Printer Friendly

Reduction of Glucocorticoid Receptor Function in Chronic Fatigue Syndrome.

1. Introduction

Chronic fatigue syndrome (CFS) has a prevalence of 2% in the UK [1, 2]. It is defined by profound, persistent, medically unexplained fatigue lasting at least 6 months, which is not caused by ongoing exertion, not significantly eased by rest, and is severe enough to cause considerable loss of function [3-5]. Alongside this are symptoms of inflammation, pain, cognitive deficits, and psychiatric and bowel problems [4]. Often, biological tests and physical examinations are unremarkable. CFS affects all ages and the peak age of onset is 20-40. Full recovery is rare [2, 6] and comorbidity with depression is common.

Many putative causes of CFS have been investigated but the absence of an agreed pathogenesis impacts the development of effective diagnostics and treatments. It is likely that multiple factors contribute which involve a number of interacting biological, environmental, and psychosocial factors [1,2, 5, 7-11].

The recognised temporal relationship between stressors and the onset and course of CFS suggests an aetiopathogenic role for systems controlling the stress response including the sympathetic nervous system and the hypothalamic-pituitary-adrenal (HPA) axis [4, 8, 12-17]. The functionality of the glucocorticoid receptor (GR), determined by its sensitivity, affinity, and density and by its interaction with transcription factors [18], is arguably the defining factor in HPA axis regulation [19] and responsible, in a large part, for basal concentrations of cortisol throughout the day. In HPA axis downregulation in response to the GR agonist, dexamethasone (dex) is the most commonly used in vivo method of determining GR function. An ex vivo technique dependent on the inhibitory effect of GR activation on cytokine release is also utilized [20, 21]. Cross-sectional studies in CFS tend to show basal hypocortisolaemia [22, 23], attenuated diurnal variation [4, 24], an attenuated response to activation by CRH or ACTH [16, 23, 25-28], an enhanced suppression by dex [10, 17, 28-30], and an enhanced dex-induced suppression of IL-6, TNF[alpha], IL-10, and IL-4 synthesis [21, 31, 32] and of peripheral blood mononuclear cell proliferation [18]. Genetic studies in CFS have shown the salience of functional single nuclear polymorphisms in NR3C1 [31, 32], which codes for GR, and have also shown hypomethylation of the 1F promotor region of this gene [33-35]. The endocrine findings however are not consistent, and elevated [36] or normal [18, 27, 25, 37] basal cortisol levels have also been reported, as has a normal cortisol response to wakening [25] and to psychosocial stress [21, 38]. There is also some evidence of an association of HPA axis dysregulation with symptom severity and poorer prognosis [15, 39] but only speculation about the mechanism through which HPA axis abnormalities may result in the symptoms of CFS; glucose supply [40], hypotension with associated reduced cerebral perfusion [41], and CRH-induced appetite and sleep disturbance have been considered [16].

The HPA axis interacts with many other systems, notably the immune system [42]. Glucocorticoids (GCs) modulate immune responses by altering gene expression, transcription, translation, and protein secretion [42], either directly, by decreasing transcription of the genes which code for cytokines, or indirectly, by inhibiting proinflammatory transcription factors [42]. GCs inhibit, with different sensitivities, cytokines such as IL-6, IL-1, and TNF (TNF the most, IL-6 the least) [42, 43]. Immune activation, with an increase in proinflammatory cytokine concentrations, including IL-6, TNF[alpha], and IL-1 [2, 44, 45], has been rather inconsistently, for example [32], demonstrated in CFS, may be secondary to insufficient glucocorticoid signaling [7], and may result in pain, fatigue, cognitive deficits, and other symptoms which are characteristic of CFS [12, 16, 38, 46].

In order to examine the nature, extent, and impact of HPA axis dysregulation in CFS, we sought to compare GR function using both in vivo and ex vivo assessment. We therefore examined the HPA axis and immune system function in a sample of patients with CFS and in healthy comparators and in participants with the systemic autoimmune condition, primary Sjogren's syndrome (pSS), who acted as disease group comparators.

2. Methods

2.1. Participants. Three groups were recruited. The study was carried out in accordance with the Declaration of Helsinki. The study design was approved by the Newcastle and North Tyneside Ethics Committee. All participants provided written informed consent. Participants were aged 22-68 years old. Exclusion criteria consisted of age < 18 years, a current or past axis I psychiatric diagnosis confirmed using the Structured Clinical Interview for DSM-IV [47, 48], and use, in the 72 hours prior to enrolment, of antihypertensives, antidepressants, or analgesics. Samples were collected as part of an MRC-funded cohort study (MRC MR/J002712/1). 48 participants with CFS (13 males (mean age 52.2) and 35 females (mean age = 44.9)) were recruited via the local CFS clinical service, all fulfilled the Fukuda diagnostic criteria, and had a mean FIS of 88 and CTQ of 32. Twenty healthy comparators (HC; 7 males (mean age = 43.1) and 13 females (mean age 44.9)) were recruited from a HC database, word of mouth, social media, and advertisement in the hospital (mean FIS = 4, CTQ = 29). HCs were age and sex matched to the patients, and attempts were made to match on activity levels using the Mean International Physical Activity Questionnaire (IPAQ) although the mean CFS IPAQ rating was "low" and the HC rating was "medium." Primary Sjogren's syndrome (pSS) patients (n = 27) fulfilled the American European Consensus Group classification [49] and were recruited from the United Kingdom pSS Registry [50].

2.2. Symptom Assessment Tools. The CFS participants completed the Fatigue Impact Scale (FIS) [51] and the Childhood Trauma Questionnaire-Short Form (CTQ-SF) [52]. The FIS quantifies individual perception of the impact that fatigue has on daily functioning (Fisk et al.) [53]. There are 40 items, each scored on a 5-point Likert scale providing a continuous scale of 0-160. It comprises three subscales looking at the impact that fatigue has on physical (10 items: motivation, effort, stamina, and coordination), psychosocial (20 items: isolation, emotions, coping, and workload), and cognitive (10 items: concentration, memory, and thinking) functioning. A higher score indicates greater fatigue. The CTQ is a 28 item self-report scale which measures the frequency and severity of childhood adversity. It consists of five factors: emotional abuse, physical abuse, sexual abuse, emotional neglect, and physical neglect, and possesses good psychometric properties [52]. Items are scored on a Likert scale, with responses ranging from 1 ("never true") to 5 ("very often true").

2.3. In Vivo Assessment of HPA Axis Function by Measurement of Cortisol Levels in Response to Low-Dose dex in CFS Compared to Healthy Controls. In the 48 CFS patients and the first 10 healthy controls (HCs), plasma samples were taken in lithium-heparin vacutainers at 30-minute intervals between 10 am and noon on two consecutive days (day 1 and day 2). At 11pm on day 1, participants took oral dex (0.5 mg). Practical consideration meant that only the baseline blood (10 am day 1 sample) were taken for the other 10 HCs. Within one hour of collection, the blood was spun at 1600 g for 10 minutes at room temperature. Aliquots of plasma were extracted and stored at -80[degrees]C until analysis. Plasma cortisol concentrations were quantified using 15 lot-matched cortisol ELISA kits, supplied by Abcam and used according to the manufacturer's protocol. The lower limit of cortisol detection was 2.44 ng/ml.

2.4. Ex Vivo Measurement of GR Function: The Glucocorticoid Receptor Response Assay. The GR response assay [20] utilised lithium-heparin-treated blood taken at 10 am on day 1. It was set up in sterile 48-well plates within 3 hours of collection. Blood was diluted 1/10 with room temperature RPMI 1640 containing penicillin-streptomycin and L-glutamine, mixed thoroughly by inversion, and then incubated for 24 hours at 37[degrees]C in a humidified atmosphere of 95% air and 5% C[O.sub.2]. There were four conditions, "null" (medium alone), "LPS" in which 50 [micro]l of a 200 /g/ml LPS solution was added to the 400 [micro]l of diluted blood (to stimulate cytokine production from cells), "dex10" in which 50 [micro]l of a 100 nanomolar (nM) dex solution was also added, and "dex100" in which a tenfold stronger dex solution was added. After incubation, the assay plates were spun in a 4[degrees]C centrifuge for 10 minutes at 1000 RPM. 300 [micro]l supernatant samples were then harvested for each condition, transferred to 0.6 ml microcentrifuge tubes, and stored at -20[degrees]C. Cytokine concentrations (TNF[alpha], IL-6, and IL-10) were determined using Cytometric Bead Array (CBA) and flow cytometry according to the manufacturer's instruction (BD Bioscience). Percentage suppression on dex 10 nM (% dex10) and on dex 100 nM (% dex100) was calculated using the following equation:

[mathematical expression not reproducible] (1)

2.5. Baseline Cytokine Levels. Gel-based specimen tubes were used to collect a further serum sample at 10 am on day 1 for baseline measurement of a range of inflammatory markers. These were spun within 3 hours at 1600 RPM for 10 minutes at room temperature. 2 x 1 ml of serum was extracted and stored at -80[degrees]C until analysis which utilized the method described above, that is, CBA and flow cytometry. Two different dilutions of serum samples were required for the measurement of different cytokines; this was therefore conducted in 2 batches for CFS (n = 45), HC (n = 19), and pSS (n = 9) participants.

2.6. Data and Statistical Analysis. Statistical tests were carried out using SPSS version 23 and the "R" statistical package. Graphs were produced using GraphPad Prism version 5.01 and MATLAB. p values are two-tailed with significance set at p < 0.05.

Repeated measures ANOVA was conducted with time (5) as within and group (2) as between factors to examine cortisol concentrations on each of the two days. Area under the curve (AUC) was also calculated using trapezoid integration, both for day 1 and day 2. Specifically, AUC with respect to ground (AUCg), considered to be a measure of overall cortisol output including baseline activity, and AUC increase (AUCi), a putative measure of the sensitivity of GR to modulation, were calculated [54]. The difference between AUCg on the 2 days was also calculated (delta AUCg). Shapiro-Wilk and QQ plots (data not shown) revealed that even after Box-Cox transformation, neither AUC nor cytokine data met the assumptions required for ANOVA; thus, nonparametric comparisons were used. Spearman correlations were conducted to examine the relationship between childhood adversity and endocrine parameters. All data, shown or not shown, is available for scrutiny upon request.

3. Results

3.1. In Vivo Assessment of HPA Axis Function by Measurement of Cortisol Levels in Response to Low-Dose dex in CFS Compared to Healthy Controls. There was a significant effect of time on day one (F = 16.61, df = 4,56, p < 0.0005) but not on day two (F = 0.87, df = 4,56, p = 0.418). No effect of group on either day (day one, F = 0.73, df = 1,56, p = 0.398; day 2, F = 0.79, df = 1,56, p = 0.378) (see Figure 1). Cortisol AUCs (g or i, day one, day two, or delta) did not differ (p > 0.2).

3.2. Ex Vivo Assessment of HPA Axis Function: The Glucocorticoid Receptor (GR) Response Assay. GR response assay blood was not taken for two CFS patients. One HC was removed due to an abnormally high null value and one CFS patient removed due to not stimulating sufficiently on LPS. Analysis was therefore conducted on samples taken from CFS (n = 40), HC (n = 19), and pSS (n = 27) participants.

In the null sample, pSS participants had higher IL-10 levels than the CFS or HC participants but there were no differences between groups for IL-6 or TNF[alpha]. LPS induced a robust cytokine response, and after LPS, group differences were evident such that, for the positive cytokines IL-6 and TNF[alpha], pSS participants had higher levels than CFS participants, who had higher levels than HCs. For IL-10, the difference was between pSS (higher) and HCs. Median cytokine levels were, on the whole, lower in the dex 10 nM samples and invariably in the dex 100 nM samples than the LPS alone samples. Percentage suppression was greater for samples incubated with dex 100 nM than those incubated with dex 10 nM. Percentage suppression with dex 10 nM or dex 100 nM was not different between the groups, except for a greater suppression in the dex 100 nM condition determined using TNF[alpha] (see Table 1).

Individual AUC and cytokine values were visualized using frequency density plots. Visual inspection suggested greater variability in patients compared to HCs with some patients showing less suppression (see Figure 2). In order to investigate this, we performed post hoc [epsilon]-tests of standard deviations derived from Bayesian hierarchical models of outcome measures under Gaussian (normal) priors. The advantage of this approach lies in its treatment of parameters as sampling variables rather than population attributes which allows us to formally compare modelled estimates of their values. Significance was assessed by comparing the 95% credible intervals of the posterior distributions. Vague priors were used for the mean and standard deviation, and analysis was carried out using the BEST package in the R statistical environment [55]. Results are displayed in Table 1. There was no difference in standard deviation of day 1 AUCg distributions. There was a significant difference in standard deviations for day 1 AUCi (p =.014), day 2 AUCg (p <.0005), and day 2 AUCi (p <.0005). There was also a significant difference in standard deviations of TNF distributions (p =.001). The difference in IL-6 was marginal (p =.054) though there was no significant difference in IL-10 distributions (p =.774).

3.3. The Relationship between Reported Adversity and HPA Axis Function in CFS and Controls. Correlation coefficients, in CFS participants, reveal a negative relationship between the CTQ score (and the emotional subscores) and cortisol AUC but no significant relationship with % suppression (after incubation with 10 nm dex). In HCs, there was no significant relationship between CTQ scores and cortisol AUC, but there was a positive relationship with percentage IL-6 suppression (see Table 2). In the combined sample, of CFS and HC (see Table 3), the significant correlations were between the CTQ total score and AUCg (positive) and IL10 (negative).

3.4. Baseline Cytokine Levels. CFS patients showed reduced production of IP-10 and IL-12/23p40 compared to HC and of IP-10, MIP1a, IL-6, and IL-1[beta] compared to pSS participants. pSS participants showed increased production of MIP1 a and IL-6 compared to HV. A violin plot was designed post hoc using "R" statistical software to visualize variance between populations and determine whether subpopulations were present and is displayed in Figure 3.

4. Discussion

We did not demonstrate a difference in cortisol levels between participants with CFS and healthy volunteers. This differs from the majority of cross-sectional HPA axis studies in this population (see "Introduction" and [10] for review). This difference may be related to the population; the sample here, for instance, was rigorously screened for comorbid depression, and Papadopoulos et al. [36] have previously demonstrated that dex-induced cortisol suppression differed only in CFS patients with comorbid depression or it may be a type II error consequent on the small sample size combined with the marked variation in cortisol levels in CFS as highlighted by the frequency density graph and the significantly greater cortisol variability in patients with CFS. The aetiopathogenic relevance of this variability is unknown but it suggests a lack of precision in cortisol regulation [56]. The heterogeneity in cortisol concentrations may suggest clinical heterogeneity within the diagnostic grouping of CFS and emphasizes the impact of disparate and competing factors on GR function including current and previous stressors, the common use of antidepressants [20, 54] (even in those who have never met criteria for major depressive disorder) [20, 57], and the impact of a primary dysregulation of proinflammatory cytokines [58].

The baseline cytokine data emphasized the status of pSS as an inflammatory disorder. The ex vivo data revealed a reduced capacity for a proinflammatory cytokine response to LPS in CFS compared with HCs (and an increased responsivity compared to the pSS participants). It further revealed that (independent of group) incubation with dex, in a dose-dependent manner, as expected, suppressed cytokine release. The percent suppression of LPS-induced TNF[alpha] release by 100 nM solution of dex was less in CFS patients than HCs. This may be suggestive of reduced GR function in CFS but any such interpretation must be made with caution as the impact of 10 nM dex did not significantly differ between CFS and HCs, neither was a significant effect seen when IL-6 or IL-10 was used as the output variable. That TNF[alpha] was most sensitive to suppression by dex accords with the existing literature [42], is congruent with the theory that GCs may preferentially inhibit Th1 over Th2 cells [59], and suggests that TNF[alpha] may be the most appropriate cytokine for GR response assay studies in CFS. The variability in percentage suppression of cytokine levels by dex is also greater in CFS than healthy or pSS controls.

In CFS participants, there was a relationship between the score on the childhood trauma questionnaire and cortisol AUCs such that higher reported levels of early adversity correlated negatively with cortisol. Interestingly, a different pattern was seen in healthy volunteers in whom reported childhood adversity associated positively with dex-induced IL-6 suppression in the absence of an effect of cortisol concentrations. When the groups were combined to maximise power, a negative relationship between reported adversity and cortisol levels and a positive relationship with GR function (here shown using IL-10 not IL-6) are revealed.

The variability in cortisol has implications for the interpretation of existing and future endocrine cohort studies in CFS because of the associated risk of type I and type II errors; our data, for instance, would suggest that the proportion of participants in a sample who experienced childhood adversity will be expected to determine the likelihood that basal hypocortisolaemia will be shown. In addition to the CTQ total score, we report here also the emotional neglect subscale, having previously argued that the pervasive nature of emotional neglect ensures that it enacts the greater sustained impact on behavioural and endocrine function [60, 61].

There can be few who argue with the notion that, in the general population, early adversity, acting for instance through methylation or other epigenetic mechanisms, impacts GR function and so GR mediated negative feedback on the HPA axis and thus cortisol synthesis [62, 63] and, further, that this has relevance for understanding the pathophysiology of mood disorders [64]. It is of interest, here, to consider the implication that this has for our understanding of the pathophysiology of CFS and for the interpretation of endocrine studies in this population. We have previously postulated that childhood adversity is not a risk factor for CFS per se, but it can appear to be because of the impact of comorbid or misdiagnosed depression [65]. It has been further conjectured that comorbid depression may commonly confound CFS studies [36] and, just as this may lead to erroneous finding of increased rates of childhood adversity in CFS, similarly, it may explain the methylation pattern [66] including in the NR3C1-1F promoter region [33], the increased GR function (shown using the DST, the dex/CRH test [28], or ex vivo measures), and the basal hypocortisolaemia which have been (inconsistently [10]) shown in previous CFS studies. In this current study, CTQ scores were not greater in the CFS participants than HCs, and it is interesting to note that the basal cortisol or GR function as determined by post-dex cortisol or dex-induced suppression of cytokine synthesis was not convincingly different.

Despite our rigorous exclusion of those who met the diagnostic criteria for depression and the lack of difference in childhood adversity reported by CFS patients compared with HCs, there was a signal that HPA axis regulation was different in CFS; the variability of pre- and post-dex cortisol levels and of dex-induced cytokine suppression was increased in CFS, the proinflammatory cytokine response to LPS was attenuated, and the TNF[alpha] suppression by the larger dex dose was greater, and, whilst we do not want to make too much of this, the graph suggested (but the stats did not back up) the possibility that post-dex cortisol was lower in CFS than HCs. Further research is needed to understand the cause and significance of this data; this will need large, well-characterised groups and will need consideration to be given to the interacting networks of biological, psychological, and social factors.

https://doi.org/10.1155/2018/3972104

Data Availability

The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Acknowledgments

This study was funded by the Medical Research Council (MR/J002712/1), by ME Research UK, and by Action for ME. Thanks are due to the participants and to Heather Slater for the help with the standard deviation graphs.

References

[1] J. L. Newton, O. Okonkwo, K. Sutcliffe, A. Seth, J. Shin, and D. E. J. Jones, "Symptoms of autonomic dysfunction in chronic fatigue syndrome," QJM: An International Journal of Medicine, vol. 100, no. 8, pp. 519-526, 2007.

[2] L. Lorusso, S. V. Mikhaylova, E. Capelli, D. Ferrari, G. K. Ngonga, and G. Ricevuti, "Immunological aspects of chronic fatigue syndrome," Autoimmunity Reviews, vol. 8, no. 4, pp. 287-291, 2009.

[3] K. Fukuda, S. E. Straus, I. Hickie, M. C. Sharpe, J. G. Dobbins, and A. Komaroff, "The chronic fatigue syndrome: a comprehensive approach to its definition and study," Annals of Internal Medicine, vol. 121, no. 12, pp. 953-959, 1994.

[4] A. S. Papadopoulos and A. J. Cleare, "Hypothalamic-pituitary-adrenal axis dysfunction in chronic fatigue syndrome," Nature Reviews Endocrinology, vol. 8, no. 1, pp. 22-32, 2012.

[5] N. Afari and D. Buchwald, "Chronic fatigue syndrome: a review," American Journal of Psychiatry, vol. 160, no. 2, pp. 221-236, 2003.

[6] R. Cairns and M. Hotopf, "A systematic review describing the prognosis of chronic fatigue syndrome," Occupational Medicine, vol. 55, no. 1, pp. 20-31, 2005.

[7] S. E. Straus, A. L. Komaroff, and H. J. Wedner, "Chronic fatigue syndrome: point and counterpoint," Journal of Infectious Diseases, vol. 170, no. 1, pp. 1-6, 1994.

[8] A. J. Parker, S. Wessely, and A. J. Cleare, "The neuroendocrinology of chronic fatigue syndrome and fibromyalgia," Psychological Medicine, vol. 31, no. 8, pp. 1331-1345, 2001.

[9] N. G. Klimas, F. R. Salvato, R. Morgan, and M. A. Fletcher, "Immunologic abnormalities in chronic fatigue syndrome," Journal of Clinical Microbiology, vol. 28, no. 6, pp. 1403-1410, 1990.

[10] C. Tomas, J. Newton, and S. Watson, "A review of hypothalamic-pituitary-adrenal axis function in chronic fatigue syndrome," ISRN Neuroscience, vol. 2013, Article ID 784520, 8 pages, 2013.

[11] S. L. Hardcastle, E. W. Brenu, S. Johnston et al., "Serum immune proteins in moderate and severe chronic fatigue syndrome/myalgic encephalomyelitis patients," International Journal of Medical Sciences, vol. 12, no. 10, pp. 764-772, 2015.

[12] B. van Houdenhove, F. Van Den Eede, and P. Luyten, "Does hypothalamic-pituitary-adrenal axis hypofunction in chronic fatigue syndrome reflect a 'crash' in the stress system?," Medical Hypotheses, vol. 72, no. 6, pp. 701-705, 2009.

[13] T. Theorell, V. Blomkvist, G. Lindh, and B. Evengard, "Critical life events, infections, and symptoms during the year preceding chronic fatigue syndrome (CFS): an examination of CFS patients and subjects with a nonspecific life crisis," Psychosomatic Medicine, vol. 61, no. 3, pp. 304-310, 1999.

[14] S. K. Lutgendorf, M. H. Antoni, G. Ironson et al., "Physical symptoms of chronic fatigue syndrome are exacerbated by the stress of Hurricane Andrew," Psychosomatic Medicine, vol. 57, no. 4, pp. 310-323, 1995.

[15] L. M. Tak and J. G. M. Rosmalen, "Dysfunction of stress responsive systems as a risk factor for functional somatic syndromes," Journal of Psychosomatic Research, vol. 68, no. 5, pp. 461-468, 2010.

[16] C. L. Raison and A. H. Miller, "When not enough is too much: the role of insufficient glucocorticoid signaling in the pathophysiology of stress-related disorders," American Journal of Psychiatry, vol. 160, no. 9, pp. 1554-1565, 2003.

[17] A. J. Cleare, "The neuroendocrinology of chronic fatigue syndrome," Endocrine Reviews, vol. 24, no. 2, pp. 236-252, 2003.

[18] J. Visser, E. Lentjes, I. Haspels et al., "Increased sensitivity to glucocorticoids in peripheral blood mononuclear cells of chronic fatigue syndrome patients, without evidence for altered density or affinity of glucocorticoid receptors," Journal of Investigative Medicine, vol. 49, no. 2, pp. 195-204, 2001.

[19] C. M. Pariante and S. L. Lightman, "The HPA axis in major depression: classical theories and new developments," Trends in Neurosciences, vol. 31, no. 9, pp. 464-468, 2008.

[20] L. A. Carvalho, M. F. Juruena, A. S. Papadopoulos et al., "Clomipramine in vitro reduces glucocorticoid receptor function in healthy subjects but not in patients with major depression," Neuropsychopharmacology, vol. 33, no. 13, pp. 3182-3189, 2008.

[21] J. Gaab, N. Rohleder, V. Heitz et al., "Enhanced glucocorticoid sensitivity in patients with chronic fatigue syndrome," Acta Neuropsychiatrica, vol. 15, no. 4, pp. 184-191, 2003.

[22] L. M. Tak, A. J. Cleare, J. Ormel et al., "Meta-analysis and meta-regression of hypothalamic-pituitary-adrenal axis activity in functional somatic disorders," Biological Psychology, vol. 87, no. 2, pp. 183-194, 2011.

[23] M. A. Demitrack, J. K. Dale, S. E. Straus et al., "Evidence for impaired activation of the hypothalamic-pituitary-adrenal axis in patients with chronic fatigue syndrome," The Journal of Clinical Endocrinology & Metabolism, vol. 73, no. 6, pp. 1224-1234, 1991.

[24] U. M. Nater, E. Maloney, R. S. Boneva et al., "Attenuated morning salivary cortisol concentrations in a population-based study of persons with chronic fatigue syndrome and well controls," The Journal of Clinical Endocrinology & Metabolism, vol. 93, no. 3, pp. 703-709, 2008.

[25] J. Gaab, D. Huster, R. Peisen et al., "Hypothalamic-pituitary-adrenal axis reactivity in chronic fatigue syndrome and health under psychological, physiological and pharmacological stimulation," Psychosomatic Medicine, vol. 64, no. 6, pp. 951-962, 2002.

[26] A. D. L. Roberts, S. Wessely, T. Chalder, A. Papadopoulos, and A. J. Cleare, "Salivary cortisol response to awakening in chronic fatigue syndrome," British Journal of Psychiatry, vol. 184, no. 2, pp. 136-141, 2004.

[27] L. V. Scott, S. Medbak, and T. G. Dinan, "Blunted adrenocorticotropin and cortisol responses to corticotropin-releasing hormone stimulation in chronic fatigue syndrome," Acta Psychiatrica Scandinavica, vol. 97, no. 6, pp. 450-457, 1998.

[28] F. v. d. Eede, G. Moorkens, W. Hulstijn et al., "Combined dexamethasone/corticotropin-releasing factor test in chronic fatigue syndrome," Psychological Medicine, vol. 38, no. 7, 2008.

[29] J. Gaab, D. Huster, R. Peisen et al., "Low-dose dexamethasone suppression test in chronic fatigue syndrome and health," Psychosomatic Medicine, vol. 64, no. 2, pp. 311-318, 2002.

[30] W. K. Jerjes, N. F. Taylor, P. J. Wood, and A. J. Cleare, "Enhanced feedback sensitivity to prednisolone in chronic fatigue syndrome," Psychoneuroendocrinology, vol. 32, no. 2, pp. 192-198, 2007.

[31] J. Visser, W. Graffelman, B. Blauw et al., "LPS-induced IL-10 production in whole blood cultures from chronic fatigue syndrome patients is increased but supersensitive to inhibition by dexamethasone," Journal of Neuroimmunology, vol. 119, no. 2, pp. 343-349, 2001.

[32] J. Visser, B. Blauw, B. Hinloopen et al., "CD4 T lymphocytes from patients with chronic fatigue syndrome have decreased interferon-gamma production and increased sensitivity to dexamethasone," The Journal of Infectious Diseases, vol. 177, no. 2, pp. 451-454, 1998.

[33] E. Vangeel, F. Van Den Eede, T. Hompes et al., "Chronic fatigue syndrome and DNA hypomethylation of the glucocorticoid receptor gene promoter 1F region: associations with HPA axis hypofunction and childhood trauma," Psychosomatic Medicine, vol. 77, no. 8, pp. 853-862, 2015.

[34] B. N. Goertzel, C. Pennachin, L. de Souza Coelho, B. Gurbaxani, E. M. Maloney, and J. F. Jones, "Combinations of single nucleotide polymorphisms in neuroendocrine effector and receptor genes predict chronic fatigue syndrome," Pharmacogenomics, vol. 7, no. 3, pp. 475-483, 2006.

[35] M. S. Rajeevan, A. K. Smith, I. Dimulescu et al., "Glucocorticoid receptor polymorphisms and haplotypes associated with chronic fatigue syndrome," Genes, Brain and Behavior, vol. 6, no. 2, pp. 167-176, 2007.

[36] A. Papadopoulos, M. Ebrecht, A. D. L. Roberts, L. Poon, N. Rohleder, and A. J. Cleare, "Glucocorticoid receptor mediated negative feedback in chronic fatigue syndrome using the low dose (0.5 mg) dexamethasone suppression test," Journal of Affective Disorders, vol. 112, no. 1-3, pp. 289-294, 2009.

[37] L. V. Scott, S. Medbak, and T. G. Dinan, "The low dose ACTH test in chronic fatigue syndrome and in health," Clinical Endocrinology, vol. 48, no. 6, pp. 733-737, 1998.

[38] J. Gaab, N. Rohleder, V. Heitz et al., "Stress-induced changes in LPS-induced pro-inflammatory cytokine production in chronic fatigue syndrome," Psychoneuroendocrinology, vol. 30, no. 2, pp. 188-198,2005.

[39] S. Torres-Harding, M. Sorenson, L. Jason et al., "The associations between basal salivary cortisol and illness symptomatology in chronic fatigue syndrome," Journal of Applied Biobehavioral Research, vol. 13, no. 3, pp. 157-180, 2008.

[40] C. Heim, U. M. Nater, E. Maloney, R. Boneva, J. F. Jones, and W. C. Reeves, "Childhood trauma and risk for chronic fatigue syndrome: association with neuroendocrine dysfunction," Archives of General Psychiatry, vol. 66, no. 1, pp. 7280, 2009.

[41] J. F. Thayer and E. Sternberg, "Beyond heart rate variability: vagal regulation of allostatic systems," Annals of the New York Academy of Sciences, vol. 1088, no. 1, pp. 361-372, 2006.

[42] T. M. O'Connor, D. J. O'Halloran, and F. Shanahan, "The stress response and the hypothalamic-pituitary-adrenal axis: from molecule to melancholia," QJM: An International Journal of Medicine, vol. 93, no. 6, pp. 323-333, 2000.

[43] J. C. Buckingham, H. D. Loxley, H. C. Christian, and J. G. Philip, "Activation of the HPA axis by immune insults: roles and interactions of cytokines, eicosanoids, and glucocorticoids," Pharmacology, Biochemistry, and Behavior, vol. 54, no. 1, pp. 285-298, 1996.

[44] J. G. Cannon, J. B. Angel, R. W. Ball, L. W. Abad, L. Fagioli, and A. L. Komaroff, "Acute phase responses and cytokine secretion in chronic fatigue syndrome," Journal of Clinical Immunology, vol. 19, no. 6, pp. 414-421, 1999.

[45] S. Gupta, S. Aggarwal, D. See, and A. Starr, "Cytokine production by adherent and non-adherent mononuclear cells in chronic fatigue syndrome," Journal of Psychiatric Research, vol. 31, no. 1, pp. 149-156, 1997.

[46] B. v. Houdenhove and P. Luyten, "Customizing treatment of chronic fatigue syndrome and fibromyalgia: the role of perpetuating factors," Psychosomatics, vol. 49, no. 6, pp. 470-477, 2008.

[47] M. B. First, R. L. Spitzer, M. Gibbon, and J. B. W. Williams, Structured Clinical Interview for DSM-IV Axis I Disorders, Research Version. Biometrics Research, New York, 1997.

[48] M. B. First, R. L. Spitzer, M. Gibbon, and W. JBW, "Structured Clinical Interview for DSM-IV Axis I disorders-Patient Edition," in SCID-I/P, Version 2.0, Biometrics Research Department, New York, NY, USA, 1995.

[49] C. Vitali, S. Bombardieri, R. Jonsson et al., "Classification criteria for Sjogren's syndrome: a revised version of the European criteria proposed by the American-European Consensus Group," Annals of the Rheumatic Diseases, vol. 61, no. 6, pp. 554-558, 2002.

[50] W.-F. Ng, S. J. Bowman, B. Griffiths, and on behalf of the UKPSSR study group, "United Kingdom Primary Sjogren's Syndrome Registry-a united effort to tackle an orphan rheumatic disease," Rheumatology, vol. 50, no. 1, pp. 32-39, 2010.

[51] J. Frith and J. Newton, "Fatigue Impact Scale," Occupational Medicine, vol. 60, no. 2, p. 159, 2010.

[52] D. P. Bernstein, J. A. Stein, M. D. Newcomb et al., "Development and validation of a brief screening version of the Childhood Trauma Questionnaire," Child Abuse & Neglect, vol. 27, no. 2, pp. 169-190, 2003.

[53] J. D. Fisk, P. G. Ritvo, L. Ross, D. A. Haase, T. J. Marrie, and W. F. Schlech, "Measuring the functional impact of fatigue: initial validation of the fatigue impact scale," Clinical Infectious Diseases, vol. 18, Supplement_1, pp. S79-S83, 1994.

[54] J. C. Pruessner, C. Kirschbaum, G. Meinlschmid, and D. H. Hellhammer, "Two formulas for computation of the area under the curve represent measures of total hormone concentration versus time-dependent change," Psychoneuroendocrinology, vol. 28, no. 7, pp. 916-931, 2003.

[55] J. K. Kruschke, "Bayesian estimation supersedes the t test," Journal of Experimental Psychology: General, vol. 142, no. 2, pp. 573-603, 2013.

[56] G. Pezzulo, F. Rigoli, and K. Friston, "Active inference, homeostatic regulation and adaptive behavioural control," Progress in Neurobiology, vol. 134, Supplement C, pp. 17-35, 2015.

[57] L. A. Carvalho, B. A. Garner, T. Dew, H. Fazakerley, and C. M. Pariante, "Antidepressants, but not antipsychotics, modulate GR function in human whole blood: an insight into molecular mechanisms," European Neuropsychopharmacology, vol. 20, no. 6, pp. 379-387, 2010.

[58] C. M. Pariante and A. H. Miller, "Glucocorticoid receptors in major depression: relevance to pathophysiology and treatment," Biological Psychiatry, vol. 49, no. 5, pp. 391-404, 2001.

[59] J. T. Visser, E. R. De Kloet, and L. Nagelkerken, "Altered glucocorticoid regulation of the immune response in the chronic fatigue syndrome," Annals of the New York Academy of Sciences, vol. 917, pp. 868-875, 2000.

[60] S. Watson, B. M. Owen, P. Gallagher, A. J. Hearn, A. H. Young, and I. N. Ferrier, "Family history, early adversity and the hypothalamic-pituitary-adrenal (HPA) axis: mediation of the vulnerability to mood disorders," Neuropsychiatric Disease and Treatment, vol. 3, no. 5, pp. 647-653, 2007.

[61] S. Watson, P. Gallagher, D. Dougall et al., "Childhood trauma in bipolar disorder," Australian & New Zealand Journal of Psychiatry, vol. 48, no. 6, pp. 564-570, 2014.

[62] I. C. G. Weaver, N. Cervoni, F. A. Champagne et al., "Epigenetic programming by maternal behavior," Nature Neuroscience, vol. 7, no. 8, pp. 847-854, 2004.

[63] P. O. McGowan, A. Sasaki, A. C. D'Alessio et al., "Epigenetic regulation of the glucocorticoid receptor in human brain associates with childhood abuse," Nature Neuroscience, vol. 12, no. 3, pp. 342-348, 2009.

[64] V. Krishnan and E. J. Nestler, "The molecular neurobiology of depression," Nature, vol. 455, no. 7215, pp. 894-902, 2008.

[65] J. E. Clark, S. L. Davidson, L. Maclachlan, J. L. Newton, and S. Watson, "Rethinking childhood adversity in chronic fatigue syndrome," Fatigue: Biomedicine, Health & Behavior, vol. 6, no. 1, pp. 20-29, 2017.

[66] W. C. de Vega, S. Herrera, S. D. Vernon, and P. O. McGowan, "Epigenetic modifications and glucocorticoid sensitivity in myalgic encephalomyelitis/chronic fatigue syndrome (ME/ CFS)," BMC Medical Genomics, vol. 10, no. 1, p. 11, 2017.

Megan Lynn [ID], (1) Laura Maclachlan, (2) Andreas Finkelmeyer, (1) James Clark, (1) James Locke, (3) Stephen Todryk, (3,4) Wan-Fai Ng, (3,5) Julia L. Newton, (3,5) and Stuart Watson [ID] (1,6)

(1) Institute of Neuroscience, Newcastle University, Newcastle upon Tyne, UK

(2) Department of Public Health and Community Medicine, University of Gothenburg, Gothenburg, Sweden

(3) Institute of Cellular Medicine and NIHR Newcastle Biomedical Research Centre, Newcastle University, Newcastle upon Tyne, UK

(4) Faculty of Health and Life Sciences, Northumbria University, Newcastle upon Tyne, UK

(5) Newcastle Hospitals NHS Foundation Trust, Newcastle upon Tyne, UK

(6) Northumberland, Tyne and Wear NHS Foundation Trust, Newcastle upon Tyne, UK

Correspondence should be addressed to Stuart Watson; stuart.watson@ncl.ac.uk

Received 11 January 2018; Revised 20 April 2018; Accepted 7 May 2018; Published 10 June 2018

Academic Editor: Ronald Gladue

Caption: Figure 1: Total Cortisol values at five time points over two days in participants with CFS and healthy controls. Bar chart showing Cortisol concentrations in CFS patients and healthy controls at five morning time points over two consecutive days. Participants administered oral dex at 11 pm on day 1. Data are shown as mean plus standard error of the mean.

Caption: Figure 2: Frequency density graphs for cortisol area under the curve data and for the suppression of cytokine release by 10 nM solution of dexamethasone. Frequency density graphs for participants with chronic fatigue syndrome and healthy controls for area under the curve (AUC) with respect to ground (g) and increase (i) and for % suppression of TNF[alpha], IL-6, and IL-10 by dex10.

Caption: Figure 3: Violin plot showing baseline cytokine and chemokine concentrations. Baseline cytokine and chemokine concentrations (in pg/ml); asterisks refer to Mann-Whitney U comparisons ** p < 0.01, * p < 0.05.
Table 1: Standard deviations for endocrine data between chronic
fatigue and healthy participants.

                       CFS                    Healthy controls

Day 1 AUCg             79.4 (63.1, 97.5)      72.1 (38.3, 113.2)
Day 1 AUCi             66.3 (52.9, 81.0)      27.5 (14.9, 43.4)
Day 2 AUCg             25.4 (20.0, 31.4)      6.8 (3.7, 10.8)
Day 2 AUCi             16.1 (12.7, 19.7)      5.3 (2.8, 8.4)
% suppression TNFa     31.2 (24.3, 39.0)      15.6 (10.5, 21.3)
% suppression IL-6     18.3 (14.2, 22.9)      13.0 (8.9, 17.9)
% suppression IL-10    47.4 (36.5, 59.3)      56.4 (38.2, 78.0)

                       Difference             p value

Day 1 AUCg             7.32 (-38.8, 48.4)     0.389
Day 1 AUCi             38.75 (16.9, 60.3)     0.014
Day 2 AUCg             18.61 (11.6, 25.8)     <0.0005
Day 2 AUCi             10.75 (5.9, 15.5)      <0.0005
% suppression TNFa     15.6 (6.1,25.1)        0.001
% suppression IL-6     5.3 (-1.2, 11.8)       0.054
% suppression IL-10    -9.1 (-33.8, 13.9)     0.774

AUCg: area under the curve with respect to ground; AUCi: area under
the curve with respect to increase. The % suppressions refer to the
percentage difference in cytokine concentration in the
dexamethasone 10 nM condition compared with the LPS condition.
Values presented are the mean values of the posterior distribution
of standard deviations with 95% credible intervals in brackets.
Differences are significant if the credible interval does not
include zero. p values refer to the proportion of the posterior
mean standard deviation difference plot which is less than or equal
to zero.

Table 2: Ex vivo glucocorticoid receptor response assay data.

                                               HC (n = 19)

         Null                                1.3 (0.4 to 2.3)
         LPS                           14515.2 (12521.4 to 15196.4)
         dex10                          13070.1 (10768 to 15175.2)
IL-6     dex10O                         7035.9 (5553.0 to 10618.0)
         % suppression dex10                3.0 (-4.5 to 12.1)
         % suppression dex100              47.1 (24.4 to 54.1)
         LPS versus dex10                         0.171
         LPS versus dex100                       <0.0005
         % suppression dex10 versus              <0.0005
         dex10O

         Null                                0.0 (0.0 to 0.0)
         LPS                              121.7 (66.6 to 260.0)
         dex10                             179.3 (100 to 264.3)
         dex10O                           248.1 (144.7 to 314.6)
IL-10    % suppression dex10              -17.5 (-59.4 to -1.0)
         % suppression dex10O            -56.1 (-110.4 to -24.5)
         LPS versus dex10                         0.027
         LPS versus dex100                        0.002
         % suppression dex10 versus               0.001
         dex10O

         Null                                0.6 (0.0 to 1.2)
         LPS                             1682.6 (895.0 to 2055.8)
         dex10                           1072.0 (659.0 to 1473.4)
         dex10O                           261.4 (129.3 to 500.6)
TNFa     % suppression dex10               29.7 (25.0 to 37.2)
         % suppression dex100              82.7 (77.9 to 87.0)
         LPS versus dex10                        <0.0005
         LPS versus dex10O                       <0.0005
         % suppression dex10 versus              <0.0005
         dex10O

                                               CFS (n = 40)

         Null                                 1.6 (0.7-5.5)
         LPS                            9517.8 (6994.4 to 12705.7)
         dex10                          9013.1 (6268.4 to 12568.1)
IL-6     dex10O                         5606.2 (3394.6 to 7914.2)
         % suppression dex10                7.1 (-3.6 to 15.2)
         % suppression dex100              40.5 (28.4 to 54.8)
         LPS versus dex10                         0.019
         LPS versus dex100                       <0.0005
         % suppression dex10 versus              <0.0005
         dex10O

         Null                                0.0 (0.0 to 0.5)
         LPS                               95.8 (58.3 to 149.5)
         dex10                            111.9 (76.7 to 151.8)
         dex10O                           135.7 (81.4 to 214.0)
IL-10    % suppression dex10               -5.3 (-31.6 to 6.4)
         % suppression dex10O             -29.2 (-95.2 to 1.9)
         LPS versus dex10                          0.01
         LPS versus dex100                        0.002
         % suppression dex10 versus               0.007
         dex10O

         Null                                0.5 (0.0 to 0.9)
         LPS                             937.5 (636.4 to 1262.9)
         dex10                            743.5 (500.0 to 900.2)
         dex10O                           246.2 (152.2 to 245.4)
TNFa     % suppression dex10               21.3 (10.6 to 37.5)
         % suppression dex100              74.9 (62.6 to 83.6)
         LPS versus dex10                        <0.0005
         LPS versus dex10O                       <0.0005
         % suppression dex10 versus              <0.0005
         dex10O

                                               PSS (n = 27)

         Null                                1.3 (0.8 tol.7)
         LPS                           16407.6 (12750.4 to 21912.2)
         dex10                         15517.0 (12445.0 to 20557.3)
IL-6     dex10O                        10757.4 (7879.5 to 12655.4)
         % suppression dex10                4.3 (-1.1 to 11.8)
         % suppression dex100              36.2 (24.7 to 48.9)
         LPS versus dex10                         0.008
         LPS versus dex100                       <0.0005
         % suppression dex10 versus              <0.0005
         dex10O

         Null                                0.4 (0.0 to 1.0)
         LPS                              149.7 (80.3 to 240.6)
         dex10                            157.7 (98.0 to 249.0)
         dex10O                           197.6 (115.2 to 250.9)
IL-10    % suppression dex10               -12.2 (-23.6 to 9.8)
         % suppression dex10O             -33.7 (-69.6 to 4.1)
         LPS versus dex10                         0.091
         LPS versus dex100                        0.006
         % suppression dex10 versus               0.003
         dex10O

         Null                                0.7 (0.0 to 1.0)
         LPS                            1872.4 (1275.1 to 2410.9)
         dex10                           1428.2 (797.5 to 1942.8)
         dex10O                           389.2 (242.3 to 611.9)
TNFa     % suppression dex10               28.9 (12.7 to 39.9)
         % suppression dex100              76.9 (67.6 to 84.5)
         LPS versus dex10                         0.001
         LPS versus dex10O                       <0.0005
         % suppression dex10 versus              <0.0005
         dex10O

                                         KW      CFS versus
                                                     HC

         Null                           0.193      0.115
         LPS                           <0.0005     0.001
         dex10                         <0.0005     0.007
IL-6     dex10O                        <0.0005     0.037
         % suppression dex10            0.806      0.626
         % suppression dex100           0.235      0.697
         LPS versus dex10
         LPS versus dex100
         % suppression dex10 versus
         dex10O

         Null                           0.007      0.615
         LPS                            0.032      0.114
         dex10                          0.029      0.020
         dex10O                         0.004      0.003
IL-10    % suppression dex10            0.375      0.249
         % suppression dex10O           0.321      0.168
         LPS versus dex10
         LPS versus dex100
         % suppression dex10 versus
         dex10O

         Null                           0.547      0.596
         LPS                           <0.0005     0.008
         dex10                          0.001      0.019
         dex10O                         0.026      0.511
TNFa     % suppression dex10            0.443      0.183
         % suppression dex100           0.080      0.034
         LPS versus dex10
         LPS versus dex10O
         % suppression dex10 versus
         dex10O

                                       CFS versus   PSS versus
                                          PSS           HC

         Null                            0.158        0.841
         LPS                            <0.0005       0.111
         dex10                          <0.0005       0.096
IL-6     dex10O                         <0.0005       0.014
         % suppression dex10             0.561        0.938
         % suppression dex100            0.172        0.113
         LPS versus dex10
         LPS versus dex100
         % suppression dex10 versus
         dex10O

         Null                            0.003        0.031
         LPS                             0.011        0.585
         dex10                           0.040        0.772
         dex10O                          0.014        0.346
IL-10    % suppression dex10             0.779        0.177
         % suppression dex100            1.000        0.183
         LPS versus dex10
         LPS versus dex100
         % suppression dex10 versus
         dex10O

         Null                            0.264        0.778
         LPS                            <0.0005       0.403
         dex10                           0.001        0.170
         dex10O                          0.006        0.129
TNFa     % suppression dex10             0.645        0.496
         % suppression dex100            0.818        0.060
         LPS versus dex10
         LPS versus dex100
         % suppression dex10 versus
         dex100

Cytokine concentrations (in pg/ml) and percent cytokine suppression
on 10 nm dexamethasone and 100 nm dexamethasone in healthy controls
(HCs), participants with chronic fatigue syndrome (CFS), and
primary Sjogren's syndrome (PSS). Comparisons are thep values for
the independent samples Kruskal-Wallis (KW) comparison of the
medians across the 3 groups (HC, CFS and PSS) and the Mann-Whitney
U comparisons between two groups (e.g. HC versus CFS). The
significance levels for the comparison of cytokine levels in the
samples treated with LPS and those treated in addition with
dexamethasone, using Related Samples Wilcoxon Signed Rank Test, are
also shown for the three groups.

Table 3: Correlation coefficients for the relationship between
childhood trauma and endocrine variables.

                                           CFS

                  AUCg         IL-6%          IL-10%      TNF[alpha] %
                  day 1     suppression    suppression    suppression
                               dex10          dex10          dex10

CTQ total        -0.40 **       0.03           0.24          -0.23
score

CTQ emotional    -0.33 *        0.07           0.29          -0.22
neglect

                                           HC

                  AUCg         IL-6%          IL-10%      TNF[alpha] %
                  day 1     suppression    suppression    suppression
                               dex10          dex10          dex10

CTQ total         0.15        0.64 **          0.44           0.02
score

CTQ emotional     0.20        0.72 **         0.53 *         -0.08
neglect

                                           Combined

                  AUCg         IL-6%          IL-10%      TNF[alpha] %
                  day 1     suppression    suppression    suppression
                               dex10          dex10          dex10

CTQ total        -0.32 *        0.17          0.29*          -0.19
score

CTQ emotional     -0.21         0.22          0.37"          -0.17
neglect

Spearman's rho values with significance level indicated by asterisk
(p ** < 0.005, * < 0.05) for relationship between childhood
adversity and endocrine data in participants with CFS, healthy
controls and in a combined group. Data analyses for each group are
in the 4 columns be10w CFS, HC and Combined.
COPYRIGHT 2018 Hindawi Limited
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2018 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Research Article
Author:Lynn, Megan; Maclachlan, Laura; Finkelmeyer, Andreas; Clark, James; Locke, James; Todryk, Stephen; N
Publication:Mediators of Inflammation
Geographic Code:4EUUK
Date:Jan 1, 2018
Words:7443
Previous Article:Haloperidol Abrogates Matrix Metalloproteinase-9 Expression by Inhibition of NF-[kappa]B Activation in Stimulated Human Monocytic Cells.
Next Article:Asymmetric Dimethyl Arginine as a Biomarker of Atherosclerosis in Rheumatoid Arthritis.
Topics:

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters