Analysis of the Relationship between Immune Dysfunction and Symptom Severity in Patients with Chronic Fatigue Syndrome/Myalgic Encephalomyelitis (CFS/ME)

Objective: Chronic Fatigue Syndrome/Myalgic Encephalomyelitis (CFS/ME) is a disabling illness, characterised by persistent, debilitating fatigue and a multitude of symptoms. Immunological alterations are prominent in CFS/ME cases, however little is known about the relationship between CFS/ME severity and the extent of immunological dysfunction. The purpose of this study was to assess innate and adaptive immune cell phenotypes and function of two groups of CFS/ME patients, bedridden (severe) and mobile (moderate).


Introduction
Chronic Fatigue Syndrome/Myalgic Encephalomyelitis (CFS/ME) is a severe physically and cognitively incapacitating illness diagnosed by symptom-specific criteria [1][2][3]. CFS/ME presents as a multifactorial illness that varies greatly in the nature of onset and severity of symptom presentation [1,2,4,5]. A key characteristic is debilitating fatigue that lasts for a period of 6 or more months that has a critical effect on a patient's daily activities are critically affected [1][2][3]6].The severity of symptoms can vary greatly in CFS/ME. For example patients with moderate symptoms are able to maintain some normal daily activities with slight reduced mobility while those severely affected by CFS/ ME experience high levels of daily fatigue and are therefore typically housebound [7].
Currently, there is no known cause for CFS/ME although research has demonstrated consistent immunological dysfunction associated with the illness [8][9][10][11][12]. We have previously been the only research group to have examined Natural Killer (NK) cell function, phenotype and receptors in housebound severe patients in comparison to a healthy control group [13]. Housebound severe patients had significantly reduced NK cell cytotoxic activity when compared with the moderately affected patients and there was an increase in the Killer Immunoglobulin-like Receptor (KIRs) KIR3DL1 in the moderate patients, highlighting that differing levels of severity may also have varying levels of immune perturbation [13].
The most consistent immunological finding in CFS/ME is significantly reduced NK cell cytotoxic activity [10,[13][14][15]. This study is one of the first to assess those housebound and severe CFS/ME patients in comparison to a moderate and mobile CFS/ME patient subgroup. Segregation of patients into moderate/mobile and severe/ houseboundCFS/MEsubgroups may elucidate further immunological markers that may explain the pathomechanism of the illness. Hence, the purpose of this study was to investigate phenotypic and functional parameters of innate (NK cells, neutrophils, monocytes, dendritic cells (DCs)) and adaptive (gamma delta (γδ), iNKT, CD8 + T cells, B cells)

Sample preparation and routine measures
A non-fasting blood sample of 50 mL was collected from the antecubital vein of participants into lithium heparinised and ethylenediamine tetraacetic acid (EDTA) tubes. Blood was collected between 8:30 am and 11:30 am and samples were analysed within 12 hours of collection. Initial full blood count assessment was undertaken by Pathology Queensland to determine levels of white blood cell and red blood cell markers.

Natural killer cell cytotoxic activity analysis
NK cell cytotoxic activity was performed as described previously [9,13]. Density gradient centrifugation using Ficoll-hypaque (Sigma, St Louis, MO) was used to isolate peripheral blood mononuclear cells (PBMCs) from whole blood. Isolated PBMCs were labelled with 0.4% PKH-26 (Sigma, St Louis, MO) and incubated with K562 cells for 4 hours at the following effector (NK cell) to target (K562) ratios, 12.5:1, 25:1 and 50:1. Cell death was analysed using Annexin-V-FITC and 7-AAD reagents (BD Biosciences, San Diego, CA) and the ability of NK cells to lyse target K562 cells was measured on the flow cytometer (Becton Dickinson Immunocytometry Systems).

Intracellular analysis
Density gradient centrifugation using Ficoll-hypaque (Sigma, St Louis, MO) was used to isolate PBMCs from EDTA whole blood. PBMCs were adjusted to 1×10 7 cells/mL and stained with monoclonal antibodies for Tregulatory cell (Treg) phenotypes, NK lytic proteins and CD8 lytic proteins as described [9,13] (Supplementary Table 1). The Treg phenotypes were assessed as PBMCs were permeablised and fixed with buffers containing diethylene glycol and formaldehyde before being stained with FOXP3. After washing with Phosphate Buffered Saline (PBS) (GibcoBiocult, Scotland), cells were analysed on the flow cytometer (Becton Dickinson Immunocytometry Systems) where the expression of FOXP3 + Tregs was determined on CD4 + CD25 + CD127 low T cells [9]. NK and CD8 T cell lytic proteins were assessed as previously described [9]. Cells were incubated for 30 mins in Cytofix then permwash was added. Perforin, granzyme A and granzyme B monoclonal antibodies were added to cells and incubated for 30 minutes in the dark at room temperature. Cells were washed and analysed on the flow cytometer where perforin, granzyme A and granzyme B expression was measured in NK and CD8 T cells.

Whole blood analysis
Appropriate antibodies (Supplementary Table 1) were added to whole blood samples and incubated for 30 minutes. Following which cells were lysed, washed, fixed and analysed on the flow cytometer. Neutrophil, monocyte, DC, B cell and γδ T cell phenotypes were assessed using appropriate antibodies (Supplementary Table 1

Data and statistical analysis
Statistical analysis was performed using SPSS statistical software version 21.0. All experimental data represented in this study are reported as plus/minus the standard error of the mean (± SEM). Comparative assessments among the three participant groups (control, moderate CFS/ME and severe CFS/ME) were performed with the analysis of variance test (ANOVA). The LSD Post Hoc test was used to determine p values of significance and statistical significance was set at an alpha criterion at p<0.05. Pearsons correlation was conducted on significant parameters to determine correlates where significance was accepted as p<0.01. Outliers were identified using a boxplot technique on SPSS whereas extreme outliers were highlighted based on lying beyond the plot's whiskers [17]. Outliers were handled by eliminating particular data points from the analysis [17].

Participants
Data were available for a total of 63 participants (22 control, 23 moderate CFS/ME and 18 severe CFS/ME). The mean ages for the control, moderate CFS/ME and severe CFS/ME were 40.14 ± 2.38, 42.52 ± 2.63 and 39.56 ± 1.51 respectively. There was no statistical difference for age between participant groups. All CFS/ME participants from the moderate and severe CFS/ME participant groups satisfied the CDC criteria for CFS/ME. The severe participant group demonstrated significantly worsened scores for symptoms in the FSS, Dr Bell's Disability Scale, the FibroFatigue Scale and the KPS (data not shown) when compared with the moderate CFS/ME group. Participant groups were also age and gender matched and there were no significant (p<0.05) differences between these parameters ( Table 1). The three participant groups consisted of predominantly females with the control, moderate CFS/ME and severe CFS/ME having 64%, 70% and 83% female participants and this was not statistically different between groups (Table 1).
Blood pressure, pulse, temperature and routine full blood counts were measured in all participants ( Table 2). There were no statistically significant differences in any clinical or routine full blood count parameters between the participant groups.

Reduced NK cytotoxic activity
In all three effector cell target ratios (12.5:1, 25:1, 50:1), there was  Age data is represented as Mean ± SEM in control (n=22), moderate CFS/ME (n=23) and severe CFS/ME (n=18). *signifies p<0.05 between participant groups. There were no significant differences in age or gender within the research groups. CFS/ME: Chronic Fatigue Syndrome/Myalgic Encephalomyelitis; Standard Error of the Mean. a significant reduction in cytotoxic activity of NK cells in moderate (p=0.001, 0.000, 0.017) and severe CFS/ME participants (p=0.000, 0.000, 0.001) ( Figure 1A). Cytotoxic activity was further reduced in the severe CFS/ME group for all three ratios although there was no statistical significance between the moderate and severe CFS/ME groups ( Figure 1A). NK cytotoxic activity at with a target cell ratio of 12.5:1 was positively correlated with the target cell ratios of 25:1 and 50:1 and the 25:1 ratio was also positively correlated to the ratio of 50:1 (p<0.01) (Supplementary Table 3).

No differences to lytic proteins in CD8 + T cells and NK cells
NK and CD8 + T cell Granzyme A, Granzyme B and perforin as well as CD25 + CD127 low CD4 + FOXP3 + Tregs were not significantly different between the groups (Data not shown).

Increased NK phenotypes and reduced KIR receptors in CFS/ ME
There was a significant increase in the number of CD56 bright CD16 -/dim and CD56 -CD16 + NK cells in severe CFS/ME compared with moderate (p=0.023, 0.049) and CD56 dim CD16 -NK cells were significantly increased in moderate CFS/ME compared with controls (p=0.045) ( Figures 1B and 1C). There was also a significant reduction in the percentage of CD56 bright CD16 -/dim CD158b + NK cells in the severe CFS/ME group compared with the moderate CFS/ME group (p=0.027) ( Figure 1D). Moderate CFS/ME patients had significantly reduced CD56 dim CD16 -CD158a/h + NK cell expression compared with the control group (p=0.003, 0.004). There was also a significant increase in the expression of CD94 + on CD56 dim CD16 -NK cells in both

Whole blood phenotypes
Plasmacytoid DCs (pDCs) were significantly higher in the moderate CFS/ME group (p=0.002). Myeloid DCs (mDCs) were not statistically different between any of the groups and CD14 -CD16 + DCs were higher in the severe CFS/ME group compared with the moderate CFS/ME (p=0.000) and control group (p=0.000) (Figure 2A). The number of memory and naïve B cells (cells/µL) was significantly increased in the severe CFS/ME group compared with the moderate CFS/ME group (p=0.025, 0.026). There was also a significantly reduced number of transitional B cells and B regulatory cells (Bregs) in the severe CFS/ME participants compared with the control CFS/ME participants (p=0.047, 0.041) ( Figure 2B). There was a significant reduction in the number of γδ1 T cells (cells/µL) with the phenotype CD45RA + CD27in the moderate and severe CFS/ME groups (p=0.007 and 0.018). γδ1 T cells (cells/µL) with the phenotype CD45RA -CD27were also significantly reduced in the moderate CFS/ME group (p=0.0142) (Figures 2C and  2D). γδ2 T cells (cells/µL) were not significantly different between the groups (Data not shown).
No significant differences in total neutrophil numbers, CD177 bright or CD177 dim neutrophils, true monocytes, proinflammatory, intermediate or classical monocytes, total CD8 + T cells or CD8 + T cell phenotypes (cells/µL) between the control, moderate CFS/ME or severe CFS/ME groups were found (p>0.05) (Data not shown).

Discussion
This is the first study to examine monocytes, neutrophils, DCs, CD8 T cells, γδ T cells, iNKT cells, Tregs and B cells in moderate CFS/ ME patients compared to severe CFS/ME patients and this is the first study to assess iNKT cells in CFS/ME. Previous literature has assessed immunological function in CFS/ME patients in relation to NK cells, monocytes, neutrophils, DCs, CD8 T cells, γδT cells, Tregs and B cells [13][14][15][18][19][20][21] however only one other study has examined immune parameters in moderate and severe CFS/ME patients [13].
This investigation confirmed previous findings that NK cell cytotoxic activity is consistently reduced in CFS/ME patients [12][13][14]20]. Reductions in cytotoxic activity may be associated with differential levels of cytotoxic molecules, including perforin, granzyme A and granzyme B which have been shown to be varied in cytotoxic NK cells and CD8 + T cells in CFS/ME [22,23]. NK cell cytotoxic activity can also be regulated by CD56 bright and CD56 dim NK cell phenotypes which have equivocal levels in CFS/ME patients [24][25][26]. This study also confirmed increases CD56 bright CD16 -/dim and CD56 -CD16 + NK cells in CFS/ME [25,26]. CD56 dim and CD56 bright NK cell phenotypes are primarily responsible for cytotoxic activity and cytokine release respectively. [27]. CD56 bright NK cells secrete immunoregulatory cytokines, such as IFN-γ, TNF-β, IL-10, IL-13 and GM-CSF which initiate an immune response and cytotoxic activities in other cells, including NK cells, B cells and DCs. CD56 bright NK cells also exhibit potent lymphokineactivated killer (LAK) activity which stimulates other lymphocytes to kill target cells, while CD56 dim NK cells are predominantly responsible for cytotoxic activity and antibody-dependent cellular cytotoxicity [27]. The increase in CD56 bright NK cells may serve as a regulatory mechanism to improve the reduced NK cytotoxic activity in CFS/ ME. This study has shown that NK cell dysfunction may vary between CFS/ME patient subgroups, with significant differences between the moderate and severe CFS/ME patients. The mechanism causing the reduced NK cytotoxic activity and increased NK phenotypes in CFS/ ME patients is unknown although potential genetic defects may affect NK cell function in CFS/ME [28].
KIR2DS1 is an activating NK cell receptor which promotes NK cytotoxic activity in a human leukocyte antigen (HLA) class-1 dependent manner [29]. In psoriasis, the KIR2DS1 gene is a strong predictor of disease where psoriasis patients demonstrate reduced KIR2DS1 genetic frequency [30]. This study was the first to find reduced expression of the KIR2DS1 in moderate and severe CFS/ME patients. This may be associated with the reduced ability of NK cells to successfully lyse target cells [24,[29][30][31]. Similarly, reductions in the corresponding inhibitory receptors KIR2DL1 and KIR2DL2/DL3 in CFS/ME may trigger reduced alloreactive KIR NK cell cytotoxic activity [29,30,[32][33][34]. Thus, changes in KIR receptors may be an important component in the CFS/ME illness mechanism. In our previous studies KIR3DL1 was increased in CFS/ME patients, however we were unable to confirm this finding and this may be related to the heterogeneity of CFS/ME [13].
CD94 is a cytokine induced receptor complex on NK cells which can also either activate or inhibit cell-mediated cytotoxicity in NK cells, dependent on NKG2 protein association which recruits SHP-1 tyrosine phosphatase and couples to tyrosine kinase [35]. CD94 NK receptors recognise HLA-E which presents HLA class 1 moleculederived peptides and inhibits NK cell-mediated lysis [36]. An increase in CD94 expression in CD56 dim CD16 -NK cells of both moderate and severe CFS/ME may be associated with an upregulated expression of HLA-E, which protects target cells from NK cell lysis and hence possibly reducing overall NK cell cytotoxic activity [24,37].
iNKT cells regulate disease conditions including type I diabetes and cancer via communication with a number of innate and adaptive immune cells [38,39]. Cellular interaction with iNKT cells results in the release of IL-10, causing reduced IL-12 in DCs and self-destructive ability in both T and B cells [38]. An increased number of iNKT cells in the severe CFS/ME participants may indicate the further promotion of iNKT cell proliferation in the bone marrow [40] and maybe related to an enhanced cell-mediated regulation of immunity in these patients [38]. The CD4 + and CD8a + iNKT cell subsets provide immunoregulation by releasing cytokines including IL-4, IL-5, IL-13, TNF-α and IFN-γ while the CD8 -CD4 -iNKT subset produces only IL-9 and IL-10 and induce cytotoxic activity [38,[41][42][43]. Reduced CD8 -CD4 -, CD8a -CD4and CD8a -CD4 + iNKT subsets in CFS/ME may reduce cytokine secretion required for maturation and activation of NK cells, DCs, B and T cells [42].
Increases in the CCR7 -iNKT cells in CFS/ME may affect the trafficking of B and T cell trafficking to secondary lymphoid organs, as CCR7 plays a role in immunosurveillance [44,45]. SLAM is a surface receptor essential for iNKT cell development and the production of cytokines, particularly IFN-γ [46][47][48]. Increased iNKTCCR7 -SLAMand reduced CCR7 -SLAM + expression in both moderate and severe CFS/ME compared to controls may be associated with a reduced ability of iNKT cells to interact and activate DC and T cells during an immune response [46]. Circulating iNKT cells also have a variable expression of the NK markers, CD56 and CD16, although little is known about the functional significance of these markers on the surface of iNKT cells [16]. Increases in CD56 and CD16 iNKT cell phenotypes in CFS/ME patients may be related to the altered expression of CD56 and CD16 NK cell phenotypes in the illness as iNKT cells are often dependent on NK cells, which are dysfunctional in CFS/ME. Similarly to NK cells, the alterations in iNKT cell markers displayed between moderate and severe CFS/ME patients may be due to NK or iNKT gene influencing an individual's susceptibility and extent of dysfunction [49].
pDCs are also responsible for modulating NK, T and B cell immune responses through antigen presentation and the release of cytokines and chemokines [50]. pDCs are particularly important in modulating and activating NK cell cytotoxic activity in response to a host viral infection through their secretion of IFN-α [51]. Increased pDCs in the moderate CFS/ME patients may be associated with increased NK cell activation and effector cell functioning. Increased pDCs in conjunction with reduced NK cytotoxic activity may highlight a reduced efficiency in cell-cell crosstalk and immune dysregulation in CFS/ME [51]. Elevated pDCs may also be linked to pDCs having an increased ability to become readily infected than mDCs, as found in HIV. This could potentially explain why there was no significant difference between mDC phenotypes in controls, moderate or severe CFS/ME [52]. Another DC phenotype, cytokine-producing CD14 -CD16 + DCs were also significantly increased in the severe CFS/ME patients compared to both controls and the moderate CFS/ME subgroup, potentially suggesting dysfunction in the secretion of inflammatory cytokines. This supports studies which have shown IL4, IL-10 and IL-12, primarily produced by CD14 -CD16 + DCs are found to be increased in CFS/ME [53][54][55][56][57]. These cytokines are important in the neutralisation of the Th1/ Th2 cytokine shift which is also altered in CFS/ME patients [56,57]. γδ T cells are sentinel cytotoxic cells involved in the elimination of bacterial infection, delayed-type hypersensitivity reactions, wound repair, antigen presentation and immunoregulation [58]. Effector memory γδT cells are responsible for cell migration to sites of inflammation and demonstrate NK-like functions such as the detection of abnormal MHC expression. Effector memory γδT cells also have potential for greater cytotoxic activity, tissue homing and rapid innatelike target recognition than central memory and naïve γδ T cell subsets [58]. Reduced γδ1 effector memory in both moderate and severe CFS/ ME and reduced γδ1 naive phenotypes in moderate CFS/ME may potentially be a reflection of the consistently reduced NK cytotoxic activity in the illness as these cells are similar to NK cytotoxic cells.
It has been previously suggested that B cell activation may be increased in CFS/ME [59]. The increased naïve and memory B cells shown in severe CFS/ME compared to moderate CFS/ME patients may be consistent with amplified B cell activation, particularly as B cells are regulated by T and NK cells which also demonstrate dysfunction in CFS/ME. The increased naïve and memory B cells were shown in only the severe CFS/ME group, potentially highlighting a significant difference between CFS/ME severity subgroups. This confirms previous studies where CFS/ME patients have demonstrated increased numbers of naive and transitional memory B cells [60]. The generation of memory B cells from naïve B cells is promoted by IL-4, IL-5, IL-13 and IL-10, secreted from CD4 + T cells. This suggests that potential increases in these cytokines, previously found in CFS/ME, may be triggering the increase in these B phenotypes [61]. Interestingly, transitional B cells are reduced in CFS/ME, indicating that the T cellmediated extrinsic signals that drive B cell progression into transitional B cells may be abnormal [62,63]. Reduced Bregs are associated with reduced interactions with pathogenic T cells via cell-cell contact which are important in suppressing inflammatory T cells as well as regulatory cytokines (such as IL-10, TGF-β) [64,65]. The anti-inflammatory cytokine IL-10 in particular, has been reported as increased in CFS/ ME patients and is often related to enhanced production and survival of B cells [66].
Overall, immunological dysfunction in CFS/ME patients occurs as a consequence of changes in cytotoxic activity, NK cell phenotypes, KIR receptors, iNKT, DCs, γδ T cells and B cell phenotypes. The severe CFS/ME subgroup of patients also experienced further immunological function in some instances. These findings suggest that immune perturbations may be further persistent in CFS/ME patients who experience more severe CFS/ME symptoms and hence may potentially dictate illness severity, as occurs in other diseases, such as rheumatoid arthritis (RA) [67][68][69].

Conclusion
This study is the first to assess a wide range of innate and adaptive immune cells in CFS/ME patients subgrouped by severity. Immune dysregulation was found in both moderate and severe patients with a consistent reduction in NK cytotoxic activity, alterations in B and iNKT cell phenotypes and an increase in CD14 -CD16 + DCs. Interestingly, CD14 -CD16 + DCs, total iNKTs and iNKT cell phenotypes differed between the moderate and severe CFS/ME patient subgroups. The findings of this study demonstrate that immune dysfunction appears to be related to the level of severity experienced by the patient hence severity subgroups may be important in identifying a specific disease mechanism in CFS/ME. Severity subgrouping of CFS/ME need to be considered in future studies as they may have implications for diagnosis and developing therapeutic strategies.