The relation of cortisol levels with hippocampus volumes under baseline and challenge conditions

BR A IN RE S EA RCH 1 1 79 ( 20 0 7 ) 7 0 –78

a v a i l a b l e a t w w w. s c i e n c e d i r e c t . c o m

w w w. e l s e v i e r. c o m / l o c a t e / b r a i n r e s

Research Report

The relation of cortisol levels with hippocampus volumes under baseline and challenge conditions Kevin D. Tessner a,⁎, Elaine F. Walker a , Shivali H. Dhruv b , Karen Hochman c , Stephan Hamann a a

Department of Psychology, Emory University, Atlanta, GA 30322, USA Graduate Division of Biological and Biomedical Sciences, Emory University, Atlanta, GA 30322, USA c Department of Psychiatry and Behavioral Sciences, Emory University School of Medicine, Atlanta, GA 30322, USA b

A R T I C LE I N FO

AB S T R A C T

Article history:

Several studies using animal models have revealed an inverse relation between gluco-

Accepted 19 May 2007

corticoid levels and hippocampus volumes. This inverse relation has been interpreted as

Available online 26 May 2007

reflecting the role of the hippocampus in modulating glucocorticoid secretion, as well as the effect of glucocorticoids on the hippocampus. The objective of this study was to

Keywords:

examine the relation between hippocampus volumes and baseline and post-challenge

Hippocampus

salivary cortisol levels in healthy young adults. A double-blind, placebo controlled design

Hydrocortisone

was used in which 14 males between 18 and 30 years of age received either 100 mg

Glucocorticoid

hydrocortisone or placebo on separate occasions approximately 1 week apart. Baseline and

Hypothalamic–pituitary–adrenal

post-challenge cortisol levels were assessed prior to and after magnetic resonance imaging.

HPA axis

Volumetric analyses of the hippocampus revealed no differences between the

MRI

hydrocortisone and placebo conditions; however, post-challenge cortisol levels were inversely associated with total and right hippocampus volumes. Cortisol levels were not associated with the volume of the hippocampus in the placebo condition (i.e. under baseline conditions). The present findings are consistent with other evidence that the hippocampus, as reflected in volume, partially determines the efficacy of negative feedback in modulating cortisol levels. © 2007 Elsevier B.V. All rights reserved.

1.

Introduction

The hippocampus plays a role in modulating activity of the hypothalamic–pituitary–adrenal (HPA) axis via a negative feedback system that involves glucocorticoid binding to receptors in the hippocampus (Herman et al., 2005; Jacobson and Sapolsky, 1991). In rodents, smaller hippocampal volume is associated with heightened glucocorticoid levels and blunted negative feedback (Hibberd et al., 2000; Meaney et al., 1996; Meaney et al., 1995). This has also been shown in the rhesus (Coe et al., 2003)

and tree shrew (Ohl et al., 2000), and a recent study of pigs exposed to chronic stress revealed that basal cortisol was negatively correlated with volume and neuron number of the hippocampal dentate gyrus on the left side, but not the right (van der Beek et al., 2004). Likewise, the density of hippocampal astrocytes was decreased in male tree shrews undergoing psychosocial stress, and these changes correlated strongly with hippocampal volume (Czeh et al., 2006). Several neuroimaging studies of human subjects, typically clinical samples, have explored the relation between

⁎ Corresponding author. Fax: +1 404 727 1284. E-mail address: [email protected] (K.D. Tessner). 0006-8993/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.brainres.2007.05.027

BR A IN RE S E A RCH 1 1 79 ( 20 0 7 ) 7 0 –7 8

glucocorticoid secretion and hippocampal volume. Again, the findings generally show an inverse relation, although some do not find this association (MacLullich et al., 2005; Vythilingam et al., 2004). In patients with dementia (Alzheimer's type), for instance, higher serum cortisol concentrations were associated with smaller hippocampal volume (Ferrari et al., 2000). Similarly, a longitudinal study revealed that in aged humans hippocampal atrophy was associated with higher basal cortisol and greater increases in cortisol over time (Lupien et al., 1998). However, MacLullich and colleagues (2005) failed to replicate this finding in a sample of healthy elderly men aged 65–70 years. In this study, the volume of the right and left hippocampus did not correlate with morning, afternoon, or post-dexamethasone plasma cortisol levels. Similarly, 24-h urinary free cortisol were not inversely associated with right and left hippocampus in middle-aged subjects with major depressive disorder, however, a negative correlation was identified in age-matched healthy subjects (Vythilingam et al., 2004). Likewise, a 6-month longitudinal study of elderly depressed subjects revealed that hippocampal volume reduction was not associated with increased cortisol levels (O'Brien et al., 2004). More recently, a study of healthy preadolescent children found no association between baseline cortisol and overall hippocampal volume, although a subregional analysis revealed significant inverse associations between baseline cortisol and the lateral aspects of the anterior, medial, and posterior portions of the hippocampus, with the most pronounced associations corresponding to the CA1 subfield (Wiedenmayer et al., 2006). Animal studies have revealed that neurons in the hippocampus contain a high density of glucocorticoid receptors that are targets for steroids (Aronsson et al., 1988; de Kloet et al., 1994, 2000), and evidence indicates that hippocampal input from these principle neurons to the bed nucleus of the stria terminalis and subsequent projections to the paraventricular nucleus of the hypothalamus is important in feedback regulation of the stress response (Petrovich et al., 2001). Negative feedback effects in the HPA axis also occurs at the level of the pituitary as well as other cortical sites including regions in the frontal cortex (Diorio et al., 1993). For example, recent studies in rats suggest that the medial frontal cortex and anterior cingulate is involved in glucocorticoid regulation (Diorio et al., 1993; Sullivan and Gratton, 2002; Brake et al., 2000), and it appears that smaller cingulate cortex volumes may be associated with HPA axis dysregulation in humans (MacLullich et al., 2006). Further, in some reports, selective lesions of the hippocampal formation have been shown to increase basal glucocorticoid secretion (Sapolsky et al., 1984; Herman et al., 1989), although other studies have failed to replicate this finding (Tuvnes et al., 2003; Bradbury et al., 1993). It is generally assumed that the functions of the hippocampus in modulating HPA activity and glucocorticoid secretion are mediated by both glucocorticoid receptors (GRs) and mineralocorticoid receptors (MRs) (Bradbury et al., 1993; Keller-Wood and Dallman, 1984; van Haarst et al., 1997). Recent evidence from rodent studies suggests a distinction between “proactive” glucocorticoid negative feedback that maintains the HPA axis under baseline conditions, and “reactive” negative feedback that modulates activity of the HPA axis following acute HPA activation (Ladd et al., 2004). Reactive negative feedback may be

71

more dependent on the hippocampus than proactive baseline modulation. The specific neural mechanisms distinguishing these two feedback processes are not fully understood, however, they are presumed to be differentially mediated by GR and MR. Further, although there is no documentation of these two processes in humans, there is reason to believe that there may be counterparts in human hippocampus feedback mechanisms (van Haarst et al., 1997). We are aware of only one published study that has examined the relation of cortisol with hippocampal volume under challenge conditions. This investigation included both young (19–30 years) and older (59–76 years) healthy male subjects, and showed that total hippocampal volume was inversely associated with 24-h urinary cortisol and basal corticotropin (ACTH) levels, after controlling for age and total cerebral volume (Wolf et al., 2002). This same study also included a condition in which the investigators also administered hydrocortisone to challenge the HPA axis. Wolf and colleagues found no relation between post-challenge ACTH and hippocampal volume, but they did not report on the relation with post-challenge cortisol levels. In summary, there is evidence of a relation between circulating glucocorticoids and hippocampal morphology. Consistent with the role of the hippocampus in modulating HPA activity, hippocampal volume is typically found to be inversely correlated with glucocorticoid levels, and this may partially reflect a reduction in negative feedback to the HPA axis in animals with reduced hippocampal volume. The present study extends previous investigations of this relation in healthy young adults. Using a double-blind, placebo-controlled paradigm, we examined the relation between hippocampal volume and cortisol levels under both placebo and challenge (hydrocortisone administration) conditions. First, it was predicted that cortisol secretion would be inversely correlated with hippocampal volume. Second, the assumption that “reactive” negative feedback is more dependent on hippocampal integrity leads to the prediction that the inverse relation between hippocampal volume and cortisol secretion will be more pronounced in the hydrocortisone challenge condition. There is no animal data that suggests that a single exposure of corticosterone in adulthood would altar the volume of the hippocampus (Sousa et al., 1998a,b). In addition, Sousa et al. (1998a,b) report that hippocampal volume reductions occur after 3 months, but not 1 month of high-dose corticosterone administration. Generally, reductions in hippocampal volumes occur as a consequence of very high doses of glucocorticoid treatment over prolonged periods of time. Therefore, we do not expect to find hippocampal volume reductions following the administration of hydrocortisone in the participants in this study.

2.

Results

2.1.

Salivary cortisol

Mean raw salivary cortisol levels for the two experimental conditions are displayed in Table 1. As shown, the administration of hydrocortisone resulted in an increase in salivary cortisol that peaks between 1 and 2 h post-administration. It is important to note that these values reflect both endogenous

72

BR A IN RE S EA RCH 1 1 79 ( 20 0 7 ) 7 0 –78

Table 1 – Descriptive statistics for raw cortisol levels on the hydrocortisone and placebo days Time (min) after drug administration

Hydrocortisone (n = 14)

Placebo (n = 14)

0.51 ± 0.28 0.70 ± 0.50 2.77 ± 3.16 14.35 ± 23.40 15.61 ± 13.80 13.84 ± 9.99 16.56 ± 6.82 14.13 ± 7.88 13.86 ± 7.52 12.81 ± 9.28

0.42 ± 0.15 0.40 ± 0.16 0.37 ± 0.25 0.38 ± 0.17 0.36 ± 0.16 0.33 ± 0.13 0.33 ± 0.12 0.31 ± 0.11 0.36 ± 0.14 0.35 ± 0.14

0 15 30 45 60 75 90 105 120 130–135

2.3. Relationships between cortisol and hippocampus volumes

Data are presented as mean ± SD (μg/dL).

cortisol and administered cortisol. The salivary cortisol values obtained in the present study are similar to those reported in other challenge studies (Groschl et al., 2002; Ward et al., 2004). Previous research on the time course of cortisol following hydrocortisone administration has yielded results consistent with the temporal gradient observed here (Abercrombie et al., 2003). To derive an aggregate index of cortisol secretion in the hydrocortisone and placebo conditions, mean time-corrected cortisol values were computed by averaging cortisol values from each individuals pre-scan cortisol samples within condition. Results from the Shapiro–Wilk test of normality indicated that average cortisol values were normally distributed in the placebo and cortisol conditions. A paired-samples t-test was conducted to evaluate whether participants had higher average cortisol after the administration of hydrocortisone compared to placebo. The results indicated that mean timecorrected cortisol in the hydrocortisone condition (mean = 4.57 ± 4.00 μg/dL) was significantly greater than in the placebo condition (mean = −4.53 ± 2.31 μg/dL, t(13) = 7.90, p b 0.001).

2.2.

To rule out the possibility that the hypothesized inverse relationships between cortisol levels and hippocampal volumes were the result of hippocampal volume reductions following hydrocortisone administration, a mixed within-subjects ANOVA was conducted. The within-subjects factor was drug condition, the order of administration was the between subjects factor, and the dependent measure was total hippocampal volume corrected for total brain volume. As predicted, there were no significant main effects of order or condition and no significant interaction effect, suggesting that a single administration of low-dose hydrocortisone does not alter the volume of the hippocampus within a 1- to 2-week period.

Hippocampus volumes

Descriptive statistics for uncorrected right, left, and total hippocampal volume derived separately for the hydrocortisone and placebo conditions are displayed in Table 2. The α level was set at 0.05 (two-tailed) for the following analyses. Using a paired-samples t-test, there was no difference in raw hippocampal volume between the placebo (5103 ± 490 mm3) and the cortisol (5147 ± 537 mm3) conditions (t(13) = 0.561, p = 0.584). When corrected for brain size, the results where the same, (t(13) = 0.899, p = 0.385).

A series of correlational analyses were performed to test the hypothesized inverse relationship between cortisol level and hippocampal volume. Given the directional hypotheses, statistical tests of the correlation coefficients were one-tailed, with p b 0.05 as the criterion. Associations between average time-corrected cortisol levels in the hydrocortisone and placebo conditions and right, left, and total hippocampal volumes corrected for total brain volume are presented in Table 3. Results reveal that average cortisol levels in the hydrocortisone challenge condition is significantly, inversely correlated with right and total hippocampal volume; r(12) = − 0.50, p = 0.03 and r(12) = − 0.50, p = 0.04, respectively. The smaller association between average cortisol in the hydrocortisone challenge condition and the volume of the left hippocampus did not reach statistical significance, r(12) = −0.44, p = 0.06. There were no significant inverse correlations between time-corrected mean cortisol levels and hippocampal volume in the placebo condition. In addition to these analyses, regression analyses were conducted with hippocampal volume as the predictor, average cortisol as the dependent variable, and age and total brain volume as covariates. As expected, after statistically controlling for the effects of age and total brain volume, hippocampal volume was a significant predictor of cortisol in the challenge condition (β = − 0.70, p = 0.03), but not the placebo condition (β = −0.07, p = 0.86).

3.

Discussion

The findings of the present study show that post-challenge cortisol levels were negatively associated with the volume of

Table 2 – Descriptive statistics for hippocampal volume by group

Cortisol (n = 14) Placebo (n = 14)

Left hippocampus

Right hippocampus

Total hippocampus

Total hippocampus corrected a

2605 ± 281 2585 ± 223 ns b

2542 ± 290 2518 ± 316 ns

5147 ± 537 5103 ± 490 ns

0.34 ± 0.04 0.33 ± 0.03 ns

Data are presented as mean ± SD (mm3). Total hippocampal volume adjusted for total brain volume. b ns (not significant) at p b 0.05. a

73

BR A IN RE S E A RCH 1 1 79 ( 20 0 7 ) 7 0 –7 8

Table 3 – Correlations between average cortisol and hippocampal volumes on the hydrocortisone and placebo days Cortisol day

Placebo day

Left HC volume Right HC volume Total HC volume Left HC volume Right HC volume Total HC volume Average cortisol levels a

−0.438

−0.502 a

−0.498 a

0.055

0.033

0.049

Significant at p b 0.05 (1-tailed).

the hippocampus. However, salivary cortisol levels were not significantly linked with hippocampal volume under baseline conditions. Thus, the pattern of results was consistent with the prediction of a more pronounced inverse relation between hippocampal volume and cortisol following hydrocortisone administration. The absence of an inverse association between baseline cortisol levels and the volume of the hippocampus in this sample of healthy young subjects is consistent with the recent report by Wiedenmayer et al. (2006). These investigators found no association between baseline cortisol levels and overall hippocampal volume in a sample of healthy children. The participants in our study, healthy young men between the ages of 18 and 30, were significantly younger than samples in previous studies that have shown a relation between baseline cortisol and hippocampal volume in humans. Thus, taken in combination with the findings from Wiedenmayer et al., the present results may indicate that the relation between baseline cortisol and hippocampal volume emerges in adulthood. Further research is needed to determine whether there are age-related changes in HPA negative feedback that are reflected in the relation between baseline cortisol and hippocampal volume. In order to gain a better understanding of the relationships among cortisol levels and hippocampus volume over time, cross-sectional studies of individuals of different ages should be conducted or longitudinal methods need to be employed. Additionally, we do not know yet whether there is a developmental change in the nature of the relation between cortisol levels and hippocampal volume in human subjects. However, a recent meta-analysis conducted by Van Petten (2004) concludes that the negative relationship between hippocampal volume and memory performance in humans may change with age. The authors found evidence that in children and young adults, the association between hippocampal volume and memory performance was negative, while in older populations the correlation becomes positive. There are far fewer studies of the relation between hippocampal volume and cortisol, thus further research is needed to determine whether there are developmental changes in this association. Our current scientific understanding of the role of the hippocampus in HPA negative feedback is limited, so at this juncture we can only speculate on the neural mechanisms underlying this relationship. One parsimonious explanation for the present findings is that the volume of the hippocampus partially determines the efficacy of negative feedback in modulating cortisol levels under challenge conditions. Smaller hippocampal volume is thus associated with less effective feedback of endogenous cortisol and/or less effective uptake of cortisol.

The study by Czeh et al. (2006), described above, suggests one scenario. In an animal model, these authors showed that stress exposure decreased both the number and volume of hippocampal astroglia, and these decreases were correlated with decreases in total hippocampal volume. Given that astroglia express both GR and MR (Hwang et al., 2006), it is plausible that the volume of the hippocampus is linked with both receptor density and distribution. However, given the localization of glucocorticoid receptors in the hippocampus, total hippocampal volume measures may not provide as much support for negative feedback to the HPA axis as a more detailed morphological approach of regions of the hippocampus that contain a large number of glucocorticoid receptors. The present study did not perform an analysis of surface morphology among regions of the hippocampal formation, as this type of detail will be important for future studies. Our finding that the relation between hippocampal volume and cortisol was restricted to the challenge condition may indicate that reduced hippocampal volume is associated with lowered density of GRs, relative to MRs. This could lead to a decrement in negative feedback to the HPA axis in the hydrocortisone challenge condition. Such an interpretation is consistent with the MR/GR balance hypothesis (de Kloet, 1991), and supported by subsequent research suggesting that an increase in the amount of MRs, relative to GRs, results in reduced sensitivity of the HPA axis to challenge conditions in rodents (Oitzl et al., 1995). Further, as noted above, research on rodents suggest a distinction between proactive and reactive negative feedback to the HPA axis (Ladd et al., 2004). The latter is detected under challenge conditions, and appears to be mediated by increased expression of MR mRNA and decreased hippocampal expression of GR mRNA. Thus, increased hippocampal MR/GR mRNA ratio is linked with relatively intact “proactive” feedback, but diminished “reactive” negative feedback. Clearly, additional research is needed to shed light on the neural mechanisms involved in these processes. In particular, studies of hippocampal MR and GR distribution and density in human subjects are needed. Finally, it is possible that the link between hippocampal volume and post-challenge cortisol levels is mediated in some way by individual differences in metabolic processes and absorption rates of oral hydrocortisone. For example, it is known that glucocorticoids have good oral bioavailability and are eliminated mainly by hepatic metabolism and renal excretion (Czock et al., 2005). In particular, absorption of hydrocortisone occurs when the glucocorticoid binds various glycoproteins like transcortin and albumin. At oral doses greater than 20 mg, however, there tends to be an increase in free glucocorticoid concentrations (Czock et al., 2005). Protein binding is biologically relevant because not only is it impor-

74

BR A IN RE S EA RCH 1 1 79 ( 20 0 7 ) 7 0 –78

tant in absorption, but only free drug can interact with receptors in the hippocampus. Therefore, individual differences in the availability of glycoproteins will ultimately have pharmacodynamic consequences on drug absorption as well as MR/GR occupancy in the hippocampus. In addition, glucocorticoid metabolism takes place in two steps, the first of which oxygen or hydrogen atoms are added, typically in the liver. Secondly, conjugation takes place through the process of glucuronidation in which the glucocorticoid is chemically bound to glucuronic acid creating a hydrophilic inactive metabolite (Czock et al., 2005). The kidney is an important organ for glucocorticoid metabolism as it converts the active cortisol into the inactive cortisone by Type 2 11β-hydroxysteriod dehydrogenase (11β-HSD2). These metabolic processes protect the mineralocorticoid receptor from the undesired effects of cortisol occupancy. It is likely that polymorphic variability in the 11β-HSD2 gene produces different rates of glucocorticoid conversion and excretion (Connell et al., 2004), and this may result in compensatory changes to complementary glucocorticoid-regulating systems (e.g. HPA axis). The use of oral administration in the present study was designed to eliminate the induction of a cortisol response by intravenous puncture. However, the methods of the present study do not rule out the potential for individual differences in cortisol due to first pass metabolism. Further research is needed that might identify and control for differences in cortisol levels due to differences in absorption rates. Another such possibility is to investigate the relationship between cortisol levels and hippocampal volume after dexamethasone challenge in humans. Hydrocortisone was used in the present study because it more closely parallels stress-induced elevations in cortisol by binding to both MRs and GRs in the hippocampus, whereas dexamethasone is a potent synthetic glucocorticoid that binds preferentially to GRs. It is noteworthy that our findings indicate that a single administration of hydrocortisone does not affect hippocampal volume, either acutely or within a 1- to 2-week period. 100 mg of hydrocortisone was administered to participants during the study, and this is considered to be similar to a moderate stress response in humans. This acute dose is much lower than the chronic high level doses and plasma concentrations obtained in rodent studies that produce hippocampus volume differences (Angelucci, 2000). Therefore, the relation between hippocampal volume and cortisol does not appear to be due to changes in hippocampal morphology over a short period of time. It seems more plausible that the link between hippocampal volume and post-challenge cortisol levels is due to the role of the hippocampus in modulating HPA activity. However, a limitation of the present study is a small sample size and limited statistical power in detecting potential volume changes as a consequence of hydrocortisone administration. Also, it is possible that the gross measure of hippocampal volume and the low spatial resolution of MRI in contrast to histology methods limits the detection of more select volume reductions could they have occurred. Therefore, it may be premature to draw inferences about the temporal course of hippocampal plasticity. Sapolsky (1994) found compensatory glial activity, increased dendritic processes and reversible dendritic atrophy in animals exposed to

moderate stress, suggesting short-term hippocampal plasticity. It is possible that effects of chronic glucocorticoid elevations are also reversible. After cortisol levels decline to normal concentrations in treated Cushing's patients, there is an increase in hippocampal volume that is accompanied by functional improvement in verbal learning (Starkman et al., 1999, 2003). It is too early to conclude whether detectable structural changes in the hippocampus may occur shortly after reductions in cortisol levels, or over more extended time periods. The present findings raise several questions for further study. First, are there age-related changes in HPA function, such that the hippocampus plays a greater role in negative feedback with age? As described above, previous studies of clinical samples and older adults have revealed an inverse relation between baseline cortisol and hippocampal volume (Ferrari et al., 2000; Lupien et al., 1998). Yet, the present study of young adults and the Wiedenmayer et al. (2006) study of children failed to detect such a relation. This may reflect developmental changes in HPA function and feedback. Second, are there differences among the subregions of the human hippocampus with respect to the association between volume and either baseline or post-challenge cortisol secretion? Based on evidence of subregional differences in the expression of MR and GR, there may be regionally specific associations between hippocampal volume and post-challenge cortisol secretion. For example, the dentate gyrus may play a more pronounced role. Finally, how is hippocampal volume linked with the densities and ratios of MR and GR expression in the human hippocampus? The answers to these questions will ultimately advance our understanding of the etiology and treatment of stressrelated disorders (Sapolsky et al., 2000).

4.

Experimental procedure

4.1.

Participants

A total of 14 right-handed male young adults ages 18–30 years (mean = 21.6 ± 3.0 years) underwent two magnetic resonance imaging (MRI) examinations at least 1 week apart (mean = 12.7 ± 11.5 days) as part of a larger randomized controlled trial investigating the effects of hydrocortisone administration on brain function. Study participants were recruited from the Emory University graduate and undergraduate population. There were seven Caucasian, three African American, and four Asian participants. All were screened for medical disorders by a physician and were found to be physically healthy. None had a history of any neurological or psychiatric disorder. All were free of medications known to affect HPA system functioning. Informed consent was obtained prior to the first study session. The study protocol was approved by the Emory University IRB.

4.2.

Procedures

A placebo-controlled, double-blind, within-subjects, crossover study design was used. All subjects underwent two MRI sessions lasting approximately 1 h. The two experimental

BR A IN RE S E A RCH 1 1 79 ( 20 0 7 ) 7 0 –7 8

conditions were placebo and hydrocortisone, with order of administration randomly assigned and counterbalanced. There were no adverse effects of the procedures. Participants received the hydrocortisone and placebo treatments on different days beginning at approximately 10:00 a.m. (range 8:54 a.m. to 2:56 p.m.; mean 10:17 a.m. ± 1.9 h). The dose and time frame for hydrocortisone administration were selected on the basis of pilot studies in our laboratory and previous research examining the effects of hydrocortisone administration to humans (Hsu et al., 2003). Specifically, a 100 mg dose of hydrocortisone has been shown to elicit a physiological cortisol response comparable to that induced by moderate surgical stress and the psychological stress of the Trier Social Stress Test (Newcomer et al., 1999). The administration of hydrocortisone has been shown to induce an increase in salivary cortisol greater than that reflected in plasma cortisol levels (Porter et al., 2002). Pretesting for this study showed that salivary cortisol levels peak between 1 and 2 h after oral administration. Subjects arrived at the General Clinical Research Center (GCRC) of the Emory University Hospital 2 h prior to scanning. They had been informed that they would receive either the inert placebo or drug prior to undergoing MRI. A licensed nurse checked vital signs (blood pressure, heart rate, temperature, height, weight, etc.) to establish that they were within normal limits. Oral placebo or hydrocortisone was then administered. On each visit, two saliva samples were obtained prior to drug/placebo administration, and 10 samples were obtained following administration. Participants were administered inert placebo or 100 mg dose of hydrocortisone taken as a pill. Saliva was then sampled at regular 15-min intervals, beginning at 130 min prior to entering the scanner. The present report is based on the average of 10 salivary cortisol samples obtained after the administration of hydrocortisone or placebo and prior to the subject entering the scanner. The timeframe for saliva sampling is depicted in Fig. 1. During the period between drug administration and scanning, a series of cognitive tasks and surveys were completed by the participants.

4.3.

Saliva assays

The saliva samples were stored in aliquots at −80 °C until assayed. Cortisol assays were conducted with the Clinical

75

Assays “GammaCoat” 1251 RIA kit (DiaSorin, Stillwater, MN). Sensitivity of the kit for saliva cortisol is 0.05 μg/dL, and interand intra-assay coefficients of variation are 6.0% and 3.5%, respectively.

4.4.

Neuroimaging

Following standard procedure, each subject was screened for any metal devices that may interfere with MRI via a self-report checklist. All imaging was conducted on a 3-T Magnetom Trio™ (Siemens Medical Solutions, Malvern, PA) whole-body MRI system with standard head-coil at the Biomedical Imaging Technology Center of the Emory University School of Medicine. Scanning was performed in a private MRI research suite located adjacent to a larger multipurpose MRI facility in the Emory University Hospital. The MRI technician was the same for all participants. Participants entered the MRI scanner approximately 2 h after drug/placebo administration. Both structural and functional images were acquired. Scanning sessions lasted approximately 1 h, and included six functional MRI trials (each approximately 5 min long) and a 9-min three-dimensional structural MRI scan using standard gradient recalled echo pulse sequences. Specifically, high-resolution T1weighted structural images were acquired with the following parameters: TR/TE = 2600/3.93 ms, flip angle = 8°, field of view = 24 cm, matrix = 224 × 256 × 176, and slice thickness = 1.00 mm. Fig. 2 displays a coronal slice of the brain at the level of the anterior hippocampus, representing the quality of the MRI images acquired for this study.

4.5.

Image analysis

The analysis of structural neuroimaging data involved importing the raw images from the scanner into MRIcro (Rorden and Brett, 2000) for conversion into ANALYZE 7.5 image format. Manual tracings of the hippocampus were completed by a trained technician (S.D.) who underwent extensive training in neuroanatomy and demonstrated proficiency in the customized software developed exclusively for region of interest (ROI) analysis of medical data (IRIS 2000, UNC-Chapel Hill). Anatomical images were displayed in 3 orthogonal views for maximum reliability and tracing of the hippocampus. All ROIs were traced in the coronal view according to a previously

Fig. 1 – Baseline cortisol samples were obtained 5 min and 1 min prior to drug administration. Samples were obtained every 15 min post drug administration until scanning, and at the end of the scanning session.

76

BR A IN RE S EA RCH 1 1 79 ( 20 0 7 ) 7 0 –78

ratings of sample images by the image technician for the present study (S.D.) were 0.976 for left hippocampus and 0.963 for the right hippocampus on four randomly selected images. Volume estimates on a random sample of four brain images, traced by the image technician on two occasions approximately 1 week apart, revealed high intra-rater reliability ranging from 0.980 to 0.997 for the right and left hippocampus, respectively.

4.6.

Fig. 2 – A representative coronal MR image of the brain at 3 T.

established protocol (Aylward et al., 1999). This method allows for the hippocampus (cornu amonnis, dentate gyrus, and subiculum) to be reliably separated from the adjacent amygdala. The total number of voxels segmented as left and right hippocampal grey matter was used to estimate total hippocampal volume. Because the size of the ventricles and adjacent brain structures like the hippocampus may be associated with overall brain volumes (Synek et al., 1976; Zatz and Jernigan, 1983), corrections were made for variations in total brain volume in analysis of hippocampal volume. The image prepossessing steps required for the calculation of total brain volume are described here. First, an automated tissue segmentation was conducted for each volume data set to classify voxels based on signal value that is most representative of gray matter, white matter, and cerebrospinal fluid (CSF) (Rorden and Brett, 2000). Tissue-classified images were then used to create a mask of the brain where the CSF around the outer edge of the cortex could be used as a predefined border between brain and skull. Using the brain mask, non-brain tissue was removed from the image volumes, and trained technicians applied standardized rules to manual define the supratentorial volume. The total number of voxels segmented by the slice by slice manual procedure was used to estimate total brain volume. In summary, total brain volume was estimated using a slice by slice masking procedure in which the skull and meniscus of the brain were automatically removed using MRIcro (Rorden and Brett, 2000) and the cerebral hemispheres, cerebellum, and brain stem were outlined manually by trained image analysts. Total hippocampal measures were corrected by total brain volume estimates, and this correction for variations in brain size was calculated by dividing the ROI by the total brain volume. This correction procedure is comparable to previously established protocols (Jack et al., 1992; Pearlson et al., 1992). The image technician for the present study was trained to a high level of reliability prior to conducting the volumetric estimates for the present study. Both intra-rater and interrater reliability for hippocampal volumetric estimation was high. Inter-rater reliabilities between ‘gold standard’ estimates of volume (consensus of multiple trained raters) and

Data analysis

Time of day of testing was in the morning, although there was variation among subjects in start time. There are normative diurnal changes in cortisol secretion, which entail a decline throughout the course of the day (Weitzman et al., 1971). In order to control for variations in time of saliva sampling, timecorrected cortisol values were derived by conducting regression analyses with time as the predictor variable and cortisol values as the dependent variable. The primary analysis of the brain images consisted of an anatomical ROI approach where volume measurements of the hippocampus were derived as the dependent measure in individual subjects and subjected to conventional ANOVA statistics. Subsequent analyses of hippocampal volume estimates were corrected for differences in brain size. Pairedsamples t-tests were used to compare cortisol values controlled for time of day, and hippocampal volumes in the hydrocortisone challenge and placebo conditions. Linear regression analyses were conducted to determine whether hippocampal volumes were predictive of cortisol levels in the hydrocortisone and placebo conditions. For the two regression equations, average cortisol in either the hydrocortisone or placebo condition was the dependent variable, and age and total brain volume were used as covariates. The magnitude of the standardized regression coefficients (β) was used to test for significance, with p b 0.05 considered significant.

Acknowledgments We acknowledge the IRIS 2000 software developed by Sean Ho under the supervision of Guido Gerig, Ph.D., free download at http://www.ia.unc.edu/dev.

REFERENCES

Abercrombie, H.C., Kalin, N.H., Thurow, M.E., Rosenkranz, M.A., Davidson, R.J., 2003. Cortisol variation in humans affects memory for emotionally laden and neutral information. Behav. Neurosci. 117, 505–516. Angelucci, L., 2000. The glucocorticoid hormone: from pedestal to dust and back. Eur. J. Pharmacol. 405, 139–147. Aronsson, M., Fuxe, K., Dong, Y., Agnati, L.F., Okret, S., Gustafsson, J.A., 1988. Localization of glucocorticoid receptor mRNA in the male rat brain by in situ hybridization. Proc. Natl. Acad. Sci. 85, 9331–9335. Aylward, E.H., Li, Q., Honeycutt, N., Warren, A.C., Pulsifer, M., Barta, P.E., et al., 1999. MRI volumes of the hippocampus and amygdala in adults with Down's syndrome with and without dementia. Am. J. Psychiatry 156, 564–568.

BR A IN RE S E A RCH 1 1 79 ( 20 0 7 ) 7 0 –7 8

Bradbury, M.J., Strack, A.M., Dallman, M.F., 1993. Lesions of the hippocampal efferent pathway (fimbria–fornix) do not alter sensitivity of adrenocorticotropin to feedback inhibition by corticosterone in rats. Neuroendocrinology 58, 396–407. Brake, W.G., Flores, G., Francis, D., Meaney, M., Srivastava, L.K., Gratton, A., 2000. Enhanced nucleus accumbens dopamine and plasma corticosterone stress responses in adult rats with neonatal excitotoxic lesions to the medial prefrontal cortex. Neuroscience 96, 687–695. Coe, C.L., Kramer, M., Czeh, B., Gould, E., Reeves, A.J., Kirschbaum, C., et al., 2003. Prenatal stress diminishes neurogenesis in the dentate gyrus of juvenile rhesus monkeys. Biol. Psychiatry 54, 1025–1034. Connell, J.M., Fraser, R., MacKenzie, S.M., Friel, E.C., Ingram, M.C., Holloway, C.D., Davies, E., 2004. The impact of polymorphisms in the gene encoding aldosterone synthase (CYP11B2) on steroid synthesis and blood pressure regulation. Mol. Cell. Endocrinol. 217, 243–247. Czeh, B., Simon, M., Schmelting, B., Hiemke, C., Fuchs, E., 2006. Astroglial plasticity in the hippocampus is affected by chronic psychosocial stress and concomitant fluozetine treatment. Neuropsychopharmacology 31, 1616–1626. Czock, D., Keller, F., Rasche, F.M., Haussler, U., 2005. Pharmacokinetics and pharmacodynamics of systemically administered glucocorticoids. Clin. Pharmacokinet. 44, 61–98. de Kloet, E.R., 1991. Brain corticosteroid receptor balance and homeostatic control. Front. Neuroendocrinol. 12, 95–164. de Kloet, E.R., Rots, N.Y., van den Berg, D.T., Oitzl, M.S., 1994. Brain mineralocorticoid receptor function. Ann. N. Y. Acad. Sci. 746, 8–20. de Kloet, E.R., Van Acker, S.A., Sibug, R.M., Oitzl, M.S., Meijer, O.C., Rahmouni, K., et al., 2000. Brain mineralocorticoid receptors and centrally regulated functions. Kidney Int. 57, 1329–1336. Diorio, D., Viau, V., Meaney, M., 1993. The role of the medial prefrontal cortex (cingulate cortex) in the regulation of hypothalamic–pituitary–adrenal responses to stress. J. Neurosci. 13, 3839–3847. Ferrari, E., Fioravanti, M., Magri, F., Solerte, S.B., 2000. Variability of interactions between neuroendocrine and immunological functions in physiological aging and dementia of the Alzheimer's type. Ann. N. Y. Acad. Sci. 917, 582–596. Groschl, M., Rauh, M., Dorr, H.G., 2002. Cortisol and 17-hydroxyprogesterone kinetics in saliva after oral administration of hydrocortisone in children and young adolescents with congenital hyperplasia due to 21-hydroxylase deficiency. J. Clin. Endocrinol. Metab. 87, 1200–1204. Herman, J.P., Schafer, M.K., Young, E.A., Thompson, R., Douglass, J., Akil, H., Watson, S.J., 1989. Evidence for hippocampal regulation of neuroendocrine neurons of the hypothalamo–pituitary–adrenocortical axis. J. Neurosci. 9, 3072–3082. Herman, J.P., Ostrander, M.M., Mueller, N.K., Figueiredo, H., 2005. Limbic system mechanisms of stress regulation: hypothalamo–pituitary–adrenocortical axis. Prog. Neuropsychopharm. Biol. Psychiatry 29, 1201–1213. Hibberd, C., Yau, J.L., Seckl, J.R., 2000. Glucocorticoids and the ageing hippocampus. J. Anat. 197, 553–562. Hsu, F., Garside, M., Massey, A., McAllister-Williams, R., 2003. Effects of singe dose of cortisol on the neural correlates of episodic memory and error processing in healthy volunteers. Psychopharmacology 167, 431–442. Hwang, I., Yoo, K., Nam, Y., Choi, J., Lee, I., Kwon, Y., Kang, T.C., Kim, Y.S., Won, M.H., 2006. Mineralocorticoid and glucocorticoid receptor expressions in astrocytes and microglia in the gerbil hippocampal CA1 region after ischemic insult. Neurosci. Res. 54, 319–327. Jack, C.R.J., Petersen, R.C., O 'Brien, P.C., Tangalos, E.G., 1992. MR-based hippocampal volumetry in the diagnosis of Alzheimer's disease. Neurology 42, 183–188.

77

Jacobson, L., Sapolsky, R., 1991. The role of the hippocampus in feedback regulation of the hypothalamic–pituitary–adrenocortical axis. Endocr. Rev. 12, 118–134. Keller-Wood, M.E., Dallman, M.F., 1984. Corticosteroid inhibition of ACTH secretion. Endocr. Rev. 5, 1–24. Ladd, C.O., Huot, R.L., Thrivikraman, K.V., Nemeroff, C.B., Plotsky, P.M., 2004. Long-term adaptations in glucocorticoid receptor and mineralocorticoid receptor mRNA and negative feedback on the hypothalamo–pituitary–adrenal axis following neonatal maternal separation. Biol. Psychiatry 55, 367–375. Lupien, S.J., de Leon, M., de Santi, S., Convit, A., Tarshish, C., Nair, N.P., et al., 1998. Cortisol levels during human aging predict hippocampal atrophy and memory deficits. Nat. Neurosci. 1, 69–73. MacLullich, A.M., Deary, I.J., Starr, J.M., Ferguson, K.J., Wardlaw, J.M., Seckl, J.R., 2005. Plasma cortisol levels, brain volumes, and cognition in healthy elderly men. Psychoneuroendocrinology 30, 505–515. MacLullich, A.M., Ferguson, K.J., Wardlaw, J.M., Starr, J.M., Deary, I.J., Seckl, J.R., 2006. Smaller left anterior cingulate cortex volumes are associated with impaired hypothalamic–pituitary–adrenal axis regulation in healthy elderly men. J. Clin. Endocrinol. Metab. 91, 1591–1594. Meaney, M.J., O'Donnell, D., Rowe, W., Tannenbaum, B., Steverman, A., Walker, M., et al., 1995. Individual differences in hypothalamic–pituitary–adrenal activity in later life and hippocampal aging. Exp. Gerontol. 30, 229–251. Meaney, M.J., Diorio, J., Francis, D., Widdowson, J., LaPlante, P., Caldji, C., et al., 1996. Early environmental regulation of forebrain glucocorticoid receptor gene expression: implications for adrenocortical responses to stress. Dev. Neurosci. 18, 49–72. Newcomer, J., Selke, G., Melson, A., Hershey, T., Craft, S., Richards, K., et al., 1999. Decreased memory performance in healthy humans induced by stress-level cortisol treatment. Arch. Gen. Psychiatry 56, 527–533. O'Brien, J.T., Lloyd, A., McKeith, I., Gholkar, A., Ferrier, N., 2004. A longitudinal study of hippocampal volume, cortisol levels, and cognition in older depressed subjects. Am. J. Psychiatry 161, 2081–2090. Ohl, F., Michaelis, T., Vollmann-Honsdorf, G.K., Kirschbaum, C., Fuchs, E., 2000. Effect of chronic psychosocial stress and long-term cortisol treatment on hippocampus-mediated memory and hippocampal volume: a pilot-study in tree shrews. Psychoneuroendocrinology 25, 357–363. Oitzl, M., van Haarst, A.D., Sutanto, W., de Kloet, E.R., 1995. Corticosterone, brain mineralocorticoid receptors (MRs) and the activity of the hypothalamic–pituitary–adrenal (HPA) axis: the Lewis rat as an example of increased central MR capacity and a hyporesponsive HPA axis. Psychoneuroendocrinology 20, 655–675. Pearlson, G.D., Harris, G.J., Powers, R.E., Barta, P.E., Camargo, E.E., Chase, G.A., et al., 1992. Quantitative changes in mesial temporal volume, regional cerebral blood flow, and cognition in Alzheimer's disease. Arch. Gen. Psychiatry 49, 402–408. Petrovich, G.D., Canteras, N.S., Swanson, L.W., 2001. Combinatorial amygdalar inputs to hippocampal domains and hypothalamic behavior systems. Brain Res. Rev. 38, 247–289. Porter, R., Barnett, N., Idey, A., McGuckin, E., O'Brien, J., 2002. Effects of hydrocortisone administration on cognitive function in the elderly. J. Psychopharmacol. 16, 65–71. Rorden, C., Brett, M., 2000. Stereotaxic display of brain lesions. Behav. Neurol. 12, 191–200. Sapolsky, R.M., 1994. Physiological relevance of glucocorticoid endangerment of the hippocampus. Ann. N. Y. Acad. Sci. 746, 294–304. Sapolsky, R.M., Krey, L.C., McEwen, B.S., 1984. Glucocorticoid-sensitive hippocampal neurons are involved in

78

BR A IN RE S EA RCH 1 1 79 ( 20 0 7 ) 7 0 –78

terminating the adrenocortical stress response. Proc. Natl. Acad. Sci. 81, 6174–6177. Sapolsky, R., Romero, L., Munck, A., 2000. How do glucocorticoids influence stress responses? Integrating permissive, suppressive, stimulatory, and preparative actions. Endocr. Rev. 21, 55–89. Sousa, N., Almeida, O.F., Holsboer, F., Paula-Barbosa, M.M., Madeira, M.D., 1998a. Maintenance of hippocampal cell numbers in young and aged rats submitted to chronic unpredictable stress. Comparison with the effects of corticosterone treatment. Stress 2, 237–249. Sousa, N., Madeira, M.D., Paula-Barbosa, M.M., 1998b. Effects of corticosterone treatment and rehabilitation on the hippocampal formation of neonatal and adult rats. An unbiased stereological study. Brain Res. 794, 199–210. Starkman, M.N., Giordani, B., Gebarski, S.S., Berent, S., Schork, M.A., Schteingart, D.E., 1999. Decrease in cortisol reverses human hippocampal atrophy following treatment of Cushing's disease. Biol. Psychiatry 46, 1595–1602. Starkman, M.N., Giordani, B., Gebarski, S., Schteingart, D.E., 2003. Improvement in learning associated with increase in hippocampal formation volume. Biol. Psychiatry 53, 233–238. Sullivan, R.M., Gratton, A., 2002. Prefrontal cortical regulation of hypothalamic–pituitary–adrenal function in the rat and implications for psychopathology: side matters. Psychoneuroendocrinology 27, 99–114. Synek, V., Reuben, J.R., Du Boulay, G.H., 1976. Comparing Evans' index and computerized axial tomography in assessing relationship of ventricular size to brain size. Neurology 26, 231–233. Tuvnes, F.A., Steffenach, H., Murison, R., Moser, M., Moser, E.I., 2003. Selective hippocampal lesions do not increase adrenocortical activity. J. Neurosci. 23, 4345–4354. van der Beek, E.M., Wiegant, V.M., Schouten, W.G., van Eerdenburg, F.J., Loijens, L.W., van der Plas, C., et al., 2004.

Neuronal number, volume, and apoptosis of the left dentate gyrus of chronically stressed pigs correlate negatively with basal saliva cortisol levels. Hippocampus 14, 688–700. van Haarst, A.D., Oitzl, M.S., de Kloet, E.R., 1997. Facilitation of feedback inhibition through blockade of glucocorticoid receptors in the hippocampus. Neurochem. Res. 22, 1323–1328. Van Petten, C., 2004. Relationship between hippocampal volume and memory ability in healthy individuals across the life span: review and meta-analysis. Neuropsychologia 42, 1394–1413. Vythilingam, M., Vermetten, E., Anderson, G.M., Luckenbaugh, D., Anderson, E.R., Snow, J., et al., 2004. Hippocampal volume, memory, and cortisol status in major depressive disorder: effects of treatment. Biol. Psychiatry 56, 101–112. Ward, A.M., Syddall, H.E., Wood, P.J., Chrousos, G.P., Phillips, D.I., 2004. Fetal programming of the hypothalamic–pituitary–adrenal (HPA) axis: low birth weight and central HPA regulation. J. Clin. Endocrinol. Metab. 89, 1227–1233. Weitzman, E., Fukushima, D., Nogeire, C., Roffwarg, H., Gallagher, T., Hellman, L., 1971. Twenty-four hour pattern of the episodic secretion of cortisol in normal subjects. J. Clin. Endocrinol. Metab. 33, 14–22. Wiedenmayer, C.P., Bansal, R., Anderson, G.M., Zhu, H., Amat, J., Whiteman, R., Peterson, B.S., 2006. Cortisol levels and hippocampus volumes in healthy preadolescent children. Biol. Psychiatry 60, 856–861. Wolf, O.T., Convit, A., de Leon, M.J., Caraos, C., Qadri, S.F., 2002. Basal hypothalamo–pituitary–adrenal axis activity and corticotropin feedback in young and older men: relationships to magnetic resonance imaging-derived hippocampus and cingulate gyrus volumes. Neuroendocrinology 75, 241–249. Zatz, L.M., Jernigan, T.L., 1983. The ventricular–brain ratio on computed tomography scans: validity and proper use. Psychiatry Res. 8, 207–214.