Skip to main content
Download PDF

Whole blood gene expression in adolescent chronic fatigue syndrome: an exploratory cross-sectional study suggesting altered B cell differentiation and survival

Journal of Translational Medicine volume 15, Article number: 102 (2017) Cite this article

Abstract

Background

Chronic fatigue syndrome (CFS) is a prevalent and disabling condition affecting adolescents. The pathophysiology is poorly understood, but immune alterations might be an important component. This study compared whole blood gene expression in adolescent CFS patients and healthy controls, and explored associations between gene expression and neuroendocrine markers, immune markers and clinical markers within the CFS group.

Methods

CFS patients (12–18 years old) were recruited nation-wide to a single referral center as part of the NorCAPITAL project. A broad case definition of CFS was applied, requiring 3 months of unexplained, disabling chronic/relapsing fatigue of new onset, whereas no accompanying symptoms were necessary. Healthy controls having comparable distribution of gender and age were recruited from local schools. Whole blood samples were subjected to RNA sequencing. Immune markers were blood leukocyte counts, plasma cytokines, serum C-reactive protein and immunoglobulins. Neuroendocrine markers encompassed plasma and urine levels of catecholamines and cortisol, as well as heart rate variability indices. Clinical markers consisted of questionnaire scores for symptoms of post-exertional malaise, inflammation, fatigue, depression and trait anxiety, as well as activity recordings.

Results

A total of 29 CFS patients and 18 healthy controls were included. We identified 176 genes as differentially expressed in patients compared to controls, adjusting for age and gender factors. Gene set enrichment analyses suggested impairment of B cell differentiation and survival, as well as enhancement of innate antiviral responses and inflammation in the CFS group. A pattern of co-expression could be identified, and this pattern, as well as single gene transcripts, was significantly associated with indices of autonomic nervous activity, plasma cortisol, and blood monocyte and eosinophil counts. Also, an association with symptoms of post-exertional malaise was demonstrated.

Conclusion

Adolescent CFS is characterized by differential gene expression pattern in whole blood suggestive of impaired B cell differentiation and survival, and enhanced innate antiviral responses and inflammation. This expression pattern is associated with neuroendocrine markers of altered HPA axis and autonomic nervous activity, and with symptoms of post-exertional malaise.

Trial registration Clinical Trials NCT01040429

Background

Chronic fatigue syndrome (CFS) is a long-lasting and disabling condition characterized by disproportional fatigue after exertions, musculoskeletal pain, headaches, cognitive impairments, and other symptoms [1, 2]. Adolescent CFS prevalence is estimated at 0.1–1.0% [3,4,5], and CFS may have detrimental effects on psychosocial and academic development [6], as well as family functioning [7].

The disease mechanisms of CFS remain poorly understood, but some studies indicate modest immunological alterations, such as low-grade systemic inflammation and attenuation of NK cell function [8,9,10]. Furthermore, the reported beneficial effect of treatment with the anti-CD20 antibody rituximab might suggest a role for B cells in the pathophysiology [11]. Studies of plasma cytokine levels have been inconclusive; findings include increased levels of interleukin (IL)-1 and tumor necrosis factor (TNF) [12], increased levels of IL-1α and IL-1β but normal levels of TNF [13], and no differences between CFS patients and healthy controls [14, 15].

Immune cell gene expression has been addressed by several studies over the last decade. However, the findings do not give a consistent picture: Kerr and co-workers reported differential expression of 88 genes in whole blood samples from CFS patients and healthy controls [16]. A similar pattern of gene expression was later found in two other CFS patient cohorts by the same research group [17]. From leukocyte samples, Light and co-workers reported an increase in expression of genes that are related to sensory, adrenergic and immune system as a response to physical exercise in CFS patients but not in healthy controls [18]. A recent review concluded that there is a larger post-exercise increase in IL-10 and Toll-like receptor 4 (TLR4) gene transcripts in CFS as compared to healthy controls [19]. Restricting the analyses to gene expression from peripheral blood mononuclear cells (PBMC) correlated with multidimensional fatigue inventory and depression scales, Fang and co-workers identified cytokine–cytokine receptor interaction as one of the most significant pathways [20]. Also studying PBMC, Gow and co-workers identified that the top upregulated genes are related to immunological processes [21]. On the other hand, a study of monozygotic twins discordant for CFS did not reveal any differences in whole blood gene expression [22], and it has been maintained that previously reported differences in gene expression were study-specific and not useful for CFS diagnostic purposes [23]. Also, attempts of relating gene expression profiles to clinical symptoms of CFS have had limited success [24]. For instance, Galbraith and co-workers investigated whole blood gene expression in three post-infective cohorts; 63 genes were identified as differentially expressed, but there were no consistent associations to clinical symptoms [25].

The reasons for these discrepancies may partly be due to the multifactorial nature of CFS, which may obscure direct correlations with molecular observations. The complex regulation of transcription, post transcriptional control and RNA metabolism may also prompt variability in gene expression studies; hence mRNA measurements are not always linearly correlated with targeted functional proteins in biological samples at varying time-points.

In addition to immune changes, some studies have found that CFS disease mechanisms are characterized by neuroendocrine alterations including enhanced sympathetic and attenuated parasympathetic cardiovascular nervous activity [26,27,28,29] and attenuation of the hypothalamus–pituitary–adrenal axis (HPA axis) [30,31,32]. These phenomena might be causally related. The complex immune influence exerted by glucocorticoids has been recognized for decades [33]; more recently, ample evidence suggests that both parasympathetic and sympathetic nervous activity promotes immunomodulation [34,35,36]. Accordingly, the “sustained arousal” model of CFS suggests that the observed immune alterations are secondary to the neuroendocrine alterations [37]. This hypothesis received some support from the observation that treatment of adolescent CFS patients with low-dose clonidine, which attenuates sympathetic and enhanced parasympathetic nervous activity through central mechanisms [38], caused a significant reduction in serum levels of C-reactive protein (CRP) [39].

To the best of our knowledge, no previous study has addressed whole blood gene expression in adolescent CFS patients, who are less burdened by comorbidity and aging processes and presumably more homogeneous than adult patients. Nor do we know of any study using high throughput sequencing (HTS) for gene expression analyses in CFS. Furthermore, no previous study has explored associations between neuroendocrine markers and gene expression in CFS. Thus, the aim of this exploratory study was twofold: (a) To map whole blood differential gene expression in adolescent CFS patients and healthy controls, and (b) To explore the associations between gene expression and neuroendocrine markers, immune markers and clinical markers within the CFS group.

Methods

CFS patients

This study is part of the NorCAPITAL-project (The Norwegian Study of Chronic Fatigue Syndrome in Adolescents: Pathophysiology and Intervention Trial; ClinicalTrials ID: NCT01040429). Details of the recruitment procedure and inclusion/exclusion criteria are described elsewhere [39]. Briefly, all hospital paediatric departments in Norway (n = 20), as well as primary care paediatricians and general practitioners, were invited to refer CFS patients aged 12–18 years consecutively to our study center. A standard form required the referral unit to confirm the result of clinical investigations considered compulsory to diagnose pediatric CFS according to national Norwegian recommendations. Exclusion criteria encompassed somatic and psychiatric co-morbidity, pharmaceutical usage (including hormone contraceptives) and being bed-ridden. Patients considered eligible to this study were summoned to a clinical encounter at our study center after which a final decision on inclusion was made.

In agreement with clinical guidelines [2, 40] and previous studies from our group [27,28,29], we applied a ‘broad’ case definition of CFS, requiring 3 months of unexplained, disabling chronic/relapsing fatigue of new onset. We did not require that patients meet any other accompanying symptom criteria.

Healthy controls

A group of healthy controls with a comparable distribution of gender and age were recruited from local schools. Controls were not matched to cases on any variable. No chronic disease and no regular use of pharmaceuticals (including hormone contraceptives) were allowed.

Study design and ethics

A 1-day in-hospital assessment included clinical examination and blood sampling and always commenced between 7.30 and 9.30 a.m. All participants were instructed to fast overnight and abstain from tobacco products and caffeine for at least 48 h. The participants were instructed to apply an ointment containing the local anesthetic lidocaine (Emla®) on the skin in the antecubital area 1 h in advance. After at least 5 min supine rest in calm surroundings, blood samples were obtained in a fixed sequence from antecubital venous puncture. A questionnaire was completed after the clinical encounter and returned in a pre-stamped envelope.

Data were collected in the period from March 2010 until October 2012. The NorCAPITAL project has been approved by the Norwegian National Committee for Ethics in Medical Research and the Norwegian Medicines Agency. Written informed consent was obtained from all participants and from parents/next-of-kin if required. Details of the design are reported elsewhere [39].

Gene expression profiling by RNA sequencing

Whole blood samples (3 mL) at baseline were collected and stored according to the protocol of the Invitrogen Tempus stabilizing reagents (Applied Biosystems, Thermo Fischer Scientific, Waltham, MA, USA). Total RNA was extracted using the Tempus Isolation kit according to manufacturer’s manual with the exception that 2 mL out of the 9 mL mixture of whole blood and reagent were extracted using a modified protocol where 3 mL blood was mixed well with 6 mL Invitrogen Tempus reagent and 2 mL of the mixture was used for RNA isolation. Removal of globin RNA was performed using the Human GLOBINclear kit (Ambion Inc., Texas, USA). The RNA sample quality was analyzed using the Lab-on-a-Chip Agilent RNA Nano kit (Agilent, Santa Clara, USA) and the Agilent 2100 Bioanalyzer platform. RNA samples with RNA integrity number (RIN) value ≥7 were used for gene expression characterization by RNA sequencing (RNA-Seq) at the Genomics Core Facilities at the Oslo University Hospital Radiumhospitalet, Norway. RNA library preparation and sequencing were performed according to the HiSeq 2500 Illumina protocol for 101 bp single-end strand-specific sequencing (Illumina Inc., San Diego, CA, USA). 130 ng of Globin depleted RNA from each sample was converted into a cDNA library using the RiboZero Gold and TruSeq Stranded mRNA Sample Prep Kit (Illumina Inc., San Diego, CA, USA). A total of 15–35 million reads were generated per sample.

Transcriptome alignment and gene expression quantification

Raw RNA reads from Illumina sequencing were assessed by the fastQC tool [41] to assess sequence quality per base, quality scores per sequence, sequence and GC content per base, sequence length distribution, sequence duplication levels, Kmer content and overrepresented sequences (which also detected the presentation of ribosomal contamination). Adapter contamination elimination and reads trimming were conducted by the fastx toolkit [42].

All reads that passed QC assessment were mapped to the human genome version GRCh38.p2 by STAR [43]. To investigate the level and uniformity of the read coverage against the human genome, we plotted mapped reads against all human chromosomes using the SeqMonk software [41].

Statistics for differential expression analyses were performed using Bioconductor tools [44] in the R environment version 3.1.2. Gene expression abundance was quantified by the Subread package [45] at the gene level. Normalization of raw read quantification and removal of variation before differential expression analyses were processed following RUVg method [46]. Differentially expressed genes (DEG) between CFS patients and controls were identified using DESeq2 package [47]. In order to correct for possible confounding background factors, age groups as scaling factor and gender input were included in the design model of DESeq2. For each DEG, a p value cut off ≤0.10 after multiple-testing adjustment by Benjamini–Hochberg [False Discovery Rate (FDR) 10%] was applied, in accordance with the DESeq2 workflow.

A heatmap of samples distance was constructed by clustering distance matrix from logarithm 2 transformed values of count data [48] using the pheatmap package of Bioconductor. Hierarchical clustering of 100 top DEGs was performed using genefilter and pheatmap packages of Bioconductor in order to measure the deviation of expression value of each sample from the average expression across all samples. The purpose is to build blocks of genes that co-vary across different samples, and clustering the amount by which each gene deviates in a specific sample from the gene’s average across all samples.

Validation of differentially expressed genes

To validate some of the genes from the DEG list, RT-qPCR was performed on the RNA material subjected to sequencing. Specific primers for each target gene were designed as to establish RT-qPCR conditions for each DEG individually (Additional file 1: Table S1). RNA was converted into cDNA by High-Capacity cDNA Reverse Transcription Kit (Life Technologies, Carlsbad, CA, US). Five nanogram cDNA was tested in duplicate reaction on a 7900 HT real-time machine (Applied Biosystems, Foster City, California, USA), using the Evagreen Sso Fast Master mix (Biorad Laboratories, CA, USA). The relative expression levels of each DEG were calculated by the 2ΔΔCt method and were normalized to the GAPDH reference gene.

Downstream data analysis

Functional annotation of genes obtained from DESeq 2 was done by uploading all DEGs into HumanMine [49]. Network visualization and Functional Enrichment Analysis was conducted through Cytoscape software 3.3. and ClueGO 2.3.2 [50]. Log2 of fold change of the expression value (after normalization) was imported into QIAGEN Ingenuity Pathways Analysis (IPA) for an Upstream Transcriptional Factor analysis as well as a mechanistic network enrichment analysis.

Previous analyses of whole blood gene expression in CFS patients [51] as well as healthy individuals [52] have revealed that co-expression of genes is a common phenomenon. Such co-expression might be the effect of neuroendocrine signaling initiating a specific expression pattern; this is in line with the “sustained arousal”-model of CFS [37]. Furthermore, a certain pattern of co-expression might be associated with specific clinical phenomena. To explore different axis of co-expression and reduce dimensionality in the present study, a factor analyses [principal component analysis (PCA) featuring varimax rotation] was applied to the DEG dataset (RNA-Seq normalized counts), in line with previous reports [51, 52]. Thereafter, the associations between factor scores and immune, neuroendocrine and clinical markers (cf. below) were explored using correlation and regression analyses. Similar association studies were also performed for some selected single gene transcriptional counts. In all these analyses, a p ≤ 0.05 was considered statistically significant; no adjustment for multiple testing was performed.

Immune markers

Serum samples from 21 CFS patients and 18 controls were used to identify levels of immunoglobulins. The immunoglobulin classes IgA, IgE, IgM and the four IgG subclasses IgG1, IgG2, IgG3 and IgG4 in serum were measured using Luminex bead-based multiplex technology with reagents from the Procartaplex Immunoassay (Affymetrix eBioscience, San Diego, USA). The concentration of each sample was determined by plotting the expected concentration of standards against fluorescence intensity. Data analysis was performed using Procartaplex Analyst 1.0 and normalization was based on the best curve fit of standards curve.

The serum concentration of C-reactive protein (CRP) was analyzed as described previously [39]. Blood samples for analyses of IL-1β, IL-6 and TNF were placed on ice; plasma was separated by centrifugation (2500×g, 10 min, 4 °C) and frozen at −80 °C until assayed using a multiplex cytokine assay (Bio-Plex Human Cytokine 27-Plex Panel; Bio-Rad Laboratories Inc., Hercules, CA, USA) as described elsewhere [15]. Hematology and biochemistry routine assays were performed at the accredited laboratory at Oslo University Hospital, Norway.

Neuroendocrine markers

As outlined in detail elsewhere [32], blood samples for plasma norepinephrine (NE) and epinephrine (E) were placed on ice; thereafter, plasma was separated by centrifugation (2250×g, 15 min, 4 °C) and assayed by high-performance liquid chromatography (HPLC) with a reversed-phase column and glassy carbon electrochemical detector (Antec, Leyden Deacade II SCC, Zoeterwoude, The Netherlands) using a commercial kit (Chromsystems, München, Germany). Plasma cortisol level was determined by routine assays at the accredited laboratory at Oslo University Hospital, Norway. Morning spot urine samples for NE and E analyses were acidified to pH 2.5 immediately after collection, and assayed with the same HPLC protocol as for plasma measurements [32]. Morning spot urine free cortisol (non-conjugated cortisol) was assayed by solid phase competitive luminescence immunoassay (LIA) (type Immulite® 2000, Siemens Healthcare Diagnostics, NY, USA). The urine levels of creatinine were analyzed using standard automatic analyzer techniques at the accredited laboratory at Oslo University Hospital, Norway.

Indices of heart rate variability (HRV) were obtained from ECG recordings of participants laying in a horizontal position and connected to the Task Force Monitor (TFM) (Model 3040i, CNSystems Medizintechnik, Graz, Austria). Methodological details are provided elsewhere [53]. Power spectral analysis of HRV was automatically provided by the TFM, returning numerical values for Low Frequency (LF) power (0.05–0.17 Hz), High Frequency (HF) power (0.17–0.4 Hz) and the LF/HF ratio. In addition, the time-domain index RMSSD (the square root of the mean square differences of successive RR-intervals) was computed. RMSSD and HF power are both considered indicative of parasympathetic heart rate modulation; LF power reflects the combined effect of sympathetic and parasympathetic heart rate control, whereas the LF/HF ratio is an index of sympathetic/parasympathetic balance [54].

Clinical markers

A CFS symptom inventory for adolescents assesses the frequency of 24 common symptoms during the preceding month, as has been described elsewhere [39]. Briefly, each symptom is rated on a 5-point Likert scale, ranging from ‘never/rarely present’ to ‘present all the time’. A composite score reflecting inflammatory symptoms was generated by taking the arithmetic mean across three single items (fever/chills, sore throat, and tender lymphatic nodes) and a composite score reflecting symptoms of post-exertional malaise was generated by taking the arithmetic mean across two single items (post-exertional fatigue and non-refreshing sleep). For both variables, the total range is from 0 to 5; higher scores imply more severe symptom burden.

The Chalder Fatigue Questionnaire (CFQ) total sum score is applied in the present study [55]; total range is from 0 to 33, where higher scores imply more severe fatigue. The Mood and Feelings Questionnaire (MFQ) consists of 34 items, each scored on a 0–2 Likert scale; thus, the total sum score is from 0 to 68 [56]. The Spielberger State-Trait Anxiety Inventory subscore reflecting trait anxiety is derived from the sum across 20 items; total range is from 20 to 80 [57]. The activPAL accelerometer device (PAL Technologies Ltd, Glasgow, Scotland) was used for monitoring of daily physical activity during 7 consecutive days [58], as described elsewhere [39].

Results

Participants

RNA was extracted from a sub-cohort of the NorCAPITAL study and a total of 60 samples with RIN value ≥7 were subjected to RNA sequencing. After removing ribosomal contamination and bad quality reads from the RNA-Seq experiment, a random sample of 29 CFS patients and 18 healthy controls (a total of 47, mean RIN value = 7.67) were analyzed further for differential gene expression quantification in the present study.

The background characteristics of the two groups are given in Table 1. In line with previously reported findings from the NorCAPITAL project [39], plasma norepinephrine, plasma epinephrine, and urine norepinephrine were significantly higher in the CFS group, as were scores of symptoms of post-exertional malaise, inflammation, fatigue, depression, and trait anxiety. The number of steps per day was significantly lower in the CFS group. Overall, the values of the different variables in the present study are comparable to the values pertaining to the entire NorCAPITAL cohort (Additional file 2: Table S2), except for urine cortisol/creatinine ratio (for which there was no across-group difference in the present study but lower among CFS patients in the entire NorCAPITAL cohort).

Table 1 Background characteristics of the chronic fatigue syndrome (CFS) group and the healthy control (HC) group in the present study
Full size table

Differentially expressed genes in whole blood between CFS patients and healthy controls

RNA-Seq produced 18–45 × 106 single end reads per sample, which was previously reported to be sufficient for transcriptome quantification [59]. The rate of unique mapping into the reference genome was 80–92%, with 50% reads mapped to exons. A percentage of the reads (3–5%) were found to be mapped to ribosomal RNAs. Multiply mapped reads, reads mapped to the sense strands and reads mapped to exon–exon boundaries were not counted. As might be expected from the inherent heterogeneity of whole-blood gene expression, there was no evident subgrouping between either patients or controls in our gene expression data before normalization. This is illustrated in Fig. 1a where the individual samples are distant from one another.

Fig. 1
figure 1

a Hierarchal clustering of all 47 samples based on the rlog value [48]. The color density at the top right panel reflects the Euclidean distance. P CFS patients, C healthy controls. b Output of variation removal of our RNA-Seq data using RUVSeq. principle component analysis (PCA) is performed without using any differently expressed genes, and demonstrates relatively good separation between CFS patients (orange) and healthy controls (green). c Relative log expression (RLE) plot shows the distribution of read counts across all samples centered around zero. The y axis corresponds to the deviation of each RLE per gene per sample compared to median RLE over all samples (x axis). (CFS patients orange. Healthy controls green)

Full size image

Normalization and differential expression analysis, with correction for age and gender factors, detected a total of 176 genes that were differentially expressed between CFS patients and healthy controls (adjusted p < 0.10) (Additional file 3: Table S3). The robustness of DEGs after normalization was confirmed by good separation between the CFS and control groups through principal component analysis (Fig. 1b) and by plotting regular log expression values compared with median of log expression across all samples (Fig. 1c).

Of the 176 DEGs, 137 were upregulated and 37 were downregulated (Fig. 2a, b; Additional file 3: Table S3). This corresponds to an observation of 78% of the DEGs being up-regulated in CFS patients as compared to 22% of the genes having a down-regulated transcriptional pattern compared to healthy controls. Although significant, the differences in normalized expression levels were small, ranging from 0.8- to 1.25-fold (linear scale) (Table 2; Additional file 3: Table S3). Among the 176 differentially expressed genes we observed nuances of expression both within the groups as well as between the two groups (Fig. 2b).

Fig. 2
figure 2

a Volcano plot showing the alignment between DESeq p values versus log2 fold changes of CFS patients against healthy controls. Red points indicate DEGs with a log2 fold change >0.2 and p < 0.0016 (Table 2). b Hierarchical clustering of all 176 differently expressed genes. The heatmap was constructed based on the deviation of gene expression levels of individual sample from averaged gene expression across all samples (Table 2). The color code for variance value is shown in the upper right corner of the panel

Full size image
Table 2 Differentially expressed immune genes, their annotated proteins, and their annotated biological processes based on gene ontologies in CFS patients as compared to healthy controls, adjusted for age and gender differences across groups and sorted according to foldchange
Full size table

A total of 12 genes were selected for further examination featuring RT-qPCR (Fig. 3). Because of the exploratory nature of this study, we wanted these selected genes to be as representative as possible for the RNA-Seq results as a whole: Three genes are related to B cells differentiation/survival (CD79A, FTL3) and B cell malignancies (BCL7A); in addition, these three genes are among the most under-expressed in the CFS group. Two genes are related to IL1 and IL17 signaling pathways (IL1RN and GLRX1, respectively). Two genes are annotated to inflammatory responses (NAMPT, CASP1). Three genes are related to innate antiviral defense (APOBEC3A, IFI16, PLSCR1). The final two genes (HK3, KCJN5) are the two most over-expressed in the CFS group. Ten of the transcripts were found to be differentially expressed in the same direction as in the RNA seq experiments; for three of the transcripts (APOBEC3A, PLSCR1, IL1RN), the fold change differences were statistically significant or close to the level of significance (p = 0.0005, p = 0.0489, p = 0.0507, respectively, Mann–Whitney test). The fold changes measured between CFS patients and healthy controls were moderate, which is in accordance with the RNA-Seq data.

Fig. 3
figure 3

RT-qPCR results of 12 selected transcripts. CFS patients and controls are plotted on the x axis and relative fold change difference normalized against GAPDH is plotted on the y axis. For three transcript, the differential expression between patients and controls were below or close to the level of significance (APOBEC3A, p = 0.0005; PLSCR1, p = 0.0498; IL1RN, p = 0.0507)

Full size image

Gene set enrichment analyses performed using Gene Ontology annotation by HumanMine and independent filtering, suggested that a large fraction of the DEGs (34 out of 176) were related to the immune system (Table 2). Five of the genes that were most down-regulated in the CFS group are associated with B cell differentiation and survival (Fig. 4, cf. above): FLT3 (encoding FLT3, a tyrosine kinase), EBF1 (encoding EBF, 1 early B cell factor 1), CD79A (encoding Igα, a co-molecule of the membrane bound B cell receptor (BCR) complex), CXCR5 (encoding CXCR5, a chemokine receptor), and TNFRSF13C (encoding BAFFR, a receptor for B cell activating factor). Conversely, many of the genes that we found to be upregulated in CFS have a role in innate immunity and inflammation. Prominent examples include CASP1 (encoding caspase 1), CLEC2B (encoding activation-induced C-type lectin), PLSCR1 (encoding phospholipid scramblase 1), IFI16 (encoding gamma-interferon-inducible protein 16), PDE1B (encoding cyclic nucleotide phosphodiesterase), IRF9 (encoding interferon regulatory factor 9), TLR8 (encoding toll-like receptor 8), and APOBEC3A (encoding a DNA editing enzyme).

Downstream data analysis

Functional enrichment by ClueGO and visualization by Cytoscape identified a network of genes related to viral genome replication in the CFS group. Also, a downstream biological analysis using Ingenuity Pathway Analyses (IPA) confirmed that genes that are important for B cell differentiation and survival were down-regulated in the CSF patients. A search in IPA for mechanistic network enrichment of the upstream transcriptional regulators identified three top genes (Additional file 4: Table S4). The top upstream regulator identified was IRF7, which has functional couplings with STAT3 or STAT6 through TNF and IFN respectively [60]. The others were transcription factors: SPI1 encodes a protein involved in myeloid and B cell lymphoid development, whereas STAT6 encodes STAT6, which is activated by IL-4 and IL-13 and is important in signal transduction in many immune cells.

Immunoglobulin classes and subclasses in CFS patients and healthy controls

As the DEGs suggested possible effects on B cell differentiation and survival among CFS patients, immunoglobulin classes and the IgG subclasses were analyzed across the two groups. Measurements of all immunoglobulin isotype fell within the linear range of the standard curve, except for one control sample in which IgG3 concentration was higher than the upper limit of detection. There were no across group differences among the serum levels of IgG1, IgG2, IgG3, IgG4, IgA, IgE, and IgM. Further characterization of B cell function in CFS could not be pursued, as viable PBMC that could be used for stimulation experiments were unavailable.

Co-expression of genes and associations with immune, neuroendocrine and clinical markers within the CFS group

The principal component analyses (PCA) of all DEGs in the CFS group revealed that a 4-factor structure would account for 70% of the total variation. Inspection of the factor loadings revealed that several of the immune process annotated genes that were most differentially expressed across groups (including genes related to B cell differentiation and survival, and innate immunity) loaded on one factor (Additional file 5: Table S5), suggesting a possible co-expression pattern. Therefore, this factor, labelled “Factor 3” in the following, was selected for further explorative analyses.

In bivariate correlation analyses, factor 3 correlated positively with serum CRP-levels, granulocyte and monocyte count, plasma cortisol levels and indices of sympathetic nervous activity. There was a negative correlation with eosinophil count and indices of parasympathetic nervous activity. Finally, there was a slight association to symptoms of post-exertional malaise (p = 0.05), but not to any other clinical markers, including symptoms of depression and anxiety as well as physical activity (steps per day).

Based on results from bivariate correlation analyses as well as theoretical considerations, a multiple regression model was explored. The final model explained 67% of Factor 3 total variance (Fig. 5). LF/HF ratio (an index of sympathetic vs parasympathetic balance), blood monocyte count, and plasma cortisol levels were positively associated with Factor 3, whereas blood eosinophil count was negatively associated with Factor 3. Furthermore, LF/HF ratio was positively associated with blood monocyte count.

Associations of individual transcripts with immune, neuroendocrine and clinical markers within the CFS group

To further explore associations between gene expression and immune, neuroendocrine and clinical markers, transcripts that loaded on Factor 3 and in addition were annotated to immune processes (cf. Table 2) were selected. Three of the selected genes (CD79A, TNFRSF13C, CXCR5) are related to B cell differentiation and survival; they loaded negatively on Factor 3 (Additional file 5: Table S5) and were also less expressed in the CFS group. Three other genes (CASP1, PLSCR1, IFI16) are related to regulation of innate immune responses; they loaded positively on Factor 3 and were also overexpressed in the CFS group.

The transcript of all the three genes related to B cell differentiation and survival tended to correlate negatively with blood neutrophil count, blood monocyte count, serum CRP, plasma cortisol, LF/HF ratio and symptoms of post-exertional malaise, and positively with blood eosinophil count and RMSSD (Additional file 6: Table S6). An opposite pattern was observed for the three genes related to innate immunity; in addition they were positively associated with urine epinephrine, but not with clinical symptoms. In multiple regression models, a homogeneous picture was observed regarding the three B cell related transcripts (Fig. 6a): there was a significant negative association to plasma cortisol levels and a significant positive association to blood monocyte count, which in turn was positively associated with LF/HF ratio. For the transcripts related to innate immunity, the picture was more heterogeneous (Fig. 6b), but all were negatively associated with eosinophil count and positively associated with plasma cortisol and urine epinephrine levels.

Discussion

The main findings of this study are: (a) A total of 176 genes are differentially expressed in whole blood across adolescent CFS patients and healthy controls after adjusting for age and gender differences (FDR 10%); in CFS, there is down-regulation of genes related to B cell differentiation and survival, and upregulation of genes related innate antiviral responses and inflammation. (b) Within the CFS group, the differentially expressed genes are associated with neuroendocrine markers of altered HPA-axis and autonomic nervous activity, as well as with symptoms of post-exertional malaise.

The down-regulated genes related to B cell differentiation and survival included the genes mentioned above: EBF1, CD79A, CXCR5, TNFRSF13C, and FLT3. The FLT3 protein acts as a cell-surface receptor and is a regulator for the differentiation, proliferation and survival of B cell progenitor cells in the bone marrow [61]. The EBF1 protein is a transcription factor that is expressed in B cells at all stages of their differentiation except for fully differentiated plasma cells [62]. The Igα encoded by CD79A is a co-molecule of the BCR complex and ensures that the signal cascade for recognition of antigen is sent. This is necessary for internalization of the BCR-antigen complex and further processing and presentation of antigen peptides on the B cell surface [63]. The chemokine receptor CXCR5 is important for migration of B cells into secondary lymphoid organs [64]. The B cell activating factor receptor (BAFFR) encoded by TNFRSF13C enhances mature B cell survival and controls peripheral B cell population [65]. Taken together, our data suggest that the efficiency of B cell differentiation is impaired and that their survival is reduced in the CFS patients (Fig. 4).

Fig. 4
figure 4

The RNA-Seq identified five down-regulated genes encoding proteins associated with B cell differentiation and survival. FLT3 encodes FLT3 (fms-related tyrosine kinase 3), which is important during the very early stages of differentiation in the bone marrow of the hematopoietic stem cell into the Pro-B cell. EBF1 encodes EBF (early B-cell factor 1), which is important during all stages of B cell differentiation except for the plasma cell. CD79A encodes Igα (immunoglobulin-associated alpha), which is a co-molecule in the membrane bound Pre-BCR and the BCR, and ensures a functional receptor. TNFRSF13C encodes BAFFR (B-cell activating factor receptor), which is important for the peripheral B cells to receive survival signal. CXCR5 encodes CXCR5 [chemokine (C-X-C motif) receptor 5], which ensures that matured B cells migrate to B cell follicles of the spleen and Peyer patches. Assuming that the down-regulation of these genes is reflected at the protein and pathway level, our data suggest that the efficiency of B cell differentiation is impaired and that their survival is reduced in the CFS. HSC hematopoietic stem cell, BCR B cell receptor, B B cell, Ig immunoglobulin

Full size image

As for upregulated innate immunity genes, a number was related to viral defence mechanisms. APOBEC3A was enriched in the negative regulation of viral genome replication together with PLSCR1 and FAM111A (a chromatin-associated DNA clamp required for proliferating cell nuclear antigen loading on replication sites). The enzyme encoded by APOBEC3A deaminates foreign DNA as part of viral clearance [66], whereas phospholipid scramblase 1 (encoded by PLSCR1) was observed to play a role in enhancement of IFN response and increase expression of antiviral genes in mice [67]. This network was in turn connected to IFI16, Gamma-interferon-inducible protein 16, which is a sensor for intracellular DNA and a mediator of IFN induction. Other genes that were found to be related to IFN signaling were the genes encoding interferon regulatory factor 9 (IRF9) and TLR8. The Interferon regulatory factor 9 is a component of the interferons stimulated gene factor 3 complex that is involved in positive regulation of type I interferon gene [68]. TLR8 is an endosomal receptor which acts against foreign ssRNAs by intracellular signalling through NF-κB or IRF7 pathways [69].

Other upregulated innate immunity genes were related to inflammation: Caspase 1 (encoded by CASP1), having a central role in the formation of inflammasomes and other inflammatory-related responses [70]; activation-induced C-type lectin (encoded by CLEC2B), which promote the cross-talk between monocytes and NK-cells [71]; and cyclic nucleotide phosphodiesterase encoded by PDE1B, which is important for the cellular response to granulocyte macrophage colony-stimulating factor [72].

Down-regulation of genes important for B cell differentiation and survival in CFS, as suggested by the present study, comply with a previous CFS studies: Recently, increased levels of the B lymphocyte activating factor of the tumor necrosis family (BAFF) was reported in adults with CFS [73]. We speculate that this might be a compensatory mechanism as BAFF is a ligand for BAFFR encoded by TNFRSF13C, which is one of the most suppressed genes among CFS patients in the present study. Taken together, these results might indicate a role for B cells in CFS pathophysiology, as is supported from studies of cellular immunology: Brenu and co-workers reported a decrease in immature B cells and an increase in memory B cells among CFS patients [74], whereas Bradley and co-workers [75] and Mensah and co-workers [76] found subtle distortions in the proportion of B cell subsets. Alterations of immunoglobulin levels in CFS have also been reported [77], but was not identified in the present material, which is not surprising given the strong propensity of compensatory mechanisms to ensure normal immunoglobulin levels in circulation despite changes in B cell function [78].

Up-regulation of genes related to innate antiviral responses has, to our knowledge, not been consistently reported previously in CFS, not even in cohorts suffering from chronic fatigue following long-lasting viral infections [25]. Our data point to functionally connected genes and pathways involved in innate immunity responses as differentially expressed in the CFS group and might suggest less efficient viral clearance or reactivation of latent viruses such as members of the herpes virus family, in the CFS group [79]. Of note, the herpes virus Epstein-Barr virus (EBV) is a well-known trigger of CFS in adolescents [80]. The possible presence of inefficient viral clearance or virus reactivation, and whether intracellular signaling cascades activated by long-lasting viral infections may be a contributor to CFS pathophysiology, warrant further studies. A model from Thorley-Lawson suggested that EBV uses a pathway similar of B cell survival and B cell differentiation in order to establish its infection, persistence and replication [81]. Loebel and co-workers assumed that a frequent EBV reactivation or impaired control of EBV was a result of the diminished EBV-specific memory B cell response in CFS patients [82]. Therefore we speculate that in some patients, CFS is characterised by persistent EBV-host interactions. Based on the observation of altered B cells differentiation and B cell survival signature, in future experiments, we aim to validate the finding by measuring B cell responsiveness to stimulator such as EBV virus antigens [viral capsid antigen (VCA) and EBV nuclear antigen 1 (EBNA-1)] alone or after exposure to the neuroendocrine hormones.

Up-regulation of genes related to inflammation in the CFS group, which is corroborated by the positive correlation between “Factor 3″ and serum CRP levels, comply with previous CFS studies reporting elevation of proinflammatory cytokines in adult CFS [12, 13, 83]. Interestingly, a recent study of gene expression in NK cells of CFS patients showed upregulation of RIPK3 [84], in line with the present data (Additional file 3: Table S3); this gene encodes a kinase that plays a vital role in inflammasomes and IL-1β signaling. However, a previous analysis of cytokine levels in the present material did not relieve any differences in CFS patients as compared to healthy controls [15]. Thus, a skewing of the immune response towards inflammation appears to be subtle, or even indirect, complying with other studies of gene expression reporting small or moderate fold changes in inflammatory related gene transcripts [16, 25].

The strong association between “Factor 3” with neuroendocrine markers within the CFS group is a novel finding. Although causal interferences cannot be made from our cross-sectional design, the results are in line with the “sustained arousal” model of CFS which suggests that immune alterations are secondary to neuroendocrine alterations [37]. This potential mechanism complies with findings in studies of neuro-immunomodulation: Sympathetic nervous activity has complex effects on B cells, monocytes and several other immune cells through adrenergic receptors that in turn promote alteration of gene expression [35]. Parasympathetic nervous activity has a well-described anti-inflammatory effect based upon gene expression alterations of spleen macrophages [34]. The glucocorticoid effects on immunity are extensive [33], and might in addition be abnormal in CFS, as some studies have indicated a fundamental alteration of glucocorticoid signaling [32, 85].

Taken together, the results of the present study might indicate a skewing of the immune responses from adaptive to innate immunity promoted by the combined effect of HPA axis alteration and sympathetic vs. parasympathetic predominance in CFS patients. We speculate that “Factor 3” in the present data set encapsulates this skewing, being negatively associated with transcripts regulating B cell differentiation and survival, and positively associated with transcripts involved in inflammation and innate antiviral defense. Interestingly, such a skewing shares some similarities with the concept of “Conserved Transcriptional Response to Adversities” (CTRA) [86]. Recent evidence suggests that CTRA is promoted by increased sympathetic nervous activity to the bone marrow, altering myeloid cell numbers and function and promoting functional glucocorticoid desensitization [87]. This complies with the present findings of autonomic nervous activity indices, blood monocyte and eosinophil counts, as well as plasma cortisol level being strongly associated with “Factor 3” (Figs. 5, 6).

Fig. 5
figure 5

Multiple regression model on the associations between neuroendocrine markers (upper row), immune markers (middle row) and co-expression of genes as captured in Factor 3 from a principal component analysis (lower row). LF/HF, B-Mono, B-Eos and P-Cort are all independently and significantly associated with Factor 3, explaining 67% of the total variance. For LF/HF, B-Mono and P-Cort, the association is positive; for B-Eos the association is negative. In addition, LF/HF is significantly associated with B-Mono. P plasma, U urine, B blood, LF/HF low-frequency/high-frequency power of heart rate (an index of sympathetic vs parasympathetic balance), RMSSD square root of the mean squared differences of subsequent RR-intervals (an index of parasympathetic activity), Cort cortisol, Epi epinephrine, Mono monocytes, Eos eosinophils, Neu neutrophils, PCA principal component analysis, B regression coefficient (unstandardized), R 2 explained variance of the dependent variable in the multiple regression model

Full size image
Fig. 6
figure 6

Multiple regression models on the associations between neuroendocrine markers, immune markers and single gene transcripts within the CFS group. a Three genes related to B cell differentiation and survival, with negative loadings of Factor 3 and down-regulated expression in the CFS-group as compared to healthy controls. b Three genes related to innate immunity, with positive loadings of Factor 3 and up-regulated expression in the CFS group as compared to healthy controls. P plasma, U urine, B blood, LF/HF low-frequency/high-frequency power of heart rate (an index of sympathetic vs parasympathetic balance), RMSSD square root of the mean squared differences of subsequent RR-intervals (an index of parasympathetic activity), Cort cortisol, Epi epinephrine, Mono monocytes, Eos eosinophils, Neu neutrophils, PCA principal component analysis, B regression coefficient (unstandardized), R 2 explained variance of the dependent variable in the multiple regression model

Full size image

The present study did not demonstrate strong associations between gene expression profiles and clinical markers; this lack of association to clinical symptoms is in line with other studies [24, 25]. Specifically, there was no correlation between gene transcripts and symptoms of inflammation within the CFS group, confirming previous findings [15]. However, the present data did suggest an association between differential gene expression and symptoms of post-exertional malaise, which is considered a hallmark of the CFS phenotype [1]. This association was primarily evident for the transcripts related to B cell differentiation and survival (Additional file 6: Table S6), an observation that warrants further studies. The lack of association between “Factor 3” and depressive symptoms, trait anxiety and steps per day suggests that the findings are not confounded by the co-existence of emotional problems nor physical inactivity in the CFS group.

Study strengths and limitations

A strength of this study is the HTS based methods combined with extensive clinical phenotyping. The background data show that the subsets of participants in the present study are comparable to the entire NorCAPITAL cohort (Additional file 2: Table S2). However, the numbers of subjects are relatively low, and the wide inclusion criteria might have obscured results pertaining to a subgroup; unfortunately, the study did not have sufficient statistical power to allow meaningful subgroup analyses. In addition, the relatively strict p value cut off of ≤0.1 (after multiple-testing adjustment) for identifying DEGs might increase the risk of type 2-errors. However, previous studies of the NorCAPITAL data set do not suggest subgroup differences [15, 32, 39, 53]. Furthermore, important background factors such as BMI, smoking status and alcohol consumption do not differ across patients and controls, reducing the risk of confounding effects [39]. There was a relatively poor correspondence between RNA seq results and RT-qPCR results; reasons for this discrepancy might be different primers and different normalization methods between RNA seq and RT-qPCR, as well as low concentration of remaining cDNA after RNA seq. Also, the study might have benefitted from a more stringent approach for selecting genes for RT-qPCR analyses. In addition, the design of the NorCAPITAL project did not allow analyses of correlation between mRNA levels and protein levels. Another limitation is that we did not assess gene expression responses to exercise or other stimuli (such as fatigue provoking mental activity), which might have provided important additional information [19]. Furthermore, the RNA seq analysis was not corrected for the different cell populations in whole blood. The investigational program in the NorCAPITAL project did neither include subtyping nor biobanking of peripheral blood cells; thus, validation of the gene expression findings with flow cytometer analyses or functional assays was not possible in the present study. Future studies should include deep phenotyping of the peripheral cell populations and analysis of their effector functions. Further studies should also be powered to allow subgroup analyses, as well as ensure robust validation of the findings.

Conclusion

Adolescent CFS is characterized by differential gene expression pattern in whole blood suggestive of impaired B cell differentiation and survival and enhanced innate antiviral responses and inflammation. This expression pattern is associated with neuroendocrine markers of altered HPA axis and autonomic nervous activity, and with symptoms of post-exertional malaise. Taken together, the results contribute to the understanding of CFS disease mechanism, which in turn is a prerequisite for development of improved diagnostic procedures and therapeutic interventions. Also, the results are in in line with the “sustained arousal”-model of CFS disease mechanisms, in which a causal relationship between neuroendocrine changes and immune alterations is suggested [37]. This possible causality, as well as the association to CFS clinical symptoms and the specific role of altered B cell function, should be explored in further studies.

Abbreviations

CFS:

chronic fatigue syndrome

NorCAPITAL:

The Norwegian Study of Chronic Fatigue Syndrome in Adolescents: Pathophysiology and Intervention Trial

HPA axis:

hypothalamus–pituitary–adrenal axis

IL:

interleukin

TFN:

tumor necrosis factor

PBMC:

peripheral blood mononuclear cells

CRP:

C-reactive protein

HTS:

high throughput sequencing

RNA-Seq:

RNA sequencing

QC:

quality control

DEG:

differentially expressed genes

RT-qPCR:

real-time quantitative polymerase chain reaction

IPA:

ingenuity pathway analysis

PCA:

principal component analysis

HPLC:

high-performance liquid chromatography

HRV:

heart rate variability

LF:

low frequency

HF:

high frequency

RMSSD:

square root of the mean square difference of successive RR-intervals

CFQ:

Chalder Fatigue Questionnaire

MFQ:

Moods and Feelings Questionnaire

BCR:

B cell receptor

BAFFR:

B cell activating factor receptor

CTRA:

conserved transcriptional response to adversities

References

  1. Institute of Medicine. Beyond myalgic encephalomyelitis/chronic fatigue syndrome: redefining an illness. Washington, DC: The National Academies Press; 2015.

    Google Scholar 

  2. Royal College of Paediatrics and Child Health. Evidence based guidelines for the management of CFS/ME (chronic fatigue syndrome/myalgic encephalopathy) in children and young adults. London: Royal College of Paediatrics and Child Health; 2004.

    Google Scholar 

  3. Crawley EM, Emond AM, Sterne JA. Unidentified chronic fatigue syndrome/myalgic encephalomyelitis (CFS/ME) is a major cause of school absence: surveillance outcomes from school-based clinics. BMJ Open. 2011;1:e000252.

    Article  PubMed  PubMed Central  Google Scholar 

  4. Jason LA, Bell DS, Rowe K, et al. A pediatric case definition for chronic fatigue syndrome. J Chronic Fatigue Syndr. 2006;13:1–44.

    Article  Google Scholar 

  5. Nijhof SL, Maijer K, Bleijenberg G, Uiterwaal CS, Kimpen JL, van der Putte EM. Adolscent chronic fatigue syndrome: prevalence, incidence, and morbidity. Pediatrics. 2011;127:e1169–75.

    Article  PubMed  Google Scholar 

  6. Kennedy G, Underwood C, Belch JJ. Physical and functional impact of chronic fatigue syndrome/myalgic encephalomyelitis in childhood. Pediatrics. 2010;125:e1324–30.

    Article  PubMed  Google Scholar 

  7. Missen A, Hollingwort W, Eaton N, Crawley E. The financial and psychological impacts on mothers of children with chronic fatigue syndrome (CFS/ME). Child Care Health Dev. 2012;38:505–12.

    Article  CAS  PubMed  Google Scholar 

  8. Bansal AS, Bradley AS, Bishop KN, Kiani-Alikhan S, Ford B. Chronic fatigue syndrome, the immune system and viral infection. Brain Behav Immun. 2012;26:24–31.

    Article  CAS  PubMed  Google Scholar 

  9. Klimas NG, Broderick G, Fletcher MA. Biomarkers for chronic fatigue. Brain Behav Immun. 2012;26:1202–10.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  10. Raison CL, Lin JM, Reeves WC. Association of peripheral inflammatory markers with chronic fatigue in a population-based sample. Brain Behav Immun. 2009;23:327–37.

    Article  CAS  PubMed  Google Scholar 

  11. Fluge Ø, Bruland O, Risa K, Storstein A, Kristoffersen EK, Sapkota D, Næss H, Dahl O, Nyland H, Mella O. Benefit from B-lymphocyte depletion using the anti-CD20 antibody rituximab in chronic fatigue syndrome: a double-blind and placebo-controlled study. PLoS ONE. 2011;6:e26358.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  12. Maes M, Twisk FN, Ringel K. Inflammatory and cell-mediated immune biomarkers in myalgic encephalomyelitis/chronic fatigue syndrome and depression: inflammatory markers are higher in myalgic encephalomyelitis/chronic fatigue syndrome than in depression. Psychother Psychosom. 2012;81:286–95.

    Article  PubMed  Google Scholar 

  13. Fletcher MA, Zeng XR, Barnes Z, Levis S, Klimas NG. Plasmas cytokines in women with chronic fatigue syndrome. J Transl Med. 2009;7:96.

    Article  PubMed  PubMed Central  Google Scholar 

  14. Vollmer-Conna U, Cameron B, Pavlonic DH, Singletary K, Davenport T, Vernon S, Reeves WC, Hickie I, Wakenfield D, Lloyd AR. Postinfective fatigue syndrome is not associated with altered cytokine production. Clin Infect Dis. 2007;45:732–5.

    Article  CAS  PubMed  Google Scholar 

  15. Wyller VB, Sørensen Ø, Sulheim D, Fagermoen E, Ueland T, Mollnes TE. Plasma cytokine expression in adolescent chronic fatigue syndrome. Brain Behav Immun. 2015;46:80–6.

    Article  CAS  PubMed  Google Scholar 

  16. Kerr JR, Petty R, Burke B, Gough J, Fear D, Sinclair LI, Mattey DL, Richards SCM, Montgomery J, Baldwin DA, Kellam P, Harrison TJ, Griffin GE, Main J, Enlander D, Nutt DJ, Holgate ST. Gene expression subtypes in patients with chronic fatigue syndrome/myalgic encephalomyelitis. J Infect Dis. 2008;197:1171–84.

    Article  CAS  PubMed  Google Scholar 

  17. Zhang L, Gough J, Christmas D, Mattey DL, Richards SCM, Main K, Enlander D, Honeybourne D, Ayres JG, Nutt DJ, Kerr JR. Microbial infections in eight genomic subtypes of chronic fatigue syndrome/myalgic encephalomyelitis. J Clin Pathol. 2010;63:156–64.

    Article  CAS  PubMed  Google Scholar 

  18. Light AR, Bateman L, Jo D, Hughen RW, Vanhaitsma TA, White AT, Light KC. Gene expression alterations at baseline and following moderate exercise in patients with chronic fatigue syndrome and fibromyalgia syndrome. J Intern Med. 2012;271:64–81.

    Article  CAS  PubMed  Google Scholar 

  19. Nijs J, Nees A, Paul L, De Kooning M, Ickmans K, Meeus M, Van Oosterwijck J. Altered immune responses to exercise in patients with chronic fatigue syndrome/myalgic encephalomyelitis: a systematic literature review. Exerc Immunol Rev. 2014;20:94–116.

    PubMed  Google Scholar 

  20. Fang H, Xie Q, Boneva R, Fostel J, Perkins R, Tong W. Gene expression profile exploration of a large dataset on chronic fatigue syndrome. Pharmacogenomics. 2006;7:429–40.

    Article  CAS  PubMed  Google Scholar 

  21. Gow JW, Hagan S, Herzyk P, Cannon C, Behan PO, Chaudhuri A. A gene signature for post-infectious chronic fatigue syndrome. BMC Med Genom. 2009;2:38.

    Article  Google Scholar 

  22. Byrnes A, Jacks A, Dahlman-Wright K, et al. Gene expression in peripheral blood leukocytes in monozygotic twins discordant for chronic fatigue: no evidence of a biomarker. PLoS ONE. 2009;4:e5805.

    Article  PubMed  PubMed Central  Google Scholar 

  23. Frampton D, Kerr J, Harrison TJ, Kellam P. Assessment of a 44 gene classifier for the evaluation of chronic fatigue syndrome from peripheral blood mononuclear cell gene expression. PLoS ONE. 2011;6:e16872.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  24. Fostel J, Boneva R, Lloyd A. Exploration of the gene expression correlates of chronic unexplained fatigue using factor analyses. Pharmacogenomics. 2006;7:441–54.

    Article  CAS  PubMed  Google Scholar 

  25. Galbraith S, Cameron B, Li H, Lau D, Vollmer-Conna U, Lloyd AR. Peripheral blood gene expression in postinfective fatigue syndrome following from three different triggering infections. J Infect Dis. 2011;204:1632–40.

    Article  CAS  PubMed  Google Scholar 

  26. Martínez-Martínez LA, Mora T, Varqas A, Fuentes-Iniestra M, Martínez-Lavin M. Sympathetic nervous system dysfunction in fibromyalgia, chronic fatigue syndrome, irritable bowel syndrome, and interstitial cystitis: a review of case–control studies. J Clin Rhematol. 2014;20:146–50.

    Article  Google Scholar 

  27. Wyller VB, Barbieri R, Saul P. Blood pressure variability and closed-loop baroreflex assessment in adolescent chronic fatigue syndrome during supine rest and orthostatic stress. Eur J Appl Physiol. 2011;111:497–502.

    Article  PubMed  Google Scholar 

  28. Wyller VB, Barbieri R, Thaulow E, Saul JP. Enhanced vagal withdrawal during mild orthostatic stress in adolescents with chronic fatigue. Ann Noninvasive Electrocardiol. 2008;13:67–73.

    Article  PubMed  Google Scholar 

  29. Wyller VB, Due R, Saul JP, Amlie JP, Thaulow E. Usefulness of an abnormal cardiovascular response during low-grade head-up tilt-test for discriminating adolescents with chronic fatigue from healthy controls. Am J Cardiol. 2007;99:997–1001.

    Article  PubMed  Google Scholar 

  30. Papadopoulos AS, Cleare AJ. Hypothalamic–pituitary–adrenal axis dysfunction in chronic fatigue syndrome. Nat Rev Endocrinol. 2011;27:22–32.

    Article  Google Scholar 

  31. Segal TY, Hindmarsh PC, Viner RM. Disturbed adrenal function in adolescents with chronic fatigue syndrome. J Pediatr Endocrinol Metab. 2005;18:295–301.

    Article  CAS  PubMed  Google Scholar 

  32. Wyller VB, Vitelli V, Sulheim D, Fagermoen E, Winger A, Godang K, Bollerslev J. Altered neuroendocrine control and association to clinical symptoms in adolescent chronic fatigue syndrome: a cross-sectional study. J Transl Med. 2016;14:121.

    Article  PubMed  PubMed Central  Google Scholar 

  33. Zen M, Canova M, Campana C, Bettio S, Nalotto L, Rampudda M, Ramonda R, Iaccarino L, Doria A. The kaleidoscope of glucocorticoid effects on immune system. Autoimmun Rev. 2011;10:305–10.

    Article  CAS  PubMed  Google Scholar 

  34. Andersson U, Tracey KJ. Neural reflexes in inflammation and immunity. J Exp Med. 2012;209:1057–68.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  35. Padro CJ, Sanders VM. Neuroendocrine regulation of inflammation. Semin Immunol. 2014;26:357–68.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  36. Thayer JF, Sternberg EM. Neural aspects of immunomodulation: focus on the vagus nerve. Brain Behav Immun. 2010;24:1223–8.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  37. Wyller VB, Malterud K, Eriksen HR. Can sustained arousal explain chronic fatigue syndrome? Behav Brain Funct. 2009;5:10.

    Article  PubMed  PubMed Central  Google Scholar 

  38. Cividjian A, Toader E, Wesseling KH, Karemaker JM, McAllen R, Quintin L. Effect of clonidine on cardiac baroreflex delay in humans and rats. Am J Physiol Regul Integr Comp Physiol. 2011;300:949–57.

    Article  Google Scholar 

  39. Sulheim D, Fargermoen E, Winger A, Andersen AM, Godang K, Müller F, Rowe PC, Saul JP, Skovlund E, Øie MG, Wyller VB. Disease mechanisms and clonidine treatment in adolescent chronic fatigue syndrome: a combined cross-sectional and randomized controlled trial. JAMA Pediatr. 2014;168:351–60.

    Article  PubMed  Google Scholar 

  40. National Institute for Health and Clinical Excellence: chronic fatigue syndrome/myalgic encephalomyelitis (or encephalopathy). Diagnosis and management of CFS/ME in adults and children. London: NICE clinical guideline; 2007, no. 53.

  41. Babraham Bioinformatics, Babraham Institute, Cambridge, UK. http://www.bioinformatics.babraham.ac.uk/projects/fastqc/. Accessed 23 Feb 2017.

  42. Hannon Lab, Cold Spring Harbor Laboratory, New York, USA. http://www.hannonlab.cshl.edu. Accessed 23 Feb 2017.

  43. Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, Batut P, Chaisson M, Gingeras TR. STAR: ultrafast universal RNA-seq aligner. Bioinformatics. 2013;29:15–21.

    Article  CAS  PubMed  Google Scholar 

  44. Gentleman RC, Carey VJ, Bates DM, Bolstad B, Dettling M, Dudoit S, et al. Bioconductor: open software development for computational biology and bioinformatics. Genome Biol. 2004;5:R80.

    Article  PubMed  PubMed Central  Google Scholar 

  45. Liao Y, Smyth GK, Shi W. The Subread aligner: fast, accurate and scalable read mapping by seed-and-vote. Nucleic Acids Res. 2013;41:e108.

    Article  PubMed  PubMed Central  Google Scholar 

  46. Risso D, Nqai J, Speed TP, Dudoit S. Normalization of RNA-seq data using factor analyses of control genes or samples. Nat Biotechnol. 2014;32:896–902.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  47. Anders S, McCarthy DJ, Chen Y, Okoniewsky M, Smyth GK, Huber W, Robinson MD. Count-based differential expression analysis of RNA sequencing data using R and bioconductor. Nat Protoc. 2013;8:1765–86.

    Article  PubMed  Google Scholar 

  48. Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15:550.

    Article  PubMed  PubMed Central  Google Scholar 

  49. HumanMine Database, University of Cambridge, UK. http://www.humanmine.org. Accessed 23 Feb 2017.

  50. Bindea G, Mlecnik B, Hackl H, Charoentong P, Tosolini M, Kirilovsky A, Fridman WH, Pages F, Trajanoski Z, Galon J. ClueGO: a cytoscape plug-into decipher functionally grouped gene ontology and pathway annotation networks. Bioinformatics. 2009;25:1091–3.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  51. Aspler AL, Bolshin C, Vernon SD, Broderick G. Evidence of inflammatory immune signaling in chronic fatigue syndrome: a pilot study of gene expression in peripheral blood. Behav Brain Funct. 2008;4:44.

    Article  PubMed  PubMed Central  Google Scholar 

  52. Preininger M, Arafat D, Kim J, Nath AP, Idaghdour Y, Brigham KL, Gibson G. Blood-informative transcripts define nine common axes of peripheral blood gene expression. PLoS Genet. 2013;9:e1003362.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  53. Fagermoen E, Sulheim D, Winger A, Andersen AM, Gjerstad J, Godang K, Rowe PC, Saul JP, Skovlund E, Wyller VB. Effects of low-dose clonidine on cardiovascular and autonomic variables in adolescents with chronic fatigue syndrome: a randomized controlled trial. BMC Pediatr. 2015;15:117.

    Article  PubMed  PubMed Central  Google Scholar 

  54. Task force of the European society of cardiology, and the North American society of pacing electrophysiology. Heart rate variability. Standards of measurement, physiological interpretation, and clinical use. Circulation. 1996;93:1043–65.

    Article  Google Scholar 

  55. Chalder T, Berelowitz G, Pawlikowska T, et al. Development of a fatigue scale. J Psychosom Res. 1993;37:147–53.

    Article  CAS  PubMed  Google Scholar 

  56. Daviss WB, Birmaher B, Melhem NA, Axelson DA, Michaels SM, Brent DA. Criterion validity of the Mood and Feelings Questionnaire for depressive episodes in clinic and non-clinic subjects. J Child Psychol Psychiatry. 2006;47:927–34.

    Article  PubMed  Google Scholar 

  57. Kendall PC, Finch AJ Jr, Auerback SM, Hooke JF, Mikulka PJ. The state-trait anxiety inventory: a systematic evaluation. J Consult Clin Psychol. 1976;44:406–12.

    Article  CAS  PubMed  Google Scholar 

  58. Grant PM, Ryan CG, Tigbe WW, Granat MH. The validation of a novel activity monitor in the measurement of posture and motion during everyday activities. Br J Sports Med. 2006;40:992–7.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  59. Mortazavi A, Williams BA, McCue K, Schaeffer L, Wold B. Mapping and quantifying mammalian transcriptomes by RNA-Seq. Nat Methods. 2008;5:621–8.

    Article  CAS  PubMed  Google Scholar 

  60. Ninq S, Pagano JS. The A20 deubiquitinase activity negatively regulates LPM1 activation of IRF7. J Virol. 2010;84:6130–8.

    Article  Google Scholar 

  61. Griffith J, et al. The structural basis for autoinhibition of FLT3 by the juxtamembrane domain. Mol Cell. 2004;13:169–78.

    Article  CAS  PubMed  Google Scholar 

  62. Vilagos B, et al. Essential role of EBF1 in the generation and function of distinct mature B cell types. J Exp Med. 2012;209:775–92.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  63. Chu PG, Arber DA. CD79: a review. Appl Immunohistochem Mol Morphol. 2001;9:97–106.

    Article  CAS  PubMed  Google Scholar 

  64. Pereira JP, et al. Finding the right niche: B-cell migration in the early phases of T-dependent antibody responses. Int Immunol. 2010;22:413–9.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  65. Kayagaki N, et al. BAFF/BLyS receptor 3 binds the B cell survival factor BAFF ligand through a discrete surface loop and promotes processing of NF-κB2. Immunity. 2002;17:515–24.

    Article  CAS  PubMed  Google Scholar 

  66. Stenglein MD, et al. APOBEC3 proteins mediate the clearance of foreign DNA from human cells. Nat Struct Mol Biol. 2010;17:222–9.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  67. Dong B, et al. Phospholipid scramblase 1 potentiates the antiviral activity of interferon. J Virol. 2004;78:8983–93.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  68. Blaszczyk K, Nowicka H, Kostyrko K, Antonczyk A, Wesoly J, Bluyssen HA. The unique role of STAT2 in constitutive and INF-induced transcription and antiviral responses. Cytokine Growth Factor Rev. 2016;29:71–81.

    Article  CAS  PubMed  Google Scholar 

  69. Cervantes JL, et al. TLR8: the forgotten relative revindicated. Cell Mol Immunol. 2012;9:434–8.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  70. Sun Q, Scott MJ. Caspase-1 as a multifunctional inflammatory mediator: noncytokine maturation roles. J Leukoc Biol. 2016;100:961–7.

    Article  CAS  PubMed  Google Scholar 

  71. Welte S, Kuttruff S, Waldhauer I, Steinle A. Mutual activation of natural killer cells and monocytes mediated by NKp80-AICL interaction. Nat Immunol. 2006;7:1334–42.

    Article  CAS  PubMed  Google Scholar 

  72. Bender AT, Beavo JA. PDE1B2 regulates cGMP and a subset of the phenotypic characteristics acquired upon macrophage differentiation from a monocyte. Proc Natl Acad Sci USA. 2006;103:460–5.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  73. Lunde S, Kristoffersen EK, Sapkota D, Risa K, Dahl O, Bruland O, Mella O, Fluge Ø. Serum BAFF and APRIL levels, T-lymphocyte subsets, and immunoglobulins after B-cell depletion using the monoclonal anti-CD20 antibody rituximab in myalgic encephalopathy/chronic fatigue syndrome. PLoS ONE. 2016;11:e0161226.

    Article  PubMed  PubMed Central  Google Scholar 

  74. Brenu EW, Huth TK, Hardcastle SL, Fuller K, Kaur M, Johnston S, Ramos SB, Staines DR, Marshall-Gradisnik SM. Role of adaptive and innate immune cells in chronic fatigue syndrome/myalgic encephalomyelitis. Int Immunol. 2014;26:233–42.

    Article  CAS  PubMed  Google Scholar 

  75. Bradley AS, Ford B, Bansal AS. Altered functional B cell subset populations in patients with chronic fatigue syndrome compared to healthy controls. Clin Exp Immunol. 2013;172:73.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  76. Mensah F, Bansal A, Berkovitz S, Sharma A, Reddy V, Leandro MJ, Cambridge G. Extended B cell phenotype in patients with myalgic encephalomyelitis/chronic fatigue syndrome: a cross-sectional study. Clin Exp Immunol. 2016;184:237–47.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  77. Guenther S, Loebel M, Mooslechner AA, Knops M, Hanitsch LG, Grabowski P, Wittke K, Meisel C, Unterwalder N, Volk HD, Scheibenbogen C. Frequent IgG subclass and mannose binding lectin deficiency in patients with chronic fatigue syndrome. Hum Immunol. 2015;76:729–35.

    Article  CAS  PubMed  Google Scholar 

  78. Levit-Zerdoun E, Becker M, Pohlmeyer R, Wilhelm I, Maity PC, Rajewsky K, Reth M, Hobeika E. Survival of Igα-deficient mature B cells requires BAFF-R function. J Immunol. 2016;196:2348–60.

    Article  CAS  PubMed  Google Scholar 

  79. Morris G, Berk M, Walder K, Maes M. The putative role of viruses, bacteria, and chronic fungal biotoxin exposure in the genesis of intractable fatigue accompanied by cognitive and physical disability. Mol Neurobiol. 2016;53:2550–71.

    Article  CAS  PubMed  Google Scholar 

  80. Katz BZ, Shiraishi Y, Mears CJ, et al. Chronic fatigue syndrome after infectious mononucleosis in adolescents. Pediatrics. 2009;124:189–93.

    Article  PubMed  PubMed Central  Google Scholar 

  81. Thorley-Lawson DA. EBV persistence—introducing the virus. Curr Top Microbiol Immunol. 2015;390:151–209.

    CAS  PubMed  PubMed Central  Google Scholar 

  82. Loebel M, Strohschein K, Gianni C, Koelsch U, Bauer S, Doebis C, Thomas S, Unterwalder N, von Baehr V, Reinke P, Knops M, Hanitsch LG, Meisel C, Volk HD, Schneibenbogen C. Deficient EBV-specific B- and T-cell responses in patients with chronic fatigue syndrome. PLoS ONE. 2014;9:e85387.

    Article  PubMed  PubMed Central  Google Scholar 

  83. Broderick G, Fuite J, Kreitz A, Vernon SD, Klimas N, Fletcher MA. A formal analysis of cytokine networks in chronic fatigue syndrome. Brain Behav Immun. 2010;24:1209–17.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  84. Chacko A, Staines DR, Johnston SC, Marshall-Gradisnik SM. Dysregulation of protein kinase gene expression in NK cells from chronic fatigue syndrome/myalgic encephalomyelitis patients. Gene Regul Syst Biol. 2016;10:85–93.

    Google Scholar 

  85. Nijhof SL, Rutten JM, Uiterwaal CS, Bleijenberg G, Kimpen JL, Putte EM. The role of hypocortisolism in chronic fatigue syndrome. Psychoneuroendocrinology. 2014;42:199–206.

    Article  CAS  PubMed  Google Scholar 

  86. Cole SW. Human social genomics. PLoS Genet. 2014;10:e1004601.

    Article  PubMed  PubMed Central  Google Scholar 

  87. Cole SW, Capitano JP, Chun K, Arevalo JM, Ma J, Cacioppo JT. Myeloid differentiation architecture of leukocyte transcriptome dynamics in perceived social isolation. Proc Natl Acad Sci USA. 2015;112:15142–7.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  88. Fukuda K, Straus SE, Hickie I, Sharpe MC, Dobbins JG, Komaroff A. The chronic fatigue syndrome: a comprehensive approach to its definition and study. Ann Intern Med. 1994;121:953–9.

    Article  CAS  PubMed  Google Scholar 

  89. Carruthers BM, Jain AK, De Meirleir KL, Peterson DL, Klimas NG, Lerner AM, et al. Myalgic encephalomyelitis/chronic fatigue syndrome: clinical working case definition, diagnostic and treatment protocols. J Chronic Fatigue Syndr. 2003;11:7–11.

    Article  Google Scholar 

Download references

Authors’ contributions

Conceived and designed the study: VBW, HN. Collected clinical data: DS, EF, AW. Analyzed the data: CBN, LA, HN, JL, MK, VBW. Interpreted the results and wrote the paper: CBN, LA, HN, JL, DS; EF, AW, MK, VBW. All authors read and approved the final manuscript.

Acknowledgements

We thank Kari Gjersum for secretary assistance; Solveig Mjelstad Olafsrud, Hatice Zeynep Nenseth, Kristin Godang, Kristine Jacobsenand Mai Britt Dahl for laboratory assistance; Sudrendra Kumar, Susanne Lorenz, Sumana Kalyansundaram and Ståle Nygård for bioinformatics support; Johannes Gjerstad, Leonardo Meza-Zepeda, Tom Eirik Mollnes and Fahri Saatcioglu for discussions on study design and results.

Competing interests

The authors declare that that they have no competing interests.

Availability of data and materials

The dataset generated and analysed during the current study is available in the Gene Expression Omnibus (GEO) repository, reference number GSE98139, web link http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE98139.

Ethics approval and consent to participate

The study was approved by the Norwegian National Committee for Ethics in Medical Research and the Norwegian Medicines Agency. Written informed consent was obtained from all participants and from parents/next-of-kin if required.

Funding

This study was funded by: The University of Oslo, The Research Council of Norway, The Southern and Eastern Norway Regional Health Authority. The funding has covered payroll and operating expenses for the involved researchers. No one of the funding sources has otherwise been involved in the study. The study has not received funding from the pharmaceutical industry or other commercial sources.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Author information

Authors and Affiliations

  1. Department of Paediatrics and Adolescent Health, Akershus University Hospital, 1478, Lørenskog, Norway

    Chinh Bkrong Nguyen & Vegard Bruun Wyller

  2. Division of Medicine and Laboratory Sciences, Medical Faculty, University of Oslo, Oslo, Norway

    Chinh Bkrong Nguyen & Vegard Bruun Wyller

  3. Institute of Clinical Medicine, Department of Clinical Molecular Biology, University of Oslo, and Akershus University Hospital, Lørenskog, Norway

    Lene Alsøe & Hilde Nilsen

  4. National Bioinformatics Infrastructure Sweden (NBIS), Science for Life Laboratory, Department of Biochemistry and Biophysics, Stockholm University, Stockholm, Sweden

    Jessica M. Lindvall

  5. Department of Paediatrics, Lillehammer County Hospital, Lillehammer, Norway

    Dag Sulheim

  6. Department of Anesthesiology and Critical Care, Oslo University Hospital, Oslo, Norway

    Even Fagermoen

  7. Institute of Nursing Sciences, Oslo and Akershus University College of Applied Sciences, Oslo, Norway

    Anette Winger

  8. Department of Microbiology, Oslo University Hospital, Oslo, Norway

    Mari Kaarbø

Authors
  1. Chinh Bkrong Nguyen
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Lene Alsøe
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Jessica M. Lindvall
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Dag Sulheim
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Even Fagermoen
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Anette Winger
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Mari Kaarbø
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Hilde Nilsen
    View author publications

    You can also search for this author in PubMed Google Scholar

  9. Vegard Bruun Wyller
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Vegard Bruun Wyller.

Additional files

Additional file 1: Table S1. Primer names and sequences for the RT-qPCR experiments.

12967_2017_1201_MOESM2_ESM.xlsx

Additional file 2: Table S2. Background characteristics of the chronic fatigue syndrome (CFS) group and the healthy control (HC) group in the present study compared with the characteristics of all CFS patients and HC in the NorCAPITAL dataset.

12967_2017_1201_MOESM3_ESM.xlsx

Additional file 3: Table S3. Differentially expressed genes and their annotated proteins or gene products in CFS patients as compared to healthy controls, adjusted for age and gender differences across groups and sorted according to foldchange.

12967_2017_1201_MOESM4_ESM.xlsx

Additional file 4: Table S4. Upstream transcriptional regulators for the observed dataset based on ingenuity pathway analyses (IPA) mechanistic network enrichment.

Additional file 5: Table S5. Principal component analysis (PCA) with varimax rotation in the CFS group.

12967_2017_1201_MOESM6_ESM.xlsx

Additional file 6: Table S6. Pearson correlation between single gene transcriptional counts and selected immune, neuroendocrine and clinical markers within the CFS group. Genes are sorted according to differential expression foldchange (column 2) as compared with healthy controls.

Rights and permissions

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Nguyen, C.B., Alsøe, L., Lindvall, J.M. et al. Whole blood gene expression in adolescent chronic fatigue syndrome: an exploratory cross-sectional study suggesting altered B cell differentiation and survival. J Transl Med 15, 102 (2017). https://doi.org/10.1186/s12967-017-1201-0

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12967-017-1201-0

Keywords

Journal of Translational Medicine

ISSN: 1479-5876

Contact us