Giloteaux et al. J Transl Med (2020) 18:387 https://doi.org/10.1186/s12967-020-02560-0 Journal of Translational Medicine RESEARCH Open Access Cytokine profiling of extracellular vesicles isolated from plasma in myalgic encephalomyelitis/chronic fatigue syndrome: a pilot study Ludovic Giloteaux1, Adam O’Neal1, Jesús Castro‑Marrero1,2, Susan M. Levine3 and Maureen R. Hanson1* Abstract Background: Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS) is a debilitating disease of unknown etiology lasting for a minimum of 6 months but usually for many years, with features including fatigue, cognitive impairment, myalgias, post-exertional malaise, and immune system dysfunction. Dysregulation of cytokine signaling could give rise to many of these symptoms. Cytokines are present in both plasma and extracellular vesicles, but little investigation of EVs in ME/CFS has been reported. Therefore, we aimed to characterize the content of extracellular vesicles (EVs) isolated from plasma (including circulating cytokine/chemokine profiling) from individuals with ME/CFS and healthy controls. Methods: We included 35 ME/CFS patients and 35 controls matched for age, sex and BMI. EVs were enriched from plasma by using a polymer-based precipitation method and characterized by Nanoparticle Tracking Analysis (NTA), Transmission Electron Microscopy (TEM) and immunoblotting. A 45-plex immunoassay was used to determine cytokine levels in both plasma and isolated EVs from a subset of 19 patients and controls. Linear regression, principal component analysis and inter-cytokine correlations were analyzed. Results: ME/CFS individuals had significantly higher levels of EVs that ranged from 30 to 130 nm in size as compared to controls, but the mean size for total extracellular vesicles did not differ between groups. The enrichment of typical EV markers CD63, CD81, TSG101 and HSP70 was confirmed by Western blot analysis and the morphology assessed by TEM showed a homogeneous population of vesicles in both groups. Comparison of cytokine concentrations in plasma and isolated EVs of cases and controls yielded no significant differences. Cytokine-cytokine correlations in plasma revealed a significant higher number of interactions in ME/CFS cases along with 13 inverse correlations that were mainly driven by the Interferon gamma-induced protein 10 (IP-10), whereas in the plasma of controls, no inverse relationships were found across any of the cytokines. Network analysis in EVs from controls showed 2.5 times more significant inter-cytokine interactions than in the ME/CFS group, and both groups presented a unique negative association. Conclusions: Elevated levels of 30-130 nm EVs were found in plasma from ME/CFS patients and inter-cytokine cor‑ relations revealed unusual regulatory relationships among cytokines in the ME/CFS group that were different from *Correspondence: mrh5@cornell.edu 1 Department of Molecular Biology and Genetics, Cornell University, 323 Biotechnology Building, 526 Campus Road, Ithaca, NY 14853, USA Full list of author information is available at the end of the article © The Author(s) 2020. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat​iveco​ mmons​.org/licen​ses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creat​iveco​mmons​.org/publi​cdoma​in/ zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Giloteaux et al. J Transl Med (2020) 18:387 Page 2 of 17 the control group in both plasma and EVs. These disturbances in cytokine networks are further evidence of immune dysregulation in ME/CFS. Keywords: Myalgic Encephalomyelitis/Chronic Fatigue Syndrome, Extracellular vesicles, Plasma, Cytokines Background passively through the bloodstream to sites distant to the Myalgic Encephalomyelitis/Chronic Fatigue Syndrome local inflammatory lesion. A recent study reported that (ME/CFS) is a life-limiting illness characterized by per- EV-associated and encapsulated cytokines were more sistent debilitating fatigue, cognitive dysfunction often stable than free cytokines and became biologically active referred to as “brain fog”, pain/myalgias, post-exertional upon interacting with sensitive cells [16]. malaise, autonomic disturbances and gastrointestinal and There is increased evidence that EVs are one of the immune system dysfunction lasting for at least 6 months, main participants in cell-to-cell communication [18] and but usually much longer as prognosis for recovery is drive inflammatory, autoimmune and infectious disease poor [1]. This symptomatology has led to the suspicion pathology [19–22]. They can be sources of proinflamma- that ME/CFS is an inflammatory disorder; therefore, tory cytokines such as Tumor Necrosis Factor-α (TNF- researchers have been investigating potential biomark- α), Interleukins -6, -1β, -8 (IL-6, IL-1β, IL-8), Monocyte ers including, oxidative stress [2, 3], energy metabolism Chemoattractant Protein-1 (MCP-1), can stimulate their [4, 5], or cytokine profiling (reviewed in [6–8]). Dysregu- production in a variety of cells, and can also induce acti- lation of cytokine profiles has been demonstrated and vation and proliferation of B and T cells as well as migra- often associated with an enhancement of pro-inflamma- tion of granulocyte into inflamed tissues, promoting tory cytokines in patients with ME/CFS [9–13]. However, inflammatory pathways in recipient cells. results are frequently inconsistent between studies and it In previous reports, circulating EVs have been shown remains difficult to determine a specific cytokine signa- to be increased in number in ME/CFS [23–25], results ture that could be implicated in the etiopathogenesis of in line with other diseases such as cancers, Alzheimer’s ME/CFS. and Parkinson’s disease [21, 26–28]. Furthermore, a Cytokines are important modulators of immune func- recent study was able to distinguish ME/CFS from idi- tion and inflammatory responses and are easily affected opathic chronic fatigue and clinical depression by ana- by several factors which could be responsible for the high lyzing the content of EVs by means of proteomics [25]. heterogeneity observed among studies. They are gener- Observations in that study were an upregulation of pro- ally considered to function as soluble factors that mediate teins involved in focal adhesion, regulation of the actin cell-to-cell communications in multicellular organisms. cytoskeletal signaling pathway and pathways relevant to Their secretion can occur in classical secretion manner Epstein-Barr virus infection. Another report showed that or via extracellular vesicles (EVs). neuronal and endocrine system pathways were among EVs are heterogeneous membrane-enclosed struc- differentially expressed miRNA in EVs in ME/CFS, which tures, spherical in shape with different origins, size and suggests that there are defects in endocrine tissue func- composition and released by cells into the extracellular tioning [23]. milieu. A wide variety of EV subtypes have been pro- Unlike our pilot study described below, the prior posed, defined by their size, biogenesis pathway, cargo, studies did not report on the cytokine composition of cellular source, and biological function, leading to a his- extracellular vesicles isolated from plasma of ME/CFS torically heterogeneous nomenclature including terms patients. We utilized an immune-profiling approach to like exosomes and ectosomes. They may be released from determine whether an abnormal set of EV-associated damaged or stressed cells and carry proteins, pathogen- cytokines could be identified in ME/CFS. associated and damage-associated molecular patterns, lipids and miRNA, which are encapsulated and thus Materials and methods protected from degrading enzymes [14]. Cytokines may Participants be selectively sorted to EVs and act as signals to regulate Cases and controls were recruited by Susan Levine, M.D. and propagate the immune and inflammatory response (Manhattan, NY) between November 19, 2013 and Octo- [15]. EV-associated cytokines are different from those ber 21, 2014. A total of 35 ME/CFS cases and 35 healthy released in a soluble form by cells of the same type [16]. controls were included in this age- and sex-matched Once secreted, EVs can be rapidly captured by dendritic case–control cross-sectional study. Patients were diag- cells, phagocytes or macrophages [17], bind neighboring nosed with ME/CFS if they met the 1994 CDC/Fukuda cells that express specific cytokine receptors, or circulate definition [29] and controls were eligible if they did not Giloteaux et al. J Transl Med (2020) 18:387 Page 3 of 17 have history of fatigue and did not meet the ME/CFS Table 1 Study population characteristics case definition. Individuals with a known acute illness or ME/CFS Controls Mann– chronic infectious disease were excluded from the study. Whitney U A number of subjects were taking psychoactive medica- test tions and were not asked to stop them to be included in Age (years) 50.6 ± 13.7 50.2 ± 17.5 p = 0.42 the study. Gender Non-fasting peripheral blood was drawn in EDTA Female 28 28 NA tubes in the morning and was stored on the same day Male 7 7 NA of collection at −80 °C until further processing. Par- BMI (kg/m2) 24.3 ± 5.3 24.3 ± 4.6 p = 0.18 ticipants’ age, sex, and age of onset of ME/CFS were Onset of disease recorded. The Bell’s disability scale [30], Short Form-36 Gradual 40% NA NA Health Survey [31] and Symptom Severity Scale (SSS) Sudden 60% NA NA were administered to each participant on the day of Bell score 35.4 ± 14.1 96.3 ± 9.5 p < 0.001 blood sample collection. A higher score on the SSS indi- SF-36 cates greater severity. As indicated in Table 1, mean of Physical function 44.1 ± 17.2 95.9 ± 7.3 p < 0.001 ages and BMI of cases and controls were within 3 years Role physical 5.0 ± 11.8 91.4 ± 26.1 p < 0.001 and the female to male ratios were equal. Written con- Pain 40.8 ± 21.6 87.3 ± 14.1 p < 0.001 sent was obtained from all participants and all protocols General health 25.9 ± 14.5 87.3 ± 15.9 p < 0.001 were approved by the Cornell University Institutional Vitality 17.9 ± 15.8 75.4 ± 17.8 p < 0.001 Review Board, approval # 1303003741. PCSa 26.8 ± 7.5 55.1 ± 5.2 p < 0.001 Social function 33.6 ± 22.6 94.8 ± 12.3 p < 0.001 Extracellular vesicle isolation and characterization Role emotional 45.7 ± 45.1 91.9 ± 24.6 p < 0.001 750 μl of plasma by precipitation using the ExoQuick™ Total extracellular vesicles (EVs) were isolated from Mental health 59.3 ± 19.8 79.7 ± 18.6 p < 0.001 MCSa 38.9 ± 10.9 53.4 ± 9.6 p < 0.001 reagent (System Biosciences, Palo Alto, CA, USA). SSSa Plasma samples from each subject were thawed on ice Fatigue 6.9 ± 2.5 1.4 ± 1.7 p < 0.001 and centrifuged at 3000×g for 15 min at room temper- Impaired memory 6.4 ± 2.8 0.8 ± 1.4 p < 0.001 ature to remove cells and debris. The supernatant was Recurrent sore throat 3.9 ± 3.3 0.4 ± 1.1 p < 0.001 transferred to a new tube, and thrombin (611 U/ml) Tender lymph nodes 3.9 ± 3.1 0.1 ± 0.4 p < 0.001 (System Bioscience, Palo Alto, CA, USA) was added and Muscle tenderness or pain 5.5 ± 2.9 1.5 ± 1.9 p < 0.001 samples were incubated for 5 min at room temperature Joint pain 5.1 ± 2.9 1.5 ± 2.0 p < 0.001 to remove fibrinogen, centrifuged at 10,000×g for 5 min, Headache 5.0 ± 3.4 0.7 ± 1.4 p < 0.001 then incubated with ExoQuick™ for 60 min at 4 °C. The and the supernatant was collected. The samples were Disturbed sleep or waking 7.6 ± 3.0 1.6 ± 2.1 p < 0.001 unrefreshed ExoQuick™/serum-like samples were then centrifuged at Post-exertional malaise 7.8 ± 2.7 0.8 ± 1.7 p < 0.001 a 12,000×g for 5 min, and the resulting pellet was resus- PCS Physical component score, MCS Mental component score, SSS Specific symptom severity (0 = none, 10 = very severe) pended in 250 ul of sterile phosphate buffered saline 1X, pH 7.4. To prevent aggregation and cryodamage, 25 mM of trehalose was added to the isolated EV fraction [32]. Purified EVs were assayed for Western blot analysis Samples were aliquoted for total protein determination, (WB). WB is used to validate the presence or absence Western blot analysis, Nanoparticle Tracking Analysis of EV protein markers in purified samples based on the (NTA), Transmission Electron Microscopy (TEM), and availability of specific antibodies CD63, CD81, HSP70 measurement (quantification) of cytokines/chemokines and TSG101 (System Biosciences, LLC, Palo Alto, CA, and growth factors. USA). Protein samples were prepared by adding 100 μl ice-cold RIPA buffer containing protease/phosphatase inhibitors to 100 μl extracted EV samples resuspended Protein quantification and western blot analysis in the appropriate buffer. EV lysates were adjusted to the Total protein quantification from EV isolates was per- same protein content (150 μg), denatured for 10 min in formed using the P ­ierceTM BCA Protein Assay kit 2X Laemmli buffer, resolved by 8–10% SDS-PAGE, and (ThermoFisher Scientific) according to manufacturer’s then proteins were transferred to PVDF membranes instructions. The assay is a detergent-compatible formu- (Amersham, GE Healthcare, USA). Membranes were lation based on bicinchoninic acid for the colorimetric blocked in 5% non-fat dry milk using TBS containing detection (A562 nm) and quantitation of total protein. Giloteaux et al. J Transl Med (2020) 18:387 Page 4 of 17 0.1% Tween-20 and then incubated with various pri- on a M­ AGPIX® Multiplexing System (Luminex Corp.) mary polyclonal antibodies (anti-CD63, anti-CD81, anti- at the Human Nutritional Chemistry Service Laboratory HSP70, anti-TSG101; dilution 1/500; Santa Cruz Biotech, at Cornell University. For each well, we used the median CA, USA) and cytochrome C antibody as a negative con- fluorescence intensity (MFI) of all beads measured for a trol for overnight at 4 °C, washed with 1X TBS-T and given analyte and averaged the MFI of the two replicates then incubated with secondary conjugated antibody for and results were accepted when the coefficient of varia- 1 h at room temperature. X-ray films were exposed in a tion (CV) was below 15%. MFIs for analytes are best for darkroom and films were developed and visualized using analysis because fluorescence enables the analysis of low and infrared Odyssey machine (LI-COR Biosciences). signals and have power for testing differences in analyte expression [33–35]. Nanoparticle tracking analysis Extracellular vesicles’ concentration and size distribu- tion were assayed in samples using a NanoSight NS300 Statistical analysis (Malvern). Samples were thawed and diluted to 1:2000 in All data was processed and analyzed using R version 3.5.1 PBS 1X and 1 ml was injected through the laser chamber (2018-07-02) via RStudio Version 1.2.5033. The inde- (NanoSight Technology, London, UK). The NanoSight pendent samples t test and non-parametric Wilcoxon- NS300 uses a source light to illuminate nanoscale par- Mann–Whitney U test were used to determine the ticles (30–800 nm) as point scatters moving under significance of differences (p < 0.05) in each subject group Brownian motion. Three recordings of 60-second digital for age, BMI, EV size and concentrations. A linear model videos of each sample were acquired and analyzed by the analysis was used to compare cytokine measurements NanoSight NTA 2.3 software to determine the size and from plasma and EVs between groups with age and BMI the concentration of nanoparticles. Results were aver- as confounding variables. Principal Component Analysis aged together. (PCA) was used to sum up and to simplify the data by reducing the dimensionality of the cytokine levels data- Transmission electron microscopy sets. Variables were first “mean-centered” and scaled to EV suspensions were visualized under a 120 kV field have standard deviation one. Then, the data correlation emission transmission electron microscope (FEI T12 matrix was calculated and eigenvalue decomposition on Spirit TEM/STEM) at the Cornell Center for Materi- the matrix was performed. Cluster tendency of the data- als Research in Ithaca, NY. Isolated EV suspensions sets was assessed using the Hopkins statistic (H) [36] by were thawed and diluted at either 1:100, 1:500, 1:1000 measuring the probability that a given data set is gener- or 1:2000 in 2% PFA overnight at 4 °C before proceeding ated by a uniform data distribution. A H value close to 1 with negative staining. Samples were applied to copper tends to indicate the data is highly clustered, and random 300-mesh Formvar coated carbon stabilized grids and data will tend to result in values close or below 0.5. PCA were allowed to adsorb to the grid for 20 min. Grids were and cluster tendency were performed using the R pack- then washed in PBS 1X and transferred to a 1% glutaral- ages FactoMineR, factoextr and clustertend. dehyde solution for one min post-fixation. Samples were Spearman’s rank correlation coefficients were also esti- then washed 8 times by floating on distilled water for mated between each cytokine and the metadata (age, 2 min. Negative staining was then achieved through plac- BMI, Bell, SF-36 and Symptom Severity Scale scores) ing the grids on a drop of 2% Aqueous Uranyl Acetate for and partial Spearman’s rank correlation were performed 10 min followed by air drying and storage in an EM grid adjusting for age and BMI. Post hoc tests were run to box. assess the differences between ME/CFS and controls using the R emmeans package. Throughout, all p-values Cytokine, chemokine and growth factor measurement were corrected using a Benjamini and Yekutieli cor- in EV and plasma samples rection to control the false discovery rate (FDR) under Due to cost considerations for cytokine assays, a subset dependency of the cytokines with a 0.05 family-wise of 19 females with ME/CFS and 19 healthy females out False Discovery Rate [37]. The relationships among the of the 70 initial samples, age and BMI-matched, were analytes were analyzed within each group allowing the used. Both purified EVs and plasma from every subject discovery of different cytokine-cytokine structures across were analyzed for cytokines, chemokines and growth the different group populations. Significantly correlated factors using a human 45-plex magnetic bead kit (R&D cytokine pairs were displayed on a network diagram Systems, Minneapolis, USA). EV samples were treated using igraph, limma and ggplot2 R packages, in which with Triton 1% to allow the release of encapsulated cytokines are represented by nodes and significantly cor- cytokines [16]. Each sample was measured in duplicate related cytokines are connected by edges. Giloteaux et al. J Transl Med (2020) 18:387 Page 5 of 17 Results Although the mean total concentration of particles/ Cohort characteristics ml of plasma (controls: 6.6 ± 4.3 ­x109, ME/CFS: 7.6 ± 5.0 Within the study population, there were 28 females and x 109; p = 0.26, Fig. 1b) and the mean concentration of 7 males in both the ME/CFS and healthy controls (HC) particles greater than 130 nm (controls: 5.0 ± 4.2 ­x108; groups (Table 1). All patients who were selected met ME/CFS: 7.1 ± 5.0 x ­108, p = 0.17, Fig. 1d) did not exhibit the 1994 Fukuda definition [29] for ME/CFS. The aver- a statistically significant difference between groups, age age was similar between both groups at 50.2 ± 17.5 the mean concentration of EVs that ranged from 30 to in controls and 50.6 ± 13.7 in patients (p = 0.42, Table 1). ­ 08 130 nm in size was statistically different (3.8 ± 2.1 x 1 8 Average Body Mass Index (BMI) was nearly identical at and 4.9 ± 2.8 ­x10 particles/mL of plasma for healthy 24.3 ± 4.6 in controls and 24.3 ± 5.3 in patients (p = 0.18, controls and ME/CFS subjects, respectively) (p = 0.04, Table 1). Sixty percent of the ME/CFS patients were able Fig. 1c). to identify an acute, often flu-like, illness that imme- Representative EV isolates from both groups were diately preceded the onset of the disease, while 40% further analyzed by transmission electron microscopy are unaware of an initiating event and consider their (TEM) on 300-mesh copper Formvar-coated grids. EV onset to be gradual (Table 1), and all patients had the isolates were clearly visualized and morphological analy- illness for more than 10 years. Bell Scale ratings were sis revealed a homogeneous population of vesicles that significantly different between groups, with scores aver- were spherical-shaped, enclosed by a lipid bilayer, and aging 35.4 ± 14.1 and 96.3 ± 9.5 for ME/CFS and con- had a size distribution that agrees with the NTA data trols, respectively (p < 0.001, Table 1). Additionally, both (Fig. 2a). the Physical and Mental Component Scores (PCS and To further confirm the presence of EVs in the samples, MCS respectively) derived from the SF-36 short survey we investigated EV fractions of 10 subjects from each were, as expected, higher in the control group (p < 0.001, group by Western blotting. The enrichment of typical EV Table 1). All symptoms from the Specific Severity Symp- markers was assessed using specific antibodies for the tom scale were higher in the ME/CFS group (Table 1). tetraspanin family proteins CD63 and CD81, the ESCRT- associated protein Tumor Susceptibility Gene 101 (TSG101), and the cytosolic heat shock protein HSP70. Characterization of extracellular vesicles Two representatives of each group are shown in Fig. 2b. We investigated whether there were differences in lev- The analysis confirmed that CD63-, CD81-, TSG101- and els and size of circulating extracellular vesicles purified HSP70-positive nanoparticle populations were recov- from plasma samples from ME/CFS patients and healthy ered. The mitochondrial protein cytochrome C was used individuals. EVs were purified by precipitation using as a negative control and its absence in EV isolates from ­ExoQuickTM reagent and then analyzed by Nanoparticle both ME/CFS patients and healthy controls demonstrate Tracking Analysis (NTA), Transmission Electron Micros- that the samples were free of cellular protein contamina- copy (TEM) and Western blotting. Protein levels of total tion (Fig. 2b). EV fractions were measured by BCA (bicinchoninic acid assay) and no significant difference was found between Cytokine content in plasma and extracellular vesicles ME/CFS and controls (21.9 mg/ml ± 1.1 and 23.2 mg/ A subset of 38 samples out of the 70 analyzed previously ml ± 0.9, respectively) (p = 0.296). Sizing and quantifica- (19 ME/CFS females and 19 females controls) were sub- tion of EVs were performed using a NanoSight NS300 jected to measurement of 45 cytokines/chemokines and instrument on all samples (N = 35 for both ME/CFS and growth factors levels through a multiplex assay on both control samples) and results are shown in Fig. 1. Accumu- plasma and purified EV samples and their levels were lating data have indicated that the EVs are highly hetero- compared using mean MFI (Table 2). Of the 45 analytes geneous and classically defined by their size and content. measured, cases and controls were not distinguished At present, at least 3 main subgroups of EVs have been overall as there were no significant differences between defined: apoptotic bodies (1– 5 µm), large microvesicles the groups after correction for multiple comparisons. (130-1000 nm), and small EVs (exosomal) (30–130 nm) Even though Interleukin-17E (IL-17E) levels in EVs and [38, 39]. All nanoparticles purified were smaller than MCP-1 levels in both plasma and EVs (Table 2) initially 500 nm, most of them being in the typical exosome size appeared significantly different, they did not remain sig- range of 30–130 nm. Nanoparticle Tracking Analy- nificant after FDR correction (Table 2). Cytokine level sis revealed that EV particles’ size means did not dif- ratios (ME/CFS vs. controls) showed that the major- fer between healthy individuals (132.7 ± 16.4 nm, range ity of the cytokines were at similar levels in both groups 112–190 nm) and ME/CFS patients (130.1 ± 12.7 nm, and type of samples with the exception of a few that were range 114–149 nm) (p = 0.65, Fig. 1a). increased or decreased in the ME/CFS group (Table 2). Giloteaux et al. J Transl Med (2020) 18:387 Page 6 of 17 a b c d Fig. 1 Sizing and quantification of Extracellular Vesicles. Size in nm (a), total concentration (b), 30–130 nm concentration (c) and > 130 nm concentration (d) of particles per ml of plasma in ME/CFS subjects and healthy controls (CTRL) as determined by Nanoparticle Tracking Analysis. The yellow dot represents the mean When cytokine levels in EVs were normalized with axis and 14.8% for the second axis and two significant either the total protein concentration measured by clusters were observed, supported by the Hopkins sta- BCA, the total concentration of particles/ml or the tistic H = 0.75 (Fig. 3c). concentration of 30-130 nm particles/ml of plasma Partial and Spearman ranked correlation analyses were measured with NTA, there were no significant differ- performed between cytokine levels measured in plasma ences between groups (data not shown). and EVs and other variables such as age, BMI, Bell Scale, Principal Component Analysis performed on the SF-36 scores, the Specific Severity Symptom scale cytokine levels explained 59.3% of the data distribu- scores and EV characteristics. In ME/CFS plasma sam- tion in plasma samples (PC-1 48.2%; PC-2 11%, Fig. 3a) ples, we found 3 significant positive correlations that and 46.8% in EV samples (PC-1 26.5%; PC-2 20.3%, were not found in the plasma from controls. IL-1RA Fig. 3b). Within plasma and EV samples, there appears positively correlated with age (r = 0.67, p = 0.002, Addi- to be no clear difference in between the ME/CFS group tional file 1: Figure S1a) and BMI (r = 0.71, p = 0.0006, and healthy group (H < 0.5, Fig. 3a, b). Comparing sam- Additional file 1: Figure S1b). The basic fibroblast fac- ple types within subjects, the percentage of variability tor FGF-β was strongly correlated with fatigue (r = 0.58, explained by each dimension was 29.9% for the first p = 0.009, Additional file 1: Figure S1c) before and after Giloteaux et al. J Transl Med (2020) 18:387 Page 7 of 17 a b Fig. 2 Characterization of Extracellular Vesicles. (a) Morphology of isolated EVs from ME/CFS was confirmed by transmission electron microscopy (a representative image is shown; scale bars: 200 nm and 500 nm) and (b) Western blot analysis of isolated EVs from healthy controls (CTRL) and ME/CFS subjects as representative samples. Thirty μg of protein was loaded in each lane and probed with specific antibodies to EV protein markers. Cytochrome C (mitochondrial marker) was used as negative control for EV and positive for PBMCs controlling for potentially confounding variables age and Network diagrams showed cytokine-cytokine correla- BMI. tions patterns that differed between ME/CFS and con- Additionally, we also observed a number of significant trols in plasma as well as in EV samples (Figs. 4 and 5). correlations between EV cytokines and other variables A total of 385 significant cytokine-cytokine interac- that were unique to a group. Interestingly, the Mac- tions were common to both groups in plasma samples rophage inflammatory proteins MIP-1β, MIP-3β and but a higher significant number of interactions was pre- Interleukin-2 (IL-2) had significant positive correlations sent solely in plasma from ME/CFS patients in compar- with the concentration of the 30-130 nm population of ison to the controls (535 and 483 respectively, Fig. 4c). particles in the control EV samples (r = 0.61, p = 0.008 Whereas in the plasma of controls no inverse relation- for MIP-1β, r = 0.67, p = 0.002 for MIP-3β and r = 0.6, ships were found across any of the cytokines (Fig. 4a p = 0.009 for IL-2, Additional file 2: Figure 2a, b, c) that and c), a total of 12 such associations (Fig. 4b, c) were were not present in the EV samples from patients. Trans- found in the plasma of case subjects with IP-10 (IP- forming Growth Factor α (TGFα) and Interleukin-33 10/GROα; IP-10/IL-3; IP-10/TNFα; IP-10/IL-2; IP-10/ were found to inversely correlate with headaches in the CD40 Ligand; IP-10/GM-CSF; IP-10/FGF-β (FGF2); ME/CFS EV samples (r = −0.66, p = 0.002 and r = −0.63, IP-10/IL1-α; IP-10/IL-1β; IP-10/TRAIL; IP-10/IL-5 and p = 0.005 for TGFα and IL-33 respectively, Additional IP-10/IL-17A) and MIP-3β with FGF2. file 2: Figure 2d, e), and these correlations were unique In EVs from controls there were > 2.5 more significant in patients with ME/CFS compared to the control group. inter-cytokine correlations than in the ME/CFS group Next, we investigated cytokine-cytokine interactions (380 vs. 139 respectively, Fig. 5c) and the two groups in plasma and EV samples from both groups. These shared a total of 98 significant inter-cytokine correla- analyses were intended to determine whether these tions. Interestingly, we found one negative correlation analytes revealed specific regulatory relationships. in each network that was different: IL-RA inversely cor- related with Eotaxin in the control group (Fig. 5a) while Giloteaux et al. J Transl Med (2020) 18:387 Page 8 of 17 Table 2 Comparison of plasma immune analytes in plasma and EVs from individuals with ME/CFS and healthy controls Plasma Extracellular vesicles Cytokine P-value Q-value CFS/controls Cytokine P-value Q-value CFS/controls CD40 ligand 0.422 0.863 1.0 CD40 ligand 0.421 0.858 1.0 EGF 0.122 0.863 0.7 EGF 0.331 0.858 0.8 Eotaxin 0.931 0.983 0.9 Eotaxin 0.301 0.858 1.1 FGF-β 1.000 1.000 1.0 FGF-β 0.553 0.858 1.0 Flt3 ligand 0.350 0.863 1.1 Flt3 ligand 0.784 0.909 1.0 Fractalkine 0.118 0.863 1.1 Fractalkine 0.584 0.858 1.0 G-CSF 0.396 0.863 1.0 G-CSF 0.503 0.858 1.0 GM-CSF 1.000 1.000 1.0 GM-CSF 0.089 0.858 1.4 Granzyme B 0.683 0.940 1.0 Granzyme B 0.855 0.933 1.0 Groα 0.773 0.940 1.0 Groα 0.796 0.909 1.1 Groβ 0.313 0.863 0.8 Groβ 0.641 0.858 0.5 IFN-α 0.521 0.937 0.9 IFN-α 0.456 0.858 0.9 IFN-β 0.661 0.940 0.9 IFN-β 0.952 0.952 1.5 IFN-γ 0.737 0.940 1.0 IFN-γ 0.727 0.909 1.1 IL-12p70 0.255 0.863 0.9 IL-12p70 0.104 0.858 1.9 IL-13 0.413 0.863 0.8 IL-13 0.563 0.858 1.0 IL-15 0.726 0.940 1.0 IL-15 0.627 0.858 1.7 IL-17 0.815 0.940 1.0 IL-17 0.386 0.858 1.3 IL-17E 0.630 0.940 0.9 IL-17E 0.049 0.858 0.6 IL-1RA 0.053 0.863 1.2 IL-1RA 0.346 0.858 0.9 IL-1α 0.115 0.863 0.9 IL-1α 0.952 0.952 0.9 IL-1β 0.661 0.940 1.0 IL-1β 0.465 0.858 0.9 IL-2 0.266 0.863 0.9 IL-2 0.202 0.858 1.6 IL-3 0.306 0.863 1.0 IL-3 0.796 0.909 1.1 IL-33 0.380 0.863 1.0 IL-33 0.236 0.858 1.2 IL-4 0.793 0.940 0.9 IL-4 0.616 0.858 0.9 IL-5 0.473 0.925 1.0 IL-5 0.594 0.858 1.1 IL-6 0.903 0.983 1.0 IL-6 0.648 0.858 1.1 IL-7 0.299 0.863 1.0 IL-7 0.378 0.858 1.8 IL-8 0.405 0.863 0.7 IL-8 0.808 0.909 1.2 IL-10 0.803 0.940 1.0 IL-10 0.784 0.909 1.3 IP-10 0.249 0.863 1.4 IP-10 0.162 0.858 1.5 MCP-1 0.008 0.338 1.5 MCP-1 0.022 0.858 1.7 MIP-1α 0.414 0.863 1.3 MIP-1α 0.475 0.858 1.3 MIP-1β 0.939 0.983 1.0 MIP-1β 0.879 0.933 0.9 MIP-3β 0.599 0.940 1.1 MIP-3β 0.891 0.933 1.2 MIP-3α 0.213 0.863 0.9 MIP-3α 0.485 0.858 0.9 PD-L1 0.726 0.940 1.0 PD-L1 0.634 0.858 1.1 PDGF-AA 0.354 0.863 0.9 PDGF-AA 0.425 0.858 0.8 PDGF-BB 0.181 0.863 0.8 PDGF-BB 0.086 0.858 0.6 RANTES 0.116 0.863 0.7 RANTES 0.221 0.858 0.7 TGF-α 0.521 0.937 1.0 TGF-α 0.504 0.858 1.1 TNF-α 0.748 0.940 1.0 TNF-α 0.466 0.858 1.2 TRAIL 0.838 0.942 1.0 TRAIL 0.218 0.858 1.3 VEGF-A 0.683 0.940 1.0 VEGF-A 0.230 0.858 1.1 p-values are shown prior correction for multiple comparison using the Benjamini and Yekutieli control for false discovery rate (q-values) at a 5% rate. Results were adjusted for age and BMI. Also indicated, are ratios of analyte levels ME/CFS vs. controls. Italics values denote statistical significance at p < 0.05 Giloteaux et al. J Transl Med (2020) 18:387 Page 9 of 17 a b c Fig. 3 Principal Component analysis of cytokine levels in ME/CFS vs. controls. In plasma (a), in EVs (b) and (c) cytokine levels in EVs vs. plasma. The Hopkins statistic values H are shown for each PCA MCP-1 was negatively correlated with IL-7 in the ME/ gastric cancer [26, 28], Alzheimer’s disease [21] or cer- CFS group (Fig. 5b). ebrovascular disease [27] in which EVs were found to be elevated in the disease state. Oxidative stress, changes in Discussion intracellular calcium level and synaptic activity are con- Extracellular vesicles were isolated from the plasma of 35 tributing factors for inducing exosome secretion [40, 41]. patients with ME/CFS and 35 healthy controls. Trans- We further investigated the cytokine content in EVs mission electronic microscopy and Western blot analyses and plasma from a subset of 19 ME/CFS and 19 healthy on isolated EVs revealed mainly an exosomal morphol- females. Using a linear model regression analysis that ogy with mean sizes ranging from 30 to 130 nm along adjusted for age and BMI on the 45 cytokines measured, with the detection of exosomal markers (CD63, CD81, we did not find any statistically significant differences HSP70 and TSG101) (Fig. 2). Contrary to two previous between groups in the plasma or EV cytokine levels. Fur- reports that found that the mean size of ME/CFS EVs was thermore, the data distribution in both plasma and EVs reduced [23, 24], we did not find any statistical significant assessed by PCA did not show any clustering that could difference in our study population. Although the total distinguish ME/CFS patients (Fig. 3a, b). Within subjects, concentration of particles per ml of plasma did not differ the cytokine levels in EVs and plasma appeared as two between patients and controls in our study, the concen- different clusters (Fig. 3c) confirming that EV and plasma tration of 30-100 nm vesicles was significantly increased are two different compartments. in the diseased group (Fig. 1c), confirming findings In a larger cohort analyzed by Hornig et al. [10], nine recently reported for ME/CFS [23–25]. These results plasma cytokine levels significantly differed between are also in line with other conditions such as breast and cases and controls. Montoya et al. [12] found only two Giloteaux et al. J Transl Med (2020) 18:387 Page 10 of 17 Fig. 4 Cytokine–cytokine correlations in plasma. Network diagrams of cytokine interactions in plasma of ME/CFS patients (a), and in controls (b). Grey edges indicate positive correlations and black bold lines negative associations. The Venn diagram (c) shows the number of significant cytokine-cytokine associations common to both groups and unique to each group. In red are the number of negative correlations. All correlations in this analysis were determined using partial Spearman’s ranked correlation adjusted for age and BMI cytokines out of 51 measured to be significantly differ- ME/CFS cases (duration ≤ 3 years) showed statisti- ent in plasma of ME/CFS patients when compared with cally significant differences in comparison to controls healthy individuals (Resistin was lower and TGF-β was for more than half of the 51 cytokines analyzed [10]. elevated). Considering the limited number of analyzed These findings were not reproduced by Montoya et al., samples in our pilot study (ME/CFS, n = 19 and controls, who did not find any cytokine to be significantly differ- n = 19) in comparison to Hornig et al. (ME/CFS, n = 298 ent between short duration cases (≤ 3 years) or long and controls, n = 348) or Montoya et al. (ME/CFS, duration cases (> 3 years) and healthy individuals [12]. n = 186 and controls, n = 38), it is not surprising that we Unfortunately, our analysis is limited as ME/CFS cases did not find any differences between groups, as a result of in this study were ill for more than 3 years, which pre- low statistical power. vented us from analyzing cytokine levels between short When comparing short duration (≤ 3 years) to long and long duration of illness. duration (> 3 years) of illness, several studies reported Despite a lack of significance, higher and lower lev- differences in plasma cytokine levels between ME/ els of several cytokines were found in plasma and EV CFS subgroups and healthy individuals. Landi and samples as shown by the ratio of cytokine levels of ME/ colleagues [42] compared the plasma cytokine levels CFS vs. controls (Table 2). In both EVs and plasma of 100 ME/CFS of long duration to 79 controls and samples, MIP-1alpha and IP-10 were both elevated found reductions in 3 out of 31 cytokines measured in the ME/CFS group but not statistically different, a (IL-7, IL-16, and Vascular Endothelial Growth Fac- trend previously observed in a report that considered tor A, VEGF-A) while Hornig et al. observed that early disease severity and compared mild ME/CFS patients Giloteaux et al. J Transl Med (2020) 18:387 Page 11 of 17 Fig. 5 Cytokine–cytokine correlations in extracellular vesicles. Network diagrams of cytokine interactions in EVs of ME/CFS patients (a), and in controls (b). Grey edges indicate positive correlations and black bold lines negative associations. The Venn diagram (c) shows the number of significant cytokine-cytokine associations common to both groups and unique to each group. In red are the number of negative correlations. All correlations in this analysis were determined using partial Spearman’s ranked correlation adjusted for age and BMI to controls [12] but opposite to another study show- [10]. Elevated levels of MCP-1 have been also observed ing significant lower levels of these cytokines in ME/ in several CNS-related pathologies and neurodegen- CFS plasma samples [10]. Interestingly, in our study, erative diseases such as in the plasma of patients with the Monocyte Chemoattractant Protein-1 (MCP-1 or Alzheimer’s disease [49, 50], amyotrophic lateral scle- CCL2) was significantly higher in both sample types rosis (ALS) [51], in the cerebrospinal fluid of patients in the ME/CFS group but lost statistical significance with ischemic stroke [52], in peripheral blood mononu- after correction for multiple comparisons (Table 2). clear cells obtained from Parkinson’s disease and ALS MCP-1 is of relevance within the central nervous sys- patients [51, 53], and in HIV-associated dementia [54], tem (CNS), as it is expressed in different parts of the schizophrenia [55] and epilepsy [56]. Furthermore, cor- brain by neurons and microglia [43–45]. MCP-1 acts relations between concentration of MCP-1 and symp- mainly as an attractor for mononuclear cells, a media- toms such as depression, anxiety and fatigue [57] or tor of inflammation, but is also involved in neuropro- severity of disease have been reported [57, 58]. Due to tection against excitotoxic injuries [46, 47]. In ME/CFS, its role in the development of the inflammatory and decreased levels of MCP-1 were observed in Cerebro- immune responses, MCP-1 could be considered as an spinal Spinal Fluid (CSF) samples in comparison to indicator that might allow the detection and quantifica- controls and these levels were higher in comparison to tion of the progression of ME/CFS. Multiple Sclerosis’ CSF samples [48]. Also, higher lev- Several EV-associated cytokine levels appeared to be els were reported in plasma obtained from ME/CFS elevated in the ME/CFS group (Table 2) but were not patients with short duration of illness versus controls statistically different probably due to our limited sample Giloteaux et al. J Transl Med (2020) 18:387 Page 12 of 17 size. Amongst them, IL-7 which was previously found to elevated cytokine levels found in EVs from our cohort of be elevated in serum samples of severely afflicted ME/ ME/CFS patients are part of a specific immune response CFS patients [59], is a hematopoietic cytokine with criti- in ME/CFS. cal functions in both B- and T-lymphocyte development. The content of different types of EVs reflect that of It is secreted by stromal cells, dendritic cells as well as the parent cells and are enriched in certain molecules, neurons and neuronal progenitor cells [60] and previous including cytokines, chemokines, functional microRNAs, studies have revealed that IL-7 promotes neuronal dif- and cell-specific antigens. EVs maintain characteristics of ferentiation [61]. It can also stimulate cytotoxic function- the antigen presenting cell from which they are derived, ing in mature T cells and NK cell proliferation, activities exposing antigen‐presenting MHC I and MHC II mole- reduced in ME/CFS [62–64]. Hardcastle et al. [59] found cules on their surface [77]. a positive correlation between Interferon-γ (IFN-γ) and We further investigated if EV or plasma cytokine lev- IL-7 in severely affected ME/CFS patients. Higher lev- els correlated with age, BMI and scores of the Bell, SF-36 els of IFNγ are associated with reduced Natural Killer and SSS forms, by performing Spearman’s and partial cell cytotoxic activity observed in this illness and there- Spearman’s rank correlation adjusting for age and BMI. fore, elevated levels of both cytokines may be involved in Sex was not included as a confounding variable in the disease severity. In turn, lower IL-7 has been correlated analysis as all our participants were females. Even though with cognitive decline during aging [65] and a reduction several cytokines correlated either positively or inversely in IL-7 in plasma previously observed in ME/CFS [42] (Additional files 1 and 2: Figs. 1, 2) there were no com- suggests a reduction in immune activation along with a monalities between EV and plasma samples. IL-1RA was potential neuropathology similar to the process of aging. significantly and positively correlated with age and BMI Another example in our study, although not significant, in plasma samples from the ME/CFS group. It is well were the elevated levels of the anti-inflammatory cytokine documented that IL-1RA, a natural antagonist to the IL-10 in EVs isolated from ME/CFS patients. In contrast, proinflammatory cytokine IL-1, increases dramatically extracellular vesicles levels of IL-10 levels were found in obese subjects [78]. Furthermore, it has been shown to be reduced in gastric cancer [26]. IL-10 receptors are that IL-1RA levels were positively correlated with serum found on different populations of brain cells including leptin levels in ME/CFS [13] and that leptin was associ- astrocytes, oligodendrocytes, and microglia [66, 67]. The ated with fatigue severity in patients with ME/CFS [12, stimulation of these receptors with IL-10 reduces synthe- 13] but also with chronic hepatitis C and irritable bowel sis of proinflammatory cytokines, allowing the survival of syndrome [79, 80]. Leptin has been identified as a major brain cells [68–70]. Its expression is elevated during the proinflammatory cytokine that induces IL-1RA secre- course of several CNS-related diseases such as multiple tion [81] and is involved in NK cell activation, and innate sclerosis, Alzheimer’s disease [71] or meningoencepha- and acquired immune responses [82]. Unfortunately, our litis [72] and promotes survival of neurons and all glial cytokine panel did not measure leptin and because of the cells. In ME/CFS, studies showed contradictory results close relationship existing between IL-1RA and leptin, about levels of IL-10, but none measured it in extracel- we can only speculate that the higher levels of IL-1RA lular vesicles. A longitudinal study reported increased observed in ME/CFS plasma (Table 2) and the signifi- blood IL-10 levels in patients [73], while others reported cant positive correlations of IL-1RA observed in the ME/ decreased [9, 74] and increased [75] levels in cerebrospi- CFS group and not in the control group may be due to nal fluid or plasma [10]. These findings are inconsistent increased inflammation in patients. and this may be due to the heterogeneity of the disease, Interestingly, in EV samples from the control group, different analytical methods, and the presence of differ- 3 cytokines strongly positively correlated with the ent patient subgroups. Compromises to brain cells may 30-130 nm extracellular vesicle population (MIP-1β, contribute to low levels of IL-10, and a decrease in CNS MIP-3β and IL-2, Additional file 2: Fig. 2a–c) but not IL-10 may be related to symptoms reported in ME/CFS in the ME/CFS group. MIPs are crucial for immune cases. The role of IL-10 in ME/CFS requires further responses towards infection and inflammation [83] and investigation. The elevated levels observed in this pilot are produced by macrophages and monocytes, and stim- are consistent with an immune activation. ulated by proinflammatory cytokines such as IL-1β [84]. While the levels of IL-2 in our plasma samples did IL-2 is a potent stimulator of T-cell proliferation and not differ between groups, they were elevated in the inhibits the development of inflammatory Th17 cells [85]. ME/CFS EVs in comparison to controls (Table 2) but The fact that these cytokines correlated with a particu- not statistically significant. IL-2 levels have been previ- lar population of EVs in the control group suggest that ously reported to be higher in CSF [48] and plasma from ME/CFS patients may be lacking specific vesicles carry- ME/CFS patients [76]. We can only speculate that these ing important proteins able to respond to inflammatory Giloteaux et al. J Transl Med (2020) 18:387 Page 13 of 17 challenges. Cytokines and chemokines can be selec- been shown to be involved in neurodegenerative disor- tively sorted to EVs into multivesicular bodies (MVBs) ders. In patients with Alzheimer’s disease, receptors for and secreted via exosomes, into microvesicles shedding IP-10 are expressed on neurons and the induction of from the plasma membrane, or into apoptotic bodies IP-10 by astrocytes and in microglia [92], along with its (AB). This selective sorting towards EV subspecies has binding on neurons and the accumulation of beta amy- been demonstrated in Type 1 diabetes in which MCP-1 loid, result in neuronal dysfunction and apoptosis [93]. was expressed in all sorts of EVs while IL-27 was solely In both multiple sclerosis and experimental autoimmune expressed in apoptotic bodies [86]. Fitzgerald et al. [16] encephalomyelitis, IP-10 is highly expressed by astrocytes observed that the number and pattern of cytokines pack- [94], and largely associated with the influx of inflamma- aged in EVs changed depending on the stimulus and tory leukocytes into neural tissue, and disease severity. suggested that the packaging of cytokines in EVs is not Abnormal levels of IP-10 have been observed in body simply the property of a particular cytokine, but rather a fluids of individuals infected with respiratory syncytial tightly controlled biological process. In pathological con- virus (RSV) [95] and contributed to LCMV (Lymphocytic ditions such as HIV or diabetes, it has been shown that choriomeningitis virus) or West Nile Virus infections, in the profile of cytokines in EVs was significantly increased which IP-10 expression was restricted mainly to neurons and that specific cytokines were associated with disease [96]. During Herpes virus infection it has been shown duration [87, 88]. that infected IP-10-deficient mice showed higher viral We may assume that in ME/CFS, inflammation may loads in the CNS, decreased numbers of natural killer induce changes in the packing of specific cytokines/ cells and C ­ D8+ T-cells [97]. Altogether, these studies chemokines into selective EVs and that their release is highlight the important role of IP-10 in mediating CNS intimately linked to the state of the releasing cells, healthy inflammation, a hallmark of ME/CFS, thus further focus versus diseased, near or at a distance. on this chemokine is needed. We further analyzed cytokine-cytokine interactions EV cytokine–cytokine correlations in ME/CFS showed by performing Spearman’s correlations for each pair a pattern that differed from controls as well, with fewer of cytokines in plasma and EV samples for both groups significant correlations (Fig. 5c). The network diagrams to find whether a disrupted profile of EV-associated or showed that both groups had a unique negative cytokine- circulating cytokines in plasma could be identified in cytokine interaction; levels of IL-1RA were inversely ME/CFS patients. Overall, there was more significant associated with eotaxin in the control group (Fig. 5a), and cytokine-cytokine correlations in the plasma from the IL-7 with MCP-1 in ME/CFS EVs. Hormig et al. found an ME/CFS group and less in the EV samples in compari- inverse relationship of IL-1RA with M-CSF, GM-CSF and son to healthy individuals (Figs. 4b and 5b). In ME/CFS IL-17 but not with eotaxin in CSF samples from ME/CFS plasma, there were no inverse correlations in the control patients [48]. group while there were 13 within the ME/CFS group. Eotaxin is a member of the C–C chemokine fam- IP-10 was highly interconnected with 12 other cytokines ily originally implicated in the selective recruitment of in the plasma from ME/CFS (Fig. 4b) and appeared to eosinophils into inflammatory sites during allergic reac- be a negative regulator. We found IP-10 to be negatively tions. It is induced by Th-2 cytokines, including inter- correlated with CD40 Ligand, but this inverse relation- leukin IL-13, IL-10, and IL-4, and is produced by B cells, ship was not observed in a previous study [10] in which endothelial cells, lymphocytes, macrophages, epithe- CD40 Ligand was the main driver of the inverse rela- lial cells, and chondrocytes [98–100]. Although it has tionships with other cytokines. Recently, Moneghetti been been thoroughly investigated in allergic reactions, and colleagues found IP-10 to play a central role in the eotaxin has been shown to be involved in a skewed cytokine network of plasma from ME/CFS patients [11], immune response toward a type-2 (Th2) profile and was but the IP-10 interactions with other cytokines that they recently associated with aging, neurogenesis and neu- reported differed from those we observed in our study rodegeneration, being able to influence neural progeni- population. tor cells, and microglia. Increased circulating levels of During neuroinflammation, NK cells and activated eotaxin have been described in fibromyalgia [101, 102] CD4 + and CD8 + T-cells are attracted to sites of inflam- and in several neuroinflammatory disorders [103] such mation, infected or tumoral areas in the presence of IP-10 as Alzheimer’s disease [104, 105], amyotrophic lateral [89]. In addition to its chemotactic properties in mediat- sclerosis, Huntington’s disease, and secondary progres- ing an influx of inflammatory leukocytes into infected or sive multiple sclerosis [105]. It has been associated with inflamed tissues, IP-10 has non chemotactic functions markers of aging and degeneration and correlated with such as inhibition of angiogenesis [90] and α-defensins- cognitive measures. Since eotaxin is capable of crossing like antibacterial properties [91]. Furthermore, IP-10 has the blood–brain barrier, it is plausible that eotaxin can Giloteaux et al. J Transl Med (2020) 18:387 Page 14 of 17 exert physiological and pathological actions in the cen- Abbreviations ME/CFS: Myalgic Encephalomyelitis/Chronic Fatigue Syndrome; EVs: Extracel‑ tral nervous system of ME/CFS patients. Hornig et al. lular vesicles; NTA: Nanoparticle tracking analysis; TEM: Transmission electron suggested the possibility that increased levels of eotaxin microscopy; BMI: Body mass index; CSF: Cerebrospinal fluid. along with dysregulation of IL-1 signaling observed in Acknowledgements ME/CFS are part of an allergic process in central com- The authors are grateful to all the participants who took part in the study and partments also seen in a CNS infections [106–108]. for providing samples and information for the study. Jesús Castro-Marrero Altogether, eotaxin seems to be very promising to fur- would like to thank the Spanish Ministry of Education, Culture and Sports (MECD) for a postdoctoral mobility grant. ther study in ME/CFS to determine its prognostic value along with careful cognitive phenotyping monitoring Authors’ contributions and neuroimaging studies to evaluate its association LG designed the experiments, processed the samples, conducted the experi‑ ments, and performed the statistical analysis with R. AO performed NTA, TEM with neurodegenerative changes. and WB analysis. JC-M contributed to EV processing and acquisition of data. Although several intercytokine network analyses SML recruited, diagnosed, and sampled the blood from the subjects. MRH were conducted by other groups in plasma and CSF contributed to study design. LG and MRH performed analysis and writing. All authors read and approved the final manuscript. [10, 11, 13, 48], there were no commonalities with our results, which may suggest again the heterogeneity of Funding the disease and the presence of different patient sub- This work was supported by grant R01AI107762 from NIH NIAID to Fabien Campagne and M.R.H and NIH NINDS U54NS105541 for support of the Cornell groups. Nevertheless, our results and those of previous ME/CFS Center. JC-M was supported partially by a mobility grant from the reports suggest unusual regulatory relationships among Spanish MECD “Salvador de Madariaga” (PRX17/00361). The funders had no cytokine in plasma and EVs, indicating a disruption of role in study design, data collection, analysis and interpretation, decision to publish, or preparation of the manuscript. intercytokine networks. Availability of data and materials Data for extracellular vesicle size, quantification, and cytokine content is avail‑ able on request to the authors. Conclusions The pilot study reported herein confirms previous results Ethics approval and consent to participate in which the numbers of EVs isolated from plasma of The institutional review board of Cornell University approved the study, and patients and controls gave written informed consent after the study protocol ME/CFS patients are elevated in comparison to a con- was fully explained. All consented to blood draw and to the availability of the trol group. We further analyzed the cytokine profiling stored samples for additional bioassays. in plasma and provide a unique report on the cytokine Consent for publication content of extracellular vesicles in ME/CFS patients. All authors reviewed and approved the final version for submission. Our analysis was based on a single plasma sample from which we isolated EVs. No significant differences were Competing interests The authors declare that they have no competing interests. found between patients and controls in both sample types. We were also not able to analyze our data based Author details 1 on gender nor duration of illness or disease. Even though Department of Molecular Biology and Genetics, Cornell University, 323 Bio‑ technology Building, 526 Campus Road, Ithaca, NY 14853, USA. 2 CFS/ME Unit, many studies have identified potential cytokine differ- Division of Rheumatology, Vall d’Hebron University Hospital Research Institute, ences between ME/CFS patients and healthy individuals, Universitat Autònoma de Barcelona, Barcelona 08035, Spain. 3 Private Practice, results have been contradictory (reviewed in [6–8]) and New York, NY, USA. this may be due to the cross-sectional nature of the stud- Received: 11 July 2020 Accepted: 3 October 2020 ies. Longitudinal study designs including larger numbers of ME/CFS subjects representing different clinical sub- groups, and classified based on disease duration and/or severity. are needed to further characterize their associa- References tion with cytokine expression. 1. Clayton EW. Beyond myalgic encephalomyelitis/chronic fatigue syn‑ drome: an IOM report on redefining an illness. JAMA. 2015;313:1101–2. 2. Fukuda S, Nojima J, Motoki Y, Yamaguti K, Nakatomi Y, Okawa N, Supplementary information Fujiwara K, Watanabe Y, Kuratsune H. A potential biomarker for fatigue: Supplementary information accompanies this paper at https​://doi. oxidative stress and anti-oxidative activity. Biol Psychol. 2016;118:88–93. org/10.1186/s1296​7-020-02560​-0. 3. Maes M, Kubera M, Uytterhoeven M, Vrydags N, Bosmans E: Increased plasma peroxides as a marker of oxidative stress in myalgic encephalo‑ myelitis/chronic fatigue syndrome (ME/CFS). Medical science monitor: Additional file 1: Figure S1. Spearman’s correlation analysis of immune international medical journal of experimental and clinical research analytes levels with metadata in plasma samples. 2011, 17:SC11. Additional file 2: Figure S2. Spearman’s correlation analysis of immune 4. Armstrong CW, McGregor NR, Lewis DP, Butt HL, Gooley PR. Meta‑ analytes levels with metadata in EV samples. bolic profiling reveals anomalous energy metabolism and oxidative stress pathways in chronic fatigue syndrome patients. Metabolomics. 2015;11:1626–39. Giloteaux et al. J Transl Med (2020) 18:387 Page 15 of 17 5. Castro-Marrero J, Cordero MD, Sáez-Francas N, Jimenez-Gutierrez C, 25. Eguchi A, Fukuda S, Kuratsune H, Nojima J, Nakatomi Y, Watanabe Aguilar-Montilla FJ, Aliste L, Alegre-Martin J. Could mitochondrial dys‑ Y, Feldstein AE. Identification of actin network proteins, talin-1 and function be a differentiating marker between chronic fatigue syndrome filamin-A, in circulating extracellular vesicles as blood biomarkers for and fibromyalgia? Antioxid Redox Signal. 2013;19:1855–60. human myalgic encephalomyelitis/chronic fatigue syndrome. Brain 6. Blundell S, Ray K, Buckland M, White P. Chronic fatigue syndrome Behav Immun. 2020;84:106–14. and circulating cytokines: a systematic review. Brain Behav Immun. 26. Im K, Baek J, Kwon WS, Rha SY, Hwang KW, Kim U, Min H. The com‑ 2015;50:186–95. parison of exosome and exosomal cytokines between young and old 7. Corbitt M, Eaton-Fitch N, Staines D, Cabanas H, Marshall-Gradisnik S. A individuals with or without gastric cancer. Int J Gerontol. 2018;12:233–8. systematic review of cytokines in chronic fatigue syndrome/myalgic 27. Jung KH, Chu K, Lee ST, Park HK, Bahn JJ, Kim DH, Kim JH, Kim M, Kun encephalomyelitis/systemic exertion intolerance disease (CFS/ME/ Lee S, Roh JK. Circulating endothelial microparticles as a marker of SEID). BMC Neurol. 2019;19:207. cerebrovascular disease. Annals Neurol. 2009;66:191–9. 8. Yang T, Yang Y, Wang D, Li C, Qu Y, Guo J, Shi T, Bo W, Sun Z, Asakawa 28. König L, Kasimir-Bauer S, Bittner A-K, Hoffmann O, Wagner B, Santos T. The clinical value of cytokines in chronic fatigue syndrome. J Transl Manvailer LF, Kimmig R, Horn PA, Rebmann V. Elevated levels of extra‑ Med. 2019;17:213. cellular vesicles are associated with therapy failure and disease progres‑ 9. Hornig M, Gottschalk C, Eddy M, Che X, Ukaigwe J, Peterson D, Lipkin W. sion in breast cancer patients undergoing neoadjuvant chemotherapy. Immune network analysis of cerebrospinal fluid in myalgic encephalo‑ Oncoimmunology. 2018;7:e1376153. myelitis/chronic fatigue syndrome with atypical and classical presenta‑ 29. Fukuda K, Straus SE, Hickie I, Sharpe MC, Dobbins JG, Komaroff A. Group tions. Transl Psychiatr. 2017;7:e1080–e1080. ICFSS: the Chronic Fatigue Syndrome: A comprehensive approach to its 10. Hornig M, Montoya JG, Klimas NG, Levine S, Felsenstein D, Bateman L, definition and study. Ann Intern Med. 1994;121:953–9. Peterson DL, Gottschalk CG, Schultz AF, Che X. Distinct plasma immune 30. Bell DS: The doctor’s guide to chronic fatigue syndrome: understand‑ signatures in ME/CFS are present early in the course of illness. Sci Adv. ing, treating, and living with CFIDS. Addison-Wesley Longman; 1994. 2015;1:e1400121. 31. Ware JE, Jr., Sherbourne CD: The MOS 36-item short-form health survey 11. Moneghetti KJ, Skhiri M, Contrepois K, Kobayashi Y, Maecker H, Davis M, (SF-36). I. Conceptual framework and item selection. Med Care 1992, Snyder M, Haddad F, Montoya JG. Value of circulating cytokine profiling 30:473-483. during submaximal exercise testing in myalgic encephalomyelitis/ 32. Bosch S, De Beaurepaire L, Allard M, Mosser M, Heichette C, Chrétien D, chronic fatigue syndrome. Sci Rep. 2018;8:2779. Jegou D, Bach J-M. Trehalose prevents aggregation of exosomes and 12. Montoya JG, Holmes TH, Anderson JN, Maecker HT, Rosenberg-Hasson cryodamage. Sci Rep. 2016;6:36162. Y, Valencia IJ, Chu L, Younger JW, Tato CM, Davis MM. Cytokine signature 33. Breen EJ, Polaskova V, Khan A. Bead-based multiplex immuno-assays associated with disease severity in chronic fatigue syndrome patients. for cytokines, chemokines, growth factors and other analytes: median Proc Natl Acad Sci. 2017;114:E7150–8. fluorescence intensities versus their derived absolute concentration 13. Stringer EA, Baker KS, Carroll IR, Montoya JG, Chu L, Maecker HT, values for statistical analysis. Cytokine. 2015;71:188–98. Younger JW. Daily cytokine fluctuations, driven by leptin, are associated 34. Breen EJ, Tan W, Khan A. The statistical value of raw fluorescence with fatigue severity in chronic fatigue syndrome: evidence of inflam‑ signal in Luminex xMAP based multiplex immunoassays. Sci Rep. matory pathology. J Transl Med. 2013;11:93. 2016;6:26996. 14. Fleshner M, Crane CR. Exosomes, DAMPs and miRNA: features of 35. Won JH, Goldberger O, Shen-Orr SS, Davis MM, Olshen RA. Signifi‑ stress physiology and immune homeostasis. Trends Immunol. cance analysis of xMap cytokine bead arrays. Proc Natl Acad Sci U S A. 2017;38:768–76. 2012;109:2848–53. 15. Barnes BJ, Somerville CC. Modulating cytokine production via select 36. Lawson RG, Jurs PC. New index for clustering tendency and its applica‑ packaging and secretion from extracellular vesicles. Front Immunol. tion to chemical problems. J Chem Inf Comput Sci. 1990;30:36–41. 2020;11:1040. 37. Benjamini Y, Krieger A, Yekutieli D. Adaptive linear step-up procedures 16. Fitzgerald W, Freeman ML, Lederman MM, Vasilieva E, Romero R, Mar‑ that control the false discovery rate. Biom. 2006;93:491–507. golis L. A system of cytokines encapsulated in extracellular vesicles. Sci 38. Kowal J, Arras G, Colombo M, Jouve M, Morath JP, Primdal-Bengtson B, Rep. 2018;8:1–11. Dingli F, Loew D, Tkach M, Théry C. Proteomic comparison defines novel 17. Montecalvo A, Shufesky WJ, Stolz DB, Sullivan MG, Wang Z, Divito SJ, markers to characterize heterogeneous populations of extracellular Papworth GD, Watkins SC, Robbins PD, Larregina AT. Exosomes as a vesicle subtypes. Proc Natl Acad Sci. 2016;113:E968–77. short-range mechanism to spread alloantigen between dendritic cells 39. Yáñez-Mó M, Siljander PRM, Andreu Z, Bedina Zavec A, Borràs FE, Buzas during T cell allorecognition. J Immunol. 2008;180:3081–90. EI, Buzas K, Casal E, Cappello F, Carvalho J, et al. Biological properties of 18. Yoon YJ, Kim OY, Gho YS. Extracellular vesicles as emerging intercellular extracellular vesicles and their physiological functions. J Extracellular communicasomes. BMB Rep. 2014;47:531. Vesicles. 2015;4:27066. 19. Boilard E, Nigrovic PA, Larabee K, Watts GF, Coblyn JS, Weinblatt ME, 40. Atienzar-Aroca S, Flores-Bellver M, Serrano-Heras G, Martinez-Gil N, Bar‑ Massarotti EM, et al. Platelets amplify inflammation in arthritis via colla‑ cia JM, Aparicio S, Perez-Cremades D, Garcia-Verdugo JM, Diaz-Llopis M, gen-dependent microparticle production. Science. 2010;327:580–3. Romero FJ. Oxidative stress in retinal pigment epithelium cells increases 20. Danzer KM, Kranich LR, Ruf WP, Cagsal-Getkin O, Winslow AR, Zhu L, exosome secretion and promotes angiogenesis in endothelial cells. J Vanderburg CR, McLean PJ. Exosomal cell-to-cell transmission of alpha Cell Mol Med. 2016;20:1457–66. synuclein oligomers. Mol Neurodegener. 2012;7:1–18. 41. Yu X, Harris SL, Levine AJ. The regulation of exosome secretion: a novel 21. Rajendran L, Honsho M, Zahn TR, Keller P, Geiger KD, Verkade P, Simons function of the p53 protein. Cancer Res. 2006;66:4795–801. K. Alzheimer’s disease β-amyloid peptides are released in association 42. Landi A, Broadhurst D, Vernon SD, Tyrrell DLJ, Houghton M. Reductions with exosomes. Proc Natl Acad Sci. 2006;103:11172–7. in circulating levels of IL-16, IL-7 and VEGF-A in myalgic encephalomy‑ 22. Słomka A, Urban SK, Lukacs-Kornek V, Żekanowska E, Kornek M. Large elitis/chronic fatigue syndrome. Cytokine. 2016;78:27–36. extracellular vesicles: have we found the holy grail of inflammation? 43. Andjelkovic AV, Kerkovich D, Shanley J, Pulliam L, Pachter JS. Expres‑ Front Immunol. 2018;9:2723. sion of binding sites for β chemokines on human astrocytes. Glia. 23. Almenar-Perez E, Sarria L, Nathanson L, Oltra E. Assessing diagnostic 1999;28:225–35. value of microRNAs from peripheral blood mononuclear cells and 44. Coughlan C, McManus C, Sharron M, Gao Z-Y, Murphy D, Jaffer S, Choe extracellular vesicles in myalgic encephalomyelitis/chronic fatigue W, Chen W, Hesselgesser J, Gaylord H. Expression of multiple functional syndrome. Sci Rep. 2020;10:2064. chemokine receptors and monocyte chemoattractant protein-1 in 24. Castro-Marrero J, Serrano-Pertierra E, Oliveira-Rodríguez M, Zaragozá human neurons. Neuroscience. 2000;97:591–600. MC, Martínez-Martínez A. Blanco-López MdC, Alegre J: circulating extra‑ 45. Gosselin RD, Varela C, Banisadr G, Mechighel P, Rostene W, Kitabgi P, cellular vesicles as potential biomarkers in chronic fatigue syndrome/ Melik-Parsadaniantz S. Constitutive expression of CCR2 chemokine myalgic encephalomyelitis: an exploratory pilot study. J Extracell receptor and inhibition by MCP-1/CCL2 of GABA-induced currents in Vesicles. 2018;7:1453730. spinal cord neurones. J Neurochem. 2005;95:1023–34. Giloteaux et al. J Transl Med (2020) 18:387 Page 16 of 17 46. Chiu K, Yeung S-C, So K-F. Chang RC-C: modulation of morphological 66. Mizuno T, Sawada M, Marunouchi T, Suzumura A. Production of changes of microglia and neuroprotection by monocyte chemoat‑ interleukin-10 by mouse glial cells in culture. Biochem Biophys Res tractant protein-1 in experimental glaucoma. Cell Mol Immunol. Commun. 1994;205:1907–15. 2010;7:61–8. 67. Molina-Holgado E, Vela JM, Arévalo-Martín A, Guaza C. LPS/IFN-γ cyto‑ 47. Madrigal JL, Leza JC, Polak P, Kalinin S, Feinstein DL. Astrocyte-derived toxicity in oligodendroglial cells: role of nitric oxide and protection by MCP-1 mediates neuroprotective effects of noradrenaline. J Neurosci. the anti-inflammatory cytokine IL-10. Eur J Neurosci. 2001;13:493–502. 2009;29:263–7. 68. Bachis A, Colangelo AM, Vicini S, Doe PP, De Bernardi MA, Brooker G, 48. Hornig M, Gottschalk G, Peterson D, Knox K, Schultz A, Eddy M, Mocchetti I. Interleukin-10 prevents glutamate-mediated cerebel‑ Che X, Lipkin W. Cytokine network analysis of cerebrospinal fluid in lar granule cell death by blocking caspase-3-like activity. J Neurosci. myalgic encephalomyelitis/chronic fatigue syndrome. Mol Psychiatr. 2001;21:3104–12. 2016;21:261–9. 69. Grilli M, Barbieri I, Basudev H, Brusa R, Casati C, Lozza G, Ongini E. Inter‑ 49. Corrêa JD, Starling D, Teixeira AL, Caramelli P, Silva TA. Chemokines leukin-10 modulates neuronal threshold of vulnerability to ischaemic in CSF of Alzheimer’s disease patients. Arq Neuropsiquiatr. damage. Eur J Neurosci. 2000;12:2265–72. 2011;69:455–9. 70. Strle K, Zhou J-H, Broussard SR, Venters HD, Johnson RW, Freund GG, 50. Westin K, Buchhave P, Nielsen H, Minthon L, Janciauskiene S, Hansson Dantzer R, Kelley KW. IL-10 promotes survival of microglia without O. CCL2 is associated with a faster rate of cognitive decline during activating Akt. J Neuroimmunol. 2002;122:9–19. early stages of Alzheimer’s disease. PLoS ONE. 2012;7:e30525. 71. Remarque E, Bollen E, Weverling-Rijnsburger A, Laterveer J, Blauw G, 51. Baron P, Bussini S, Cardin V, Corbo M, Conti G, Galimberti D, Scarpini Westendorp R. Patients with Alzheimer’s disease display a pro-inflam‑ E, Bresolin N, Wharton SB, Shaw PJ. Production of monocyte chem‑ matory phenotype. Exp Gerontol. 2001;36:171–6. oattractant protein-1 in amyotrophic lateral sclerosis. Muscle Nerve. 72. Deckert M, Soltek S, Geginat G, Lütjen S, Montesinos-Rongen M, Hof H, 2005;32:541–4. Schlüter D. Endogenous interleukin-10 is required for prevention of a 52. Losy J, Zaremba J. Monocyte chemoattractant protein-1 is increased hyperinflammatory intracerebral immune response in Listeria monocy‑ in the cerebrospinal fluid of patients with ischemic stroke. Stroke. togenes meningoencephalitis. Infect Immun. 2001;69:4561–71. 2001;32:2695–6. 73. ter Wolbeek M, van Doornen LJ, Kavelaars A, van de Putte EM, Sched‑ 53. Reale M, Iarlori C, Thomas A, Gambi D, Perfetti B, Di Nicola M, Onofrj lowski M, Heijnen CJ. Longitudinal analysis of pro-and anti-inflamma‑ M. Peripheral cytokines profile in Parkinson’s disease. Brain Behav tory cytokine production in severely fatigued adolescents. Brain Behav Immun. 2009;23:55–63. Immun. 2007;21:1063–74. 54. Kelder W, McArthur JC, Nance-Sproson T, McClernon D, Griffin DE. 74. Peterson D, Brenu E, Gottschalk G, Ramos S, Nguyen T, Staines D, β-Chemokines MCP-1 and RANTES are selectively increased in Marshall-Gradisnik S. Cytokines in the cerebrospinal fluids of patients cerebrospinal fluid of patients with human immunodeficiency virus– with chronic fatigue syndrome/myalgic encephalomyelitis. Mediat. associated dementia. Ann Neurol. 1998;44:831–5. Inflamm. 2015;2015:929720. 55. Drexhage R, Padmos R, de Wit H, Versnel M, Hooijkaas H, van der Lely 75. Natelson BH, Weaver SA, Tseng C-L, Ottenweller JE. Spinal fluid abnor‑ A-J, van Beveren N, de Rijk R, Cohen D. Patients with schizophrenia malities in patients with chronic fatigue syndrome. Clin Diagn Lab show raised serum levels of the pro-inflammatory chemokine CCL2: Immunol. 2005;12:52–5. association with the metabolic syndrome in patients? Schizophr Res. 76. Cheney PR, Dorman SE, Bell DS. Interleukin-2 and the chronic fatigue 2008;102:352–5. syndrome. Ann Intern Med. 1989;110:321–321. 56. Wu Y, Wang X, Mo X, Xi Z, Xiao F, Li J, Zhu X, Luan G, Wang Y, Li Y. 77. Robbins PD, Morelli AE. Regulation of immune responses by extracel‑ Expression of monocyte chemoattractant protein-1 in brain tissue of lular vesicles. Nat Rev Immunol. 2014;14:195–208. patients with intractable epilepsy. Clin Neuropathol. 2008;27:55. 78. Meier CA, Bobbioni E, Gabay C, et al. IL-1 receptor antagonist serum 57. Lindqvist D, Hall S, Surova Y, Nielsen HM, Janelidze S, Brundin L, levels are increased in human obesity: a possible link to the resistance Hansson O. Cerebrospinal fluid inflammatory markers in Parkinson’s to leptin? J Clinical Endocrinol Metab. 2002;87:1184–8. disease–associations with depression, fatigue, and cognitive impair‑ 79. Piche T, Gelsi E, Schneider S, Hebuterne X, Giudicelli J, Ferrua B, Laffont ment. Brain Behav Immun. 2013;33:183–9. C, Benzaken S, Hastier P, Montoya M. Fatigue is associated with high 58. Comini-Frota ER, Teixeira AL, Angelo JP, Andrade MV, Brum DG, circulating leptin levels in chronic hepatitis C. Gut. 2002;51:434–9. Kaimen-Maciel DR, Foss NT, Donadi EA. Evaluation of serum levels 80. Piche T, Huet P-M, Gelsi E, Barjoan EM, Cherick F, Caroli-Bosc FX, Héb‑ of chemokines during Interferon-β treatment in multiple sclerosis uterne X, Tran A. Fatigue in irritable bowel syndrome: characterization patients. CNS Drugs. 2011;25:971–81. and putative role of leptin. Eur J Gastroenterol Hepatol. 2007;19:237–43. 59. Hardcastle SL, Brenu EW, Johnston S, Nguyen T, Huth T, Ramos S, 81. Hosoi T, Okuma Y, Nomura Y. Leptin induces IL-1 receptor antago‑ Staines D, Marshall-Gradisnik S. Serum immune proteins in moderate nist expression in the brain. Biochem Biophys Res Commun. and severe chronic fatigue syndrome/myalgic encephalomyelitis 2002;294:215–9. patients. Int J Med Sci. 2015;12:764. 82. Procaccini C, La Rocca C, Carbone F, De Rosa V, Galgani M, Matarese G. 60. Moors M, Vudattu N, Abel J, Krämer U, Rane L, Ulfig N, Ceccatelli S, Leptin as immune mediator: interaction between neuroendocrine and Seyfert-Margolies V, Fritsche E, Maeurer M. Interleukin-7 (IL-7) and IL-7 immune system. Dev Comp Immunol. 2017;66:120–9. splice variants affect differentiation of human neural progenitor cells. 83. Ren M, Guo Q, Guo L, Lenz M, Qian F, Koenen RR, Xu H, Schilling Genes Immun. 2010;11:11–20. AB, Weber C, Ye RD. Polymerization of MIP-1 chemokine (CCL3 and 61. Mehler MF, Rozental R, Dougherty M, Spray DC, Kessler JA. Cytokine CCL4) and clearance of MIP-1 by insulin-degrading enzyme. EMBO J. regulation of neuronal differentiation of hippocampal progenitor 2010;29:3952–66. cells. Nature. 1993;362:62–5. 84. Maurer M, Von Stebut E. Macrophage inflammatory protein-1. Int J 62. Curriu M, Carrillo J, Massanella M, Rigau J, Alegre J, Puig J, Garcia- Biochem Cell Biol. 2004;36:1882–6. Quintana AM, Castro-Marrero J, Negredo E, Clotet B. Screening NK-, 85. Hoyer KK, Dooms H, Barron L, Abbas AK. Interleukin-2 in the B-and T-cell phenotype and function in patients suffering from development and control of inflammatory disease. Immunol Rev. Chronic Fatigue Syndrome. J Transl Med. 2013;11:68. 2008;226:19–28. 63. Lloyd AR, Wakefield D, Boughton CR, Dwyer JM. Immunologi‑ 86. Giri KR, De Beaurepaire L, Jegou D, Lavy M, Mosser M, Dupont A, Fleuris‑ cal abnormalities in the chronic fatigue syndrome. Med J Aust. son R, Dubreil L, Collot M, Van Endert P: Molecular and functional diver‑ 1989;151:122–4. sity of distinct subpopulations of extracellular vesicles from stressed 64. Maher KJ, Klimas NG, Fletcher MA. Chronic fatigue syndrome is pancreatic beta cells: implications for autoimmunity. BioRxiv 2020. associated with diminished intracellular perforin. Clin Exp Immunol. 87. Konadu KA, Chu J, Huang MB, Amancha PK, Armstrong W, Powell MD, 2005;142:505–11. Villinger F, Bond VC. Association of cytokines with exosomes in the 65. Andrew D, Aspinall R. Age-associated thymic atrophy is linked to a plasma of HIV-1–seropositive individuals. J Infect Dis. 2015;211:1712–6. decline in IL-7 production. Exp Gerontol. 2002;37:455–63. 88. Tokarz A, Kuśnierz-Cabala B, Szuścik I, Kapusta M, Konkolewska M, Żurakowski A, Georgescu A, Stępień E. Extracellular vesicles participate Giloteaux et al. J Transl Med (2020) 18:387 Page 17 of 17 in the transport of cytokines and angiogenic factors in diabetic patients 99. Adar T, Shteingart S, Ya’acov AB. Shitrit AB-G, Goldin E: from airway with ocular complications. Folia Med. Cracov. 2015;55:35–48. inflammation to inflammatory bowel disease: eotaxin-1, a key regulator 89. Loetscher M, Loetscher P, Brass N, Meese E, Moser B. Lymphocyte-spe‑ of intestinal inflammation. Clinical Immunol. 2014;153:199–208. cific chemokine receptor CXCR3: regulation, chemokine binding and 100. Mochizuki M, Bartels J, Mallet AI, Christophers E, Schröder J-M. IL-4 gene localization. Eur J Immunol. 1998;28:3696–705. induces eotaxin: a possible mechanism of selective eosinophil recruit‑ 90. Angiolillo AL, Sgadari C, Taub DD, Liao F, Farber JM, Maheshwari S, ment in helminth infection and atopy. J Immunol. 1998;160:60–8. Kleinman HK, Reaman GH, Tosato G. Human interferon-inducible 101. García JJ, Cidoncha A, Bote ME, Hinchado MD, Ortega E. Altered protein 10 is a potent inhibitor of angiogenesis in vivo. J Exp Med. profile of chemokines in fibromyalgia patients. Ann Clin Biochem. 1995;182:155–62. 2014;51:576–81. 91. Cole AM, Ganz T, Liese AM, Burdick MD, Liu L, Strieter RM. Cutting edge: 102. Zhang Z, Cherryholmes G, Mao A, Marek C, Longmate J, Kalos M, IFN-inducible ELR − CXC chemokines display defensin-like antimicrobial Amand RPS, Shively JE. High plasma levels of MCP-1 and eotaxin pro‑ activity. J Immunol. 2001;167:623–7. vide evidence for an immunological basis of fibromyalgia. Exp Biol Med. 92. Xia MQ, Bacskai BJ, Knowles RB, Qin SX, Hyman BT. Expression of the 2008;233:1171–80. chemokine receptor CXCR3 on neurons and the elevated expression of 103. Roy-O’Reilly M, Ritzel RM, Conway SE, Staff I, Fortunato G, McCullough its ligand IP-10 in reactive astrocytes: in vitro ERK1/2 activation and role LD: CCL11 (Eotaxin-1) levels predict long-term functional outcomes in in Alzheimer’s disease. J Neuroimmunol. 2000;108:227–35. patients following ischemic stroke. Transl Stroke Res. 2017. 8:578-584. 93. Sui Y, Stehno-Bittel L, Li S, Loganathan R, Dhillon NK, Pinson D, Nath A, 104. Bettcher BM, Fitch R, Wynn MJ, Lalli MA, Elofson J, Jastrzab L, Mitic L, Kolson D, Narayan O, Buch S. CXCL10-induced cell death in neurons: Miller ZA, Rabinovici GD, Miller BL. MCP-1 and eotaxin-1 selectively role of calcium dysregulation. Eur J Neurosci. 2006;23:957–64. and negatively associate with memory in MCI and Alzheimer’s disease 94. Ransohoff RM, Hamilton TA, Tani M, Stoler MH, Shick HE, Major JA, Estes dementia phenotypes. Alzheimer’s Dementia. 2016;3:91–7. ML, Thomas DM, Tuohy VK. Astrocyte expression of mRNA encoding 105. Huber AK, Giles DA, Segal BM, Irani DN. An emerging role for eotaxins in cytokines IP-10 and JE/MCP-1 in experimental autoimmune encephalo‑ neurodegenerative disease. Clinical Immunology. 2018;189:29–33. myelitis. FASEB J. 1993;7:592–600. 106. Boulware DR, Bonham SC, Meya DB, Wiesner DL, Park GS, Kambugu A, 95. Dyer KD, Percopo CM, Fischer ER, Gabryszewski SJ, Rosenberg HF. Pneu‑ Janoff EN, Bohjanen PR. Paucity of initial cerebrospinal fluid inflamma‑ moviruses infect eosinophils and elicit MyD88-dependent release of tion in cryptococcal meningitis is associated with subsequent immune chemoattractant cytokines and interleukin-6. Blood J Am Soc Hematol. reconstitution inflammatory syndrome. J Infect Dis. 2010;202:962–70. 2009;114:2649–56. 107. Chang EE, Yen CM. Eosinophils chemoattracted by eotaxin from 96. Klein RS, Lin E, Zhang B, Luster AD, Tollett J, Samuel MA, Engle M, cerebrospinal fluid of mice infected with Angiostrongylus cantonensis Diamond MS. Neuronal CXCL10 directs CD8 + T-cell recruitment and assayed in a microchamber. Kaohsiung J Med Sci. 2004;20:209–15. control of West Nile virus encephalitis. J Virol. 2005;79:11457–66. 108. Kielian T. Microglia and chemokines in infectious diseases of the nerv‑ 97. Wuest TR, Carr DJ. Dysregulation of CXCR3 signaling due to CXCL10 ous system: views and reviews. Front Biosci. 2004;9:50. deficiency impairs the antiviral response to herpes simplex virus 1 infection. J Immunol. 2008;181:7985–93. 98. Pope SM, Brandt EB, Mishra A, Hogan SP, Zimmermann N, Matthaei KI, Publisher’s Note Foster PS, Rothenberg ME. IL-13 induces eosinophil recruitment into Springer Nature remains neutral with regard to jurisdictional claims in pub‑ the lung by an IL-5–and eotaxin-dependent mechanism. J Allergy Clini‑ lished maps and institutional affiliations. cal Immunol. 2001;108:594–601. Ready to submit your research ? Choose BMC and benefit from: • fast, convenient online submission • thorough peer review by experienced researchers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations • maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions