships.
Subjects with AD were younger than control subjects (t = 2.14, p < 0.05), reflecting the presence of seven very-early-onset familial AD subjects in the AD group. Of the 64 subjects with AD included in the analysis, 13 had the APOE-ε4/ε4 genotype, 34 had the APOE-ε3/ε4 genotype, and 17 had the APOE-ε3/ε3 genotype. Two AD subjects had the APOE-ε2/ε4 genotype and two had the APOE-ε2/ε3 genotype; because of the very low number of AD subjects with the ε2/ε3 and ε2/ε4 genotypes, these four subjects were not included in the analysis. Of the 34 normal control subjects, 11 had the APOE-ε3/ε4 genotype, 17 had the APOE-ε3/ε3 genotype, and 6 had the APOE-ε2/ε3 genotype; all control subjects were included in the analysis. AD subjects with the APOE-ε3/ε3 genotype were younger (59 ± 2 years) than AD subjects with either the APOE-ε3/ε4 (69 ± 1 years) or the APOE-ε4/ε4 (72 ± 1 years) genotype (F[2,61] = 11.3, p < 0.0001). Also, age at AD onset was lower in APOE-ε3/ε3 subjects (53 ± 3 years) than in either APOE-ε3/ε4 (63 ± 1 years) or APOE-ε4/ε4 (67 ± 1 years) subjects (F[2,60] = 10.62, p < 0.0001). These differences were a function of the APOE-ε3/ε3 group having five of seven subjects with early-onset familial AD carrying the presenilin-1 or presenilin-2 mutations. There was no relationship between age and APOE genotype in the normal control subjects (F[2,31] = 0.56, p = 0.6). There was no relationship between APOE genotype and duration of AD or MMSE score.
CSF cortisol and genotype relationships.
CSF cortisol concentration was higher in the AD group than the normal older control group (0.82 ± 0.03 versus 0.73 ± 0.03 ng/mL; t = −2.14, p < 0.05, unpaired, two tailed). Because CSF cortisol level was positively correlated with age within both AD (r = 0.36, p < 0.01) and control (r = 0.51, p < 0.01) groups, ANCOVA for CSF cortisol level by APOE genotype was performed with age as a co-variate. Within subjects with AD, there was a dose effect of the APOE-ε4 allele on CSF cortisol concentration (F[2,60] = 4.36, p < 0.02) such that APOE-ε4/ε4 homozygotes had higher CSF cortisol level than APOE-ε3/ε4 heterozygotes, who in turn had higher CSF cortisol level than APOE-ε3/ε3 homozygotes (all post hoc pairwise comparisons by Fisher’s PLSD test, p < 0.05) (see figure 1). There was also a trend toward a stepwise effect of APOE genotype on CSF cortisol in control subjects, such that APOE-ε3/ε4 > ε3/ε3 > ε2/ε3 (F[2,30] = 3.31, p = 0.05) (see figure 1). Because comparison of CSF cortisol concentrations between AD and control subject groups within the APOE-ε3/ε4 and APOE-ε3/ε3 genotypes revealed no differences, subject groups were combined for further analysis. When groups were combined, two-way ANOVA revealed a marked effect of APOE genotype on CSF cortisol such that APOE-ε4/ε4 > ε3/ε4 > ε3/ε3 > ε2/ε3 (F[3,93] = 7.8, p = 0.0001) (see figure 2). Post hoc pairwise comparisons revealed differences (all p < 0.05) in CSF cortisol concentration among all APOE genotypes with the exception of APOE-ε3/ε3 versus APOE-ε2/ε3. There was no effect of gender on CSF cortisol and no interaction of gender and APOE genotype.
Figure 1. CSF cortisol concentrations by APOE genotype in 64 patients with AD (open circles) and 34 nondemented older control subjects (filled circles). There was a significant effect of the APOE-ε4 allele on CSF cortisol concentration in AD subjects (APOE-ε4/ε4 > ε3/ε4> ε3/ε3; F[2,60] = 4.36, p < 0.02, all post hoc pairwise comparisons by Fisher’s protected least significant difference test, p < 0.05). There was a trend toward a stepwise effect of APOE genotype on CSF cortisol level in control subjects (APOE-ε3/ε4 > ε3/ε3 > ε2/ε3; F[2,30] = 3.31, p = 0.05).
Figure 2. CSF cortisol concentrations by APOE genotype in 64 patients with AD and 34 nondemented older control subjects combined. CSF cortisol concentrations differ significantly by genotype: APOE-ε4/ε4 > ε3/ε4 > ε3/ε3 > ε2/ε3 (F[3,93] = 7.8, p = 0.0001). All post hoc pairwise comparisons were at p < 0.05 (Fisher’s protected least significant difference test) with the exception of APOE-ε3/ε3 versus APOE-ε2/ε3.
CSF and plasma cortisol relationships.
Plasma cortisol determinations were available for 26 of the 34 control subjects and 44 of the 64 AD subjects. Plasma and CSF cortisol levels were positively correlated in all subjects (r = 0.39, p = 0.001). Plasma cortisol concentrations and relationship between CSF and plasma cortisol did not differ between AD and control subject groups or by APOE genotype.
CSF cortisol and cognitive function.
There was no correlation between MMSE scores and CSF cortisol level within the subjects with AD (r = 0.172, p = 0.19). There was a trend toward an inverse correlation between CSF cortisol and MMSE scores within the cognitively normal older subjects (r = −0.327, p = 0.06). There were no differences in CSF cortisol concentration by CDR score among AD subjects (F[4,55] = 1.82, p = 0.14).
Discussion.
These results demonstrate an effect of APOE genotype on CSF cortisol concentrations in older persons similar to the effect of APOE genotype on the risk for expressing AD in later life.14,15⇓ Subjects with the APOE-ε4 allele, which increases risk for AD in a dose-dependent manner,14,25⇓ had a dose-dependent enhancement of CSF cortisol level. Subjects with the APOE-ε2 allele, which decreases the risk for AD,15 had the lowest CSF cortisol concentrations. These findings demonstrate effects of APOE genotype on HPA axis activity that may have neurobiological implications for AD.
ApoE is abundant in steroidogenic tissues, and in the human adrenal cortex, apoE is synthesized at a relative rate equal to or greater than that observed in the liver.26 ApoE in the adrenal cortex is thought to play a role in lipid redistribution among cells and regulation of the utilization of cholesterol for steroid production.27,28⇓ Human APOE gene expression in cultured Y1 adrenocortical cells results in a marked decrease in both basal and adrenocorticotrophic hormone-induced steroidogenesis29,30⇓ and alterations in cholesterol metabolism favoring storage over utilization.31 Characterization by high-resolution two-dimensional gel analysis has indicated that isoforms of human adrenocortical apoE correspond exactly to the specific isoforms of plasma apoE.26 These findings suggest that the association of the APOE allele with CSF cortisol concentration that we have found may be the result of differing efficacies of the apoE-ε2, ε3, and ε4 isoforms in inhibiting adrenal steroidogenesis. Direct investigation of the isoform-specific effects of apoE on adrenal steroidogenesis is necessary to confirm this hypothesis.
Increased HPA axis activity in AD, manifested by increased cortisol concentrations in plasma and urine1-8⇓⇓⇓⇓⇓⇓⇓ and resistance to suppression of plasma cortisol by the synthetic glucocorticoid dexamethasone,32,33⇓ is perhaps the most consistently demonstrated physiologic dysregulation in this disorder.34 That elevated glucocorticoid concentrations induced by stress produce hippocampal neurodegenerative changes in the rat and tree shrew9-11⇓⇓ and that aging-associated hippocampal neurodegeneration can be prevented by adrenalectomy in the rat35 provide rationale for the hypothesis that excess cortisol in human aginglowers the threshold for the expression of neurodegeneration in AD. Because APOE genotype was not reported in earlier studies demonstrating increased HPA activity in AD,1-8⇓⇓⇓⇓⇓⇓⇓ it is unknown if increased HPA axis activity in those AD samples studied was a function of increased APOE-ε4 frequency.
These results confirm and extend to living patients the observation in AD of increased cortisol concentrations in CSF obtained post mortem.16 The concentration of cortisol in CSF provides a better estimate of CNS exposure to cortisol than does the concentration in blood as CSF cortisol is in direct equilibrium with brain extracellular fluid. Furthermore, only ≈5% of plasma cortisol is the free, biologically active form, unbound to corticosteroid-binding globulin or albumin;17 total plasma cortisol levels therefore do not necessarily accurately reflect free cortisol concentration. CSF corticosteroid-binding globulin concentrations are ≈0.3% of those in blood, and very little CSF cortisol is bound to proteins.17,18⇓ Also, rapid penetration of free cortisol from blood into CSF, but slower clearance of cortisol from CSF than from blood, results in higher and more prolonged cortisol elevations in CSF than in serum after episodic increases in HPA axis activity.19
The lack of a relationship between cognitive function and CSF cortisol level in the subjects with AD makes it unlikely that higher CSF cortisol level reflects disease severity; rather, it is more consistent with the increased frequency of the APOE-ε4 allele and decreased frequency of the APOE-ε2 allele within the AD group. A recent study reported a similar lack of correlation between plasma cortisol and cognitive function in patients with AD.36 CSF cortisol tended to be inversely correlated with MMSE scores within the very constricted range of MMSE scores (26–30) in the cognitively normal older control subjects, consistent with the subtle cognitive deficits noted and recently reported in nondemented older persons carrying the APOE-ε4 allele.37,38⇓
Although in nondemented persons, the APOE-ε4 allele appears to accelerate the pathobiological process that culminates in AD25 and is associated with increased cognitive decline37 and hippocampal volume loss,39 these effects of the APOE-ε4 allele are not observed once AD becomes apparent. It has been repeatedly reported in most40-44⇓⇓⇓⇓ but not all45 longitudinal studies of AD progression and APOE genotype that there is no effect of the APOE-ε4 allele on the rate of AD progression. The lack of effect of the ε4 allele on the rate of progression in AD remains an unexplained paradox regardless of the mechanisms responsible for the effect of the ε4 allele to increase the risk of AD. If high CSF cortisol concentration were involved in the mechanisms by which a nondemented person “converts” to AD, one would expect that high CSF cortisol level in a nondemented person destined to develop AD would persist after AD becomes clinically evident. This issue can be resolved only by prospective longitudinal studies.
Several mechanisms have been hypothesized to contribute to the role of APOE genotype in the pathobiology of AD.46 For example, apoE 1) acts as a chaperone mediating β-amyloid protein aggregation, 2) binds to β-amyloid protein and mediates clearance via the apoE receptor, 3) is an injury response protein, and 4) modulates neurite extension. The current results together with those from animal and other human studies raise the possibility that APOE genotype effects on the HPA axis may be involved in the pathobiology of AD. The APOE-ε4 allele may produce increases in cortisol similar to the aging-dependent increases in corticosterone observed in APOE knockout mice. These APOE knockout mice also manifest CNS neurodegenerative changes.47,48⇓ A recent longitudinal study in older nondemented humans demonstrated a relationship between elevated plasma cortisol concentrations, reduced hippocampal volume, and cognitive impairment.39 Taken together, these results provide rationale for longitudinal studies of nondemented older persons with the ε3/ε4 genotype. Such studies could determine if nondemented ε3/ε4 persons with high CSF cortisol concentration are at greater risk for AD than ε3/ε4 persons with low CSF cortisol concentration. If so, such a result would support involvement of APOE genotype effects on CSF cortisol concentrations in the pathobiological processes associated with AD.
Acknowledgments
Supported by the Department of Veterans Affairs, National Institute on Aging (NIA AGO5136 and AGO8419), and the Joan Alhadeff Research Foundation.
Acknowledgment
The authors acknowledge the excellent technical assistance of Molly Wamble, Rebekah Rein, Robert Beckham III, and Elizabeth Colasurdo.
- Received August 14, 2000.
- Accepted January 6, 2001.
References
Letters: Rapid online correspondence
- Reply to Sass et al
- Increased CSF cortisol in AD is a function of APOE genotype
REQUIREMENTS
If you are uploading a letter concerning an article:
You must have updated your disclosures within six months: http://submit.neurology.org
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.
You May Also be Interested in
Hemiplegic Migraine Associated With PRRT2 Variations A Clinical and Genetic Study
Dr. Robert Shapiro and Dr. Amynah Pradhan