The clinical value of cytokines in chronic fatigue syndrome

TNF-α is a proinflammatory molecule with antitumor and antiviral effects and is thought to play a role in the pathogenesis of acquired immune deficiency syndrome and multiple sclerosis. It is the most commonly studied cytokine in the majority of CFS studies. However, the relationship between TNF-α and CFS remains controversial. Chao et al. established a cellular model of CFS by simulating peripheral blood mononuclear cells (PBMCs) with lipopolysaccharides. They reported increased levels of TNF-α in the model [24]. Another study found the similar result in non-adherent lymphocytes [25]. As for in vivo evidence, Milrad et al. reported that poor sleep quality was associated with increased levels of TNF-α and symptom severity in patients with CFS [3]. The authors reported that TNF-α levels are positively correlated with depressive symptoms in patients with CFS [26]; however, conflicting results were reported by few studies. Lidbury et al. found no significant differences in the circulating TNF-α levels between patients with CFS and controls [27]. Groven et al. reported that patients with CFS have higher TNF-α levels than healthy controls; however, the difference was not statistically significant (p = 0.056). TNF-α levels exhibited a weaker association with depression in patients with CFS than in healthy controls [28]. However, these studies did not corroborate with the results that TNF-α levels are positively correlated with CFS symptoms, probably owing to differences in the experimental conditions and study design, such as in vivo data used, small sample size, and different diagnostic criteria. Thus, we support the hypothesis that serum TNF-α is a relevant marker for the diagnosis of CFS.

Interferons (IFNs) can be categorized into two subtypes, namely, type I interferons, which include IFN-α and IFN-β, and type II interferons, which include IFN-γ. Both subtypes influence the development of CFS. Interestingly, in some patients undergoing treatment with IFN-α/β, a primary complaint is “severe fatigue” [29]; this is regarded as a major side effect of IFN-α/β. This may be an indirect evidence of a relationship between IFNs and CFS. In 1997, Vojdani reported upregulated circulating IFN-α levels accompanied by activated protein kinase R and induction of apoptosis in patients with CFS [30]. A subsequent study reported that the dysregulation of type I interferon pathway results in sustained upregulation of 2′-5′-oligoadenylate synthetase, which possibly contributes to the development of fatigue symptoms associated with CFS [29]. IFN-γ triggers several antiviral mechanisms related to chronic infection [31]. A comparative study demonstrated enhanced and persisting circulating TNF-α and IFN-γ levels in patients with CFS, similar to that observed in patients with acute B19 infection [32]. Another study reported significantly higher IFN-γ levels in PBMCs isolated from patients with CFS than from healthy individuals [18]. However, the immunopathology of CFS is not static, and the cytokine profile changes along the course of the disease. Differences in IFN-γ have been reported in patients with both short-and long-term disease. Markedly elevated IFN-γ levels have been associated with the early phase of CFS [33]. Moreover, IFN-γ is closely associated with CFS severity. Hardcastle reported that IFN-γ levels are higher in patients with severe CFS than in those with moderate disease [34]. Further, Montoya et al. reported that IFN-γ levels exhibit a significant linear upward correlation with CFS severity [22]. The above-mentioned evidence indicates that IFN-γ, rather than IFN-α/β, is a good biomarker in the early stage of CFS and that it can be used for clinical decisions regarding CFS severity.

TGF-β is a multifunctional cytokine involved in several biological activities, such as cell-cycle control, hematopoiesis, angiogenesis, chemotaxis, and immune responses. Many TGF-β isoforms have proven to be associated with immune and neuroendocrine regulation in patients with CFS. TGF-β1 plays an inhibitory role in macrophages and NK cells, in the proliferation of T or B cells, and in the maturation of cytotoxic T lymphocytes [57]. TGF-β3 has been suggested to partly mediate the association between plasma cortisol and downregulation of expression of some B cell genes [58]. A systematic review concluded that TGF-β levels were enhanced in patients with CFS in most studies [21]. Montoya et al. reported higher TGF-β levels in patients with CFS than in healthy controls, independent of CFS severity [22]. In addition, activin B, as a member of the TGF-β family, has been suggested as a novel biomarker due to its significantly increased levels in CFS patients [27]. However, recently, few studies have reported unchanged TGF-β levels in patients with CFS [58, 59]. The relationship between TGF-β and CFS and the value of alterations in the TGF-β levels require further investigation in the future.

In this section, we discussed the potential value of using circulating cytokines as diagnostic biomarkers for CFS. The main strengths of such an approach lie in the easy sample collection, sensitivity, and reduced invasiveness. However, several limitations must be considered with respect to this approach: (1) Because of the blood–brain barrier (BBB), changes in peripheral cytokines sometimes may not represent the changes in the central nervous system (CNS). In patients with CFS, activated cytokines might be confined to the brain [60]. (2) Measurement of circulating cytokines is technically complicated. Although cytokine indices are sensitive, they are easily affected by many factors; therefore, proper sample handling is critical. (3) Systematic errors may arise when using circulating cytokines: for example, Roerink et al. have reported that most cytokines remain in the intercellular environment, where actual cytokine levels (particularly those of IL-1) are generally below the threshold of detection [60], but in a later study of the same authors found that the normalized protein expression value of IL-12p40 and CSF-1 was significantly higher in patients with CFS [61]. Therefore, we believe that the values of circulating cytokines must be defined as “indirect” and “auxiliary.” Thus far, they cannot be employed as “independent diagnostic biomarkers” for CFS and should be used in conjunction with other indices for the diagnosis of CFS. We conclude that further investigation of cytokines in the CNS of CFS patients is indispensable.

CFS is a disorder that involves the CNS. Neuropsychological symptoms, such as depression and anxiety, are commonly observed in patients with CFS. A previous meta-analysis has reported that the primary cognitive problems in patients with CFS are deficits of attention, memory, and reaction time [62]. This result is in agreement with the findings of a study that involved 307 cases [63]. Clinically, cognitive behavior therapy (CBT) has been used to treat CFS patients [64]. Brain imaging research has also demonstrated CNS abnormalities in patients with CFS. Reduced gray matter volumes have been well documented in patients with CFS [65, 66]. A longitudinal MRI study has found a significant reduction in white matter volumes in the left inferior fronto-occipital fasciculus in patients with CFS [67]. These findings confirm the role of central neural mechanisms in the onset of CFS.

These alterations in brain structure and function can be explained by a variety of complex mechanisms related to CNS inflammation [68, 69]. It is well documented that inflammation in CNS plays a key role in the pathophysiology of CFS [56]. Nakatomi et al. employed positron emission tomography to investigate the inflammatory changes occurring in the brain during CFS. A wide range of inflammatory changes were observed in the cingulate cortex, thalamus, midbrain, hippocampus, amygdala, and pons in patients with CFS. Additionally, this study pointed out the range of cerebral inflammation was highly associated with the severity of neuropsychological symptoms [70]. Chan et al. also reported that abnormal hippocampal neurogenesis is related to depression in CFS [71].

All the retrieved evidence suggests that inflammation in CNS is responsible for the pathophysiological changes in CFS, with regional symptoms being coherent with involved brain areas.

In normal states, lymphocytes and their products are rarely found in the parenchyma of CNS because of the BBB action. However, in inflammatory states, BBB is damaged due to the effects of pro-inflammatory cytokines [72]. IL-1β exclusively attacks the paracellular barrier through the breakdown and translocation of tight junction proteins, and TNF-α targets the transcellular processes mediated by caveolae. A number of reactive oxygen and nitrogen species, accompanied by inflammatory reaction, also trigger BBB damage [73]. In the context of a damaged BBB, circulating cytokines can enter the CNS and activate glial cells (astrocytes, microglia, and oligodendrocytes) in the brain, promoting the secretion of cytokines, such as TNF-α, IL-I, and IL-6, in CNS (Fig. 2) [74]. One psycho-neuro-immunological hypothesis is that once the vagus nerve is infected, it becomes very sensitive to proinflammatory cytokines; it then communicates this cytokine exposure to the brain, thus, initiating the inflammatory components of CFS [75, 76].

It is commonly known that CSF can reflect biochemical changes in the extracellular brain space. Because of the limitations of peripheral cytokines. It is reasonable to consider whether exploring cytokine changes in CSF can be used as a potential diagnostic biomarker, as well as an important tool to understand the underlying mechanisms of inflammation in CNS related to CFS. Many studies reported changes in pro-inflammatory biomarkers in CSF of patients with CFS. Early in 1991, Lloyd et al. have compared cytokine levels between patients with CFS and normal controls, revealing that IFN-α levels in CSF were enhanced in CFS patients [77]. Natelson et al. investigated 11 CSF cytokines in patients with CFS. They found that the levels of granulocyte-macrophage colony-stimulating factor reduced in the patients, while the levels of IL-10 increased. These results support the hypothesis that some symptoms of CFS occur due to immune system dysfunction within CNS [78]. However, subsequent studies yielded conflicting results. In a pilot study examining the role of 27 CSF cytokines in patients with CFS, the authors found that only IL-10 significantly decreased in the patients, and no significant differences were observed in other cytokines when compared with healthy controls. This observation was attributed to a weakened anti-inflammatory role of IL-10 in CNS [79]. Hornig et al. compared the levels of CFS cytokines among patients with CFS, patients with multiple sclerosis, and healthy controls. They found a markedly disturbed immune signature in patients with CFS, including disrupted IL-1 signaling and increased eotaxin and CXCL10 levels [80]. Additionally, not all CFS cases have a typical presentation; CSF cytokine levels exhibit different associations between typical and atypical CFS [81]. Results between different studies were heterogeneous due to the different experimental conditions and designs. Since the status of BBB, circulating cytokines, cytokines in CNS, as well as their interactions, are quite complex and different in different case, it is easy to understand the difficulty to use such cytokines in CSF as an independent index for CFS. Indeed, although the current body of evidence cannot support the use of cytokines in CSF as an independent biomarker for CFS. We believe that analysis of the cytokines in CSF, and combined consideration of the changes of circulating cytokines, must be helpful to understand the pathophysiological states of one CFS patient (such as inflammatory states of CNS, the state of BBB, etc.). Although the invasive nature of the procedure required to analyze the cytokine profile in CSF (lumbar puncture) may limit its clinical applications, comprehensive consideration the changes of CNS and circulating cytokines might be a selective strategy for diagnosis/assessment of CFS, which should be verified in the future.