Genetic and environmental effects on diurnal dehydroepiandrosterone sulfate concentrations in middle-aged men

Genetic and environmental effects on diurnaldehydroepiandrosterone sulfate concentrations inmiddle-aged men

Elizabeth C. Prom-Wormley a,*, Timothy P. York a, Kristen C. Jacobson b,Lindon J. Eaves a, Sally P. Mendoza c, Dirk Hellhammer d, Nicole Maninger c,Seymour Levine g, Sonia Lupien h, Michael J. Lyons i, Richard Hauger e,f,Hong Xian j, Carol E. Franz e, William S. Kremen e,f

aVirginia Institute for Behavioral and Psychiatric Genetics, Virginia Commonwealth University, United StatesbDepartment of Psychiatry, University of Chicago, United StatescCalifornia National Primate Research Center, University of California Davis, United StatesdDepartment of Psychobiology, University of Trier, GermanyeDepartment of Psychiatry, University of California San Diego, United StatesfVA San Diego Healthcare System, United StatesgDepartment of Psychiatry, University of California Davis, United StateshMental Health Research Centre Fernand Seguin, Hopital Louis-H Lafontaine, Universite de Montreal, CanadaiDepartment of Psychology, Boston University, United StatesjDepartment of Internal Medicine, Washington University, United States

Received 12 August 2010; received in revised form 27 January 2011; accepted 29 March 2011

Psychoneuroendocrinology (2011) 36, 1441—1452

KEYWORDSDHEAS;Diurnal concentrations;Genetics;Twin study;Aging;Men

Summary

Background: Dehydroepiandrosterone sulfate (DHEAS) is important for its association withimmune system function and health outcomes. The characterization of the genetic and environ-mental contributions to daily DHEAS concentrations is thus important for understanding thegenetics of health and aging.Methods: Saliva was collected from 783 middle-aged men (389 complete pairs and 5 unpairedtwins) as part of the Vietnam Era Twin Study of Aging. Samples were taken at multiple specifiedtime points across two non-consecutive days in the home and one day at the study sites. A twinmodeling approach was used to estimate genetic and environmental contributions for time-specific and average DHEAS concentrations.Results: There was a consistent diurnal pattern for DHEAS concentrations in both at-home andday-of-testing (DOT) measures, which was the highest at awakening and decreased slightly

* Corresponding author at: Virginia Institute for Psychiatric and Behavioral Genetics, Virginia Commonwealth University, PO Box 980126,Richmond, VA 23298-0126, United States. Tel.: +1 804 828 8154; fax: +1 804 827 0018.

E-mail addresses: [email protected], [email protected] (E.C. Prom-Wormley).

a va i l a ble at ww w. sc ie nce di r ect . com

j our na l h omepa g e: www.e l se v ie r.c om/l oca te/ psyne ue n

0306-4530/$ — see front matter. Published by Elsevier Ltd.
doi:10.1016/j.psyneuen.2011.03.018
http://dx.doi.org/10.1016/j.psyneuen.2011.03.018

mailto:[email protected]

mailto:[email protected]

http://dx.doi.org/10.1016/j.psyneuen.2011.03.018

throughout the day. Heritability estimates were significant for measures at 10 am, 3 pm andbedtime for the in-home days and at 10 am and 3 pm on the DOT, ranging between 0.37 and 0.46.Conclusions: The significant heritability estimates later in the day reflect time-specific geneticeffects for DHEAS, compared with prior twin and family designs studies which frequently usedaveraged morning-only measures. Additive genetic influences on DHEAS concentrations wereconsistent between at-home and DOT measures.Published by Elsevier Ltd.

1442 E.C. Prom-Wormley et al.

Introduction

Dehydroepiandrosterone (DHEA) and its sulfated metabolite(DHEAS) are the most abundant products secreted by thezona reticularis of the adrenal cortex (Orentreich et al.,1984). Like cortisol, DHEA and DHEAS are secreted fromthe adrenal cortex in response to adrenocorticotrophin(ACTH) stimulation (Pavlov et al., 1986; Parker et al.,1996). DHEA is the precursor for approximately 50% of theandrogens produced in adult men (Brown et al., 2002),underscoring the importance of this hormone in the produc-tion of sex steroids. DHEA is converted to the more stableDHEAS by DHEA sulfotransferase (HST, SULT2A1) and DHEASeasily becomes DHEA via steroid sulfatase (STS) (Krobothet al., 1999). DHEA is considered to be a biologically activehormone, and many studies focus on the association betweenDHEA and health-related outcomes. However, DHEAS mayhave a distinct role from DHEA in the etiology of disease,particularly in the regulation of the immune system (Radfordet al., 2010) and in neuroprotection (Maninger et al., 2009).Further, DHEAS is thought to be the circulating storage poolfor DHEA because DHEAS has a half-life of 10—20 h, while thehalf-life of DHEA is 1—3 h (Rosenfeld et al., 1975). Similarly,the clearance rate of DHEAS is much slower than that of DHEA(Longcope, 1996). There is a need to study the causes of theindividual variation of diurnal DHEAS production, which isanticipated to be caused by genetic and environmentalinfluences because it is the most abundant steroid hormonein the body, it is necessary for sex hormone synthesis, and isassociated with diseases related to aging.

The roles between DHEAS in the etiology ofchronic disorders related to aging

Low DHEAS concentrations have been associated with cor-onary artery disease, cardiovascular disease, non-insulindependent diabetes mellitus, rheumatoid arthritis, systemiclupus erythematosis, pemiphigoid/pemphigus, and HIV/AIDS,indicating the relationship between DHEAS concentrationsand the immune system (Chen and Parker, 2004). Both DHEASand cortisol modulate the immune system, although DHEASgenerally acts to enhance while glucocorticoids suppressimmune function (Butcher et al., 2005; Chen and Parker,2004). DHEAS has recently been reported to have immunos-timulatory effects, increasing superoxide generation inprimed human neutrophils in response to pathogens (Radfordet al., 2010). Additionally, DHEAS has anti-inflammatoryeffects through the inhibition of NF-kB activation (Iwasakiet al., 2004).

Lower concentrations of salivary DHEAS have been asso-ciated with depression (Barrett-Connor et al., 1999; Corpe-chot et al., 1981; Fabian et al., 2001; Goodyer et al., 1996,

2000; Takebayashi et al., 1998; Michael et al., 2000). Further,DHEAS has anxiolytic, anti-convulsant and sedative-hypnoticactions (Zinder and Dar, 1999). DHEAS is a ‘‘neurosteroid’’and as such is synthesized in the central nervous system denovo (Baulieu, 1981; Majewska, 1995). Increases in synapticconcentrations of DHEAS inhibit neuronal GABA-induced cur-rents and results in excitatory neurotransmission (Krobothet al., 1999; Debonnel et al., 1996). Modulation of neuro-transmission is related to brain function, which in turn con-tributes to different neuropsychological states (Majewska,2002).

DHEAS concentrations rise throughout childhood and peakin early adulthood followed by an age-dependent decline(Rainey and Nakamura, 2008). The lowest concentrations ofDHEAS occur between the ages of 65—70, pointing the pos-sible relevance of these hormones in age-related illness(Maninger et al., 2009). Few studies have tested associationsbetween genes related to DHEAS/DHEA function and disease.Further, of the studies testing the relationships betweengenes related to DHEAS concentrations and disease, no sig-nificant associations have been reported (Boger-Megiddoet al., 2008; Karlson et al., 2009). This may be due to anincomplete understanding of the relative importance of therole of genetic and environmental effects on diurnal regula-tion of DHEAS concentrations, particularly in aging adults.

Genetic epidemiology of DHEAS production

To date, seven family studies have estimated the impact ofgenetic and environmental factors on DHEAS (Table 1). Thefamily study design takes advantage of the familial correla-tions between individuals across generations (i.e.: parent—child, siblings, spouses, and grandparent—grandchild) toestimate the degree to which a trait is due to familialaggregation, which includes additive genetic effects andthose due to the shared (family) environment. In the absenceof additional information, nuclear family data and sibling-only samples are unable to resolve familial resemblance intogenetic and environmental effects since family membersshare both genetic and familial environments (Kendler andNeale, 2009; Rice and Borecki, 2001). Consequently, herit-ability estimates from these types of studies refer to themaximal effect of genes on a trait, also known as maximalheritability. One nuclear family study of 184 familiesreported an average heritability of 66% in families of AfricanAmerican descent and 58% in families of European Americandescent for baseline DHEAS concentrations (An et al., 2001).Another study of 348 families, reported a pooled, maximalheritability of 45% for DHEAS concentrations, unadjusted forsex differences. Maximal heritability estimates were found todiffer by gender in this sample, with estimates of 29% and 74%in men and women, respectively (Rice et al., 1993). A study

Table 1 Summary of heritability studies on DHEAS.

Author/date Studydesign

Sample Agerange

Ethnicity Collection Specimen Adjusted covariates Estimatedheritability

Rotter et al.(1985)

F N = 178Families = 26Males = 82Females = 96

10—90 EA One sample in hospital Bloodserum

Age, sex 65%

Rice et al.(1993)


NR EA One sample in clinic Bloodserum

Age, sex, effects of luteal phase Males = 29%Females = 74%Pooled = 45%

Jaquish et al.(1996a)


16—82 MA One morning sample in clinicafter 12-h fast

Bloodserum

Age, sex, contraceptive use,menopausal status, serum levelsof sex hormone binding globulin,% protein ingested/week, %EtOHingested/week, BMI, waist-hipratio, subcapsular to tricepsratio, HDL3, triglycerides, ApoB

39%

Mitchell et al.(1996)



Bloodserum

29%

Jaquish et al.(1996b)



Wholeblood

Age, sex 43%

An et al.(2001)

F N = 768Families = 184

17—65 AAEA

Two morning samples in labafter 12-h fast

Bloodserum

Age 66% (AA)58% (EA)

Yildiz et al.(2006)

F N = 131 (all female)Sister groups = 62

NR EA One morning sample between8—10:30 am in clinic after fastduring follicular phase ofmenstrual cycle

Bloodserum

BMI 43% (proband+ all sisters)

Meikle et al.(1988)

T 20 MZM pairs20 DZM pairs

20—60 EA Three pooled samples collectedbetween 8—9:30 am

Bloodplasma

Age, smoking, alcoholuse, exercise, BMI

58%

Meikle et al.(1997)

T 126 MZM pairs88 DZM pairs

25—75 EA Three pooled samples collectedbetween 8—10:30 am

Bloodplasma

None 26%

Nestler et al.(2002)

T 438 MZF pairs160 MZM pairs224 DZF pairs84 DZM pairs223 DZO pairs

NR EA One sample with time samplewas collected

Bloodserum

Age, sex 57% (<50 year)60% (50+ year)

F, family design; T, twin design; EA, European American Ethnicity; MA, Mexican American Ethnicity; MZM, monozygotic males; MZF, monozygotic females; DZM, dizygotic males; DZF, dizygoticfemales; DZO, opposite sex twin pair; NR, not reported.

Genetic

and

enviro

nmental

effe

cts on

DHEAS

1443


of women with polycystic ovary disease and their sisters(N = 62 sister pairs) reported a heritability 43% for probandsand their sisters (Yildiz et al., 2006).

Heritability estimates, or the degree to which additivegenetic influences contribute to DHEAS concentrations canbe estimated from extended (across several generations)family and twin studies. Extended family studies of DHEASranging in size from 25 to 42 families reported heritabilityestimate of DHEAS from 29% to 65% (Jaquish et al., 1996a,b;Mitchell et al., 1996; Rotter et al., 1985). In addition toestimating heritability, twin studies are able to estimate andtest for the relative importance of both common (often dueto shared familial environments) environmental and uniqueenvironmental effects (Eaves, 1982). To date, only three twinstudies ranging in size from 40 to 1129 pairs have reportedheritability estimates of 26—60% for DHEAS concentrations(Meikle et al., 1988, 1997; Nestler et al., 2002). Further,these studies report moderate variance due to unique envir-onmental effects.

Previous studies estimating the heritability of DHEAS havegenerally relied on a single measurement at any pointthroughout the day since the half-life of DHEAS can rangefrom 7 to 20 h and DHEAS concentrations are expected tohave low daily variation (Goodyer et al., 1996; Rosenfeldet al., 1975). Most studies have utilized one (Jaquish et al.,1996a,b; Mitchell et al., 1996; Rice et al., 1993; Rotter et al.,1985) or two (An et al., 2001) measures of DHEAS concentra-tions early in the day because they are highest in the morning(Nieschlag et al., 1973; Rosenfeld et al., 1975). Two studiestook multiple measures from each participant across a smallwindow in the morning, usually ranging from 8 to 10:30 amand used the pooled samples for analysis (Meikle et al., 1988,1997). To date, no known studies of DHEAS concentrationshave estimated the magnitude of genetic effects on DHEASconcentrations over multiple samples throughout the day.Therefore, measures at a single time point or over a smallperiod of time as in previous studies may be inadequate tounderstand the genetic and environmental contributions onDHEA/DHEAS and chronic disorders.

Prior genetic studies of DHEAS have used blood plasma toestimate heritability. However, salivary measurement ofDHEAS provides a unique opportunity to estimate geneticand environmental contributions to DHEAS for comparisonagainst prior studies using blood. Additionally, the ability tosystematically measure salivary DHEAS concentrations acrossthe day is advantageous in large-scale data collection anddoes not alter HPA function compared to intravenous sam-pling methods. DHEAS is a charged molecule, and is activelytransported through the neutral lipid membranes of thesalivary cells using a large family of organic anion transportpolypeptides (Konttinen et al., 2010; Pomari et al., 2009).Despite low concentrations of salivary DHEAS compared toplasma (Vining et al., 1983), these amounts are high enoughfor use as a result of high DHEAS concentrations in blood(Kroboth et al., 1999). Prior studies comparing salivary andplasma concentrations of DHEAS have shown strong associa-tions between saliva and blood (r = 0.738, p < 0.0001) (Lacet al., 1999).

In order to inform future studies of DHEAS and chronicdisorders related to aging as well as to improve geneticassociation studies using DHEAS measures, it is importantto estimate the genetic and environmental effects contribut-

ing to the individual differences of diurnal DHEAS concentra-tions. The present study measured salivary DHEASconcentrations across two non-consecutive days in the homeand one day in a testing facility in a large twin sample ofmiddle-aged men. Five samples were collected at specifictimes on each day. This design allows for the examination ofdifferences in the genetic and environmental influences onDHEAS both within and across days.

Methods

Research participants

The present study consists of 783 individuals comprising 193MZ and 196 DZ pairs (389 complete pairs) and 5 unpairedtwins who provided saliva samples for hormone measures aspart of the Vietnam Era Twin Study of Aging (VETSA); theVETSA has been described in detail elsewhere (Kremen et al.,2006). VETSA participants are men enrolled in the VietnamEra Twin (VET) Registry which comprises a sample of MZ andDZ twin pairs who served in the United States military duringthe Vietnam era (1965—1975), although the majority did notserve in combat or in Vietnam (Eisen et al., 1987; Hendersonet al., 1990). The VETSA twin pairs were randomly selectedfrom a pool of 3322 VET Registry twin pairs who had parti-cipated in a study of psychological health in 1992.

The VETSA sample consisted of 1237 twins between theages of 51 and 59 at the time of recruitment and bothmembers of a pair had to agree to participate. Twins traveledeither to University of California, San Diego or Boston Uni-versity for a day-long series of interviews and physical andcognitive assessments. In cases in which a twin could nottravel (N = 26 individuals, 3.5%) research assistants con-ducted assessments at a facility close to the twin’s home.The VETSA Hormone Study was later initiated to examine therole of hormonal regulation and its association with cognitiveaging in participants in the primary VETSA study and consistsof a smaller sub-sample of participants (N = 783) (Franzet al., 2010).

Participants completed two days of saliva collection athome, prior to the day of testing and then sent the salivasamples via overnight mail to the University of California,Davis to be assayed. In addition to the two days of salivacollection at home, participants provided saliva samplesduring the day of testing. IRB approval was obtained at bothsites, and all participants provided signed informed consent.A combination of DNA testing, previously obtained question-naire and blood group methods was used to determine zyg-osity (Eisen et al., 1989; Nichols and Bilbro, 1966; Peeterset al., 1998).

Procedures

Saliva collectionParticipants were contacted six weeks prior to the day oftesting (DOT) to establish the two ‘‘typical’’ working daysseparated by one day (preferably Tuesday/Thursday to avoidsampling on the beginning and end of the work week) onwhich they would provide the at-home saliva samples. At-home sample collection took place two to three weeks beforethe DOT in order to avoid the disruption of schedules that can

Genetic and environmental effects on DHEAS 1445

be caused by travel. Participants were asked the time theyusually woke up in the morning in order to individualize theirschedules and to set times on reminder watches. Saliva kitswere mailed to participants and participants were called theday prior to starting sampling to ensure that the reminderwatch was turned on, instructions were understood, and thesampling kit was placed by their bed for the morning sample.The saliva kit included all supplies: labeled 4.5 ml Cryotubesaliva vial, Trident original sugarless gum, straws to facilitatedrooling into each vial, tissues, instructions, a daily log, pen,a reminder watch, and a storage container with a MEMS 6TM

(Aardex) track cap for detecting compliance with the proto-col. All materials coming in contact with saliva, including thegum, have been checked for the presence of hormones orcompounds that could compromise the assays under thesupervision of SPM.

Participants provided samples at wake-up, 30 min afterwake-up, 10:00 am, 3:00 pm, and 9:00 pm or bedtime. Timeswere selected on the basis of appropriate times for themeasurement of cortisol, which was also measured as partof the VETSA Hormone Study. Diurnal DHEAS is secreted syn-chronously with cortisol, although the patterns of hormoneconcentrations differ throughout the day. The highest concen-trations of cortisol occur 30—60 min after awakening, whilethe highest concentrations of DHEAS occurring at awakeningfollowed by a gradual decline throughout the day (Huckle-bridge et al., 2005; Nieschlag et al., 1973). At the specifiedtime, participants selected the appropriate vial and providedapproximately 2.25 ml of saliva. If necessary, the participantchewed original Trident gum to stimulate saliva and removedthe gum prior to providing the sample. Previous testing by oneof the investigators (SPM) showed that this particular gum didnot alter DHEAS concentrations. Once a saliva sample wasprovided, participants were instructed to close the vial tightly,place it in the storage bottle, and complete the log entry forthat time period. Each opening of the storage bottle wasautomatically logged by the track cap. Participants were askedto keep the samples refrigerated and were provided with aninsulated lunch bag to keep the supplies together. The remin-der watch was programmed to go off at each scheduled time;however, the time protocol was carefully explained to parti-cipants (verbally and in writing) to allow for normal variationsin schedules. Reminder watch times were individualized sothat all participants provided samples at equivalent times(typical awakening, awake + 30 min, awake + 4 h, awa-ke + 9 h, and bedtime).

Saliva samples were also collected on the DOT. Subjectsreceived their supplies when they arrived at the hotel the daybefore. As in the at-home sampling, the first sample wascollected at awakening the morning of the day of testing atthe hotel, then half an hour after awakening at hotel, thenapproximately 10:00 am and 3:00 pm in the laboratory, andbedtime at the hotel. Test day protocols were standardizedacross sites. Participants completed log entries followingeach sample and provided a detailed description of whatthey ate for lunch.

DHEAS assaysSamples were centrifuged at 3000 rpm for 20 min to separatethe aqueous component from mucins and other suspendedparticles. Salivary concentrations of DHEAS were estimatedin duplicate using commercial radioimmunoassay kits (Diag-

nostics Systems Laboratories, Webster, TX) according tomanufacturer’s instructions. Assay sensitivity (least detect-able dose) was 0.0206 ng/ml. Intra- and inter-assay coeffi-cients of variation were 5.36% and 5.82%, respectively.Samples where the DHEAS concentration was greater than3 standard deviations from the mean were considered to beoutliers and coded as missing (N = 49 samples). All 17 samplesfrom each participant were analyzed in the same assay; oneto three individuals were included in the same assay batch.Assays were performed without knowledge of the zygosity ofthe participant.

Scores were imputed for missing values only if the parti-cipant had no more than one missing value on a day. Missingdata were imputed by calculating the full samples’ meanDHEAS change between the time point with the missing valueand the adjacent time point. For all time points exceptawakening, the time point prior to the missing value wasused. The mean DHEAS change for those two points was thenadded (or subtracted) from the individual participant’s non-missing time point to get the imputed value for the missingtime point in question. For example, if a participant wasmissing a value for the 3 pm measure, the full samples’ meanchange cortisol from 10 am to 3 pm hours was calculated. Thisvalue was then subtracted from the participant’s own 10 amvalue to obtain the imputed 3 pm value. The distributions ofDHEAS values were skewed, and as such were natural logtransformed prior to data analysis in order to produce normaldistributions.

At-home DHEAS concentrations were averaged at corre-sponding times on day one and day two to create a singlevalue for analysis; which was supported by the high inter-correlations observed between daily measures of the sametime point, which ranged from 0.73 to 0.78 ( p < 0.0001).Correlations between DHEAS samples at each time point andage were also small, but some were significant; these corre-lations ranged from r = �0.02 ( p = 0.42) to r = �0.15( p < 0.0001). Further, there were no significant associationsbetween salivary DHEAS concentrations and common covari-ates including BMI, current insulin use, current smoking, orcurrent use of hypertensive medication. Subsequent analysesadjusted for the effect of age on the mean concentrations ofDHEAS.

Data analysis

Determination of study sample representativenessThe participants included in this study were compared withthe participants in the VETSA who were not included todetermine whether the sub-sample used was representativeof the larger VETSA population. Age, body mass index, andeducational attainment were compared using a Wilcoxonrank-sum test. Current smoking, current self-reporteddepression, ever receiving a depression diagnosis by a phy-sician, and current use of any prescribed medications werecompared using the x2-test.

Use of a Markov Chain Monte Carlo approach toestimate genetic and environmental effects in thepresence of batch and age effectsThe classic twin study is based on the comparison of similaritiesbetween monozygotic (MZ) and dizygotic (DZ) twin pairs. MZtwins are expected to be more similar compared to DZ pairs


because MZ twins share 100% of their genes, while dizygotic(DZ) twins share on average 50% of their genes. Therefore, agreater difference in measures of similarity between MZ andDZ pairs suggests the presence of additive genetic effects on atrait. Twin modeling approaches estimate proportions of thetotal variance due to genetic and environmental sources,specifically: (1) additive genetic influences (A); (2) commonenvironmental influences (C) which represents life experi-ences shared between twin pairs, making them more alike;and (3) unique environmental influences (E) which representsexperiences that make siblings different (e.g., having a spousedie). The estimate of unique environmental influences alsoincludes measurement error.

Estimation of genetic and environmental contributionsWe controlled for the possibility of extraneous effects thatcan arise in laboratory measures due to instances of membersof a twin pair that were assayed simultaneously on the sameassay instead of randomly across all assays. In the context oftwin studies, these so-called ‘‘batch’’ effects will increasewithin pair similarity as a consequence of more similarlaboratory conditions and can be mistakenly attributed tothe effects of the shared environment. The pattern of cor-relations between unrelated individuals in the same batch,and related individuals (twin pairs) in different batches wasused to estimate the random effects of differences betweenbatches. Correlations for MZ and DZ twin pairs could then beestimated without the confounded effects of differencesbetween batches. We also controlled for the effect of ageon DHEAS concentrations since increasing age was signifi-cantly associated with lower DHEAS concentrations for sometime points.

The simultaneous estimation of the components of var-iances due to batch and age differences and differencesbetween and within twin pairs was done within a Bayesianframework using Markov Chain Monte Carlo (MCMC) methodsin the freely available package WinBUGS (Speigelhalteret al., 2004). We denote the estimated DHEAS concentrationof the jth twin of the ith pair as Yijk, where the subscript kindicates the batch in which the value was assayed such thatassays in the same batch have the same value of k.

We then let : Yi jk ¼ m þ ti j þ ai þ bk

where m represents the measured mean DHEAS concentra-tion, tij represents the random differences between twins,conceivably correlated between pairs, and ai represents therandom differences across ages between pairs, and bk repre-sents the random differences between batches. For eachtype of twin (MZ and DZ) we assume that the twin effects tij,are bivariate normal with standard deviation st and intraclasscorrelation rMZ for MZ and rDZ for DZ pairs, respectively.Batch and age effects are assumed to have a normal distri-bution ðN½0; s2

bk� and N½0; s2ai�Þ.

When the MCMC algorithm converges, it yields successivesamples from the posterior distribution that may be used toestimate summary statistics such as confidence intervals ofmodel parameters. We assumed a relative uninformativenormal prior for m. We assumed broad uniform priors on st

and sb, following a suggestion of Spiegelhalter et al. (2003).The twin correlations for MZ and DZ pairs were sampledinitially from a uniform prior distribution over the range0—0.9.

Random effects were simulated for each pair (pi), age (ai)and batch (bk) using the appropriate between-pair compo-nent of variance for each twin type. Individual twin effects(tij) were simulated to have a normal distribution ðN½pI; s2

tw �Þ,where the intra-pair variance is s2

tw ¼ s2t ð1 � rÞ, r being the

intraclass correlation for MZ or DZ twins as appropriate.Ancillary summary statistics were computed from successivesamples of rMZ for and rDZ together with their confidenceintervals. These statistics include estimates of broad senseheritability (H), an indicator of the proportion of totalpopulation variance that is due to genetic variation andthe shared environment (C), an estimate of the relativecontribution of the variance shared between twins. Herit-ability was estimated using Holzinger’s H, H = 2(rMZ � rDZ)and the contribution of the shared environment was esti-mated as C = 2rDZ � rMZ. Samples from the posterior distri-bution of H and C allowed for the estimation of theirconfidence intervals. Note that estimates C may be negativeif there are large non-additive genetic effects. Similarly,estimates of H may be negative when no significant geneticeffects are present. Thus, this approach to the estimation ofH and C avoids the biases inherent in constraining either to begreater than zero in the more familiar maximum likelihoodcomponents of variance approach. A copy of the WinBUGScode, together with illustrative data structure and initialvalues may be obtained from the first author.

Results

Sample representativeness

As expected from the later start of the VETSA Hormone Studycompared to the full VETSA sample, the average age ofparticipants included in this sub-sample was older (55.92;SD = 2.58) compared to VETSA participants who were not inthe neuroendocrine study (54.63; SD = 2.05); this is a small,but significant difference ( p < 0.0001). There were no sig-nificant differences for BMI ( p = 0.15) or educational attain-ment ( p = 0.21) between those who were part of the sub-sample and larger VETSA sample. Further, there were nosignificant differences between the groups for the preva-lence of any prescribed drug use ( p = 0.40), current smoking( p = 1.00), current self-reported depression ( p = 0.63), orany depression diagnosed by physician ( p = 0.77).

Summary statistics

There was a consistent diurnal pattern for salivary DHEASconcentrations in both at-home and DOT measures (Figs. 1and 2). Repeated-measures contrasts were used to compareaverage unadjusted concentrations of DHEAS as well as age-adjusted concentrations from each time point to the next(i.e.: awake vs. awake + 30, and awake + 30 vs. 10 am).DHEAS concentrations were significantly higher at awakeningand at awake + 30 min ( p < 0.0001) for both unadjusted andage-adjusted measures. Measures taken between10 am andbedtime did not differ significantly from one another. DHEASconcentrations between at-home and DOT measures acrossspecific times were generally consistent, ranging from 0.57 to0.83 ( p < 0.0001) in unadjusted measures and 0.61—0.75( p < 0.0001) for age-adjusted measures.

Univaria

te ge

netic

analysis

of

DHEAS

concentra

tions

There

was

significan

t heritab

ility fo

r th

e 10

am,

3 pm

and

ove

rall daily

mean

conce

ntratio

ns

of

unad

juste

d DHEAS

dur-

ing

the

at-home

days

(Table

2)

and

for

the

DOT

(Table

3).Heritab

ility estim

ates

for

these

measu

res

range

d betw

een

0.32 an

d 0.47,

with

95% co

nfidence

intervals

as lo

w as

0.01 to

as h

igh as

0.81. Sign

ifican

t ad

ditive

genetic

effe

cts w

ere

detecte

d fo

r th

e 10

am,

3 p

m an

d b

edtim

e age

-adjuste

dmeasu

res

durin

g the

at-h

ome

days

(Tab

le 4).

Sign

ifican

theritab

ility was

detecte

d for

the

10

am

an

d 3

pm

age

-ad

juste

d m

easu

res

on

the

DOT

(Table

5). H

eritab

ility esti-

mate

s fo

r th

ese

measu

res

range

d betw

een

0.37 an

d 0.46,

with

95% co

nfidence

intervals

as lo

w as

0.01 to

as high

as

Bedtim

e3P

M10A

MA

wake+30

Aw

ake

98765432

Measurem

ent Times

Unadjusted Mean Levels of DHEAS (ng/ml)

Figu

re 1

Unad

juste

d m

ean

conce

ntra

tions

of

log-tra

nsfo

rmed

DHEAS

acro

ss at-h

ome

day

measu

res.

Figu

re 2

Unad

juste

d m

ean

conce

ntra

tions

of

log-tra

nsfo

rmed

DHEAS

acro

ss te

sting

day

measu

res.

mates of genetic and environmental effects for unadjusted at-home testing measures of DHEAS.

AS rMZ 95% CI rDZ 95% CI A 95% CI C 95% CI E

0.32 (0.18; 0.44) 0.27 (0.14; 0.40) 0.09 (�0.26; 0.44) 0.22 (�0.07; 0.51) 0.680.39 (0.27; 0.50) 0.28 (0.12; 0.42) 0.22 (�0.12; 0.59) 0.17 (�0.16; 0.46) 0.610.47 (0.36; 0.57) 0.31 (0.18; 0.44) 0.32 (0.01; 0.63) 0.15 (�0.12; 0.41) 0.530.46 (0.34; 0.56) 0.22 (0.08; 0.36) 0.47 (0.12; 0.81) �0.01 (�0.31; 0.27) 0.540.35 (0.22; 0.47) 0.19 (0.06; 0.33) 0.31 (�0.05; 0.66) 0.04 (�0.25; 0.33) 0.650.48 (0.36; 0.58) 0.29 (0.14; 0.41) 0.38 (0.07; 0.71) 0.1 (�0.20; 0.35) 0.52

n; A, estimate of broad sense heritability, 2(rMZ � rDZ); C, estimate of shared environmental effects, 2 � rMZ � rDZ; E, estimate of

Genetic

and

enviro

nmental

effe

cts on

DHEAS

Table 2 Summary statistics and variance components esti

Days 1 and 2 Mean DHEAS levelsin ng/ml (SD)

Mean time of DHEcollection (SD)

Awake 6.54 (0.05) 0631 h (2.25)Awake + 30 5.59 (0.05) 0684 h (2.26)10 am 5.12 (0.05) 1019 h (1.56)3 pm 5.21 (0.05) 1502 h (1.66)Bedtime 5.19 (0.05) 2040 h (3.63)Home overall 5.53 (0.05)

All bold numbers indicate significant estimates at p < 0.05.rMZ, monozygotic twin correlation; rDZ, dizygotic twin correlatiounique environmental effects 1 � r .
MZ
1447

Table 3 Summary statistics and variance components estimates of genetic and environmental effects for unadjusted day of testing measures of DHEAS adjusted for age.

Test day Mean DHEASlevels in ng/ml (SD)

Mean time ofDHEAS collection (SD)

rMZ 95% CI rDZ 95% CI A 95% CI C 95% CI E

Awake 6.39 (0.05) 0589 h (0.78) 0.38 (0.24; 0.50) 0.26 (0.09; 0.40) 0.24 (�0.13; 0.62) 0.14 (�0.19; 0.43) 0.62Awake + 30 5.54 (0.05) 0644 h (0.87) 0.39 (0.27; 0.50) 0.27 (0.14; 0.40) 0.23 (�0.10; 0.58) 0.16 (�0.14; 0.43) 0.6110 am 5.37 (0.05) 0973 h (0.58) 0.35 (0.22; 0.46) 0.16 (0.03; 0.30) 0.37 (0.02; 0.73) �0.03 (�0.32; 0.27) 0.653 pm 5.37 (0.05) 1493 h (0.98) 0.39 (0.27; 0.51) 0.20 (0.08; 0.34) 0.38 (0.02; 0.72) 0.01 (�0.26; 0.31) 0.61Bedtime 5.25 (0.05) 2058 h (4.67) 0.32 (0.15; 0.45) 0.16 (0.02; 0.32) 0.31 (�0.05; 0.69) 0.01 (�0.30; 0.31) 0.68Test overall 5.59 (0.04) 0.45 (0.33; 0.55) 0.26 (0.13; 0.38) 0.38 (0.05; 0.69) 0.07 (�0.19; 0.33) 0.55

rMZ, monozygotic twin correlation; rDZ, dizygotic twin correlation; A, estimate of broad sense heritability, 2(rMZ � rDZ); C, estimate of shared environmental effects, 2 � rMZ � rDZ; E, estimate ofunique environmental effects 1 � rMZ.

Table 4 Summary statistics and variance components estimates of genetic and environmental effects for at-home testing measures of DHEAS adjusted for age.

Measure Mean DHEASlevels in ng/ml (SD)

Mean time ofDHEAS collection (SD)

rMZ 95% CI rDZ 95% CI A 95% CI C 95% CI E

Waking 6.63 (0.05) 0631 h (2.25) 0.33 (0.19; 0.45) 0.30 (0.16; 0.41) 0.07 (�0.29; 0.42) 0.26 (�0.03; 0.52) 0.67Awake + 30 5.71 (0.06) 0684 h (2.26) 0.39 (0.27; 0.5) 0.27 (0.13; 0.41) 0.24 (�0.11; 0.59) 0.15 (�0.14; 0.44) 0.6110 am 5.24 (0.05) 1019 h (1.56) 0.44 (0.31; 0.55) 0.23 (0.04; 0.37) 0.41 (0.07; 0.78) 0.03 (�0.33; 0.32) 0.563 pm 5.33 (0.05) 1502 h (1.66) 0.44 (0.32; 0.55) 0.21 (0.09; 0.34) 0.46 (0.13; 0.78) �0.02 (�0.29; 0.25) 0.56Bedtime 5.32 (0.05) 2040 h (3.63) 0.35 (0.21; 0.47) 0.16 (0.05; 0.29) 0.37 (0.01; 0.7) �0.02 (�0.28; 0.26) 0.65Mean 5.93 (0.05) 0.45 (0.33; 0.56) 0.29 (0.11; 0.42) 0.31 (�0.02; 0.71) 0.14 (�0.24; 0.41) 0.55

Bold values indicate significant estimate; rMZ, monozygotic twin correlation; rDZ, dizygotic twin correlation; A, estimate of broad sense heritability, 2(rMZ � rDZ); C, estimate of shared environmentaleffects, 2 � rMZ � rDZ; E, estimate of unique environmental effects 1 � rMZ.

1448

E.C.

Prom-W

orm

ley

et

al.

Table

5Su

mmarystatistics

andva

rian

ceco

mponents

estim

atesofge

netican

denvironmentaleffectsforday

oftestingmeasuresofDHEASad

justedforag

e.

Measure

MeanDHEAS

leve

lsin

ng/

ml(SD)

Mean

timeof

DHEASco

llection(SD)

r MZ

95%CI

r DZ

95%CI

A95

%CI

C95

%CI

E

Wak

ing

6.41(0.05)

058

9h(0.78)

0.35

(0.22;

0.47

)0.22

(0.09;

0.35

)0.26

(�0.10

;0.61

)0.09

(�0.20

;0.37

)0.65

Awake

+30

5.57(0.06)

0644h(0.87)

0.41

(0.28;

0.52

)0.27

(0.13;

0.38

)0.28

(�0.04

;0.61

)0.12

(�0.15

;0.38

)0.59

10am

5.40(0.05)

097

3h(0.58)

0.34

(0.21;

0.46

)0.16

(0.03;

0.29

)0.37

(0.02;

0.69

)�0.03

(�0.28

;0.26

)0.66

3pm

5.42

(0.05)

149

3h(0.98)

0.38

(0.24;

0.5)

0.19

(0.06;

0.33

)0.38

(0.03;

0.71

)0

(�0.27

;0.29

)0.62

Bedtime

5.31(0.05)

2058

h(4.67)

0.34

(0.19;

0.46

)0.18

(0.05;

0.31

)0.32

(�0.05

;0.67

)0.01

(�0.26

;0.30

)0.66

Mean

5.75

(0.05)

0.42

(0.3;0.53

)0.26

(0.12;

0.39

)0.32

(�0.02

;0.66

)0.10

(�0.19

;0.37

)0.58

Bold

valuesindicatesign

ifica

ntestim

ate;r M

Z,monozygo

tictw

inco

rrelation;r D

Z,dizygotictw

inco

rrelation;A,estim

ateofbroad

sense

heritability,2(r M

Z�

r DZ);C,estim

ateofsharedenvironmental

effects,

2�

r MZ�

r DZ;E,estim

ateofuniqueenvironmentaleffects1�

r MZ.


0.78. The heritability estimates of mean DHEAS concentra-tions on both at-home days and for DOT were no longersignificant after adjusting for age.

For both unadjusted and age-adjusted DHEAS concentra-tions, the heritability estimates for awake and awake + 30 -min measures were not significant either at-home or for DOTmeasures. The variance due to the shared environment wasnot significant for any at-home or DOT measures. The major-ity of the variance of diurnal DHEAS concentrations was dueto unique environmental influences (0.52—0.68).

Discussion

This is the first study to characterize the diurnal concentra-tions of salivary DHEAS across multiple salivary time-points inboth at-home and in-lab settings in a community-basedsample of middle-aged men. The pattern and variation ofDHEAS concentrations is consistent across days, with signifi-cantly higher concentrations occurring at awakening andawake + 30 min compared to later in the day. This has beenreported in similar studies of humans using blood with smallersample sizes (Nicolau et al., 1985; Rosenfeld et al., 1975;Carlstrom et al., 2002; Zhao et al., 2003) as well as non-human primates (Maninger et al., 2010). Additionally, highcorrelations across the days indicated stable and constantconcentrations of diurnal DHEAS. Consequently, salivary con-centrations may reliably reflect patterns of DHEAS in bloodand in turn provide insight into the steroidogenic activity ofthe zona reticularis region of the adrenal cortex.

To date, this is the largest population-based twin study toestimate the genetic and environmental contributions acrossmultiple salivary measures of diurnal DHEAS concentrationsusing multiple days and environmental settings. SalivaryDHEAS is a heritable measure, with genetic effects account-ing for 37—46% of the total variance for the late morning (10am) and afternoon (3 pm) age-adjusted measures on at-homeand DOT. These estimates and their 95% confidence intervalsfall within the range of those previously reported from familyand twin studies of DHEAS (26—66%) (Meikle et al., 1988;Nestler et al., 2002). These heritability estimates areexpected to reflect the role genetic effects in an agingpopulation which may differ across development. For exam-ple, in a small pilot study of adult twin men with an averageage of 32, Meikle et al. (1988) reported an estimate ofheritability of 58%. In a follow-up study of a larger sampleof twin men with average age of approximately 54.5 years,heritability was estimated at 26% (Meikle et al., 1997).

The consistency of mean DHEAS concentrations betweenat-home and DOT days as well as for the magnitudes of thegenetic and environmental effects for 10 am and 3 pmmeasures suggests that DHEAS is not responsive to mildstressors similar to those experienced during the day oftesting. These mild stressors included study involvement ina laboratory setting and the long-distance travel required toparticipate. DHEAS has historically been considered a goodmarker of individual adrenal cortex function and is expectedto reflect chronic rather than acute response to the environ-ment (Baulieu, 1996). Recent studies have reported signifi-cant differences in DHEAS concentrations in individualsexposed to chronic stressors such as burnout and caring forpatients with Alzheimer’s disease (Jeckel et al., 2010). Most


studies of the HPA axis have focused on cortisol secretion.However, DHEAS concentrations may also be an indicator ofHPA axis activity in response to chronic stressors. A recentstudy of rhesus monkeys found that DHEAS concentrationsincreased in response to a chronic (repeated) stressor (Man-inger et al., 2010). The current results may reflect thegenetic and environmental effects of the response relatedto exposure to chronic stressors and may help to identifyadditional candidate genes related to HPA axis activity.

The significant heritability estimates for the afternoonrather than morning measures or the average daily concen-trations suggest that some aspects of DHEAS production maybe under greater genetic control than others. DHEAS con-centrations are not expected to be highly variable through-out the day (Kroboth et al., 1999). Given the demonstrateddiurnal rhythm of DHEAS in this and other studies as well assignificant heritability estimates for afternoon measures, theuse of multiple measures across the day may better identifyimportant time-specific genetic and environmental effectsrelated to daily concentrations of this hormone for use ingenetic association studies rather than single or morning-onlymeasures (Boger-Megiddo et al., 2008; Karlson et al., 2009).The variation associated with diurnal DHEAS concentrationsand associated time-specific genetic effects identify a needto study the patterns of DHEAS concentrations across the dayrather than only for a specific time-point. Additionally, eva-luation of diurnal regulation of DHEAS across developmentmay be an important risk factor for disease. Prior studies ofDHEAS across the lifespan indicated that lifetime trajectoriesof DHEAS concentrations, rather than baseline DHEAS con-centrations alone were associated with higher mortality inolder adults (Cappola et al., 2009) as well as cardiovasculardisease (Sanders et al., 2010). Consequently, these resultsencourage the characterization diurnal DHEAS profiles andthe estimation of genetic and environmental contributions toprofiles across the lifespan.

These results should be evaluated in the light of thefollowing limitations. First, this is a sample of middle-agedCaucasian men and these findings may not generalize towomen or other ethnic populations. Additionally, becauseDHEAS concentrations are age-specific, these results may notgeneralize to other periods of development. Nevertheless,the range of these moderate heritability estimates and their95% confidence intervals suggests that they are within therange of other studies with participants across greater ageranges and consisting of men and women. Further, theseestimates serve as a baseline for the genetic and environ-mental effects on DHEAS concentrations throughout lateradulthood because this population has been collected as partof a study of aging. Second, no adjustments were made toaccount for associations between DHEAS and currentaffected status for common chronic diseases since therewere no significant associations between DHEAS concentra-tions and common risk factors of chronic disease such as BMI,current insulin use, current smoking, or current use ofhypertensive medication. However, it is expected that thosewith a current diagnosis may have altered concentrations ofDHEAS. Third, chewing gum was recently reported to have aneffect on concentrations of salivary testosterone, estradioland immunoglobulin A assays although the mechanism bywhich this effect is not understood (van Anders, 2010). Twopossibilities explaining this effect were suggested: (1) chew-

ing gum stimulates analyte production and (2) chewing guminteracts with assay results. No data was collected on thisaspect of hormone production and consequently the effectsof chewing gum on mean concentrations of DHEAS were notincluded. However, chewing gum represents non-systematicerror and its effects would be included in the estimate ofunique environmental effects. Further, it is unlikely thatgenetic influences on chewing gum will bias estimates onthe variance due genetic effects for salivary DHEAS. Fourth,no blood DHEAS concentrations were collected for compar-ison against the salivary measures. Consequently, theseresults will require replication in other populations with bothblood and salivary samples for to compare estimates ofgenetic and environmental effects.

These results have demonstrated patterns between sali-vary samples of diurnal DHEAS concentrations to be similar tothose of blood. DHEAS obtained from saliva is heritable whichis anticipated to encourage future population-based studiesof DHEAS. Further, the heritability of DHEAS is higher for mid-morning and early afternoon measures, suggestive of time-specific genetic influences on DHEAS. Future research ofDHEAS has the potential to improve understanding of diseaseetiology and may span across several concentrations ofinquiry including: (1) the use of cross-sectional data todetermine associations of diurnal DHEAS concentrations withspecific outcomes as well as the presence of genetic variancethat is common to both the outcome and DHEAS, (2) theanalysis of longitudinal data to determine whether thegenetic contributions of DHEAS concentrations vary acrossage and (3) the analysis of DHEAS concentrations and itsrelationship with functionally related hormones such as cor-tisol and testosterone to understand the biological pathwaybetween DHEAS, other sex hormones and related outcomes.

Role of funding source

Funding for this study was provided by NIMH Training grantMH-20030 and National Institutes of Health (NIA) grantsAG018386, AG022381 and AG018384. These funding agencieshad no additional role in the study design, data collection,data analysis and interpretation, writing of the manuscript,or in the decision to submit the paper for publication.

Conflict of interest

The authors declare no conflicts of interest.

Acknowledgements

The US Department of Veteran Affairs has provided financialsupport for the development and maintenance of the Viet-nam Era Twin (VET) Registry. The authors gratefully acknowl-edge the continued cooperation and participation of the VETRegistry and their families.

References

An, P., Rice, T., Gagnon, J., Hong, Y., Leon, A.S., Skinner, J.S.,Wilmore, J.H., Bouchard, C., Rao, D.C., 2001. Race differencesin the pattern of familial aggregation for dehydroepiandrosterone


sulfate and its responsiveness to training in the HERITAGE FamilyStudy. Metabolism 50, 916—920.

Barrett-Connor, E., von Muhlen, D., Laughlin, G.A., Kripke, A., 1999.Endogenous levels of dehydroepiandrosterone sulfate, but notother sex hormones, are associated with depressed mood in olderwomen: the Rancho Bernardo Study. J. Am. Geriatr. Soc. 47,685—691.

Baulieu, E.-E., 1981. Steroid hormones in the brain: several mechan-isms. In: Fuxe, J.A., Gustafson, L., Wetterberg, (Eds.), SteroidHormone Regulation of the Brain. Pergamon, Elmsford, NY, pp.3—14.

Baulieu, E.E., 1996. Dehydroepiandrosterone (DHEA): a fountain ofyouth? J. Clin. Endocrinol. Metab. 81, 3147—3151.

Boger-Megiddo, I., Weiss, N.S., Barnett, M.J., Goodman, G.E., Chen,C., 2008. V89L polymorphism of the 5alpha-reductase type II gene(SRD5A2), endogenous sex hormones, and prostate cancer risk.Cancer Epidemiol. Biomarkers Prev. 17, 286—291.

Brown, R.C., Liu, Y., Papadopoulos, V., 2002. DHEA: biosynthesis,regulation and function in the central nervous system. In: Morfin,R. (Ed.), DHEA and the Brain. Taylor and Francis, Paris, pp. 65—80.

Butcher, S.K., Killampalli, V., Lascelles, D., Wang, K., Alpar, E.K.,Lord, J.M., 2005. Raised cortisol:DHEAS ratios in the elderly afterinjury: potential impact upon neutrophil function and immunity.Aging Cell 4, 319—324.

Cappola, A.R., O’Meara, E.S., Guo, W., Bartz, T.M., Fried, L.P.,Newman, A.B., 2009. Trajectories of dehydroepiandrosteronesulfate predict mortality in older adults: the cardiovascularhealth study. J. Gerontol. A: Biol. Sci. Med. Sci. 64, 1268—1274.

Carlstrom, K., Karlsson, R., von Schoultz, B., 2002. Diurnal rhythmand effects of oral contraceptives on serum dehydroepiandros-terone sulfate (DHEAS) are related to alterations in serum albu-min rather than to changes in adrenocortical steroid secretion.Scand. J. Clin. Lab. Invest. 62, 361—368.

Chen, C.C., Parker Jr., C.R., 2004. Adrenal androgens and theimmune system. Semin. Reprod. Med. 22, 369—377.

Corpechot, C., Robel, P., Axelson, M., Sjovall, J., Baulieu, E.-E.,1981. Characterization and measurement of dehydroepiandros-terone sulfate in rat brain. Proc. Natl. Acad. Sci. U.S.A. 78,4704—4707.

Debonnel, G., Bergeron, R., de Montigny, C., 1996. Potentiation bydehydroepiandrosterone of the neuronal response to N-methyl-D-aspartate in the CA3 region of the rat dorsal hippocampus: aneffect mediated via sigma receptors. J. Endocrinol. 150 (Suppl.),S33—S42.

Eaves, L.J., 1982. The utlity of twins. In: Anderson, V.E., Hauser,W.A., Penry, J.K., Sing, C.F. (Eds.), Genetic Basis of the Epilep-sies. Raven Press, New York, pp. 249—276.

Eisen, S.A., True, W.R., Goldberg, J., Henderson, W., Robinette,C.D., 1987. The Vietnam Era Twin, VET. Registry: method ofconstruction. Acta. Genet. Med. Gemellol., Roma. 36, 61—66.

Fabian, T.J., Dew, M.A., Pollock, B.G., Reynolds III, C.F., Mulsant,B.H., Butters, M.A., Zmuda, M.D., Linares, A.M., Trottini, M.,Kroboth, P.D., 2001. Endogenous concentrations of DHEA andDHEA-S decrease with remission of depression in older adults.Biol. Psychiatry 50, 767—774.

Franz, C.E., York, T.P., Eaves, L.J., Mendoza, S.P., Hauger, R.L.,Hellhammer, D.H., Jacobson, K.C., Levine, S., Lupien, S.J.,Lyons, M.J., Prom-Wormley, E., Xian, H., Kremen, W.S., 2010.Genetic and environmental influences on cortisol regulationacross days and contexts in middle-aged men. Behav. Genet..

Goodyer, I.M., Herbert, J., Altham, P.M.E., Pearson, J., Secher, S.M.,Shiers, H.M., 1996. Altered diurnal rhythms in salivary cortisoland dehydroepiandosterone (DHEA) at presentation. Psychol.Med. 26, 245—256.

Goodyer, I.M., Herbert, J., Tamplin, A., Altham, P.M., 2000. First-episode major depression in adolescents. Affective, cognitive andendocrine characteristics of risk status and predictors of onset.Br. J. Psychiatry 176, 142—149.

Henderson, W.G., Eisen, S.E., Goldberg, J., True, W.R., Barnes, J.E.,Vitek, M., 1990. The Vietnam Era Twin Registry: a resource formedical research. Public Health Rep. 105, 368—373.

Hucklebridge, F., Hussain, T., Evans, P., Clow, A., 2005. The diurnalpatterns of the adrenal seroids cortisol and dehydroepiandroster-one (DHEA) in relation to awakening. Psychoneuroendocrinology30, 51—57.

Iwasaki, Y., Asai, M., Yoshida, M., Nigawara, T., Kambayashi, M.,Nakashima, N., 2004. Dehydroepiandrosterone-sulfate inhibitsnuclear factor-kappaB-dependent transcription in hepatocytes,possibly through antioxidant effect. J. Clin. Endocrinol. Metab.89, 3449—3454.

Jaquish, C.E., Blangero, J., Haffner, S.M., Stern, M.P., Maccluer,J.W., 1996a. Quantitative genetics of dehydroepiandrosteronesulfate and its relation to possible cardiovascular disease riskfactors in Mexican Americans. Hum. Hered. 46, 301—309.

Jaquish, C.E., Mahaney, M.C., Blangero, J., Haffner, S.M., Stern,M.P., Maccluer, J.W., 1996b. Genetic correlations between lipo-protein phenotypes and indicators of sex hormone levels inMexican Americans. Atherosclerosis 122, 117—125.

Jeckel, C.M., Lopes, R.P., Berleze, M.C., Luz, C., Feix, L., Argimon,I.I., Stein, L.M., Bauer, M.E., 2010. Neuroendocrine and immu-nological correlates of chronic stress in ‘strictly healthy’ popula-tions. Neuroimmunomodulation 17, 9—18.

Karlson, E.W., Chibnik, L.B., McGrath, M., Chang, S.C., Keenan, B.T.,Costenbader, K.H., Fraser, P.A., Tworoger, S., Hankinson, S.E.,Lee, I.M., Buring, J., De, V.I., 2009. A prospective study ofandrogen levels, hormone-related genes and risk of rheumatoidarthritis. Arthritis Res. Ther. 11, R97.

Kendler, K.S., Neale, M.C., 2009. ‘‘Familiality’’ or heritability. Arch.Gen. Psychiatry 66, 452—453.

Konttinen, Y.T., Stegaev, V., Mackiewicz, Z., Porola, P., Hanninen, A.,Szodoray, P., 2010. Salivary glands — ‘‘an unisex organ’’? Oral Dis.16, 577—585.

Kremen, W.S., Thompson-Brenner, H., Leung, Y.M., Grant, M.D.,Franz, C.E., Eisen, S.A., Jacobson, K.C., Boake, C., Lyons,M.J., 2006. Genes, environment, and time: The Vietnam Era TwinStudy of Aging, VETSA. Twin. Res. Hum. Genet. 9, 1009—1022.

Kroboth, P.D., Salek, F.S., Pittenger, A.L., Fabian, T.J., Frye, R.F.,1999. DHEA and DHEA-S: a review. J. Clin. Pharmacol. 39, 327—348.

Lac, G., Marquet, P., Chassain, A.P., Galen, F.X., 1999. Dexametha-sone in resting and exercising men. II. Effects on adrenocorticalhormones. J. Appl. Physiol. 87, 183—188.

Longcope, C., 1996. Dehydroepiandrosterone metabolism. J. Endo-crinol. 150 (Suppl.), S125—S127.

Majewska, M.D., 1995. Neuronal actions of dehydroepiandrosterone.Possible roles in brain development, aging, memory, and affect.Ann. N. Y. Acad. Sci. 774, 111—120.

Majewska, M.D., 2002. DHEA: hormone of youth and resilience-still amaverick. In: Morfin, R. (Ed.), DHEA and the Brain. Taylor andFrancis, Paris, pp. 65—80.

Maninger, N., Capitanio, J.P., Mason, W.A., Ruys, J.D., Mendoza, S.P.,2010. Acute and chronic stress increase DHEAS concentrations inrhesus monkeys. Psychoneuroendocrinology 35, 1055—1062.

Maninger, N., Wolkowitz, O.M., Reus, V.I., Epel, E.S., Mellon, S.H.,2009. Neurobiological and neuropsychiatric effects of dehydro-epiandrosterone (DHEA) and DHEA sulfate (DHEAS). Front. Neu-roendocrinol. 30, 65—91.

Meikle, A.W., Stephenson, R.A., Lewis, C.M., Wiebke, G.A., Middle-ton, R.G., 1997. Age, genetic, and nongenetic factors influencingvariation in serum sex steroids and zonal volumes of the prostateand benign prostatic hyperplasia in twins. Prostate 33, 105—111.

Meikle, A.W., Stringham, J.D., Woodward, M.G., Bishop, D.T., 1988.Heritability of variation of plasma cortisol levels. Metabolism 37,514—517.

Michael, A., Jenaway, A., Paykel, E.S., Herbert, J., 2000. Alteredsalivary dehydroepiandrosterone levels in major depression inadults. Biol. Psychiatry 48, 989—995.


Mitchell, B.D., Kammerer, C.M., Blangero, J., Mahaney, M.C., Rain-water, D.L., Dyke, B., Hixson, J.E., Henkel, R.D., Sharp, R.M.,Comuzzie, A.G., VandeBerg, J.L., Stern, M.P., Maccluer, J.W.,1996. Genetic and environmental contributions to cardiovascularrisk factors in Mexican Americans. The San Antonio Family HeartStudy. Circulation 94, 2159—2170.

Nestler, J.E., Whitfield, J.B., Williams, T.Y., Zhu, G., Condon, J.,Kirk, K.M., Heath, A.C., Montgomery, G.W., Martin, N.G., 2002.Genetics of serum dehydroepiandrosterone sulfate and its rela-tionship to insulin in a population-based cohort of twin subjects.J. Clin. Endocrinol. Metab. 87, 682—686.

Nichols, R.C., Bilbro, W.C.J., 1966. The diagnosis of twin zygosity.Acta. Genet. Stat. Med. 16, 265—275.

Nicolau, G.Y., Haus, E., Lakatua, D.J., Bogdan, C., Sackett-Lund-een, L., Popescu, M., Berg, H., Petrescu, E., Robu, E., 1985.Circadian and circannual variations of FSH, LH, testosterone,dehydroepiandrosterone-sulfate (DHEA-S) and 17-hydroxy pro-gesterone (17 OH-Prog) in elderly men and women.Endocrinologie 23, 223—246.

Nieschlag, E., Loriaux, D.L., Ruder, H.J., Zucker, I.R., Kirschner,M.A., Lipsett, M.B., 1973. The secretion of dehydroepiandroster-one and dehydroepiandrosterone sulphate in man. J. Endocrinol.57, 123—134.

Orentreich, N., Brind, J.L., Rizer, R.L., Vogelman, J.H., 1984. Agechanges and sex differences in serum dehydroepiandrosteronesulfate concentrations throughout adulthood. J. Clin. Endocrinol.Metab. 59, 551—555.

Parker Jr., C.R., Azziz, R., Potter, H.D., Boots, L.R., 1996. Adrenalandrogen production in response to adrenocorticotropin infusionsin men. Endocr. Res. 22, 717—722.

Pavlov, E.P., Harman, S.M., Chrousos, G.P., Loriaux, D.L., Blackman,M.R., 1986. Responses of plasma adrenocorticotropin, cortisol,and dehydroepiandrosterone to ovine corticotropin-releasinghormone in healthy aging men. J. Clin. Endocrinol. Metab. 62,767—772.

Peeters, H., Van Gestel, S., Vlietinck, R., Derom, C., Derom, R.,1998. Validation of a telephone zygosity questionnaire in twins ofknown zygosity. Behav. Genet. 28, 159—163.

Pomari, E., Nardi, A., Fiore, C., Celeghin, A., Colombo, L., Dalla,V.L., 2009. Transcriptional control of human organic anion trans-porting polypeptide 2B1 gene. J. Steroid Biochem. Mol. Biol. 115,146—152.

Radford, D.J., Wang, K., McNelis, J.C., Taylor, A.E., Hechenberger,G., Hofmann, J., Chahal, H., Arlt, W., Lord, J.M., 2010. Dehdyr-oepiandrosterone sulfate directly activates protein kinase C-betato increase human neutrophil superoxide generation. Mol. Endo-crinol. 24, 813—821.

Rainey, W.E., Nakamura, Y., 2008. Regulation of the adrenal andro-gen biosynthesis. J. Steroid Biochem. Mol. Biol. 108, 281—286.

Rice, T., Borecki, I.B., 2001. Familial resemblance and heritability.In: Rao, D.C., Province, M.A. (Eds.), Genetic Dissection of Com-plex Traits. Academic Press, London, pp. 35—43.

Rice, T., Sprecher, D.L., Borecki, I.B., Mitchell, L.E., Laskarzewski,P.M., Rao, D.C., 1993. The Cincinnati Myocardial Infarction andHormone Family Study: family resemblance for dehydroepian-drosterone sulfate in control and myocardial infarction families.Metabolism 42, 1284—1290.

Rosenfeld, R.S., Rosenberg, B.J., Fukushima, D.K., Hellman, L.,1975. 24-Hour secretory pattern of dehydroisoandrosteroneand dehydroisoandrosterone sulfate. J. Clin. Endocrinol. Metab.40, 850—855.

Rotter, J.I., Wong, F.L., Lifrak, E.T., Parker, L.N., 1985. A geneticcomponent to the variation of dehydroepiandrosterone sulfate.Metabolism 34, 731—736.

Sanders, J.L., Boudreau, R.M., Cappola, A.R., Arnold, A.M., Robbins,J., Cushman, M., Newman, A.B., 2010. Cardiovascular disease isassociated with greater incident dehydroepiandrosterone sulfatedecline in the oldest old: the cardiovascular health study all starsstudy. J. Am. Geriatr. Soc. 58, 421—426.

Speigelhalter, D., Thomas, A., Best, N., Lunn, D., 2004. WinBUGS(Version 1.4.1) (computer software).

Spiegelhalter, D.J., Thomas, A., Best, N.G., 2003. WinBUGS Version1.4 User Manual. .

Takebayashi, M., Kagaya, A., Uchitomi, Y., Kugaya, A., Muraoka, M.,Yokota, N., Horiguchi, J., Yamawaki, S., 1998. Plasma dehydro-epiandrosterone sulfate in unipolar major depression. Shortcommunication. J. Neural Transm. 105, 537—542.

van Anders, S.M., 2010. Chewing gum has large effects on salivarytestosterone, estradiol, and secretory immunoglobulin A assays inwomen and men. Psychoneuroendocrinology 35, 305—309.

Vining, R.F., McGinley, R.A., Symons, R.G., 1983. Hormones in saliva:mode of entry and consequent implications for clinical interpre-tation. Clin. Chem. 29, 1752—1756.

Yildiz, B.O., Goodarzi, M.O., Guo, X., Rotter, J.I., Azziz, R., 2006.Heritability of dehydroepiandrosterone sulfate in women withpolycystic ovary syndrome and their sisters. Fertil. Steril. 86,1688—1693.

Zhao, Z.Y., Xie, Y., Fu, Y.R., Li, Y.Y., Bogdan, A., Touitou, Y., 2003.Circadian rhythm characteristics of serum cortisol and dehydro-epiandrosterone sulfate in healthy Chinese men aged 30 to 60years. A cross-sectional study. Steroids 68, 133—138.

Zinder, O., Dar, D.E., 1999. Neuroactive steroids: their mechanism ofaction and their function in the stress response. Acta Physiol.Scand. 167, 181—188.

Genetic and environmental effects on diurnal dehydroepiandrosterone sulfate concentrations in middle-aged men

Documents