Myalgic Encephalomyelitis/Chronic Fatigue Syndrome

Psychoneuroendocrinology 113 (2020) 104578

Contents lists available at ScienceDirect

Psychoneuroendocrinology

journal homepage: www.elsevier.com/locate/psyneuen

The role of low-grade inﬂammation in ME/CFS (Myalgic Encephalomyelitis/ Chronic Fatigue Syndrome) - associations with symptoms T

Martin A Jonsjöa,b,f,*, Gunnar L Olssona,b, Rikard K Wicksellc, Kjell Alvingd, Linda Holmströma,e, Anna Andreassonf,g a Behavior Medicine, Karolinska University Hospital, Stockholm, Sweden b Department of Physiology and Pharmacology, Karolinska Institutet, Stockholm, Sweden c Dept. of Clinical Neuroscience, Karolinska Institutet, Stockholm, Sweden d Dept. of Women's and Children's Health, Uppsala University, Uppsala, Sweden e Dept. of Women's and Children's Health, Karolinska Institutet, Stockholm, Sweden f Stress Research Institute, Stockholm University, Stockholm, Sweden g Dept. of Medicine Solna, Karolinska Institutet, Stockholm, Sweden

ARTICLE INFO ABSTRACT

Keywords: Background: Patients with Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS) often present with a Myalgic encephalomyelitis range of flu-like symptoms resembling sickness behavior as well as widespread pain and concentration deficits. Chronic fatigue syndrome The aim of this study was to explore the association between inflammatory markers previously shown to be ME/CFS related to fatigue severity in ME/CFS and common ME/CFS symptoms post-exertional fatigue, impaired cog- Cytokines nitive processing, musculoskeletal pain and recurrent flu-like symptoms, and the moderating effect of sex on Symptoms these associations. Methods: 53 adult patients diagnosed with ME/CFS at a specialist clinic were included in the study. Fasting blood plasma was analyzed using the Olink Proseek Multiplex Inflammation panel (β-NGF, CCL11, CXCL1, CXCL10, IL-6, IL-7, IL-8, IL-10, IL-18, TGF-α, TGF-β-1 and SCF) and BioRad Human Cytokine Type 1 assay (TNF- α). Participants rated the average severity of symptoms (0–10) based on the 2011 International Consensus Criteria of ME/CFS during a structured clinical interview. Associations between inflammatory markers and symptom severity were analyzed using bivariate correlations and moderated regression analyses bootstrapped with 5000 repetitions. Results and conclusions: Only β-NGF was associated with the fatigue severity measure. However, higher levels of CCL11, CXCL10, IL-7, TNF-α and TGF-β-1 were significantly associated with higher levels of impaired cognitive processing and musculoskeletal pain, and sex was a significant moderator for CXCL10, IL-7 and TGF-β-1. Future studies should investigate the relationship between inflammatory markers and key symptoms in ME/CFS in a longitudinal design in order to explore if and for whom low-grade inflammation may contribute to illness development.

1. Introduction recurrent flu-like symptoms pertaining to immune activation (e.g. malaise, tender lymph nodes and feverishness), as well as musculoskeletal Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS) is pain and neurocognitive impairment (Carruthers et al., 2003). characterized by a large heterogeneity in symptoms, severity and So far, no consistent illness-specific findings have been shown in treatment response. It has been suggested that this may be due to dif- clinically used inflammatory markers, such as erythrocyte sedimenta- ferences in pathophysiological and etiological mechanisms (Aslakson tion rate or C-reactive protein. However, currently used clinical tests et al., 2009). ME/CFS is characterized by longstanding unclear fatigue may fail to detect pathophysiological processes in ME/CFS such as interfering with functioning (personal, social and work-related dis- neurological infections and inflammation (Davis and Tyler, 2005; ability) (2). Importantly, patients present with an array of additional Ghose et al., 2019) and an increasing number of studies show various

⁎ Corresponding author at: Functional Area Medical Psychology, Functional Unit Behavior Medicine, Karolinska University Hospital Solna, QA25, 171 76, Stockholm, Sweden. E-mail address: [email protected] (M.A. Jonsjö). https://doi.org/10.1016/j.psyneuen.2019.104578 Received 5 July 2019; Received in revised form 4 December 2019; Accepted 20 December 2019 0306-4530/ © 2020 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/). M.A. Jonsjö, et al. Psychoneuroendocrinology 113 (2020) 104578 imbalances in levels of cytokines in ME/CFS, both in blood and in overnight fast. Samples were stored in room temperature before cen- cerebrospinal fluid (Hornig et al., 2016, 2015; Montoya et al., 2017; trifugation at 2000 rpm for 10 min, and was stored at -70 degrees Russell et al., 2016). In a recent study by Montoya et al., a significant Celsius until analyses. Blood plasma was analyzed using the Olink relationship between high levels of inflammatory markers and fatigue Proseek Multiplex Inflammation panel measuring 92 different analytes severity in ME/CFS was found, however more research is needed to (Olink, Uppsala, Sweden). investigate if and for whom longstanding low-grade inflammatory For this study, 13 inflammatory markers were chosen based on processes are related to symptoms in ME/CFS (Montoya et al., 2017). previous research. Three sickness behavior-related inflammatory mar- Furthermore, there is a scarcity of studies investigating the relationship kers (tumor necrosis factor (TNF)-α, interleukins (IL)-6 and (IL)-8) and between low-grade inflammation and other primary symptoms in ME/ 10 inflammatory markers previously shown to be related to fatigue CFS besides fatigue. severity in ME/CFS; IL-7, IL-10, IL-18, Eotaxin (CCL11), C-X-C motif The interaction between the immune system and the brain might chemokine (CXCL)1, CXCL10, Latency-associated peptide transforming contribute to the understanding of symptom variation in ME/CFS. In growth factor beta (LAP-TGF-β)-1, Transforming growth factor (TGF)- short, when the immune system is activated, inflammatory cytokines α, Stem cell factor (SCF) and Beta-nerve growth factor (β-NGF) were are released and soon the afflicted person will present with sickness included. The Human Cytokine Type 1 assay (BioRad, Hercules, CA, behavior such as fatigue, malaise, anxiety, anhedonia, worsened mood, USA) was used to ascertain levels of TNF-α as this marker was not reduced social interaction and increased pain sensitivity (Dantzer et al., detected in half of the samples using the Olink panel. All other markers 2008). The sickness behavior response has been considered a key me- had adequate detectability. chanism in the recovery from acute illness, while persistent sickness behavior may contribute to the debilitating symptoms in some patients 2.2.2. ME/CFS symptoms & illness duration with ME/CFS. To date, the relationships between low-grade in- Data on illness duration and symptoms were collected as part of the flammation and the commonly reported post-exertional, flu-like, mus- visit to the ME/CFS specialist clinic. A structured clinical interview with culoskeletal and neurocognitive symptoms in ME/CFS are still unclear. each participant was carried out by a senior physician. As part of the The aim of this study was to preliminarily explore the association interview, participants rated the average severity of symptoms, based between inflammatory markers and common symptoms in ME/CFS. on the 2011 International Consensus Criteria (Carruthers et al., 2011), Specifically, we wanted to investigate (a) the association between on a numerical rating scale from 0 (not present) to 10 (worst imagin- sickness behavior-related inflammatory markers and recurrent flu-like able level). Patients also reported duration of the illness, which was symptoms in ME/CFS, and (b) the associations between inflammatory cross-referenced through medical records. Symptoms were selected markers previously shown to be related to fatigue severity in ME/CFS based on (a) sickness behavior similarity (recurrent flu-like symptoms, and common ME/CFS symptoms post-exertional fatigue, impaired i.e. general malaise) and (b) previous ME/CFS research and our clinical cognitive processing and musculoskeletal pain, and (c) the moderating experience of the symptoms reported by patients as most debilitating effect of biological sex on these associations, as there is a lack of studies (post-exertional fatigue; impaired cognitive processing, i.e. slowed in ME/CFS research exploring sex differences. We also wanted to pro- thought and difficulty concentrating, and; musculoskeletal pain). vide data on inflammatory markers and symptoms as a supplement in order to enable analyses on the relationships between inflammatory 2.3. Statistical analyses markers and symptoms in ME/CFS in larger materials with data from multiple research groups and studies. All data were analyzed using SPSS version 25 (IBM Corp. IBM SPSS statistics for Macintosh, version 25. Armonk, NY: IBM Corp, 2017). 2. Methods Patient characteristics, inflammatory markers and symptoms were analyzed using descriptive statistics. Bivariate Spearman correlations 2.1. Participants and procedure were utilized for investigating (a) the association between sickness behavior-related inflammatory markers (TNF-α, IL-6 and IL-8) and re- 53 adult patients diagnosed with ME/CFS were included in the current flu-like symptoms and; (b) the association between in- study. Participants were recruited consecutively as part of the diag- flammatory markers implicated in ME/CFS (IL-7, IL-10, IL-18, CCL11, nostic assessment at a tertiary specialist clinic after referral, from 2013 CXCL1, CXCL10, LAP-TGF-beta-1, TGF-α, SCF and β -NGF) and post- to 2016, as described previously (Jonsjö et al., 2019). After referral, exertional fatigue, impaired cognitive processing and musculoskeletal participants were assessed by a physician and a psychologist in order to pain. The participants biological sex was explored as a moderator of the exclude other potential causes for symptomatology. Patients were ex- association between inflammatory markers (predictor) and symptoms cluded from the study if: (a) somatic comorbidity which could explain (outcome) using the PROCESS SPSS Macro version 3.3, model 1, for symptomatology was present; (b) they were non-adherent during the moderated regression analyses bootstrapped with 5000 repetitions assessment phase; (c) they did not speak Swedish; (d) they had ongoing (Hayes, 2018). Inflammatory markers were standardized so that the treatment known to modulate immune functioning; (f) they had psy- coefficient corresponds to the difference in symptom rating between chiatric comorbidity meeting criteria for schizophrenia, bipolar dis- two subjects whose levels of inflammatory marker differ by 1 SD order or major depressive disorder. All participants fulfilled the 1994 (hereafter referred to as b coefficient). CDC and 2011 ICC criteria for ME/CFS. All participants presented with Data on all inflammatory markers from the Olink panel, BioRad data longstanding unexplained fatigue, post-exertional malaise and on TNF-α and reported symptom ratings are provided as supplementary symptom exacerbation as well as prolonged recovery period after even information accompanying this paper at the Dryad data repository mild to moderate mental, physical or emotional effort. The study was (https://doi.org/10.5061/dryad.f1vhhmgsb). approved by the Regional Ethical Review Board in Stockholm, Sweden (Dnr 2015/370-31/4) and all participants gave written informed con- 3. Results and discussion sent. Participants (mean age 48.1 years, SD 10.4) were mainly women 2.2. Measures (66.1 %), with a mean illness duration of 12.2 (SD = 7.6) years. Higher levels of TNF-α was significantly associated with recurrent flu-like 2.2.1. Inflammatory markers symptoms (mean = 6.2, SD = 2.6; Spearman’s ρ = 0.38, p < .01; Blood samples were taken at the Karolinska University Hospital Moderated regression analysis independent of sex b = 0.74, p = .01), Solna laboratory in the morning, between 8 a.m. and 12 a.m., after an an association that was not moderated by sex (b=0.20,p= .78). This is

2 M.A. Jonsjö, et al. Psychoneuroendocrinology 113 (2020) 104578

Table 1 Spearman correlations between inﬂammatory markers previously shown to be related to fatigue severity in ME/CFS and symptoms.

Post-exertional fatigue Impaired cognitive processing Musculoskeletal pain

Mean (SD) 7.69 (2.09) 6.64 (1.82) 5.46 (2.61) ρ p ρ p ρ p β-NGF .21 0.17 .30* 0.04 .31* 0.03 CCL11 .17 0.27 .21 0.15 .10 0.50 CXCL1 −.09 0.55 −.17 0.25 .34* 0.02 CXCL10 −.10 0.49 −.03 0.85 .25 0.10 IL7 −.05 0.75 .04 0.79 .29 0.05 IL10 −.09 0.55 −.24 0.10 .03 0.87 IL18 −.12 0.43 −.06 0.69 .01 0.96 LAP TGF-β-1 −.12 0.42 −.07 0.67 .34* 0.02 SCF .02 0.92 .14 0.34 −.06 0.70 TGF-α .06 0.71 .03 0.86 .17 0.27

consistent with previous studies in the sickness behavior literature, for pain only for men (bIL-7=4.25, p < .01 and bTGF-β=2.71, p < .001), example in experimental studies where elevations of pro-inflammatory indicating potential differences in the effect of low-grade inflammation cytokines and sickness symptoms after stimulation of the immune for musculoskeletal pain between women and men in ME/CFS. Higher system are repeatedly demonstrated (Dantzer et al., 2008). No sig- levels of β-NGF was associated with more musculoskeletal pain without nificant associations were seen between IL-6 (ρ=-0.20, p = .18; b=- evidence for a moderation effect of sex (Table 2). 0.56, p = .09) and IL-8 (ρ = 0.07, p = .64; b = 0.19, p = .60) and flu- Comparing results from studies investigating levels of inflammatory like symptoms. markers is difficult, as variation in levels of cytokines in peripheral Higher levels of β-NGF was significantly associated with post-exer- blood can be affected by several factors including behavior, sleep, BMI, tional fatigue (Table 1 and 2), in both men and women. Sex moderated smoking, time of day and method of analysis (Nilsonne et al., 2016; the association between CXCL10 and post-exertional fatigue (Table 2). VanElzakker et al., 2019). There is a lack of studies investigating the Further analysis demonstrated a significant negative relationship for association between inflammatory markers and common symptoms, women (b=-0.89, p = .05) and a non-significant positive relationship other than fatigue, in the ME/CFS patient population. In contrast to the for men (b=0.82, p = .11). Higher levels of β-NGF and CCL11 were study by (Montoya et al. (2017), we found only one significant asso- significantly associated with impaired cognitive processing, no mod- ciation between inflammatory markers, β-NGF, and severity of post- eration by sex was found for these associations (Table 2). Notably, for exertional fatigue. In the present study, average severity of post-exer- musculoskeletal pain, IL-7 and TGF-β both demonstrated significant tional fatigue was used as a single measure. In the Montoya study, total main and interaction effects (Table 2). Further analyses showed a sig- scores on the MFI-20 were used to classify patients in fatigue severity nificant association between higher levels of the markers and higher groups (mild; moderate; severe), and the use of different measures of

Table 2 Moderated regression analyses of inﬂammatory markers previously shown to be related to fatigue severity in ME/CFS and symptoms.

Post-exertional fatigue Impaired cognitive processing Musculoskeletal pain

ba pbp bp β-NGF .58 .01 β-NGF .49 .03 β-NGF .95 .002 Sex .59 .40 Sex −.09 .85 Sex 1.54 .05 Interaction −1.16 .07 Interaction −.61 .25 Interaction −.83 .22 CCL11 .39 .20 CCL11 .56 .04 CCL11 .40 .28 Sex .97 .24 Sex .09 .86 Sex 1.86 .03 Interaction −.98 .25 Interaction .53 .28 Interaction −.86 .29 CXCL1 −.22 .58 CXCL1 −.22 .48 CXCL1 .55 .22 Sex .75 .33 Sex −.08 .89 Sex .97 .17 Interaction −.47 .65 Interaction −.84 .15 Interaction −1.37 .13 CXCL10 −.38 .27 CXCL10 −.04 .90 CXCL10 .23 .62 Sex .82 .29 Sex .02 .98 Sex 1.71 .04 Interaction −1.70 .01 Interaction −.26 .59 Interaction −.09 .93 IL-7 .59 .31 IL-7 .40 .15 IL-7 1.19 .005 Sex .17 .79 Sex −.30 .56 Sex .53 .41 Interaction −1.28 .51 Interaction −.49 .55 Interaction −4.44 < .001 IL-10 −.07 .77 IL-10 −.53 .21 IL-10 −.10 .76 Sex .78 .34 Sex −.04 .94 Sex 1.72 .04 Interaction −.53 .24 Interaction −1.05 .10 Interaction −.97 .10 IL-18 −.09 .70 IL-18 .03 .90 IL-18 .12 .74 Sex .84 .34 Sex .02 .96 Sex 1.74 .05 Interaction −.86 .16 Interaction −.08 .86 Interaction −.26 .72 TGF-β −.48 .17 TGF-β −.22 .35 TGF-β 1.15 .004 Sex .73 .34 Sex .00 .99 Sex 1.61 .03 Interaction −.22 .83 Interaction .19 .72 Interaction −2.26 .004 SCF −.16 .51 SCF .07 .77 SCF −.20 .57 Sex .73 .34 Sex .00 .99 Sex 1.63 .055 Interaction −.26 .65 Interaction −.11 .79 Interaction −.08 .92 TGF-α −.07 .76 TGF-α −.05 .79 TGF-α .32 .36 Sex .72 .36 Sex −.05 .93 Sex 1.92 .04 Interaction −.10 .83 Interaction .18 .64 Interaction −1.13 .20

a The b-coeﬃcient corresponds to the diﬀerence in symptom rating between two subjects with cytokine values 1 SD apart.

3 M.A. Jonsjö, et al. Psychoneuroendocrinology 113 (2020) 104578 fatigue make comparisons between the studies difficult (Montoya et al., Declaration of Competing Interest 2017). Furthermore, the natural variation of cytokines in peripheral blood may also explain the lack of associations. In addition, the rela- The authors declare that they have no conflict of interest. tively low power in the present study might have led to associations being missed. References The present study is limited regarding the cross-sectional design, small sample size likely leading to type II errors, and large number of Aslakson, E., Vollmer-Conna, U., Reeves, W.C., White, P.D., 2009. Replication of an inflammatory markers measured. It should be noted that the analyses empirical approach to delineate the heterogeneity of chronic unexplained fatigue. fi Popul. Health Metr. 7, 17. are not adjusted for multiple comparisons, and several of the ndings Carruthers, B.M., van de Sande, M.I., De Meirleir, K.L., Klimas, N.G., Broderick, G., are likely due to type 1 errors, why the results should be interpreted Mitchell, T., et al., 2011. Myalgic encephalomyelitis: international consensus criteria. with caution until replicated in future studies. Patients in the present J. Intern. Med. 270 (4), 327–338. Carruthers, B.M., Jain, A.K., De Meirleir, K.L., Peterson, D.L., Klimas, N.G., Lerner, A.M., study had a relatively high degree of functioning on the ME/CFS et al., 2003. Myalgic Encephalomyelitis/Chronic fatigue syndrome. J. Chronic spectrum (i.e. not homebound) which could reduce the variance in Fatigue Syndr. 11 (1), 7–115. inflammatory markers and symptoms. Future studies on relations be- Dantzer, R., O’Connor, J.C., Freund, G.G., Johnson, R.W., Kelley, K.W., 2008. From in- fl tween low-grade inflammation and key symptoms in ME/CFS should ammation to sickness and depression: when the immune system subjugates the brain. Nat. Rev. Neurosci. 9 (1), 46–56. investigate dimensions of fatigue and other symptoms of high relevance Davis, L.E., Tyler, K.L., 2005. Molecular diagnosis of CNS viral infections. J. Neurol. to ME/CFS as multifaceted phenomena that may vary both between and Neurosurg. Psychiatr. 76 (1), 10. within individuals (Karshikoff et al., 2017). Furthermore, diverging Ghose, C., Ly, M., Schwanemann, L.K., Shin, J.H., Atab, K., Barr, J.J., et al., 2019. The virome of cerebrospinal fluid: viruses where we once thought there were none. Front. results may indicate that ME/CFS consists of a heterogenous illness Microbiol. 10, 2061. population, and longstanding low-grade inflammation may have a dif- Hayes, A.F., 2018. Introduction to Mediation, Moderation, and Conditional Process ferential influence on symptoms across individuals. A longitudinal Analysis: A Regression-based Approach, second ed. The Guilford Press, New York, NY. study with multiple measurements will improve investigations of in- Hornig, M., Gottschalk, G., Peterson, D.L., Knox, K.K., Schultz, A.F., Eddy, M.L., et al., flammatory markers and their importance for symptom development in 2016. Cytokine network analysis of cerebrospinal fluid in myalgic encephalomyelitis/ ME/CFS. In order to identify potential disease phenotypes, there is a chronic fatigue syndrome. Mol. Psychiatry 21 (2), 261–269. Hornig, M., Montoya, J.G., Klimas, N.G., Levine, S., Felsenstein, D., Bateman, L., et al., need to build large materials with data from multiple studies, including 2015. Distinct plasma immune signatures in ME/CFS are present early in the course data on illness duration which has been implicated as important for the of illness. Sci. Adv. 1 (1). variation of inflammatory markers in ME/CFS (Hornig et al., 2015). We Jonsjö, M.A., Wicksell, R.K., Holmström, L., Andreasson, A., Olsson, G.L., 2019. Acceptance & commitment therapy for ME/CFS (Chronic Fatigue Syndrome) – a anticipate that the data from the present study will be used by other feasibility study. J. Contextual Behav. Sci. 12, 89–97. researchers for this purpose. Karshikoff, B., Sundelin, T., Lasselin, J., 2017. Role of inflammation in human fatigue: relevance of multidimensional assessments and potential neuronal mechanisms. Front. Immunol. 8, 21. Montoya, J.G., Holmes, T.H., Anderson, J.N., Maecker, H.T., Rosenberg-Hasson, Y., Funding Valencia, I.J., et al., 2017. Cytokine signature associated with disease severity in chronic fatigue syndrome patients. Proc. Natl. Acad. Sci. U.S.A. 114 (34), – This work was supported by the research fund (Risk Hälsa) at E7150 E7158. Nilsonne, G., Lekander, M., Akerstedt, T., Axelsson, J., Ingre, M., 2016. Diurnal variation Skandia Insurance Company, Ltd, Sweden. The funder had no role in of circulating Interleukin-6 in humans: a meta-analysis. PLoS One 11 (11), e0165799. study design, data collection, analysis, interpretation, decision to pub- Russell, L., Broderick, G., Taylor, R., Fernandes, H., Harvey, J., Barnes, Z., et al., 2016. lish, or preparation of the manuscript. The funder has no financial in- Illness progression in chronic fatigue syndrome: a shifting immune baseline. BMC Immunol. 17, 3. terest, in Sweden or internationally, in classifying ME/CFS patients as VanElzakker, M.B., Brumfield, S.A., Lara Mejia, P.S., 2019. Neuroinflammation and cy- psychiatric. tokines in myalgic Encephalomyelitis/Chronic fatigue syndrome (ME/CFS): a critical review of research methods. Front. Neurol. 9 (1033).