Determinants of the NF-κB response to acute psychosocial stress in humans

Brain, Behavior, and Immunity 23 (2009) 742–749

Contents lists available at ScienceDirect

Brain, Behavior, and Immunity journal homepage: www.elsevier.com/locate/ybrbi

Named Series: Health, Psychology and Immunity

Determinants of the NF-jB response to acute psychosocial stress in humans Jutta M. Wolf a,*, Nicolas Rohleder a, Angelika Bierhaus b, Peter P. Nawroth b, Clemens Kirschbaum c a b c

Brandeis University, Department of Psychology, MS 062, Waltham, MA 02454, Canada University of Heidelberg, Department of Medicine I and Clinical Chemistry, Germany Dresden University of Technology, Department of Psychology, Germany

a r t i c l e

i n f o

Article history: Received 3 June 2008 Received in revised form 14 August 2008 Accepted 17 September 2008 Available online 26 September 2008 Keywords: NF-jB activity Acute psychosocial stress Age Perceived stress Cortisol Norepinephrine IL-6 production

a b s t r a c t Previous research has shown that psychosocial stress is associated with an increased activity of the transcription factor nuclear factor-kappaB (NF-jB), a major inducer of inflammatory genes. While considerable individual variation has been noted, factors contributing to this variation have not been described so far. Therefore, 29 healthy participants (35.8 ± 12 yrs) were exposed to the Trier Social Stress Test. Blood was collected before and repeatedly afterward for determination of NF-jB activity, leukocyte subset numbers, cortisol, norepinephrine, and in vitro-stimulated IL-6 production. Additionally, age, sex, and ratings of perceived chronic and acute stress were assessed. Regression analyses revealed that older participants showed a lower NF-jB stress response compared to younger adults (b = .42, p = .026). Higher NF-jB stress responses were associated with lower cortisol stress responses (b = .37, p = .05), higher pre-stress IL-6 production (b = .38, p = .043), and high chronic in combination with low acute stress, or vice versa (b = .61, p = .06). Norepinephrine and sex were not associated with NF-jB stress responses (all p P .13). In summary, the present study shows for the first time in human psychosocial stress the negative association of cortisol and NF-jB. This parallels results from in vitro studies. Our finding of lower NF-jB stress responses in older age and in people with high chronic and acute stress might be interpreted as an adaptive dampening of NF-jB activity. In the absence of longitudinal data, however, this interpretation remains speculative. Ó 2008 Elsevier Inc. All rights reserved.

1. Introduction Stress is known to activate major biological systems, including the hypothalamus-pituitary adrenal (HPA) axis and the sympatho-adrenal-medullary (SAM) system. (Dickerson and Kemeny, 2004; Goldstein, 2000; Kvetnansky and McCarty, 2000). Stress-related changes in cortisol and catecholamine levels follow a specific pattern, such that catecholamines are released within seconds after stressor onset and return to baseline levels shortly after stressor termination, while the release of cortisol shows a time-lag of approximately 10 min and takes up to 1 h after stress cessation to return to baseline levels (Sapolsky et al., 2000). These phased responses are thought to be adaptive and beneficial for the organism. This concerns especially their effects on the immune system, where cortisol is thought to terminate immune processes—activated among others by stress-induced increases in catecholamines (DeRijk et al., 1994)—in order to protect the organism from an overactive immune system (Besedovsky and del Rey, 1996, 2000; Sapolsky et al., 2000). Consequently, deviations from these ex* Corresponding author. Fax: +1 781 736 3291. E-mail address: [email protected] (J.M. Wolf). 0889-1591/$ - see front matter Ó 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.bbi.2008.09.009

pected response patterns have been associated with different disease processes or susceptibilities towards different diseases (Chrousos, 1998), such that, for example, a decreased HPA axis activity is found in individuals with fibromyalgia (Chikanza et al., 1992) and a blunted cortisol response in combination with an elevated catecholamine response in atopic diseases (Buske-Kirschbaum et al., 2002). Cortisol and catecholamines exert their immune effects via specific signal transduction pathways. In detail, cortisol binds in the cytoplasm to its glucocorticoid receptor (GR) and this hormone– receptor complex then translocates to the nucleus, where it binds to glucocorticoid bindings elements (GRE) or negative GREs (nGRE), thereby activating or inhibiting the transcription of responsive genes. Catecholamines exert their effects by binding to adrenoceptors (AR), which can be classified into three major groups, a1-, a2-, and b-AR types (Hasko and Szabo, 1998). Norepinephrine predominantly activates a-ARs and b1-AR and is a weak stimulator of b2-AR, whereas epinephrine is a strong stimulator of b-ARs (Motulsky and Insel, 1982). ARs directly activate G-proteins, with different types of ARs coupling to different G-proteins and thus initiating different signal transduction pathways. For example, b-ARs couple to Gs proteins, which activates adenylate cyclase

J.M. Wolf et al. / Brain, Behavior, and Immunity 23 (2009) 742–749

(AC). AC in turn increases intracellular cyclic adenosine monophosphate (cAMP), which activates protein kinase A (PKA). Contrary, coupling of a2-AR to Gi proteins inhibits AC and thus subsequently the formation of cAMP. The a1-AR coupling to Gq proteins activates another intracellular effector, namely phospholipase C (PLC), which increases inositol triphosphate (IP3) and diacylglycerol (DAG). DAG then activates protein kinase C (PKC) and IP3 mobilizes Ca2+ from intracellular stores. The latter is further linked to the Ca2+/calmodulin (Ca2+/CaM) pathway, which—like PKA—subsequently transfers signals to the nucleus (Elenkov et al., 2000). While activation of the above signal transduction pathways can have a wide range of effects, they also converge at one point. Eventually they all influence the activity of the transcription factor nuclear factor-kappaB (NF-jB). The designation NF-jB refers to the most frequently occurring and ubiquitously expressed heterodimeric complex between two members of the NF-jB/Rel family of proteins, the p50 and the p65 (Rel A) subunit (Caamano and Hunter, 2002; Wulczyn et al., 1996). NF-jB is found in virtually every cell of the immune system and regulates a great number of genes, including genes of growth factors (e.g., granulocyte/macrophage colony stimulating factor), pro-inflammatory cytokines (especially interleukin-1 (IL-1), IL-2, IL-6, IL-8, and tumor necrosis factor alpha (TNF-a)), or cell adhesion molecules (e.g., vascular cell adhesion molecule; for review see McKay and Cidlowski, 1999). It is thought that many of the immune-inhibitory functions of cortisol are due to cortisol interfering with NF-jB activity, either by inducing expression of the NF-jB inhibitory protein IjB-a (Auphan et al., 1995; Scheinman et al., 1995), by GRs interacting with the p65 subunit and thereby repressing NF-jB DNA binding activity (De Bosscher et al., 1997, 2000; Ray and Prefontaine, 1994), or by competing for limited amounts of essential coactivators (Lee et al., 1998; Sheppard et al., 1998). In parallel to glucocorticoids, several lines of evidence exist of catecholamine signal transduction pathways interfering with activity of NF-jB as well. Haraguchi et al. suggested that all three pathways, i.e., the cAMP/PKA, the PKC, and the Ca2+/CaM pathway, modulate NF-jB activity by regulating its phosphorylation status (Haraguchi et al., 1995). Furthermore, elevated levels of cAMP are known to inhibit NF-jB activity by inhibiting the binding of NF-jB to the NF-jB DNA binding site (Chen and Rothenberg, 1994; Neumann et al., 1995; Tsuruta et al., 1995). The cAMP/PKA pathway also induces impaired nuclear translocation and DNA binding of p65 (Neumann et al., 1995; Paliogianni et al., 1993). Alternatively, this pathway may inhibit NFjB transcription by phosphorylating the transcription factor CREB (cAMP response element-binding protein), which then competes with NF-jB p65 for limited amounts of CREB-binding protein (CBP) (Parry and Mackman, 1997). Contrary, NF-jB activity can be stimulated via the Ca2+/CaM pathway enhancing inactivation of the inhibitory protein IjB-a (Frantz et al., 1994). Given the tight interaction between stress hormone and NF-jB signaling pathways, we hypothesized that NF-jB activity should vary in response to stress. We were able to show that NF-jB activity indeed increased in response to the Trier Social Stress Test (TSST) in healthy young men (Bierhaus et al., 2003). These findings have been replicated in patients with major depression (Pace et al., 2006). As we further showed in subsequent animal and in vitro studies, norepinephrine in physiological concentrations (but not epinephrine) was able to increase NF-jB activity (Bierhaus et al., 2003). Although these findings showed that psychosocial stress impacts a biological pathway with major health implications, little is still known about the determinants of this inflammatory stress response. No study has yet tested in vivo the hypotheses that NF-jB activity, and thus, peripheral inflammation, is activated by catecholamines, and negatively controlled by glucocorticoids. Although these might remain impossible to test in humans, not even correlational studies are available that tested associations compatible with these hypotheses.

743

Little is also known about the role of demographic variables, i.e., sex and age differences in NF-jB stress responses as well as the role of perceived acute and chronic stress. The current study therefore aimed at testing potential determinants of the NF-jB stress response and recovery. Specifically, we hypothesized that the NF-jB stress response would be higher in older participants and in women, based on the findings of age-related increases of inflammatory activity (Ershler and Keller, 2000) and on documented sex differences in inflammatory responses (O’Connor et al., 2007). We further hypothesized that perceived stress, both acutely and chronically, would be associated with higher NF-jB stress responses, the former based on findings from acute stress studies (e.g., Bierhaus et al., 2003), the latter based on findings of altered inflammatory regulation in chronic stress (Miller et al., 2002). In both cases, we also aimed at testing the effect of the interaction between age and sex as well as the interaction between acute and chronic stress on NF-jB stress responses, since various reports exists on, for example, endocrine stress responses being influenced by the interaction between age and sex (Kudielka et al., 2004), and on, for example, GR and b2-AR gene expression being influenced by the interaction of acute and chronic stress (Miller and Chen, 2006). Furthermore, we set out to test the hypothesis that norepinephrine responses would be positively and cortisol responses inversely associated with the NF-jB stress response, as outlined above. Since reports exist that changes measured in NF-jB activity may be due to changes in the cell type composition of the assessed sample (Richlin et al., 2004), we additionally aimed at testing whether the NF-jB stress response is associated with stress-induced changes in leukocyte subsets. Lastly, we also assessed in vitro-stimulated IL-6 production. As pointed out above, NF-jB is centrally involved in many inflammatory processes, among others by inducing one of the most important pro-inflammatory cytokines, namely IL-6. In vitro-stimulated IL-6 production, in turn, is a parameter which is frequently used to assess inflammatory activity (e.g., Miller et al., 2002). Given the tight connection between the two parameters and both parameters being indicators of inflammatory activity, we aimed at comparing the results of in vivo stimulation of the relevant pathways by psychosocial stress via assessing changes in NF-jB activity and the in vitro stimulation of the relevant pathways by assessing one important outcome, i.e., the stimulated production of IL-6. Lastly, we aimed at testing whether NF-jB recovery, instead of the NF-jB stress response, is predicted by any of the above proposed determinants. 2. Study design and methods 2.1. Subjects A total of 29 adults were recruited through advertisements in newspapers and by flyers. All subjects underwent a comprehensive medical examination for past or current health problems. Exclusion criteria were any psychiatric, endocrine, cardiovascular, or other chronic disease, as well as medication with psychoactive drugs, b-blockers, or glucocorticoids. The sample consisted of 17 women and 12 men with a mean age of 35.8 yrs (SD = 12.2; range = 20–59 yrs) and a mean BMI of 23.3 kg/m2 (SD = 3.84; range = 17.8–37.1). Men and women did not significantly differ with regard to BMI and age (t = 0.35; p = 0.73 and t = –0.045; p = 0.96, respectively). Three female participants reported to be habitual smokers. All participants were Caucasians. 2.2. Experimental protocol Participants reported to the laboratory between 13:00 h and 15:00 h and were examined by a physician for past and current health problems. A venous blood catheter was inserted and partic-

744

J.M. Wolf et al. / Brain, Behavior, and Immunity 23 (2009) 742–749

ipants were asked to fill out a questionnaire asking about their perceived chronic stress level (see below). After a resting period of 45 min, a blood sample and a saliva sample were collected. Subsequently, participants were exposed to the psychosocial stress test ‘‘Trier Social Stress Test” (TSST), which consists of a 3 min preparation period, a 5 min free speech, and a 5 min mental arithmetic task in front of an audience (Kirschbaum et al., 1993). After stress exposure, participants rated the stressfulness of the TSST situation, and additional blood and saliva samples were collected. In detail, a total of three saliva samples for determination of free cortisol and a total of three blood samples for determination of NF-jB DNA binding activity, interleukin-6 (IL-6) production, and leukocyte subset numbers were collected at baseline (1 min pre-TSST), 10 min post-TSST, and 60 min post-TSST. For determination of norepinephrine, blood samples were taken at baseline, 1 min after TSST, and 20 min after TSST, since the acute stress response as well as the recovery in norepinephrine is expected to be earlier and faster than for the other parameters (Sapolsky et al., 2000). Participants received a payment of 50 Euros. The study protocol was approved by the ethics committee of the University of Duesseldorf, and written informed consent was obtained from all participants. 2.3. Stress assessment 2.3.1. Chronic stress Chronic stress was assessed using a German translation of the Perceived Stress Scale (PSS), which is a widely used self-reported questionnaire with adequate reliability and validity asking about how stressful, overwhelming, and uncontrollable a person has found his/her life during the previous month (Cohen et al., 1983; Cohen and Williamson, 1988). The items are answered on a 0 = never to 4 = very scale, with higher sum scores indicating greater perceived stress. 2.3.2. Acute stress To assess perceived acute stress, participants were asked to rate how stressful they perceived the TSST situation on a Visual Analogue Scale (VAS) ranging from 0 = not at all to 100 = very immediately after the end of the stress test. The VAS has been used in a large number of previous studies employing the TSST and has proven useful in assessing acute stressfulness of the situation (Kirschbaum et al., 1999). 2.4. Biochemical analyses 2.4.1. Cortisol For free cortisol determination, saliva samples were collected at three time points before and after TSST using the Salivette device (Sarstedt, Nümbrecht, Germany). Samples were frozen immediately and stored at 20° C until analysis. Free cortisol levels in saliva were measured using a commercially available chemiluminescence immunoassay (IBL, Hamburg, Germany). The intra- and inter-assay coefficients of variation (CV) were below 8%. 2.4.2. Norepinephrine To determine norepinephrine levels, blood was drawn into tubes containing EDTA and plasma was immediately separated at 4 °C and stored at 80 °C. Plasma samples were sent to the ‘Laboratory for Stress-Monitoring’ in Göttingen, Germany, where norepinephrine was determined by high-performance liquid chromatography with electrochemical detection as previously described (Smedes et al., 1982). The lower detection limit of this method was 0.25 pg/ml. 2.4.3. Stimulated interleukin-6 production For determination of IL-6 production, blood was drawn into tubes containing heparin, diluted 10:1 with saline, and coincubated with

lipopolysaccharide (LPS; Escherichia coli, Sigma, Deisenhofen, Germany). In detail, 400 ll diluted whole blood was added to 50 ll of LPS (final concentration 30 ng/ml) and 50 ll of saline. After 18 h of incubation at 37 °C and 5% CO2, plates were centrifuged for 10 min at 2000g and 4 °C, and plasma supernatants were collected and stored at 80 °C until assayed. IL-6 in plasma supernatants was measured using enzyme-linked immunosorbent assays (BD Pharmingen, San Diego, CA, USA). The intra- and inter-assay CVs were below 10%. 2.4.4. Leukocyte subsets Blood was collected in EDTA tubes and the number of leukocytes, lymphocytes, monocytes, and granulocytes were determined by a Coulter AcTdiff cell counter (Beckman-Coulter, Krefeld, Germany). 2.4.5. NF-jB DNA binding activity To determine the NF-jB DNA binding activity, blood was drawn into tubes containing sodium citrate. After density-gradient centrifugation (Ficoll, Biochrom, Berlin, Germany) and three subsequent wash steps, 3  106 PBMCs were used for nuclear protein preparation. Nuclear extracts were then assayed for transcription factorbinding activity by electrophoretic mobility shift assay (EMSA) using NF-jB consensus oligonucleotides (Promega, Mannheim, Germany) as previously described (Bierhaus et al., 2001; Hofmann et al., 1999). For each subject, baseline NF-jB activity (1 min preTSST) was used as reference and all post-TSST values are given as percent change from baseline, since intra-individual changes, and not absolute levels, are found to be associated with clinical outcomes, such as sepsis survival rate (Boehrer et al., 1997). 2.5. Statistical analysis Data were analyzed using the Statistical Package for the Social Science for Mac OS X Version 11.0.4 (SPSS Institute, Chicago, IL). In the following, NF-jB stress response will refer to changes in NF-jB activity from baseline to 10 min post-TSST and NF-jB recovery to changes from 10 min post-TSST to 60 min post-TSST. The same two indices will be used with regard to cortisol, IL-6 production, and cell numbers. For norepinephrine, however, stress response will refer to the changes from baseline to 1 min post-TSST and recovery to the changes from 1 min post-TSST to 20 min post-TSST. For all analyses, p-values of p < .05 were considered significant. Analyses followed a three-step approach: First, the demographic variables age and sex were tested for associations with questionnaire data, hormone measures, IL-6 production, and cell numbers using Pearson correlations and Student’s t-tests. Second, repeated measures ANOVAs were computed to test for stress-induced changes in the major biological variables. Next, associations between the above variables and NF-jB indices were analyzed. In detail, to test for relationships between demographic variables and NF-jB indices, hierarchical regression analyses were computed. These analyses tested whether age, sex, or their interactions predicted NF-jB stress responses or recovery using recommended procedures for testing interaction effects in multiple regression analyses (Aiken and West, 1991). Within regression models, NF-jB stress response or recovery was entered as dependent variable and predicted by the proposed determinants. Regression analyses predicting the NF-jB stress response included age and sex in step 1, and the interaction between age and sex in step 2. Regression predicting NF-jB recovery included NF-jB activity 10 min post-TSST in step 1, age and sex in step 2, and the interaction between age and sex in step 3. The same set of regressions was computed for stress perception ratings and their interaction (i.e., acute stress, chronic stress, acute-by-chronic stress). Next, regression analyses were computed including baseline levels, stress responses, or stress recovery in cortisol, norepinephrine, IL-6

745

J.M. Wolf et al. / Brain, Behavior, and Immunity 23 (2009) 742–749 Table 1 Demographic, questionnaire, and biological data of study participants (mean, standard deviation, and range)

a

Chronic stress Acute stressb Cortisol (nmol/l) Norepinephrine (pg/ml)c IL-6 (ng/ml) Leukocytesd Lymphocytesd Monocytesd Granulocytesd a b c d

1 min before stress

10 min after stress

60 min after stress

Change over time

22.3 ± 7.8 (11–47) 71.8 ± 22.0 (13–100) 11.04 ± 1.65 431.3 ± 32.54 58.14 ± 3.78 5.73 ± 0.21 1.86 ± 0.1 0.36 ± 0.02 3.52 ± 0.16

21.67 ± 2.56 663.98 ± 41.35 60.37 ± 4.68 5.86 ± 0.21 2.02 ± 0.11 0.38 ± 0.02 3.46 ± 0.16

11.33 ± 1.39 437.36 ± 35.72 61.66 ± 3.75 5.89 ± .21 1.92 ± 0.11 0.38 ± 0.03 3.60 ± 0.18

F(2, 56) = 33.6 (p < 0.001) F(2, 56) = 66.9 (p < 0.001) F(2, 56) = 1.2 (p = 0.32) F(2, 56) = 1.9 (p = 0.16) F(2, 56) = 4.1 (p = 0.022) F(2, 56) = 2.3 (p = 0.11) F(2, 56) = 1.9 (p = 0.15)

Perceived Stress Scale. Visual Analogue Scale (0–100). Time points: 1 min before stress, 1 min after stress, and 20 min after stress. 103 cells/ml.

production, or cell type numbers in step 1 testing for associations with NF-jB stress responses and in step 2 testing for associations with NF-jB recovery. Lastly, where appropriate, we computed Sobel tests to test for mediation effects (MacKinnon et al., 2002).

34.52 yrs of age (see Fig. 1). Neither sex nor the age-by-sex interaction was associated with the NF-jB stress response (b = .06, t = 0.36, p = .72; b = .02, t = 0.07, p = .94, respectively). Furthermore, age and sex did not predict NF-jB stress recovery (all p P .45).

3. Results

3.3.2. Perceived chronic stress and perceived acute stress Using hierarchical regressions to test for effects of perceived acute and perceived chronic stress ratings on NF-jB indices, no significant main effects were found for either one of the two parameters (b = .22, t = 0.94, p = .36; b = .17, t = 0.75, p = .46, respectively). However, there was a marginal effect for the relationship between their interaction and NF-jB stress responses in a shape comparable to an inverted U-function (b = .61, t = 1.98, p = .06). NF-jB stress responses were highest if participants either scored low on chronic stress, but rated the acute laboratory task as highly stressful, or, if they reported to be under high chronic stress, but did not perceive the acute task as stressful. Lower NF-jB responses were found in participants who reported low chronic stress and were not stressed by the TSST, and by those who rated chronic and acute stress to be high. For illustration purposes, Fig. 2 shows NF-jB stress responses grouped into participants with low vs. high chronic and acute stress ratings using median splits. An additional two-way ANOVA testing the impact of these acute and chronic stress categories revealed a trend towards an acute-by-chronic stress interaction as well (F(1, 18) = 3.89; p = 0.06), thus confirming the results of the regression model. No significant associations were found with regard to perceived stress and NF-jB recovery (all p P .68).

3.1. Preliminary analyses We first tested for associations between the demographic variables age and sex with questionnaire data, hormone measures, IL-6 production, as well as cell subset numbers. Analyses revealed a positive association of baseline norepinephrine levels with age (r(27) = .44, p = .02) and an inverse relationship of baseline IL-6 production with age (r(27) = .44, p = .02). None of the remaining variables were associated with age (all p P .12). Furthermore, there was a trend for males showing higher stress responses in IL-6 production than females and, despite comparable cortisol stress responses, a steeper subsequent decrease in cortisol levels than females (t(27) = 1.87, p = .07; t(27) = 1.81, p = .08, respectively; all other p P .16). With regard to cell numbers, females showed a higher number of leukocytes (t(27) = 2.49, p = .019), due to marginally higher lymphocyte numbers (t(27) = 1.87, p = .07) and higher granulocytes numbers (t(27) = 1.83, p = .08). 3.2. Acute stress responses To test whether the acute stress paradigm induced significant changes in the major biological study variables, a set of repeated measures ANOVAs was computed and results as well as means and standard deviations are presented in Table 1. Briefly, salivary cortisol and plasma norepinephrine increased significantly in response to stress (both time effects: p < 0.001). Lymphocyte numbers also increased significantly after the TSST (time effect: p = 0.022), while the remaining leukocyte subsets as well as LPS-stimulated IL-6 production did not change significantly over time (all p > 0.11). 3.3. Determinants of the NF-jB stress response 3.3.1. Age and sex Using hierarchical regression analyses, we found associations between age and NF-jB stress response, such that younger participants showed increases in NF-jB activity in response to stress, while older participants showed no response or even the opposite pattern, i.e., decreases in NF-jB activity in response to stress (b = .42, t = 2.37, p = .026),1 with the NF-jB stress response being 0 at 1 All analyses were additionally computed on a subset of n = 26 participants remaining after excluding three women who were habitual cigarette smokers. None of the results were changed in significance or directionality.

3.3.3. Stress hormones We next tested whether baseline and stress-related hormone levels would be associated with NF-jB response and recovery. Hierarchical linear regressions did not reveal any associations between NF-jB parameters and norepinephrine baseline levels, stress-related increases in norepinephrine levels, or norepinephrine recovery (all p P .13). However, a significant relationship between stress responses in cortisol and stress responses in NF-jB activity was found, such that stronger increases in cortisol stress responses were associated with less pronounced stress responses in NF-jB activity (b = .37, t = 2.05, p = .05, DR2 = .13; see Fig. 3). No significant relationships emerged for baseline cortisol levels or cortisol recovery, neither with regard to NF-jB stress responses nor with regard to NF-jB recovery (all p P .48). 3.3.4. Immune measures Further regression analyses revealed a significant association between baseline in vitro IL-6 production and NF-jB stress responses. As shown in Fig. 4, higher IL-6 production at baseline in vitro predicted more pronounced increases in NF-jB activity in response to

746

J.M. Wolf et al. / Brain, Behavior, and Immunity 23 (2009) 742–749

Fig. 1. Stress-induced increases in NF-jB DNA binding activity vary with age.

Fig. 3. Stress-induced increases in NF-jB DNA binding activity vary with the magnitude of the cortisol stress response.

Fig. 2. Stress-induced increases in NF-jB DNA binding activity vary with acute and chronic stress perception (shown are means and standard errors; groups by median split).

stress in vivo (b = .38, t = 2.12, p = .043, DR2 = .14). No additional significant relationships were found, neither for IL-6 stress response (p = .62) nor for IL-6 recovery (p = .25). The same was true for associations between IL-6 production and NF-jB recovery (all p P .18). Next, we used hierarchical linear regressions to test whether total leukocyte numbers or specific subsets were associated with changes in NF-jB activity. None of the cell type parameters (i.e., baseline, stress response, recovery) showed a relationship with stress responses in NF-jB activity (all p P .21) or recovery in NFjB activity (p P .13). 3.4. Mediation analyses The above analyses revealed four parameters to be associated with NF-jB stress responses: Age, interaction of acute and chronic stress perception, cortisol stress responses, and baseline

Fig. 4. Stress-induced increases in NF-jB DNA binding activity vary with baseline in vitro IL-6 production.

IL-6 production. To gain further insight about the relationships of these different determinants, we tested whether interaction of acute and chronic stress perception, cortisol stress responses, or baseline IL-6 production would mediate the relationships between age and NF-jB stress responses. Sobel tests (Preacher and Hayes, 2004) revealed a z-value of 0.07 for acute-by-chronic stress interaction (p = .94), a z-value of 0.98 for cortisol stress responses (p = .33), and a z-value of 1.06 for baseline IL-6 production (p = .29), suggesting that age effects on NF-jB stress responses were not mediated by any of these three variables. Lastly, we tested whether the association between stress-induced increases in cortisol and NF-jB would be mediated by baseline IL-6 production, however, the Sobel test did not support this hypothesis (z = 0.82, p = .41).

J.M. Wolf et al. / Brain, Behavior, and Immunity 23 (2009) 742–749

4. Discussion The present study investigated potential determinants of the NF-jB response to psychosocial stress in humans. We found four variables to be associated with NF-jB stress responses: age, the interaction of acute and chronic stress, cortisol stress responses, and baseline IL-6 production. In detail, participants who showed a strong NF-jB stress response were younger, perceived high acute stress in combination with low chronic stress, or vice versa, showed a lower cortisol stress response, or produced high IL-6 levels in response to an in vitro stimulation. Conversely, participants who showed a blunted NF-jB stress response or a decrease in NF-jB activity were older, perceived both high acute as well as high chronic stress, showed high cortisol stress response, or a diminished response to in vitro stimulation in IL-6 levels. Since none of the mediation analyses yielded significant results, each of the four factors will be discussed separately. Out of the demographic variables tested, age came out as a strong predictor of NF-jB stress responses. In young participants, we observed the same pattern as described earlier, i.e., a strong stress-induced increase in NF-jB activity (Bierhaus et al., 2003). This increase in NF-jB activity may be interpreted as an adaptive immune activation in response to stress (Besedovsky and del Rey, 1996, 2000; Sapolsky et al., 2000). However, with increasing age, NF-jB responses became lower, and in older participants, NF-jB activity even decreased in response to stress. These decreases are an unexpected result. Based on the fact that low-grade inflammation increases with age (Ershler, 1993), and inflammation is a central pathophysiological process in age-related diseases such as cardiovascular disease, type2 diabetes, or Alzheimer’s disease (Ershler and Keller, 2000; Harris et al., 1999; Liu et al., 2007; Pradhan et al., 2001), we had hypothesized that older people would have exaggerated, instead of downregulated NF-jB stress responses. It is therefore an open question whether the stress-induced down-regulation observed here represents an adaptive or a maladaptive response. One could speculate that a missing NF-jB response is maladaptive, because it is incompatible with the idea that the immune system is prepared by acute stress to better respond to pathogens (Sapolsky et al., 2000). On the other hand, a missing NF-jB response may also be adaptive, since it prevents negative health effects of increased peripheral inflammation. In addition, it is important to point out that our participants were on average 36 years old, which makes them considerably younger than the populations typically investigated in studies of age-related inflammation (Ershler, 1993). Thus, longitudinal studies including a greater age range are needed to test the long-term health consequences of high vs. low NF-jB responses to acute stress. Besides the associations between the demographic variable age and NF-jB activity, we were also interested in psychological determinants of the NF-jB stress response, specifically perceived chronic and acute stress. We found that participants who rated both their perceived acute and their perceived chronic stress as low, did not show an increase in NF-jB activity in response to acute stress. Only individuals who reported one of the two, either high chronic stress or high acute stress, showed a strong NF-jB stress response. Interestingly, participants reporting high levels of acute stress in addition to high levels of chronic stress failed to mount NF-jB stress responses. This finding is puzzling at first sight, because one would rather expect additive effects of acute and chronic stress. Instead, these findings might indicate an inability of the organism to cope with the double threat of acute and chronic stress exposure. Again, this could be interpreted as adaptive, because acute stress would not further increase low-grade inflammation in consequence of already existing chronic stress (Kiecolt-Glaser et al., 2003). On the other hand, this could be interpreted as maladaptive, and indicate an exhaustion of relevant bio-

747

logical systems in the presence of overwhelming stress levels. Miller and Chen (2006) have investigated expression of receptors for cortisol and norepinephrine (i.e., GRs and b2-adrenergic receptor) and found that only the combination of acute and chronic stress was associated with expression patterns that were interpreted as potentially detrimental. However, whether the associations reported here are adaptive or maladaptive in the long run remains to be investigated. Thirdly, we found that free cortisol stress responses are negative determinants of stress-induced NF-jB activity, such that higher cortisol responses were associated with lower NF-jB stress responses. Although the negative relationship between cortisol and inflammation is well-documented in vitro (e.g., Waage et al., 1990), this is, to the best of our knowledge, the first study to document an inverse relationship of cortisol and NF-jB responses to psychosocial stress in humans in vivo. This emphasizes the importance of investigating the effects of psychosocial stress on signal transduction pathways not only in animals or in vitro, but also in vivo in humans. Our present finding further supports the general hypothesis that HPA axis stress responses are important factors in controlling inflammatory responses (e.g., Besedovsky and del Rey, 1996, 2000; Sapolsky et al., 2000), and might help explain the associations found between blunted HPA axis responses and inflammatory conditions, such as atopic diseases in humans (BuskeKirschbaum et al., 2002). The fourth determinant of NF-jB stress responses revealed by the present study was the stimulated production of IL-6. Higher Il-6 production at baseline was associated with increased excitability of the cells by acute stress in vivo, as indicated by NF-jB stress responses. This is a rather interesting finding, because the association found here adds support to the use of in vitro cytokine stimulation assays. These have repeatedly been criticized as a least physiological alternative, because they test only a specific subgroup of cells, and most mitogens used induce unspecific activation (see e.g., Vedhara et al., 1999). However, the high correlation between IL-6 production and NF-jB stress response found here argues in favor of measuring in vitro-stimulated IL-6 production. The present finding indicates that mitogen-stimulated inflammatory cytokine production might be used as a marker for the individual potential to mount an inflammatory response to acute psychosocial stress, thereby underscoring the validity of this method for studies in the field of health psychology. Some of the hypothesized determinants tested did not come out as significant predictors of NF-jB stress responses. Firstly, we did not find sex differences in NF-jB responses. Based on the fact that inflammatory and autoimmune diseases are more frequent in women (Beeson, 1994; Whitacre, 2001), we had hypothesized exaggerated NF-jB stress responses in women. However, the present finding suggests that modulation of NF-jB activity by estrogen and progesterone (Kalkhoven et al., 1996) may play a less important role during acute stress exposure, when other hormones influencing NF-jB activity, such as catecholamines and cortisol, are significantly elevated. This finding might also be the result of our strategy to select only healthy participants. Future studies will be needed to investigate patients suffering from autoimmune and inflammatory diseases. Secondly, we did not observe an association between norepinephrine and NF-jB activity. Based on our previous report of norepinephrine inducing NF-jB activity (Bierhaus et al., 2003), one could have expected norepinephrine and NF-jB stress responses to be related. We can think of several reasons for this negative finding. The first is timing. Although we have carefully chosen the time points to capture peaks of the norepinephrine and the NF-jB stress response, we might have failed to assess the specific characteristic of the norepinephrine response that stimulates NF-jB activity.

748

J.M. Wolf et al. / Brain, Behavior, and Immunity 23 (2009) 742–749

That specific time point might have been during the TSST, a time at which we were unable to obtain blood. Another reason might be that in contrast to in vitro studies, in the present in vivo situation, the increase in norepinephrine was not tested alone, but was part of a well-orchestrated hormonal stress response (Sapolsky et al., 2000). Thus, norepinephrine effects (both, on the receptor level and by post-receptor signaling) on NF-jB activity might be masked by, for example, cortisol effects or other mediators secreted during the stress response. Thirdly, we did not find any relationships between stress-induced leukocyte redistribution and NF-jB activity. This is an important finding, because it has been suggested that changes in NF-jB activity are due to cell distribution effects, such that during stress, more cell subsets with higher NF-jB activity will be found peripherally compared to baseline levels and thus NF-jB activity will increase without any actual changes in DNA binding activity itself causing this effect. In support of this hypothesis, Richlin et al. (2004) found that increases in NF-jB activity in response to physical exercise were caused by natural killer (NK) cell redistribution (Richlin et al., 2004). Our results are not in accordance with this hypothesis, but in contrast to Richlin et al. (2004), we were not able to assess changes in NK cells. Hence, we cannot exclude that the changes in NF-jB activity found here are mediated by changes in NK cells. However, Pace et al. (2006) measured NF-jB activity and NK cell changes in response to the TSST, and did not find NK cell changes being responsible for the NF-jB response (Pace et al., 2006). The major difference between these two studies is that Richlin et al. (2004) used an exercise test, which did not activate the HPA axis, and Pace et al. (2006) employed the TSST, which induced profound cortisol responses. One could therefore argue that the hypothesis of NK cell redistribution being responsible for stress-induced changes in NF-jB activity, is better suited to explain changes in response to stressors that activate the sympathetic nervous system (SNS) alone, but not to stressors that activate both, SNS and HPA axis, such as the TSST. However, the range of cell subsets assessed in the current study was very limited and we thus cannot rule out that other cell subsets vary in their NF-jB activity and thus redistribution effects nevertheless explain the current changes in NF-jB activity. Finally, we did not find any determinants of NF-jB recovery. One could have hypothesized that cortisol responses should predict NF-jB down-regulation, mediated by cortisol-induced expression of IjB-a, which would further inhibit NF-jB translocation to the nucleus (Auphan et al., 1995; Scheinman et al., 1995). However, the observation period of the present study of 1 h after stress might have been too short to observe such genomic effects. Preliminary evidence, however, suggest I-jB increases to be detectable 60 min after psychosocial stress (data not shown). However, whether these results are true for other jB-responsive genes as well and whether such early increases would be functionally relevant is unclear. This consideration further does not explain why cortisol did not act through the second pathway, i.e., through direct physical interaction of the GR–cortisol complex with NF-jB at specific DNA binding sites (McKay and Cidlowski, 1998). Nevertheless, the independence of NF-jB recovery from cortisol might indicate that down-regulation of NF-jB post-stress is a self-regulatory process that shuts down after cessation of the activating stimulus (Arenzana-Seisdedos et al., 1997). Future studies using longer observation periods will have to show whether cortisol through genomic effects might prepare the organism for subsequent stressors either by pro-active inhibition or by sensitization as formulated by Sapolsky et al. (2000). The results presented here need to be interpreted in the light of some limitations. Most importantly, the number of participants

investigated was relatively low, not only given that the sample covers a larger age range and is comprised of women and men, but also compared to the number of analyses computed. Furthermore, acute stress was assessed by a Visual Analogue Scale. Using an established stress inventory might have allowed us to disentangle specific components of acute stress perception, such as social threat. However, the VAS in the present form has been used in a large number of TSST studies and proven to be a reasonably valid tool for assessment of perceived acute stress. Lastly, in future studies it would be interesting to also assess the unstimulated inflammatory state of the participants by measuring plasma IL-6 or Creactive protein levels. In sum, the present study addressed the intriguing question of which variables determine NF-jB stress responses. Investigating these associations is important, since NF-jB plays a central role in activation of inflammatory responses (Bierhaus et al., 2006; McKay and Cidlowski, 1999). It is crucial to understand the role of psychosocial stress in such processes, because inflammation is a major pathophysiological factor in some of the most debilitating chronic diseases in humans, such as cardiovascular diseases, type2 diabetes, and Alzheimer’s disease (Bierhaus et al., 2006; Ershler and Keller, 2000; Liu et al., 2007; Pradhan et al., 2001; Ross, 1999). Improving our understanding of the determinants of NFjB stress responses can thus help to characterize additional risk factors for these diseases and eventually, to develop effective preventive strategies. Acknowledgments Funding for this study was provided by grants from the German Research Foundation to C.K. (DFG; Ki 537/9-1 and Ki 537/14-1) and a grant from the Lautenschläger-Foundation for the study of Diabetes (LSD; to P.P.N.). J.M.W. and N.R. were supported by fellowships from the German Research Foundation (DFG; WO 1209/3-1; RO 2353/4-1). References Aiken, L., West, S., 1991. Multiple Regression: Testing and Interpreting Interaction Effects. Sage, Thousand Oaks, CA. Arenzana-Seisdedos, F., Turpin, P., Rodriguez, M., Thomas, D., Hay, R.T., Virelizier, J.L., Dargemont, C., 1997. Nuclear localization of I kappa B alpha promotes active transport of NF-kappa B from the nucleus to the cytoplasm. J. Cell Sci. 110 (Pt. 3), 369–378. Auphan, N., DiDonato, J.A., Rosette, C., Helmberg, A., Karin, M., 1995. Immunosuppression by glucocorticoids: inhibition of NF-kappa B activity through induction of I kappa B synthesis. Science 270, 286–290. Beeson, P.B., 1994. Age and sex associations of 40 autoimmune diseases. Am. J. Med. 96, 457–462. Besedovsky, H.O., del Rey, A., 1996. Immune-neuro-endocrine interactions: facts and hypotheses. Endocr. Rev. 17, 64–102. Besedovsky, H.O., del Rey, A., 2000. The cytokine-HPA axis feed-back circuit. Z. Rheumatol. 59 (Suppl. 2), II/26–30. Bierhaus, A., Humpert, P.M., Nawroth, P.P., 2006. Linking stress to inflammation. Anesthesiol. Clin. 24, 325–340. Bierhaus, A., Schiekofer, S., Schwaninger, M., Andrassy, M., Humpert, P.M., Chen, J., Hong, M., Luther, T., Henle, T., Kloting, I., Morcos, M., Hofmann, M., Tritschler, H., Weigle, B., Kasper, M., Smith, M., Perry, G., Schmidt, A.M., Stern, D.M., Haring, H.U., Schleicher, E., Nawroth, P.P., 2001. Diabetes-associated sustained activation of the transcription factor nuclear factor-kappaB. Diabetes 50, 2792–2808. Bierhaus, A., Wolf, J., Andrassy, M., Rohleder, N., Humpert, P.M., Petrov, D., Ferstl, R., von Eynatten, M., Wendt, T., Rudofsky, G., Joswig, M., Morcos, M., Schwaninger, M., McEwen, B., Kirschbaum, C., Nawroth, P.P., 2003. A mechanism converting psychosocial stress into mononuclear cell activation. Proc. Natl. Acad. Sci. USA 100, 1920–1925. Boehrer, H., Qiu, F., Zimmermann, T., Zhang, Y., Jllmer, T., Mannel, D., Bottiger, B.W., Stern, D.M., Waldherr, R., Saeger, H.D., Ziegler, R., Bierhaus, A., Martin, E., Nawroth, P.P., 1997. Role of NFkappaB in the mortality of sepsis. J. Clin. Invest. 100, 972–985. Buske-Kirschbaum, A., Geiben, A., Höllig, H., Morschhäuser, E., Hellhammer, D., 2002. Altered responsiveness of the hypothalamus-pituitary-adrenal axis and the sympathetic adrenomedullary system to stress in patient with atopic dermatitis. J. Clin. Endocrinol. Metab. 87, 4245–4251.

J.M. Wolf et al. / Brain, Behavior, and Immunity 23 (2009) 742–749 Caamano, J., Hunter, C.A., 2002. NF-kappaB family of transcription factors: central regulators of innate and adaptive immune functions. Clin. Microbiol. Rev. 15, 414–429. Chen, D., Rothenberg, E.V., 1994. Interleukin 2 transcription factors as molecular targets of cAMP inhibition: delayed inhibition kinetics and combinatorial transcription roles. J. Exp. Med. 179, 931–942. Chikanza, I.C., Petrou, P., Kingsley, G., Chrousos, G., Panayi, G.S., 1992. Defective hypothalamic response to immune and inflammatory stimuli in patients with rheumatoid arthritis. Arthritis Rheum. 35, 1281–1288. Chrousos, G.P., 1998. Stress as a medical and scientific idea and its implications. Adv. Pharmacol. 42, 552–556. Cohen, S., Kamarck, T., Mermelstein, R., 1983. A global measure of perceived stress. J. Health Soc. Behav. 24, 385–396. Cohen, S., Williamson, G., 1988. Perceived stress in a probability sample of the US. In: Spacapam, S., Oskamp, S. (Eds.), The Social Psychology of Health: Claremont Symposium on Applied Social Psychology. Sage, Newbury Park, CA. De Bosscher, K., Schmitz, M.L., Vanden Berghe, W., Plaisance, S., Fiers, W., Haegeman, G., 1997. Glucocorticoid-mediated repression of nuclear factorkappaB-dependent transcription involves direct interference with transactivation. Proc. Natl. Acad. Sci. USA 94, 13504–13509. De Bosscher, K., Vanden Berghe, W., Vermeulen, L., Plaisance, S., Boone, E., Haegeman, G., 2000. Glucocorticoids repress NF-kappaB-driven genes by disturbing the interaction of p65 with the basal transcription machinery, irrespective of coactivator levels in the cell. Proc. Natl. Acad. Sci. USA 97, 3919– 3924. DeRijk, R.H., Boelen, A., Tilders, F.J., Berkenbosch, F., 1994. Induction of plasma interleukin-6 by circulating adrenaline in the rat. Psychoneuroendocrinology 19, 155–163. Dickerson, S.S., Kemeny, M.E., 2004. Acute stressors and cortisol responses: a theoretical integration and synthesis of laboratory research. Psychol. Bull. 130, 355–391. Elenkov, I.J., Wilder, R.L., Chrousos, G.P., Vizi, E.S., 2000. The sympathetic nerve—an integrative interface between two supersystems: the brain and the immune system. Pharmacol. Rev. 52, 595–638. Ershler, W.B., 1993. Interleukin-6: a cytokine for gerontologists. J. Am. Geriatr. Soc. 41, 176–181. Ershler, W.B., Keller, E.T., 2000. Age-associated increased interleukin-6 gene expression, late-life diseases, and frailty. Annu. Rev. Med. 51, 245–270. Frantz, B., Nordby, E.C., Bren, G., Steffan, N., Paya, C.V., Kincaid, R.L., Tocci, M.J., O’Keefe, S.J., O’Neill, E.A., 1994. Calcineurin acts in synergy with PMA to inactivate I kappa B/MAD3, an inhibitor of NF-kappa B. EMBO J. 13, 861–870. Goldstein, D.S., 2000. Sympathetic nervous system. In: Fink, G. (Ed.), Encyclopedia of Stress. Academic Press, San Diego, pp. 558–565. Haraguchi, S., Good, R.A., Day, N.K., 1995. Immunosuppressive retroviral peptides: cAMP and cytokine patterns. Immunol. Today 16, 595–603. Harris, T.B., Ferrucci, L., Tracy, R.P., Corti, M.C., Wacholder, S., Ettinger Jr., W.H., Heimovitz, H., Cohen, H.J., Wallace, R., 1999. Associations of elevated interleukin-6 and C-reactive protein levels with mortality in the elderly. Am. J. Med. 106, 506–512. Hasko, G., Szabo, C., 1998. Regulation of cytokine and chemokine production by transmitters and co-transmitters of the autonomic nervous system. Biochem. Pharmacol. 56, 1079–1087. Hofmann, M.A., Schiekofer, S., Isermann, B., Kanitz, M., Henkels, M., Joswig, M., Treusch, A., Morcos, M., Weiss, T., Borcea, V., Abdel Khalek, A.K., Amiral, J., Tritschler, H., Ritz, E., Wahl, P., Ziegler, R., Bierhaus, A., Nawroth, P.P., 1999. Peripheral blood mononuclear cells isolated from patients with diabetic nephropathy show increased activation of the oxidative-stress sensitive transcription factor NF-kappaB. Diabetologia 42, 222–232. Kalkhoven, E., Wissink, S., van der Saag, P.T., van der Burg, B., 1996. Negative interaction between the RelA(p65) subunit of NF-kappaB and the progesterone receptor. J. Biol. Chem. 271, 6217–6224. Kiecolt-Glaser, J.K., Preacher, K.J., MacCallum, R.C., Atkinson, C., Malarkey, W.B., Glaser, R., 2003. Chronic stress and age-related increases in the proinflammatory cytokine IL-6. Proc. Natl. Acad. Sci. USA 100, 9090–9095. Kirschbaum, C., Kudielka, B.M., Gaab, J., Schommer, N.C., Hellhammer, D.H., 1999. Impact of gender, menstrual cycle phase, and oral contraceptives on the activity of the hypothalamus-pituitary-adrenal axis. Psychosom. Med. 61, 154–162. Kirschbaum, C., Pirke, K.M., Hellhammer, D.H., 1993. The ‘Trier Social Stress Test—a tool for investigating psychobiological stress responses in a laboratory setting. Neuropsychobiology 28, 76–81. Kudielka, B.M., Buske-Kirschbaum, A., Hellhammer, D.H., Kirschbaum, C., 2004. HPA axis responses to laboratory psychosocial stress in healthy elderly adults, younger adults, and children: impact of age and gender. Psychoneuroendocrinology 29, 83–98. Kvetnansky, R., McCarty, R., 2000. Adrenal medulla. In: Fink, G. (Ed.), Encyclopedia of Stress. Academic Press, San Diego, pp. 63–70. Lee, S.K., Kim, H.J., Na, S.Y., Kim, T.S., Choi, H.S., Im, S.Y., Lee, J.W., 1998. Steroid receptor coactivator-1 coactivates activating protein-1-mediated

749

transactivations through interaction with the c-Jun and c-Fos subunits. J. Biol. Chem. 273, 16651–16654. Liu, S., Tinker, L., Song, Y., Rifai, N., Bonds, D.E., Cook, N.R., Heiss, G., Howard, B.V., Hotamisligil, G.S., Hu, F.B., Kuller, L.H., Manson, J.E., 2007. A prospective study of inflammatory cytokines and diabetes mellitus in a multiethnic cohort of postmenopausal women. Arch. Intern. Med. 167, 1676–1685. MacKinnon, D.P., Lockwood, C.M., Hoffman, J.M., West, S.G., Sheets, V., 2002. A comparison of methods to test mediation and other intervening variable effects. Psychol. Methods 7, 83–104. McKay, L.I., Cidlowski, J.A., 1998. Cross-talk between nuclear factor-kappa B and the steroid hormone receptors: mechanisms of mutual antagonism. Mol. Endocrinol. 12, 45–56. McKay, L.I., Cidlowski, J.A., 1999. Molecular control of immune/inflammatory responses: interactions between nuclear factor-kappa B and steroid receptorsignaling pathways. Endocr. Rev. 20, 435–459. Miller, G.E., Chen, E., 2006. Life stress and diminished expression of genes encoding glucocorticoid receptor and beta2-adrenergic receptor in children with asthma. Proc. Natl. Acad. Sci. USA 103, 5496–5501. Miller, G.E., Cohen, S., Ritchey, A.K., 2002. Chronic psychological stress and the regulation of pro-inflammatory cytokines: a glucocorticoid-resistance model. Health Psychol. 21, 531–541. Motulsky, H.J., Insel, P.A., 1982. Adrenergic receptors in man: direct identification, physiologic regulation, and clinical alterations. N. Engl. J. Med. 307, 18–29. Neumann, M., Grieshammer, T., Chuvpilo, S., Kneitz, B., Lohoff, M., Schimpl, A., Franza Jr., B.R., Serfling, E., 1995. RelA/p65 is a molecular target for the immunosuppressive action of protein kinase A. EMBO J. 14, 1991–2004. O’Connor, M.F., Motivala, S.J., Valladares, E.M., Olmstead, R., Irwin, M.R., 2007. Sex differences in monocyte expression of IL-6: role of autonomic mechanisms. Am. J. Physiol. Regul. Integr. Comp. Physiol. 293, R145–R151. Pace, T.W., Mletzko, T.C., Alagbe, O., Musselman, D.L., Nemeroff, C.B., Miller, A.H., Heim, C.M., 2006. Increased stress-induced inflammatory responses in male patients with major depression and increased early life stress. Am. J. Psychiatry 163, 1630–1633. Paliogianni, F., Kincaid, R.L., Boumpas, D.T., 1993. Prostaglandin E2 and other cyclic AMP elevating agents inhibit interleukin 2 gene transcription by counteracting calcineurin-dependent pathways. J. Exp. Med. 178, 1813–1817. Parry, G.C., Mackman, N., 1997. Role of cyclic AMP response element-binding protein in cyclic AMP inhibition of NF-kappaB-mediated transcription. J. Immunol. 159, 5450–5456. Pradhan, A.D., Manson, J.E., Rifai, N., Buring, J.E., Ridker, P.M., 2001. C-reactive protein, interleukin 6, and risk of developing type 2 diabetes mellitus. JAMA 286, 327–334. Preacher, K.J., Hayes, A.F., 2004. SPSS and SAS procedures for estimating indirect effects in simple mediation models. Behav. Res. Methods Instrum. Comput. 36, 717–731. Ray, A., Prefontaine, K.E., 1994. Physical association and functional antagonism between the p65 subunit of transcription factor NF-kappa B and the glucocorticoid receptor. Proc. Natl. Acad. Sci. USA 91, 752–756. Richlin, V.A., Arevalo, J.M., Zack, J.A., Cole, S.W., 2004. Stress-induced enhancement of NF-kappaB DNA-binding in the peripheral blood leukocyte pool: effects of lymphocyte redistribution. Brain Behav. Immun. 18, 231–237. Ross, R., 1999. Atherosclerosis—an inflammatory disease. N. Engl. J. Med. 340, 115–126. Sapolsky, R.M., Romero, L.M., Munck, A.U., 2000. How do glucocorticoids influence stress responses? Integrating permissive, suppressive, stimulatory, and preparative actions. Endocr. Rev. 21, 55–89. Scheinman, R.I., Cogswell, P.C., Lofquist, A.K., Baldwin Jr., A.S., 1995. Role of transcriptional activation of I kappa B alpha in mediation of immunosuppression by glucocorticoids. Science 270, 283–286. Sheppard, K.A., Phelps, K.M., Williams, A.J., Thanos, D., Glass, C.K., Rosenfeld, M.G., Gerritsen, M.E., Collins, T., 1998. Nuclear integration of glucocorticoid receptor and nuclear factor-kappaB signaling by CREB-binding protein and steroid receptor coactivator-1. J. Biol. Chem. 273, 29291–29294. Smedes, F., Kraak, J.C., Poppe, H., 1982. Simple and fast solvent extraction system for selective and quantitative isolation of adrenaline, noradrenaline and dopamine from plasma and urine. J. Chromatogr. 231, 25–39. Tsuruta, L., Lee, H.J., Masuda, E.S., Koyano-Nakagawa, N., Arai, N., Arai, K., Yokota, T., 1995. Cyclic AMP inhibits expression of the IL-2 gene through the nuclear factor of activated T cells (NF-AT) site, and transfection of NF-AT cDNAs abrogates the sensitivity of EL-4 cells to cyclic AMP. J. Immunol. 154, 5255–5264. Vedhara, K., Fox, J.D., Wang, E.C., 1999. The measurement of stress-related immune dysfunction in psychoneuroimmunology. Neurosci. Biobehav. Rev. 23, 699–715. Waage, A., Slupphaug, G., Shalaby, R., 1990. Glucocorticoids inhibit the production of IL6 from monocytes, endothelial cells and fibroblasts. Eur. J. Immunol. 20, 2439–2443. Whitacre, C.C., 2001. Sex differences in autoimmune disease. Nat. Immunol. 2, 777–780. Wulczyn, F.G., Krappmann, D., Scheidereit, C., 1996. The NF-kappa B/Rel and I kappa B gene families: mediators of immune response and inflammation. J. Mol. Med. 74, 749–769.