Glucocorticoid receptor dimerization is required for proper ... - Nature

9 downloads 0 Views 2MB Size Report
Oct 23, 2012 - ORIGINAL ARTICLE. Glucocorticoid receptor dimerization is required for proper recovery of LPS-induced inflammation, sickness behavior and.
Molecular Psychiatry (2013) 18, 1006–1017 & 2013 Macmillan Publishers Limited All rights reserved 1359-4184/13 www.nature.com/mp

ORIGINAL ARTICLE

Glucocorticoid receptor dimerization is required for proper recovery of LPS-induced inflammation, sickness behavior and metabolism in mice MN Silverman1, P Mukhopadhyay2, E Belyavskaya1, LH Tonelli3, BD Revenis1, JH Doran1, BE Ballard1, J Tam2, P Pacher2 and EM Sternberg1,4 Endogenous glucocorticoids are essential for mobilizing energy resources, restraining inflammatory responses and coordinating behavior to an immune challenge. Impaired glucocorticoid receptor (GR) function has been associated with impaired metabolic processes, enhanced inflammation and exaggerated sickness and depressive-like behaviors. To discern the molecular mechanisms underlying GR regulation of physiologic and behavioral responses to a systemic immune challenge, GRdim mice, in which absent GR dimerization leads to impaired GR–DNA-binding-dependent mechanisms but intact GR protein–protein interactions, were administered low-dose lipopolysaccharide (LPS). GRdim-LPS mice exhibited elevated and prolonged levels of plasma corticosterone (CORT), interleukin (IL)-6 and IL-10 (but not plasma tumor necrosis factor-a (TNFa)), enhanced early expression of brain TNFa, IL-1b and IL-6 mRNA levels, and impaired later central TNFa mRNA expression. Exaggerated sickness behavior (lethargy, piloerection, ptosis) in the GRdim-LPS mice was associated with increased early brain proinflammatory cytokine expression and late plasma CORT levels, but decreased late brain TNFa expression. GRdim-LPS mice also exhibited sustained locomotor impairment in the open field, body weight loss and metabolic alterations measured by indirect calorimetry, as well as impaired thermoregulation. Taken together, these data indicate that GR dimerization-dependent DNA-binding mechanisms differentially regulate systemic and central cytokine expression in a cytokine- and time-specific manner, and are essential for the proper regulation and recovery of multiple physiologic responses to low-dose endotoxin. Moreover, these results support the concept that GR protein–protein interactions are not sufficient for glucocorticoids to exert their full anti-inflammatory effects and suggest that glucocorticoid responses limited to GR monomer-mediated transcriptional effects could predispose individuals to prolonged behavioral and metabolic sequelae of an enhanced inflammatory state. Molecular Psychiatry (2013) 18, 1006–1017; doi:10.1038/mp.2012.131; published online 23 October 2012 Keywords: cytokine; depression; hippocampus; HPA axis; hypothalamus; thermoregulation

INTRODUCTION Endogenous glucocorticoids play an important role in regulating homeostatic processes under basal and challenge conditions, including metabolism, immune function and behavior.1,2 The ultimate effect of glucocorticoids is established at the level of the glucocorticoid receptor (GR). Unrestrained inflammation and hypothalamic–pituitary–adrenal (HPA) axis activity, as well as metabolic alterations, are commonly observed in certain subsets of depressed patients and may be due to glucocorticoid resistance via impaired GR function.3–6 This phenomenon may be of particular relevance to depression in the medically ill, which is characterized by an enhanced inflammatory state.7–10 Innate immune cytokines can influence virtually every pathophysiologic domain relevant to depression, including neuroendocrine function, neurotransmitter metabolism, glucose metabolism, regional brain activity and, ultimately, behavior.11–13 Insight into the molecular mechanisms that underlie the action of GR in regulating inflammation and its behavioral and metabolic sequelae might 1

help explain the high comorbidity between depression, metabolic disorders and chronic inflammatory disorders. Upon exposure to an infectious or inflammatory challenge, proinflammatory cytokines, such as tumor necrosis factor-a (TNFa), interleukin (IL)-1b and IL-6, contribute to the progression of the immune response and also act on the central nervous system (CNS) to coordinate an adaptive sickness syndrome (acutephase response), which includes fever production, a characteristic set of sickness behaviors and stimulation of the HPA axis with the resultant release of glucocorticoids.14–17 Although fever is adaptive because of its antimicrobial and immunostimulating nature, it is metabolically costly.18,19 To conserve energy, animals shift their motivational state toward a behavioral depression, including reduced locomotor activity and food intake and increased sleep.15 Glucocorticoids, as well as proinflammatory cytokines, facilitate the mobilization of energy sources to meet the metabolic demands associated with heat production and immune activation.13,20–22 Glucocorticoids also feed back on the immune

Section on Neuroendocrine Immunology and Behavior, National Institute of Mental Health, National Institutes of Health, Bethesda, MD, USA; 2Laboratory of Physiologic Studies, National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Bethesda, MD, USA; 3Mood and Anxiety Program, Department of Psychiatry, University of Maryland School of Medicine, Baltimore, MD, USA and 4Arizona Center for Integrative Medicine, University of Arizona College of Medicine, Tucson, AZ, USA. Correspondence: Dr EM Sternberg, Arizona Center for Integrative Medicine, University of Arizona College of Medicine, P.O. Box 245153, Tucson, AZ, 85724, USA. E-mail: [email protected] Received 12 March 2012; revised 26 July 2012; accepted 15 August 2012; published online 23 October 2012

Glucocorticoid regulation of LPS-induced sickness MN Silverman et al

1007 system to protect the host from the detrimental consequences of an overactive inflammatory response.23–27 Indeed, upon exposure to an inflammatory challenge, adrenalectomized animals or those treated with a GR antagonist exhibit exaggerated peripheral and central inflammatory responses,28–31 sickness behavior32–35 and fever,36,37 and in extreme cases of unresolved inflammation, septic shock-induced mortality.38,39 Several features of sickness behavior overlap with symptoms of depression, particularly neurovegetative symptoms (that is, psychomotor slowing, fatigue, anorexia, weight loss, hypersomnia). Sickness behavior is usually acute and considered adaptive for the recovery from infection and usually subsides along with the resolution of inflammation. However, sustained levels of inflammation can contribute to prolonged expression of these behaviors and in some conditions, culminate in the maladaptive development of depression.8–11 Similarly, the sustained catabolic effects of an enhanced inflammatory state may contribute to an ‘inflammatory’ metabolic syndrome.13 Indeed, sickness behaviors/neurovegetative symptoms are a prominent feature of immunotherapy (for example, interferon-a)-induced depression and in those with metabolic syndrome.40,41 The molecular mechanisms by which an impaired GR may be a vulnerability factor for inflammation-induced behavioral and metabolic dysfunction requires further elucidation and is the focus of the present study. The GR is a ligand-dependent transcription factor that belongs to the nuclear hormone receptor super family. The two main molecular pathways by which GR controls transcription are by directly binding to DNA and through protein–protein interactions with other transcription factors. Direct binding of GR to DNA includes binding to positive or negative glucocorticoid response elements (GREs) in the promoter regions of genes, leading to transcriptional increases (that is, anti-inflammatory and metabolic molecules) or decreases (that is, negative feedback on the HPA axis). Simple or ‘classical’ GREs consist of two palindromic GRE half-sites, separated by 3 base pairs, and are associated with GR dimerization-dependent transcription of genes. Complex GREs can include multiple GRE half-sites or GRE half-sites in conjunction with response elements for other transcription factors (composite GREs), and do not require GR dimerization. GR monomers also indirectly influence the transcription of genes through protein– protein interactions with other transcription factors (that is, inhibiting the transcriptional activity of nuclear factor-kB and activator protein-1 on proinflammatory molecules), without binding to DNA themselves.42–46 Whereas the metabolic actions of glucocorticoids are believed to be predominantly mediated by GR dimerization-dependent DNA-binding mechanisms, the extent to which these mechanisms contribute to the anti-inflammatory effects of GR remains an issue of debate.47,48 To dissect the relative importance of GR–DNA-binding-dependent mechanisms versus protein–protein interactions in GR regulation of the acutephase response to a systemic inflammatory challenge, we administered low-dose lipopolysaccharide (LPS) to GRdim mice, in which a point mutation in the GR–DNA-binding domain results in deficient GR dimerization. Hence, GR–DNA-binding-dependent transcriptional regulation is impaired, but GR protein–protein interactions with other transcription factors remain intact.49 A recent study using the GRdim mice has shown that endogenous glucocorticoids require GR dimerization to effectively inhibit systemic inflammatory responses to high-dose LPS and, consequentially, GRdim mice show enhanced lethality in LPSinduced septic shock.50 In the present study, we evaluated systemic and CNS cytokine production in GRdim mice over a 3-day time course following administration of low-dose, sublethal LPS to determine whether these mice are more susceptible to the HPA axis, inflammatory, behavioral and metabolic alterations characteristic of nonseptic inflammatory disorders. We show that GR dimerization, and hence intact GR–DNA-binding-dependent & 2013 Macmillan Publishers Limited

mechanisms, are essential for the proper regulation and recovery of LPS-induced corticosterone (CORT), peripheral and brain cytokines, sickness behavior and various metabolic parameters. In addition, we show that GR dimerization-dependent DNAbinding mechanisms differentially regulate systemic and central cytokine expression in a cytokine- and time-dependent manner. Taken together, these results support the concept that GR protein–protein interactions are not sufficient for glucocorticoids to exert their full anti-inflammatory effects, and suggest that endogenous glucocorticoid responses limited to GR monomermediated transcriptional effects could predispose individuals to prolonged behavioral and metabolic sequelae of an enhanced inflammatory state.

MATERIALS AND METHODS Animals Male mice, 8–10 weeks old, were housed at the National Institutes of Health (NIH) for at least 1 week before use and were maintained on a 12 h light/dark cycle with lights on at 6:30 AM. The generation of the GRdim mice was described previously.49 GRdim mouse breeder pairs (129 genetic background) were obtained from Dr Gu¨nther Schu¨tz (German Cancer Research Centre, Heidelberg, Germany) and breeding colonies were established at Taconic Farms (Rockville, MD, USA). 129S6.B6Nr3c1tm3GscN24 mice were generated from male and female homozygous GRdim breeder pairs. Age-matched wild-type (WT) controls (129SvEv) were purchased separately from Taconic Farms. Food and water were available ad libitum. All procedures were conducted in accordance with the NIH Guide for the Care and Use of Experimental Animals.

LPS challenge Mice were injected intraperitoneally with either LPS (0.25 mg/kg) or sterile saline at 10 ml per g body weight. LPS (L2880; Escherichia coli, strain 055:B5; Sigma-Aldrich, St Louis, MO, USA) was reconstituted to 1 mg/ml with sterile saline. Aliquots were stored at  80 1C until further diluted before injection. To find a sublethal dose of LPS for both genotypes, we initially tested a standard dose of LPS (1 mg/kg) known to cause upregulation of central cytokines and sickness behavior with 100% survival in normal mice.51,52 As in these previous studies, 100% of our WT-LPS mice survived throughout the 72-h test period. However, 0% of the GRdim-LPS mice survived after 48 h. We, therefore, chose to use a dose of 0.25 mg/kg LPS for this study, where the majority of GRdim mice survived throughout the test period (72%) and the WT mice exhibited a reliable expression of sickness behavior. Data from all mice during survival were included in analysis. Mice were killed if they reached the clinical end point of a sickness behavior score of 11 (see below).

Experimental design Four groups of experiments were carried out in this study, each with a different set of mice. Experiment 1. Examined peripheral and central neuroendocrine and cytokine responses (plasma CORT, TNFa, IL-6 and IL-10 levels; hippocampal and hypothalamic TNFa, IL-1b, IL-6 and IL-10 mRNA; n ¼ 8–12 per group); home cage sickness behavior scores were also recorded in these mice. Experiment 2. Examined home cage sickness behavior, skin temperature, body weight loss and locomotor activity in the open field test (n ¼ 8–12 per group). Experiment 3. Examined metabolic parameters (oxygen consumption (VO2), carbon dioxide production (VCO2), respiratory quotient (RQ), carbohydrate oxidation, fat oxidation and total energy expenditure (TEE)) by indirect calorimetry (n ¼ 6 per group). Experiment 4. Examined core body temperature (Tc) via telemetry (n ¼ 4 per group). Molecular Psychiatry (2013), 1006 – 1017

Glucocorticoid regulation of LPS-induced sickness MN Silverman et al

1008 Plasma and tissue collection

Metabolic chambers: indirect calorimetry and locomotor activity

Blood and brains were collected at various time points after injection (2 h (10 AM), 6 h (2 PM), 12 h (8 PM), 24–48–72 h (10 AM)). Trunk blood was collected after decapitation into EDTA tubes, kept on ice and centrifuged at 4 1C, 3000 r.p.m. for 15 min. Plasma was aliquoted and stored at  80 1C until assayed for hormone/cytokine levels. Brains were removed from the skull and the hippocampus and hypothalamus were dissected over ice, flash frozen on dry ice and stored at  80 1C until ready for RNA isolation.

VO2 and VCO2 were monitored by indirect calorimetry (Oxymax, Columbus Instruments, Columbus, OH, USA), using sealed, transparent chambers (8.5  5.25  4.5 inches). Each chamber housed 1 mouse. Eight mice were run simultaneously. Mice had free access to food and water during their 12 h light/dark cycle with lights on at 9 AM. Conditioned fresh air at 21±0.5 1C and 55±5% relative humidity was pumped into the chambers at 0.6 l/min. Mice were acclimated to the chambers for 24 h before treatment and VO2 and VCO2 measurements were recorded every 8 min on a computer over the next 72 h after saline or LPS injection. O2 and CO2 content were measured in the efflux air. RQ was calculated as VCO2/VO2, with the values of 1.0 and 0.7 indicating 100% carbohydrate oxidation (CHO) and 100% fat oxidation (FO), respectively. TEE was calculated as VO2  (3.815 þ 1.232  RQ), normalized to (body mass)0.75, and expressed as kcal/h/ kg0.75. FO and CHO were calculated as follows: FO ¼ 1.69  VO2–1.69  VCO2 and CHO ¼ 4.57  VCO2–3.23  VO2 and expressed as g/d/kg0.75. Each chamber was also equipped with a two-dimensional (xy) infrared beam system (OPTO-M3, Columbus Instruments, Columbus, OH, USA) to record locomotor activity.

Hormone and cytokine assays Plasma CORT was measured by enzyme immunoassay (OCTEIA Corticosterone HS, Immunodiagnostic Systems, Fountain Hills, AZ, USA; limit of detection ¼ 0.17 ng/ml). Plasma TNFa, IL-6 and IL-10 were measured by sandwich ELISA (Quantikine murine kit, R&D Systems, Minneapolis, MN, USA; limit of detection ¼ 5.1, 1.6 and 4.0 pg/ml, respectively). Samples were run in duplicate.

Reverse transcription and real-time PCR Total RNA was isolated from the hippocampus and hypothalamus using TRIZOL (Invitrogen, Carlsbad, CA, USA), followed by DNase I treatment (Ambion-Applied Biosystems, Carlsbad, CA, USA). RNA was further purified with an RNeasy mini kit (QIAGEN, Valencia, CA, USA) and eluted in 30 ml RNase free water. Yields were determined by spectrophotometer at l 260 nm. Then, 1 mg of each RNA sample was reverse transcribed using the iScript complementary DNA Synthesis kit (Bio-Rad, Hercules, CA, USA). Real-time PCR was performed on the resulting complementary DNA using a Bio-Rad iCycler iQ for 40 cycles (denatured at 95 1C for 15 s, annealed at 55 1C for 30 s and extended at 72 1C for 30 s) and the SYBR Green PCR Master Mix (Bio-Rad). Samples were normalized to b-actin and run in duplicate. Amplifications without template were included as negative controls. Amplicon homogeneity was confirmed by melting curve analyses for all samples. The following forward (F) and reverse (R) primers were used: TNFa (F: 50 -TCTCATTCCTGCTTGTGGCAG-30 ; R: 50 -TCCACTTGGTGGTTTGCT ACG-30 ) IL-1b (F: 50 -AAAAAAGCCTCGTGCTGTCG-30 ; R: 50 -GTCGTTGCTTGGTTC TCCTTG-30 ) IL-6 (F: 50 -TCCATCCAGTTGCCTTCTTG-30 ; R: 50 -TTCCACGATTTCCCAGAG AAC-30 ) IL-10 (F: 50 -GGGTTGCCAAGCCTTATCGG-30 ; R: 50 -ATCACTCTTCACCTGCTCC ACTGC-30 ) b-actin (F: 50 -TGCACCACCAACTGCTTAG-30 ; R: 50 -GGATGCAGGGATG ATGTTC-30 ).

Sickness parameters Sickness behavior score. Sickness behavior was measured at 6, 24, 48 and 72 h after injection in the home cage. A total sickness behavior score was based on an 11-point system, determined by the sum of scores for three parameters: piloerection (1–3), ptosis (squinty/watery eyes; 1–4) and lethargy (1–4), with a score of 1 within each parameter considered healthy. Body weight. Mice were weighed twice a day to collect AM and PM body weight measurements, with PM measurements taken just before lights off. Skin temperature. The QuickTEMP non-contact thermometer (Linear Laboratories, Fremont, CA, USA) was used to measure skin temperature twice a day (AM and PM) via an infrared energy detector.

Open field: locomotor activity Horizontal locomotor activity was assessed using the open field test between 2 and 4 PM at 6, 24, 48 and 72 h after injection. The open field chamber consisted of a square plexiglass enclosure (40  40 cm) with opaque walls and grids divided into 16 equal squares on the floor. Two mice were tested simultaneously (one in each box). Mice were placed in the apparatus at the bottom right corner and allowed to explore the area for 10 min. Each test session was recorded and analyzed at a later time. After testing, mice were returned to their home cage. Open field videos were analyzed using ANY-maze software (Stoelting, Wood Dale, IL, USA) utilizing the color subtraction method to track mice. Total horizontal distance was measured from a point calculated to be the center of the mouse’s total body area. Molecular Psychiatry (2013), 1006 – 1017

Telemetry Mice were implanted intraperitoneally with telemetry transmitters (Data Sciences International, St Paul, MN, USA) to measure Tc under ambulatory conditions. All surgeries were performed under aseptic conditions and with the mice under isoflurane (inhalant) anesthesia. After a 10-day postsurgery recovery period, temperature was continuously recorded from 72 h before to 7 days after LPS administration.

Statistical analysis Values are presented as mean±s.e.m. A two-way analysis of variance was used to evaluate genotype (WT vs GRdim) and treatment (saline vs LPS) effects within each time point (across time points for the open field data.) The Bonferroni method was used for post hoc tests of significant differences between group means. The level of significance was set at Pr0.05, and all tests of significance were two tailed. *Treatment difference within genotype (LPS vs saline). #Genotype difference within treatment (GRdim vs WT). */#Po0.05, **/##Po0.01, ***/###Po0.001.

RESULTS GRdim mice show prolonged LPS-induced plasma CORT, IL-6 and IL-10 responses, but not TNFa responses GRdim mice displayed greater basal CORT levels than WT mice across the circadian cycle (2, 24, 48 and 72 h (10 AM sac)-average of 5.8  ; 6 h (2 PM sac)-2.1  ; and 12 h (8 PM sac)-9.3  greater than WT; Figure 1a). In response to LPS, WT and GRdim mice showed similar early rises (2 and 6 h) in plasma CORT levels. However, GRdim-LPS mice exhibited greater CORT levels than WTLPS mice from 12 to 72 h, showing a delayed return to baseline relative to WT mice (Figure 1a). (Taken into consideration that GRdim mice displayed greater basal CORT levels, the area under the curve between LPS-induced CORT levels and their respective basal values indicated that GRdim mice displayed on overall 1.4  greater response compared with WT mice (1533 vs 1087 ng/ml  time).) LPS-induced plasma cytokine responses were differentially regulated by glucocorticoids. WT and GRdim mice showed similar early rises (2 h) in plasma TNFa responses to LPS, which returned to baseline levels by 12 h (Figure 1b). In contrast, GRdim-LPS mice exhibited increased plasma IL-6 at 2 h and prolonged elevations in plasma IL-6 and IL-10 for up to 24 h relative to WT-LPS mice (Figures 1c and d). (Plasma cytokine levels were undetectable in saline-treated mice.) GRdim mice show exaggerated early brain inflammatory responses to LPS, but impaired later TNFa responses Out of the four cytokine transcripts measured in the brain (TNFa, IL-1b, IL-6 and IL-10), LPS-induced IL-1b mRNA was the most highly expressed in both the hippocampus and hypothalamus, the & 2013 Macmillan Publishers Limited

Glucocorticoid regulation of LPS-induced sickness MN Silverman et al

1009

Figure 1. GRdim mice show prolonged lipopolysaccharide (LPS)-induced plasma corticosterone (CORT), interleukin (IL)-6 and IL-10 responses, but not tumor necrosis factor-a (TNFa) responses. (a) Trunk blood was collected from mice and plasma CORT was measured by enzyme immunoassay (EIA; n ¼ 8–12 mice per group). (b–d) TNFa, IL-6 and IL-10 were measured by enzyme-linked immunosorbent assay (ELISA; n ¼ 8–12 mice per group). Data are presented as mean±s.e.m. *Treatment difference within genotype (LPS vs saline). #Genotype difference within treatment (GRdim vs wild type (WT)). */#Po0.05, **/##Po0.01, ***/###Po0.001.

latter showing the greatest expression in both genotypes. IL-10 mRNA could not be detected in either brain region. WT expression of LPS-induced TNFa mRNA in the hippocampus and hypothalamus showed a biphasic response, where peak levels were expressed at 2 and 24 h, the latter being the time of greatest expression. TNFa mRNA responses in GRdim-LPS mice also showed a biphasic response, with the 2 h peak exceeding that of the WT, but the 24 h peak being impaired. LPS-induced IL-1b mRNA levels peaked at 2 h in the hippocampus and hypothalamus, in both genotypes, where GRdim-LPS mice showed exaggerated responses at early time points (2 and 6 h). WT expression of LPS-induced IL-6 mRNA also peaked at 2 h in the hippocampus and hypothalamus. However, GRdim-LPS mice showed peak IL-6 mRNA responses at 2 h in the hypothalamus and at 6 h in the hippocampus, exceeding WT-LPS IL-6 mRNA levels at both respective time points (Figures 2a–f). GRdim mice show sustained LPS-induced sickness behavior, locomotor activity impairment and body weight loss WT-LPS mice exhibited peak expression of home cage sickness behavior (piloerection, ptosis and lethargy) and impairment in open field locomotor activity at 6 h. GRdim-LPS mice exhibited sustained sickness behavior and locomotor impairment, peaking at 24 h and showing a delayed recovery compared with WT-LPS mice (Figures 3a and b). Interestingly, GRdim-Saline mice displayed less basal locomotor activity compared with WT-Saline mice (71% of WT). Greatest LPS-induced genotype differences occurred at 24 h (6 h: WT-62% of saline, GRdim-56% of saline; 24 h: WT-76% of saline, GRdim-38% of saline). Total sickness behavior scores positively correlated with brain cytokine expression (hippocampal and hypothalamic TNFa, IL-1b and IL-6 mRNA) at 6 h after LPS and with plasma CORT levels at later time points (12, 24 and 48 h) across both genotypes. Interestingly, sickness behavior scores were negatively correlated with hippocampal and hypothalamic expression of TNFa mRNA at 24 h after LPS across both WT and GRdim mice (Figure 4). Under baseline conditions, saline-treated WT and GRdim mice displayed a normal circadian rhythm in their body weight, with & 2013 Macmillan Publishers Limited

AM increases following their active period in the dark phase and PM decreases following their inactive period in the light phase. Baseline body weights across the circadian rhythm did not differ between genotypes. In response to LPS, both WT and GRdim mice displayed a disruption in their diurnal weight patterns up to 48 h (D2-AM) (Figure 3c). GRdim-LPS mice showed a delayed recovery in weight gain, where LPS-WT mice exhibited peak weight loss at 24 h (D1) and returned to baseline by 72 h (D3), and GRdim-LPS mice showed sustained peak weight loss up to 48 h (D2) and recovery by 96 h (D4) (Figure 3d). GRdim mice exhibit altered metabolism and thermoregulation Substrate utilization was analyzed by indirect calorimetry. Both genotypes displayed a basal circadian rhythm in which RQ (VCO2/ VO2), TEE, carbohydrate oxidation and ambulatory activity were elevated during the active/dark phase and reduced during the inactive/light phase, whereas fat oxidation showed the reverse cycle. Genotype differences were apparent under basal conditions, where GRdim-Saline mice showed reduced VO2 inhaled, VCO2 exhaled, TEE, carbohydrate oxidation and ambulatory activity relative to WT-Saline mice during their inactive and active periods. RQ and fat oxidation parameters were similar under basal conditions. In response to LPS, WT mice showed reductions in RQ (decrease in both VO2 and VCO2), carbohydrate oxidation, TEE and ambulatory activity, and an increase in fat oxidation. GRdimLPS mice showed an exacerbation of these changes, in amplitude and duration, indicating more sustained metabolic alterations. All metabolic changes were most prominent within the first 24 h (D0PM) for both genotypes and returned to baseline by 48 h (D2-AM) in the WT-LPS mice (Figures 5a–g). As part of the normal circadian rhythm, Tc, as measured by telemetry, was elevated during the active/dark phase and reduced during the inactive/light phase. Before LPS injection, WT and GRdim mice exhibited similar Tc values in amplitude and circadian rhythm. A peak elevation in Tc of 1.4 1C was evident in the WT-LPS mice by 8 h after injection (still during the light phase) and returned to normal dark phase levels by 12 h after injection. In contrast, GRdim-LPS mice exhibited an early transient rise in Tc Molecular Psychiatry (2013), 1006 – 1017

Glucocorticoid regulation of LPS-induced sickness MN Silverman et al

1010

Figure 2. GRdim mice show exaggerated early brain inflammatory responses to lipopolysaccharide (LPS), but impaired later tumor necrosis factor-a (TNFa) responses. Real-time-PCR was performed on hippocampal (a–c) and hypothalamic (d–f ) regions of the brain for TNFa, interleukin (IL)-1b and IL-6 mRNA expression (n ¼ 8–12 mice per group). Data are presented as mean±s.e.m. for the fold change in mRNA expression relative to the wild-type (WT)-Saline group. *Treatment difference within genotype (LPS vs saline). #Genotype difference within treatment (GRdim vs WT). */#Po0.05, **/##Po0.01, ***/###Po0.001.

Figure 3. GRdim mice show sustained lipopolysaccharide (LPS)-induced sickness behavior, locomotor activity impairment and body weight loss. (a) Total sickness score was based on an 11-point system, determined by the sum of scores for three parameters: piloerection (1–3), ptosis (squinty/watery eyes; 1–4) and lethargy (1–4), with a score of 1 within each parameter considered healthy (n ¼ 31–39 mice per group). (b) Distance traveled during a 10-min open field test is shown (n ¼ 8–12 mice per group, combined saline ¼ 39–42 animals). (c) Mice were weighed in the AM and PM (n ¼ 28–34 mice per group). Injections were administered on day 0 (D0) between AM and PM data points (arrow). (d) Delta body weight from baseline are presented as: after injection AM weight  before injection (D0) AM weight. Data are presented as mean±s.e.m. *Treatment difference within genotype (LPS vs saline). #Genotype difference within treatment (GRdim vs wild type (WT)). */#Po0.05, **/##Po0.01, ***/###Po0.001. Molecular Psychiatry (2013), 1006 – 1017

& 2013 Macmillan Publishers Limited

Glucocorticoid regulation of LPS-induced sickness MN Silverman et al

1011

Figure 4. Biological correlations with sickness behavior scores. Sickness behavior positively correlated with hippocampal (a–c) and hypothalamic (d–f ) cytokine mRNA expression (Pearson’s r, Po0.05) at 6 h and (g–i) plasma corticosterone (CORT) levels (Pearson’s r, Po0.001) at 12, 24 and 48 h after lipopolysaccharide (LPS) across wild-type (WT) and GRdim mice. IL, interleukin. However, sickness behavior negatively correlated with hippocampal (j) and hypothalamic (k) tumor necrosis factor-a (TNFa) mRNA expression (Pearson’s r, Po0.01 and Po0.05, respectively) at 24 h after LPS.

(2–4 h after injection), but were impaired in the sustained rise in Tc during the remainder of the light phase immediately following LPS injection and in the normal circadian rise of Tc during the first dark phase. GRdim-LPS mice regained their normal circadian rhythm by 24 h after injection (Figure 5h). Basal skin temperature was the same for all groups. WT-LPS mice did not display any significant changes in skin temperature. However, GRdim mice exhibited a peak reduction in skin temperature (  1.6 1C) at 24 h (D1-AM) after LPS (Figure 5i). This period of impaired thermoregulation corresponded to the time of greatest sickness behavior expression and peak impairment of locomotor activity. DISCUSSION We show here for the first time that mice deficient in GR dimerization, and therefore impaired in GR–DNA-binding-dependent transcriptional regulation, are more sensitive to the neuroendocrine, inflammatory, behavioral and metabolic effects of low-dose LPS. We also show for the first time that GR dimerization-dependent mechanisms differentially regulate cytokine expression in the periphery versus the brain, in a cytokineand time-dependent manner. GRdim-LPS mice exhibited elevated and prolonged levels of plasma CORT, IL-6 and IL-10 (but not plasma TNFa), enhanced early expression of brain TNFa, IL-1b and IL-6 mRNA levels and impaired later central TNFa mRNA expression. Home cage sickness behavior (lethargy, piloerection, ptosis) and impaired locomotor activity in the open field peaked later in the GRdim-LPS mice (24 h) compared with WT-LPS mice (6 h). Exaggerated sickness behavior in the GRdim-LPS mice was associated with increased early hippocampal and hypothalamic proinflammatory cytokine expression and late plasma CORT levels, but decreased late brain TNFa expression. GRdim-LPS mice also exhibited sustained body weight loss and metabolic alterations, including reduced RQ, TEE and carbohydrate oxidation, and enhanced fat oxidation compared with WT-LPS mice. Impaired & 2013 Macmillan Publishers Limited

thermoregulation, as evidenced by the inability to maintain a fever and normal skin temperature, was also displayed by GRdimLPS mice. These results indicate that intact GR dimerizationdependent DNA-binding mechanisms are essential and GR monomer-mediated transcriptional effects, such as protein– protein interactions—although may be important—are not sufficient for the proper regulation and recovery of HPA axis, inflammatory, behavioral and metabolic responses to a low-dose endotoxin challenge. HPA axis GR negative feedback on various levels of the HPA axis is regulated by both GR dimerization-dependent and -independent mechanisms.49 In agreement with previous studies, the GRdim mice in our study exhibited greater basal plasma CORT levels across the circadian rhythm53 and similar early LPS-induced CORT responses compared with WT mice.50,54 However, in contrast to similar high-dose LPS-induced CORT responses in the GRdim and WT mice at later time points,50 GRdim mice treated with low-dose LPS (in this study) exhibited exaggerated plasma CORT responses at later time points with a delayed return to baseline. The difference in the kinetics of the CORT response to low- versus high-dose LPS appears to be in the earlier recovery of CORT levels in the WT mice treated with low-dose LPS. A possible mechanism contributing to the sustained CORT levels observed in the GRdim mice may be their prolonged cytokine expression.55 Indeed, LPSinduced plasma IL-6 levels were positively correlated with plasma CORT levels at 12 h, the time point when CORT responses start to diverge across both genotypes (r ¼ 0.81, Po0.001, data not shown; plasma cytokine levels were not detected in salinetreated WT or GRdim mice). These results suggest that GR dimerization-dependent DNA-binding mechanisms are required for glucocorticoid negative feedback of CORT release under basal conditions as well as for the termination of CORT responses to an Molecular Psychiatry (2013), 1006 – 1017

Glucocorticoid regulation of LPS-induced sickness MN Silverman et al

1012

Figure 5. GRdim mice exhibit altered metabolism and thermoregulation. (a–f ) Indirect calorimetry parameters, oxygen consumption (VO2), carbon dioxide production (VCO2), respiratory quotient, total energy expenditure (TEE), carbohydrate (CH) oxidation and fat oxidation, as well as (g) ambulatory activity, were measured for 3 days after injection (n ¼ 6 mice per group). Each AM/PM data point is the average of a 12-h light (AM) or dark (PM) phase of the light cycle. Injections were administered during the day 0 (D0) AM period (arrow). (h) Core body temperature (Tc) was recorded continuously by abdominally implanted telemetry sensors (n ¼ 4 mice per group). Each data point is the average of every second within a 2-h period. Light (white boxes) and dark (black boxes) phases of the light cycle are shown above the graph. (i) Skin temperature was recorded by an infrared laser-based thermometer (n ¼ 28–39 mice per group) in the AM and PM. Injections were administered on D0 between AM and PM data points (arrow). Data are presented as mean±s.e.m. *Treatment difference within genotype (lipopolysaccharide (LPS) vs saline). #Genotype difference within treatment (GRdim vs wild type (WT)). */#Po0.05, **/##Po0.01, ***/###Po0.001.

LPS challenge. This sustained CORT response may be because of impaired negative feedback at the level of the HPA axis and/or cytokine production. Inflammation The role of glucocorticoids in keeping peripheral and central inflammatory responses in check has been well established;23–27 however, the mechanism by which it does so requires further elucidation. The predominant role of GR protein–protein interactions with other transcription factors (for example, nuclear factorkB and activator protein-1) in mediating the anti-inflammatory effects of GR is mainly supported by in vitro studies.44,49,54 Under in vivo conditions, the dependence of the anti-inflammatory effects of GR on GR dimerization-dependent DNA-binding mechanisms has been shown to differ according to the type of immune challenge.50,54,56,57 Our data indicate that the role of GR– DNA-binding-dependent mechanisms in regulating LPS-induced systemic inflammatory responses is cytokine specific. In agreement with previous studies using high-dose LPS,50 the kinetic profile of low-dose LPS-induced plasma TNFa responses was similar in GRdim and WT mice. In contrast to this profile, GRdim-LPS mice (both high and low doses) exhibited greater and prolonged plasma IL-6 and IL-10 levels relative to WT-LPS mice, indicating impaired recovery in these latter cytokine responses. Our study is the first to show that GR dimerization-dependent DNA-binding mechanisms differentially regulate cytokine expression in the Molecular Psychiatry (2013), 1006 – 1017

periphery versus the brain, and even within the brain itself. In contrast to the similar plasma TNFa levels observed in GRdim and WT-LPS mice, exaggerated hippocampal and hypothalamic TNFa mRNA expression was observed in the GRdim-LPS mice at early time points. Early expression of IL-1b and IL-6 mRNA in the brain was also exaggerated in the GRdim-LPS mice relative to WTLPS mice. Interestingly, GRdim-LPS mice showed impaired central TNFa mRNA responses at the time when WT-LPS mice exhibited a second peak in TNFa mRNA expression (24 h). This may be because of the exaggerated plasma CORT levels in GRdim-LPS mice at later time points, exerting a negative effect through GR dimerization-independent mechanisms. Alternatively, these data suggest that GR dimerization-dependent DNA-binding mechanisms are necessary to promote the second TNFa mRNA peak in the WT-LPS mice in the face of permissive low plasma CORT levels. A bi-phasic central TNFa mRNA response in WT-LPS mice has been shown by others,51 where the early 2-h peak has been associated with cytokine expression in endothelial cells of the brain vasculature and the later 24-h peak with the progressive activation of intraparenchymal cells.58 It is also possible that the second peak of TNFa in the WT-LPS mice could be anti-inflammatory and neuroprotective through TNFa binding to the TNF-R2 receptor rather than the proinflammatory TNF-R1 receptor.59,60 Overall, these data suggest that LPS-induced cytokine responses are tissue and time specific. Even within the same tissue (that is, the brain), different cell types and/or receptor subtypes may mediate early versus late cytokine & 2013 Macmillan Publishers Limited

Glucocorticoid regulation of LPS-induced sickness MN Silverman et al

1013 expression. Differential transcriptional effects of glucocorticoids could be related to cell-specific expression of transcription factors, cofactors, GRE architecture as well as GR itself.42,44,61,62 Although it appears that GR monomer-mediated protein–protein interactions with other transcription factors are sufficient for keeping LPS-induced plasma TNFa responses in check, our data show that GR dimerization-dependent DNA binding mechanisms are required for glucocorticoid negative feedback on plasma IL-6 and IL-10 responses, as well as on early brain cytokine responses to LPS (TNFa, IL-1b and IL-6 mRNA). A possible mechanism for the cytokine-specific inhibition by GR dimers is their binding to negative GREs in the promoter regions of cytokine genes.45,63,64 Of note, no known GRE has been identified in the TNFa promoter.65 Alternatively, glucocorticoids upregulate the transcription of genes that inhibit cytokine mRNA stability,42,66,67 and hence reduced transcription of these genes may lead to increased cytokine mRNA stability and expression. Enhanced cytokine responses in the GRdim-LPS mice could also be due to impaired upregulation of antiinflammatory molecules.47 Because IL-10 is considered an antiinflammatory cytokine and has been shown to express a putative GRE in its promoter,68 the greater plasma IL-10 levels in the GRdimLPS mice were not expected. However, glucocorticoids have been shown to increase or decrease LPS-induced IL-10 secretion (in vitro), depending on glucocorticoid dose.69 In addition, enhanced levels of IL-10 may be a compensatory mechanism for the enhanced expression of proinflammatory cytokines in the GRdim-LPS mice. The finding of undetectable levels of IL-10 mRNA expression in the brain is in agreement with studies showing that LPS-induced expression of central anti-inflammatory cytokines is limited in comparison with proinflammatory cytokine expression.70 However, glucocorticoids have been shown to upregulate other anti-inflammatory molecules in the brain, such as IkBa and IL-1R2 receptor.71,72 The second wave of LPS-induced TNFa expression in the WT brain has been shown to be dependent on Toll-like receptor-2 (TLR-2) upregulation.73 Although the inflammatory signaling of TLR activation has traditionally been associated with host defense mechanisms, TLR activation in the brain also plays a role in neuroprotection.74,75 Interestingly, glucocorticoids, on their own and in synergy with proinflammatory cytokines, have been shown to increase TLR2 expression.76–78 GR-mediated transcription of TLR2 is dependent on both indirect (via another GR target gene) and direct (via a putative GRE in its proximal promoter region) mechanisms.43,77,78 Therefore, the impaired central TNFa mRNA expression at 24 h in the GRdim-LPS mice might be due to reduced GR-mediated transcription of TLR-2. Sickness behavior A functional consequence that might be predicted as a result of enhanced early proinflammatory cytokine expression in the GRdimLPS mice is exaggerated sickness behavior.14 Indeed, mice or rats with impaired glucocorticoid signaling, either via adrenalectomy or administration of a GR antagonist, exhibit exaggerated LPSinduced sickness behaviors (that is, reduced locomotor activity and social exploration and weight loss).32,33,35 Upon LPS administration, GRdim mice in our study exhibited greater and sustained home cage sickness behavior (piloerection, ptosis and lethargy), locomotor activity impairment and weight loss relative to WT mice. Under noninflammatory conditions, previous studies have shown that impaired glucocorticoid and GR levels lead to reduced basal locomotor activity and food intake, whereas transgenic mice with impaired GR expression exhibit similar body weights compared with WT controls.79–81 Although GRdim-Saline mice in our study displayed similar body weights compared with WT-Saline mice, as well as a normal circadian rhythm in body weight, they exhibited reduced basal locomotor activity, as measured by the open field test and an infrared beam system in the metabolic chambers (particularly in the active-PM period). This & 2013 Macmillan Publishers Limited

latter finding is in contrast to previous reports in the GRdim mice.53 Differential basal behavioral phenotypes between studies may be because of differences in the genetic background of the mice. Our data suggest that GR dimerization-dependent DNA-binding mechanisms play an important role in the regulation of basal locomotor activity (at least in mice of the 129SvEv strain) and in the recovery from LPS-induced sickness behaviors. Taking into consideration the role of GR negative feedback on the expression of corticotropin-releasing hormone (CRH), altered CRH levels may contribute to reductions in locomotor activity, food intake and body weight.82,83 Basal levels of CRH mRNA expression in the PVN of the hypothalamus are unchanged in the GRdim mice.49 However, hypothalamic CRH induction during an immune challenge, as well as extrahypothalamic CRH expression, have yet to be measured in the GRdim mice. In further support of the relation between exaggerated early central cytokines and behavioral responses, hippocampal and hypothalamic cytokine expression positively correlated with total sickness behavior scores at 6 h after LPS, when sickness behavior was high in both genotypes. However, at 24 h, when GRdim-LPS mice displayed peak sickness behavior and that of WT-LPS mice was already on the decline, sickness behavior scores no longer correlated with brain cytokine expression, but with plasma CORT levels. Sustained peak sickness behavior at 24 h in the GRdim-LPS mice may be because of alterations in downstream factors induced by the exaggerated early CNS cytokine response or the sustained plasma CORT response. Such factors may include changes in neurotransmitter metabolism or growth factor expression, which have been associated with depressive-like behaviors.11,12 If the 24-h TNFa peak in the WT-LPS mice is associated with an anti-inflammatory/neuroprotective response through TNF-R2 signaling (rather than proinflammatory TNF-R1 signaling),59,60 the impaired 24-h TNFa peak in the GRdim-LPS mice may also contribute to their slower recovery in sickness behaviors. Indeed, hippocampal and hypothalamic expression of TNFa mRNA was inversely correlated with sickness behavior scores at 24 h after LPS across both genotypes, in contrast to the positive correlation at earlier time points. Metabolism In addition to their potent anti-inflammatory effects, glucocorticoids are known to promote a catabolic state of metabolism, providing substrates for increased liver gluconeogenesis and hence increased glucose availability and utility.48 Previous studies have shown that GRdim mice exhibit impaired upregulation of glucocorticoid-induced transcription of liver gluconeogenic enzymes (for example, tyrosine aminotransferase),49,84 indicating that gluconeogenesis requires GR dimerization-dependent DNAbinding mechanisms. Using indirect calorimetry, glucocorticoids and intact GR expression have been shown to be essential for normal RQ and TEE.79,81 We further characterized the metabolic profile of the GRdim mice, showing reduced basal VO2 inhaled, VCO2 exhaled, TEE and carbohydrate oxidation relative to WTSaline mice. In contrast, there were no differences in the basal RQ (VCO2/VO2) or fat oxidation between GRdim and WT mice. Although the GRdim-Saline mice exhibited reductions in amplitude for most metabolic parameters, their circadian rhythm was maintained. In addition to GR–DNA-bindingdependent mechanisms directly upregulating the transcription of gluconeogenic enzymes, they may indirectly regulate metabolic processes via the upregulation of catecholamine biosynthetic enzymes. Transgenic mice with impaired GR expression contain lower levels of norepinephrine in metabolically active tissues, such as brown adipose tissue.81 Therefore, another possible mechanism by which gluconeogenesis may be impaired in the GRdim mice is through reduced sympathetic nervous system activity. Molecular Psychiatry (2013), 1006 – 1017

Glucocorticoid regulation of LPS-induced sickness MN Silverman et al

1014 A catabolic state of metabolism can also be promoted by proinflammatory cytokines.21 Therefore, in addition to having a direct effect on metabolic processes, impaired glucocorticoid signaling can alter metabolic pathways by enhancing inflammatory signaling.5 In our study, both LPS-treated WT and GRdim mice showed a reduction in RQ, carbohydrate oxidation and TEE, whereas they displayed an elevation in fat oxidation, relative to saline-treated mice. A reduced RQ is indicative of an increased dependence on fat for energy metabolism in the face of reduced glucose utilization.79 Decreased food intake (anorexia), although not directly measured in this study, may contribute to reduced carbohydrate availability and the body weight loss observed in our LPS-treated mice. In addition, increased catabolic processes, especially in muscle (proteolysis), may lead to reduced lean mass, and hence body weight loss in LPS-treated mice.85 The reduction in TEE could be attributed to the observed impairment in locomotor activity. GRdim-LPS mice exhibited exaggerated and sustained metabolic alterations relative to WT-LPS mice, suggesting that GR dimerization-dependent DNA-binding mechanisms are required for various metabolic processes to occur under basal conditions and to be kept in check during an immune challenge. Thermoregulation It appears that the GR dimerization-dependent DNA-binding mechanisms that regulate basal metabolism are not required to maintain basal Tc, as the circadian rhythm and amplitude of Tc were similar in the GRdim and WT saline-treated mice. Systemic inflammation is usually accompanied by changes in body temperature, either fever or hypothermia.16 In addition to glucocorticoids’ role in mobilizing energy to facilitate thermogenesis, they also act as antipyretics by blocking the synthesis of pyrogenic cytokines and eicosanoids and by upregulating the transcription of anti-inflammatory molecules.36,37,86 Upon LPS administration, WT mice in our study exhibited an elevation in Tc and a constant skin temperature, whereas GRdim-LPS mice exhibited impaired thermoregulation, as evidenced by their inability to maintain a fever and display of reduced skin temperature relative to WT-LPS mice. In the face of impaired heat production in the GRdim-LPS mice, the observation of reduced skin temperature (because of increased shunting of blood from peripheral tissues to internal organs) and prolonged sickness behaviors, such as piloerection and locomotor impairment, may serve as compensatory mechanisms to reduce heat loss from the body’s core.15 Indeed, although impaired heat production was most apparent within the first 24 h after LPS injection, peak sickness behaviors and reductions in skin temperature occurred in the GRdim-LPS mice at 24 h. These data indicate that GR dimerization-dependent DNA-binding mechanisms are necessary for the proper regulation of inflammation-induced thermoregulatory processes. The observation that GRdim-LPS mice were impaired in their ability to maintain a fever, rather than exhibiting exaggerated fever responses, suggests that the role of GR dimerization-dependent DNA-binding mechanisms in mobilizing energy to facilitate thermogenesis is more prominent than its antipyretic effects. As fever is the result of a balance between pyrogenic and cryogenic mechanisms, the impaired fever observed in the GRdim-LPS mice may also be due to enhanced early expression of cryogenic factors, such as TNFa and IL-10.86,87 Moreover, under conditions of impaired gluconeogenesis, and hence deficient energy resources for thermogenesis, a transient state of hypothermia—rather than an energetically costly fever— may be an adaptive mechanism to conserve energy.19

CONCLUSION In summary, GRdim mice are more sensitive to the neuroendocrine, inflammatory, behavioral and metabolic effects of LPS, indicating Molecular Psychiatry (2013), 1006 – 1017

that GR dimerization-dependent DNA-binding mechanisms are essential for the proper regulation and recovery of multiple physiologic responses to a low-dose endotoxin challenge. These data also support the concept that GR dimerization-independent mechanisms are not sufficient for glucocorticoids to exert their full anti-inflammatory effects. As a consequence, endogenous glucocorticoid responses or exogenous glucocorticoid therapy limited to GR monomer-mediated transcriptional effects, such as protein– protein interactions, could predispose individuals to prolonged behavioral and metabolic sequelae of an enhanced inflammatory state. These include a continued state of sickness/depressive behaviors, such as lethargy/fatigue, reduced locomotor activity and anorexia and metabolic alterations favoring a sustained catabolic state (especially increased proteolysis in skeletal muscle required for increased gluconeogenesis in the liver), which can lead to muscle wasting, fat gain (cachectic obesity), insulin resistance (selective inhibition of fuel storage in liver, adipose tissue and muscle), osteopenia (reduced calcium absorption in bone) and anemia (reduced iron in blood) to provide sustained fuel to the activated immune system.13 Therefore, many features traditionally attributed to elevated glucocorticoids could actually be caused by impaired GR function, and hence glucocorticoid resistance and ensuing enhanced inflammation.5 Indeed, the use of antidepressants has been shown to attenuate multiple aspects of the acute-phase response, including LPS-induced CORT responses, sickness behaviors and hypothermia.88,89 These protective effects may be mediated by the antidepressant’s ability to increase GR expression/function90 and reduce proinflammatory cytokine expression/signaling.9,12 Genetic and environmental/physiological factors can contribute to impaired GR function. Although it may be unlikely that humans with depression have GR mutations that affect dimerization per se, GR variants have been described that confer reduced GR function. These variants have been associated with reduced dexamethasone suppression of CORT, increased vulnerability to inflammatory conditions/autoimmune diseases and increased risk of depression.91 Inflammation can also affect GR function. In addition to an enhanced inflammatory state being evident in the medically ill, whether because of infection or noninfectious chronic inflammatory disorders,7–10 elevated proinflammatory cytokine signaling has also been associated with chronic psychosocial stress.92–95 Proinflammatory cytokines have been shown to impair GR function via a variety of mechanisms: (1) reduced expression and/or affinity; (2) reduced translocation from the cytoplasm to the nucleus, which could be affected by its phosphorylation state (via mitogen-activated protein kinase pathways); (3) increased interaction with inflammatory-related transcription factors, such as nuclear factor-kB or activator protein-1 (jun/fos); and (4) increased interaction with the dominant-negative isoform, GRb.26,96 Because impaired GR function can contribute to enhanced inflammatory processes and enhanced inflammatory signaling can contribute to impaired GR signaling, a vicious feed-forward cycle perpetuating reduced inhibitory feedback on the HPA axis and inflammatory responses may ensue. Therefore, this bi-directional phenomenon has clinical relevance regarding increased susceptibility to the development of neuroendocrine (HPA axis)- and inflammatoryrelated disorders (including depression) and the development of resistance to glucocorticoid therapy in treating inflammatory conditions. With regard to chronic stress being a common risk factor for depression and morbidity, both animal and human studies support the role of glucocorticoid resistance as an underlying mechanism in psychosocial stress-associated enhancement of peripheral inflammation, thereby indicating the importance of the anti-inflammatory effects of GR.92,95,97 However, recent animal studies have shown that some psychological stress paradigms and stress levels of glucocorticoids increase central inflammatory responses to a subsequent immune challenge, and that this & 2013 Macmillan Publishers Limited

Glucocorticoid regulation of LPS-induced sickness MN Silverman et al

1015 enhanced inflammation is mediated by GR activation.98 The importance of the timing of the glucocorticoid surge relative to the immune challenge is supported by the findings that exogenous administration of stress levels of CORT before a systemic LPS challenge enhances proinflammatory cytokine expression and signaling in the brain (and periphery), whereas CORT administration after LPS challenge exerts anti-inflammatory effects.76,99 Consistent with this, the sensitizing effect of psychogenic stress on subsequent LPS-induced neuroinflammation has been shown to be mediated through GR-dependent priming of microglia (see Frank et al.100 and references therein). On the other hand, rats with endogenously higher basal CORT levels exhibit reduced LPS-induced GR upregulation concomitant with increased peripheral and CNS inflammatory responses and reduced CNS antiinflammatory responses relative to rats with lower basal CORT levels,101 suggesting that impaired GR function may contribute to enhanced inflammatory responses. Similarly, GRdim mice start off with higher basal CORT levels relative to their WT counterparts, and our data indicate that an intact GR is crucial for preventing the overexpression of early LPS-induced central proinflammatory cytokines and sickness behavior. Whereas the aforementioned studies only measured early (2–4 h) LPS-induced cytokine expression, our data are the first to suggest that an intact GR may be important to promote the induction of later central cytokine responses. It is possible that later TNFa responses may be neuroprotective by reducing inflammation and allowing for the recovery of sickness behavior. Therefore, both impaired and enhanced GR function may mediate stress and/or glucocorticoidinduced enhancement of cytokine responses in different contexts of glucocorticoid exposure (type of glucocorticoid, concentration, duration of exposure, timing relative to immune challenge), immune response type, activation state of cells, and tissue/cell specificity.97,98,102,103 Further studies are needed to identify the molecular mechanisms contributing to the exaggerated inflammation and sustained expression of sickness behavior and metabolic alterations observed in the GRdim-LPS mice. Putative factors include alterations in the central expression of anti-inflammatory molecules, neurotransmitter-related enzymes or metabolites, growth factors and CRH. In addition, future studies should investigate potential changes in metabolic enzymes in metabolically active tissue (gluconeogenic enzymes in the liver, lypolytic enzymes in adipose tissue and proteolytic enzymes in muscle), sympathetic nervous system activity, glucose and insulin levels/sensitivity and the expression/function of GR. By further characterizing the molecular mechanisms by which GR regulates inflammation and its behavioral and metabolic sequelae, our data and future studies can inform the pathophysiology of and potential risk factors for clinical conditions associated with HPA axis dysfunction and prolonged inflammation (for example, autoimmune disease, cardiovascular disease, type II diabetes, metabolic syndrome, osteoporosis, depression, chronic fatigue, fibromyalgia,) and may also explain the high comorbidity between these syndromes.5,7,12,13,104 Moreover, these data can better inform the development of glucocorticoid-related therapeutic agents for inflammatory disorders and their downstream correlates. Numerous selective glucocorticoid receptor agonists or dissociated GR ligands are being designed to favor the GR monomer conformation, thereby retaining the anti-inflammatory actions traditionally thought to be associated with GR monomer-mediated transrepression (via protein–protein interactions), and minimizing the metabolic side effects associated with GR dimerization-dependent transactivation (via GR–DNA binding).46,48,105,106 However, our findings show that GR dimerization-dependent DNA-binding mechanisms are important in keeping peripheral and central inflammation in check. Therefore, when designing dissociative glucocorticoids, one may need to consider the cost/benefit ratio between losing some anti-inflammatory potency (and hence gaining the behavioral & 2013 Macmillan Publishers Limited

consequences) and attenuating other adverse side effects, such as hyperglycemia. CONFLICT OF INTEREST The authors declare no conflict of interest.

ACKNOWLEDGEMENTS This work was supported by the NIH (intramural funds from the National Institute of Mental Health (NIMH) to EM Sternberg and from the National Institute on Alcohol Abuse and Alcoholism (NIAAA) to P Pacher). Additional support was received by the Henry M Jackson Foundation (Bethesda, MD, USA—MNS salary) and the Center for Neuroscience and Regenerative Medicine (Uniformed Services University of the Health Sciences, Bethesda, MD, USA). We thank Dr Gu¨nther Schu¨tz (German Cancer Research Centre, Heidelberg, Germany) for providing homozygous GRdim mouse breeder pairs to Taconic Farms (Rockville, MD, USA). We also thank Laura Tucker (Uniformed Services University of the Health Sciences, Center for Neuroscience and Regenerative Medicine, Mouse Behavioral Assessment Core, Bethesda, MD, USA) for instruction on the ANY-maze software used to analyze our open field data and Dr Sandor Batkai (Laboratory of Physiologic Studies, NIAAA, NIH, Bethesda, MD, USA) for setting up the metabolic chambers.

REFERENCES 1 Chrousos GP, Kino T. Glucocorticoid action networks and complex psychiatric and/or somatic disorders. Stress 2007; 10: 213–219. 2 Sapolsky RM, Romero LM, Munck AU. How do glucocorticoids influence stress responses? Integrating permissive, suppressive, stimulatory, and preparative actions. Endocr Rev 2000; 21: 55–89. 3 Holsboer F. The corticosteroid receptor hypothesis of depression. Neuropsychopharmacology 2000; 23: 477–501. 4 McIntyre RS, Soczynska JK, Konarski JZ, Woldeyohannes HO, Law CW, Miranda A et al. Should depressive syndromes be reclassified as ‘‘metabolic syndrome type II’’? Ann Clin Psychiatry 2007; 19: 257–264. 5 Raison CL, Miller AH. When not enough is too much: the role of insufficient glucocorticoid signaling in the pathophysiology of stress-related disorders. Am J Psychiatry 2003; 160: 1554–1565. 6 Zunszain PA, Anacker C, Cattaneo A, Carvalho LA, Pariante CM. Glucocorticoids cytokines and brain abnormalities in depression. Prog Neuropsychopharmacol Biol Psychiatry 2011; 35: 722–729. 7 Licinio J, Wong ML. The role of inflammatory mediators in the biology of major depression: central nervous system cytokines modulate the biological substrate of depressive symptoms, regulate stress-responsive systems, and contribute to neurotoxicity and neuroprotection. Mol Psychiatry 1999; 4: 317–327. 8 Maes M, Yirmyia R, Noraberg J, Brene S, Hibbeln J, Perini G et al. The inflammatory & neurodegenerative (I&ND) hypothesis of depression: leads for future research and new drug developments in depression. Metab Brain Dis 2009; 24: 27–53. 9 Pollak Y, Yirmiya R. Cytokine-induced changes in mood and behaviour: implications for ‘depression due to a general medical condition’, immunotherapy and antidepressive treatment. Int J Neuropsychopharmacol 2002; 5: 389–399. 10 Raison CL, Capuron L, Miller AH. Cytokines sing the blues: inflammation and the pathogenesis of depression. Trends Immunol 2006; 27: 24–31. 11 Dantzer R, O’Connor JC, Freund GG, Johnson RW, Kelley KW. From inflammation to sickness and depression: when the immune system subjugates the brain. Nat Rev Neurosci 2008; 9: 46–56. 12 Miller AH, Maletic V, Raison CL. Inflammation and its discontents: the role of cytokines in the pathophysiology of major depression. Biol Psychiatry 2009; 65: 732–741. 13 Straub RH, Cutolo M, Buttgereit F, Pongratz G. Energy regulation and neuroendocrine-immune control in chronic inflammatory diseases. J Intern Med 2010; 267: 543–560. 14 Dantzer R. Cytokine-induced sickness behaviour: a neuroimmune response to activation of innate immunity. Eur J Pharmacol 2004; 500: 399–411. 15 Hart BL. Biological basis of the behavior of sick animals. Neurosci Biobehav Rev 1988; 12: 123–137. 16 Romanovsky AA, Almeida MC, Aronoff DM, Ivanov AI, Konsman JP, Steiner AA et al. Fever and hypothermia in systemic inflammation: recent discoveries and revisions. Front Biosci 2005; 10: 2193–2216. 17 Silverman MN, Pearce BD, Biron CA, Miller AH. Immune modulation of the hypothalamic-pituitary-adrenal (HPA) axis during viral infection. Viral Immunol 2005; 18: 41–78.

Molecular Psychiatry (2013), 1006 – 1017

Glucocorticoid regulation of LPS-induced sickness MN Silverman et al

1016 18 Hasday JD, Fairchild KD, Shanholtz C. The role of fever in the infected host. Microbes Infect 2000; 2: 1891–1904. 19 Romanovsky AA, Szekely M. Fever and hypothermia: two adaptive thermoregulatory responses to systemic inflammation. Med Hypotheses 1998; 50: 219–226. 20 Beisel WR. Metabolic response to infection. Annu Rev Med 1975; 26: 9–20. 21 Chang HR, Bistrian B. The role of cytokines in the catabolic consequences of infection and injury. JPEN J Parenter Enteral Nutr 1998; 22: 156–166. 22 Schobitz B, Reul JM, Holsboer F. The role of the hypothalamic-pituitaryadrenocortical system during inflammatory conditions. Crit Rev Neurobiol 1994; 8: 263–291. 23 Besedovsky HO, del Rey A. Immune-neuro-endocrine interactions: facts and hypotheses. Endocr Rev 1996; 17: 64–102. 24 Glezer I, Rivest S. Glucocorticoids: protectors of the brain during innate immune responses. Neuroscientist 2004; 10: 538–552. 25 McEwen BS, Biron CA, Brunson KW, Bulloch K, Chambers WH, Dhabhar FS et al. The role of adrenocorticoids as modulators of immune function in health and disease: neural, endocrine and immune interactions. Brain Res Brain Res Rev 1997; 23: 79–133. 26 Silverman MN, Sternberg EM. Neuroendocrine-immune interactions in rheumatoid arthritis: mechanisms of glucocorticoid resistance. Neuroimmunomodulation 2008; 15: 19–28. 27 Webster JI, Sternberg EM. Role of the hypothalamic-pituitary-adrenal axis, glucocorticoids and glucocorticoid receptors in toxic sequelae of exposure to bacterial and viral products. J Endocrinol 2004; 181: 207–221. 28 Goujon E, Parnet P, Laye S, Combe C, Dantzer R. Adrenalectomy enhances pro-inflammatory cytokines gene expression, in the spleen, pituitary and brain of mice in response to lipopolysaccharide. Brain Res Mol Brain Res 1996; 36: 53–62. 29 Nadeau S, Rivest S. Glucocorticoids play a fundamental role in protecting the brain during innate immune response. J Neurosci 2003; 23: 5536–5544. 30 Silverman MN, Macdougall MG, Hu F, Pace TW, Raison CL, Miller AH. Endogenous glucocorticoids protect against TNF-alpha-induced increases in anxiety-like behavior in virally infected mice. Mol Psychiatry 2007; 12: 408–417. 31 Sternberg EM, Hill JM, Chrousos GP, Kamilaris T, Listwak SJ, Gold PW et al. Inflammatory mediator-induced hypothalamic-pituitary-adrenal axis activation is defective in streptococcal cell wall arthritis-susceptible Lewis rats. Proc Natl Acad Sci USA 1989; 86: 2374–2378. 32 Goujon E, Parnet P, Aubert A, Goodall G, Dantzer R. Corticosterone regulates behavioral effects of lipopolysaccharide and interleukin-1 beta in mice. Am J Physiol 1995; 269(1 Part 2): R154–159. 33 Johnson RW, Propes MJ, Shavit Y. Corticosterone modulates behavioral and metabolic effects of lipopolysaccharide. Am J Physiol 1996; 270(1 Part 2): R192–198. 34 Pezeshki G, Pohl T, Schobitz B. Corticosterone controls interleukin-1 beta expression and sickness behavior in the rat. J Neuroendocrinol 1996; 8: 129–135. 35 Wang D, Lin W, Pan Y, Kuang X, Qi X, Sun H. Chronic blockade of glucocorticoid receptors by RU486 enhances lipopolysaccharide-induced depressive-like behaviour and cytokine production in rats. Brain Behav Immun 2011; 25: 706–714. 36 Coelho MM, Luheshi G, Hopkins SJ, Pela IR, Rothwell NJ. Multiple mechanisms mediate antipyretic action of glucocorticoids. Am J Physiol 1995; 269(3 Part 2): R527–535. 37 Morrow LE, McClellan JL, Conn CA, Kluger MJ. Glucocorticoids alter fever and IL-6 responses to psychological stress and to lipopolysaccharide. Am J Physiol 1993; 264(5 Part 2): R1010–1016. 38 Kapcala LP, Chautard T, Eskay RL. The protective role of the hypothalamicpituitary-adrenal axis against lethality produced by immune, infectious, and inflammatory stress. Ann NY Acad Sci 1995; 771: 419–437. 39 Ruzek MC, Pearce BD, Miller AH, Biron CA. Endogenous glucocorticoids protect against cytokine-mediated lethality during viral infection. J Immunol 1999; 162: 3527–3533. 40 Capuron L, Fornwalt FB, Knight BT, Harvey PD, Ninan PT, Miller AH. Does cytokine-induced depression differ from idiopathic major depression in medically healthy individuals? J Affect Disord 2009; 119: 181–185. 41 Capuron L, Su S, Miller AH, Bremner JD, Goldberg J, Vogt GJ et al. Depressive symptoms and metabolic syndrome: is inflammation the underlying link? Biol Psychiatry 2008; 64: 896–900. 42 Barnes PJ. Corticosteroid effects on cell signalling. Eur Respir J 2006; 27: 413–426. 43 Chinenov Y, Rogatsky I. Glucocorticoids and the innate immune system: crosstalk with the toll-like receptor signaling network. Mol Cell Endocrinol 2007; 275: 30–42. 44 De Bosscher K, Vanden Berghe W, Haegeman G. Cross-talk between nuclear receptors and nuclear factor kappaB. Oncogene 2006; 25: 6868–6886. 45 Dostert A, Heinzel T. Negative glucocorticoid receptor response elements and their role in glucocorticoid action. Curr Pharm Des 2004; 10: 2807–2816.

Molecular Psychiatry (2013), 1006 – 1017

46 Newton R, Holden NS. Separating transrepression and transactivation: a distressing divorce for the glucocorticoid receptor? Mol Pharmacol 2007; 72: 799–809. 47 Clark AR. Anti-inflammatory functions of glucocorticoid-induced genes. Mol Cell Endocrinol 2007; 275: 79–97. 48 Schacke H, Docke WD, Asadullah K. Mechanisms involved in the side effects of glucocorticoids. Pharmacol Ther 2002; 96: 23–43. 49 Reichardt HM, Kaestner KH, Tuckermann J, Kretz O, Wessely O, Bock R et al. DNA binding of the glucocorticoid receptor is not essential for survival. Cell 1998; 93: 531–541. 50 Kleiman A, Hubner S, Rodriguez Parkitna JM, Neumann A, Hofer S, Weigand MA et al. Glucocorticoid receptor dimerization is required for survival in septic shock via suppression of interleukin-1 in macrophages. FASEB J 2012; 26: 722–729. 51 Andre C, O’Connor JC, Kelley KW, Lestage J, Dantzer R, Castanon N. Spatiotemporal differences in the profile of murine brain expression of proinflammatory cytokines and indoleamine 2,3-dioxygenase in response to peripheral lipopolysaccharide administration. J Neuroimmunol 2008; 200: 90–99. 52 Frenois F, Moreau M, O’Connor J, Lawson M, Micon C, Lestage J et al. Lipopolysaccharide induces delayed FosB/DeltaFosB immunostaining within the mouse extended amygdala, hippocampus and hypothalamus, that parallel the expression of depressive-like behavior. Psychoneuroendocrinology 2007; 32: 516–531. 53 Oitzl MS, Reichardt HM, Joels M, de Kloet ER. Point mutation in the mouse glucocorticoid receptor preventing DNA binding impairs spatial memory. Proc Natl Acad Sci USA 2001; 98: 12790–12795. 54 Reichardt HM, Tuckermann JP, Gottlicher M, Vujic M, Weih F, Angel P et al. Repression of inflammatory responses in the absence of DNA binding by the glucocorticoid receptor. EMBO J 2001; 20: 7168–7173. 55 Silverman MN, Miller AH, Biron CA, Pearce BD. Characterization of an interleukin6- and adrenocorticotropin-dependent, immune-to-adrenal pathway during viral infection. Endocrinology 2004; 145: 3580–3589. 56 Grose R, Werner S, Kessler D, Tuckermann J, Huggel K, Durka S et al. A role for endogenous glucocorticoids in wound repair. EMBO Rep 2002; 3: 575–582. 57 Tuckermann JP, Kleiman A, Moriggl R, Spanbroek R, Neumann A, Illing A et al. Macrophages and neutrophils are the targets for immune suppression by glucocorticoids in contact allergy. J Clin Invest 2007; 117: 1381–1390. 58 Rivest S, Lacroix S, Vallieres L, Nadeau S, Zhang J, Laflamme N. How the blood talks to the brain parenchyma and the paraventricular nucleus of the hypothalamus during systemic inflammatory and infectious stimuli. Proc Soc Exp Biol Med 2000; 223: 22–38. 59 Marchetti L, Klein M, Schlett K, Pfizenmaier K, Eisel UL. Tumor necrosis factor (TNF)-mediated neuroprotection against glutamate-induced excitotoxicity is enhanced by N-methyl-D-aspartate receptor activation. Essential role of a TNF receptor 2-mediated phosphatidylinositol 3-kinase-dependent NF-kappa B pathway. J Biol Chem 2004; 279: 32869–32881. 60 Veroni C, Gabriele L, Canini I, Castiello L, Coccia E, Remoli ME et al. Activation of TNF receptor 2 in microglia promotes induction of anti-inflammatory pathways. Mol Cell Neurosci 2010; 45: 234–244. 61 Lefstin JA, Yamamoto KR. Allosteric effects of DNA on transcriptional regulators. Nature 1998; 392: 885–888. 62 Miller AH, Spencer RL, Pearce BD, Pisell TL, Azrieli Y, Tanapat P et al. Glucocorticoid receptors are differentially expressed in the cells and tissues of the immune system. Cell Immunol 1998; 186: 45–54. 63 Ray A, LaForge KS, Sehgal PB. On the mechanism for efficient repression of the interleukin-6 promoter by glucocorticoids: enhancer, TATA box, and RNA start site (Inr motif) occlusion. Mol Cell Biol 1990; 10: 5736–5746. 64 Zhang G, Zhang L, Duff GW. A negative regulatory region containing a glucocorticosteroid response element (nGRE) in the human interleukin-1beta gene. DNA Cell Biol 1997; 16: 145–152. 65 Rhoades KL, Golub SH, Economou JS. The regulation of the human tumor necrosis factor alpha promoter region in macrophage, T cell, and B cell lines. J Biol Chem 1992; 267: 22102–22107. 66 Amano Y, Lee SW, Allison AC. Inhibition by glucocorticoids of the formation of interleukin-1 alpha, interleukin-1 beta, and interleukin-6: mediation by decreased mRNA stability. Mol Pharmacol 1993; 43: 176–182. 67 Smoak K, Cidlowski JA. Glucocorticoids regulate tristetraprolin synthesis and posttranscriptionally regulate tumor necrosis factor alpha inflammatory signaling. Mol Cell Biol 2006; 26: 9126–9135. 68 Eskdale J, Kube D, Tesch H, Gallagher G. Mapping of the human IL10 gene and further characterization of the 50 flanking sequence. Immunogenetics 1997; 46: 120–128. 69 Franchimont D, Martens H, Hagelstein MT, Louis E, Dewe W, Chrousos GP et al. Tumor necrosis factor alpha decreases, and interleukin-10 increases, the sensitivity of human monocytes to dexamethasone: potential regulation of the glucocorticoid receptor. J Clin Endocrinol Metab 1999; 84: 2834–2839.

& 2013 Macmillan Publishers Limited

Glucocorticoid regulation of LPS-induced sickness MN Silverman et al

1017 70 Wong ML, Bongiorno PB, Rettori V, McCann SM, Licinio J. Interleukin (IL) 1beta, IL-1 receptor antagonist, IL-10, and IL-13 gene expression in the central nervous system and anterior pituitary during systemic inflammation: pathophysiological implications. Proc Natl Acad Sci USA 1997; 94: 227–232. 71 Goujon E, Laye S, Parnet P, Dantzer R. Regulation of cytokine gene expression in the central nervous system by glucocorticoids: mechanisms and functional consequences. Psychoneuroendocrinology 1997; 22(Suppl 1): S75–80. 72 Quan N, He L, Lai W, Shen T, Herkenham M. Induction of IkappaBalpha mRNA expression in the brain by glucocorticoids: a negative feedback mechanism for immune-to-brain signaling. J Neurosci 2000; 20: 6473–6477. 73 Laflamme N, Echchannaoui H, Landmann R, Rivest S. Cooperation between tolllike receptor 2 and 4 in the brain of mice challenged with cell wall components derived from gram-negative and gram-positive bacteria. Eur J Immunol 2003; 33: 1127–1138. 74 Glezer I, Simard AR, Rivest S. Neuroprotective role of the innate immune system by microglia. Neuroscience 2007; 147: 867–883. 75 Hanisch UK, Johnson TV, Kipnis J. Toll-like receptors: roles in neuroprotection? Trends Neurosci 2008; 31: 176–182. 76 Frank MG, Miguel ZD, Watkins LR, Maier SF. Prior exposure to glucocorticoids sensitizes the neuroinflammatory and peripheral inflammatory responses to E. coli lipopolysaccharide. Brain Behav Immun 2010; 24: 19–30. 77 Hermoso MA, Matsuguchi T, Smoak K, Cidlowski JA. Glucocorticoids and tumor necrosis factor alpha cooperatively regulate toll-like receptor 2 gene expression. Mol Cell Biol 2004; 24: 4743–4756. 78 Homma T, Kato A, Hashimoto N, Batchelor J, Yoshikawa M, Imai S et al. Corticosteroid and cytokines synergistically enhance toll-like receptor 2 expression in respiratory epithelial cells. Am J Respir Cell Mol Biol 2004; 31: 463–469. 79 Hamelink CR, Currie PJ, Chambers JW, Castonguay TW, Coscina DV. Corticosterone-responsive and -unresponsive metabolic characteristics of adrenalectomized rats. Am J Physiol 1994; 267(3 Part 2): R799–804. 80 Linthorst AC, Karanth S, Barden N, Holsboer F, Reul MH. Impaired glucocorticoid receptor function evolves in aberrant physiological responses to bacterial endotoxin. Eur J Neurosci 1999; 11: 178–186. 81 Richard D, Chapdelaine S, Deshaies Y, Pepin MC, Barden N. Energy balance and lipid metabolism in transgenic mice bearing an antisense GCR gene construct. Am J Physiol 1993; 265(1 Part 2): R146–150. 82 Reul JM, Labeur MS, Wiegers GJ, Linthorst AC. Altered neuroimmunoendocrine communication during a condition of chronically increased brain corticotropinreleasing hormone drive. Ann NY Acad Sci 1998; 840: 444–455. 83 Sternberg EM, Glowa JR, Smith MA, Calogero AE, Listwak SJ, Aksentijevich S et al. Corticotropin releasing hormone related behavioral and neuroendocrine responses to stress in Lewis and Fischer rats. Brain Res 1992; 570: 54–60. 84 Frijters R, Fleuren W, Toonen EJ, Tuckermann JP, Reichardt HM, van der Maaden H et al. Prednisolone-induced differential gene expression in mouse liver carrying wild type or a dimerization-defective glucocorticoid receptor. BMC Genomics 2010; 11: 359. 85 Ogimoto K, Harris Jr. MK, Wisse BE. MyD88 is a key mediator of anorexia, but not weight loss, induced by lipopolysaccharide and interleukin-1 beta. Endocrinology 2006; 147: 4445–4453. 86 Roth J. Endogenous antipyretics. Clin Chim Acta 2006; 371: 13–24. 87 Steiner AA, Romanovsky AA. Leptin: at the crossroads of energy balance and systemic inflammation. Prog Lipid Res 2007; 46: 89–107.

& 2013 Macmillan Publishers Limited

88 Yirmiya R. Endotoxin produces a depressive-like episode in rats. Brain Res 1996; 711: 163–174. 89 Yirmiya R, Pollak Y, Barak O, Avitsur R, Ovadia H, Bette M et al. Effects of antidepressant drugs on the behavioral and physiological responses to lipopolysaccharide (LPS) in rodents. Neuropsychopharmacology 2001; 24: 531–544. 90 Anacker C, Zunszain PA, Carvalho LA, Pariante CM. The glucocorticoid receptor: pivot of depression and of antidepressant treatment? Psychoneuroendocrinology 2011; 36: 415–425. 91 Manenschijn L, van den Akker EL, Lamberts SW, van Rossum EF. Clinical features associated with glucocorticoid receptor polymorphisms. An overview. Ann NY Acad Sci 2009; 1179: 179–198. 92 Avitsur R, Powell N, Padgett DA, Sheridan JF. Social interactions, stress, and immunity. Immunol Allergy Clin North Am 2009; 29: 285–293. 93 Black PH. The inflammatory response is an integral part of the stress response: implications for atherosclerosis, insulin resistance, type II diabetes and metabolic syndrome X. Brain Behav Immun 2003; 17: 350–364. 94 Kiecolt-Glaser JK, Glaser R. Depression and immune function: central pathways to morbidity and mortality. J Psychosom Res 2002; 53: 873–876. 95 Miller GE, Chen E, Sze J, Marin T, Arevalo JM, Doll R et al. A functional genomic fingerprint of chronic stress in humans: blunted glucocorticoid and increased NF-kappaB signaling. Biol Psychiatry 2008; 64: 266–272. 96 Pace TW, Hu F, Miller AH. Cytokine-effects on glucocorticoid receptor function: relevance to glucocorticoid resistance and the pathophysiology and treatment of major depression. Brain Behav Immun 2007; 21: 9–19. 97 Rohleder N. Acute and chronic stress induced changes in sensitivity of peripheral inflammatory pathways to the signals of multiple stress systems—2011 Curt Richter Award Winner. Psychoneuroendocrinology 2012; 37: 307–316. 98 Sorrells SF, Sapolsky RM. An inflammatory review of glucocorticoid actions in the CNS. Brain Behav Immun 2007; 21: 259–272. 99 Munhoz CD, Sorrells SF, Caso JR, Scavone C, Sapolsky RM. Glucocorticoids exacerbate lipopolysaccharide-induced signaling in the frontal cortex and hippocampus in a dose-dependent manner. J Neurosci 2010; 30: 13690–13698. 100 Frank MG, Thompson BM, Watkins LR, Maier SF. Glucocorticoids mediate stressinduced priming of microglial pro-inflammatory responses. Brain Behav Immun 2012; 26: 337–345. 101 Perez-Nievas BG, Madrigal JL, Garcia-Bueno B, Zoppi S, Leza JC. Corticosterone basal levels and vulnerability to LPS-induced neuroinflammation in the rat brain. Brain Res 2010; 1315: 159–168. 102 Dhabhar FS. Enhancing versus suppressive effects of stress on immune function: implications for immunoprotection and immunopathology. Neuroimmunomodulation 2009; 16: 300–317. 103 Galon J, Franchimont D, Hiroi N, Frey G, Boettner A, Ehrhart-Bornstein M et al. Gene profiling reveals unknown enhancing and suppressive actions of glucocorticoids on immune cells. FASEB J 2002; 16: 61–71. 104 Silverman MN, Heim CM, Nater UM, Marques AH, Sternberg EM. Neuroendocrine and immune contributors to fatigue. PMR 2010; 2: 338–346. 105 De Bosscher K, Van Craenenbroeck K, Meijer OC, Haegeman G. Selective transrepression versus transactivation mechanisms by glucocorticoid receptor modulators in stress and immune systems. Eur J Pharmacol 2008; 583: 290–302. 106 Kleiman A, Tuckermann JP. Glucocorticoid receptor action in beneficial and side effects of steroid therapy: lessons from conditional knockout mice. Mol Cell Endocrinol 2007; 275: 98–108.

Molecular Psychiatry (2013), 1006 – 1017