Because CFS often has an acute flu-like presentation, can follow severe viral infection (Hotopf, Noah, and Wessely 1996; White, Thomas, Amess, Grover, Kangro, and Clare 1995) and often presents with fever, sore throat and tender lymph nodes, a major hypothesis as to its cause is that it represents a form of chronic smoldering infection, perhaps by one of the family of herpes viruses. Unfortunately, our own work has not found support for the hypothesis that active viral infection (Natelson 2001; Wallace, Natelson, Gause, and Hay 1999) has an important role in CFS.
The early idea that CFS represented a form of chronic Epstein-Barr infection was dropped when data were reported indicating that elevated EBV titers B reflecting prior infection are not uncommon in healthy people (Gold et al. 1990). Adherents for chronic infection by other agents including HHV-6, mycoplasma, and Astealth viruses continue to advocate their beliefs, but convincing data remain to be seen. Infection can certainly trigger the onset of CFS. Rates of ~9% of developing CFS have been reported following infectious mononucleosis (White et al. 1998), Lyme disease and severe viral infection (Hotopf et al. 1996). Thus, postviral fatigue exists, but persistence of an infectious agent has not been demonstrated.
If persistent infection is not the cause, another hypothesis is that it is infection triggers abnormal processes in the immune system. A number of papers have reported immune activation in CFS (for review, see Strober (1994)) but it is not clear whether these change are the cause or the result of changes brought about by the CFS such as inactivity, disturbed sleep and chronic stress. Some data do support some underlying immunological problem: first, some CFS patients appear to have an antibody against a specific nuclear antigen (Von Mikecz, Konstantinov, Buchwald, Gerace, and Tan 1997); second, patients have a dysregulated 2,5 RNase L antiviral defense pathway (Suhadolnik et al. 1999); and third, treatment with an immune active agent, mismatched RNA, may reduce disability (Strayer et al. 1994) (a study to replicate this outcome has recently been completed). The immune dysregulation hypothesis was further supported by reports from two prominent groups that found evidence for immunological activation in CFS (Landay, Jessop, Lennette, and Levy 1991; Straus et al. 1993), as well as studies that suggest cytokines such as IL-6 and TNF-a could produce many of the symptoms of CFS. However, as we will point out below, these studies used broad groups of study patients with chronic fatigue and compared their data to control subjects who may have been very active. Importantly, activity is known to up-regulate the immune system and thus differences in activity might explain these results.
Reported studies of sleep in CFS patient have not considered the well-established effects of cytokines on sleep (Krueger, Fang, Taishi, Chen, Kushikata, and Gardi 1998; Kubota, Kushikata, Fang, and Krueger 2000; Vitkovic, Bockaert, and Jacque 2000). Some cytokines prolong non-REM (NREM) sleep, while others seem to shorten or interrupt it. Imbalances in the effects of these two sorts of cytokines may be involved in sleep disturbances in CFS. Of the cytokines in the sleep-producing group, the most extensively studied and best established are IL-1P and TNF-a, which are proinflammatory cytokines that produce fever in addition to sleep (Mullington, Korth, Hinze-Selch, Schreiber, Galanos, and Pollmacher 1998). The febrile effects of these cytokines can be blocked without altering their soporific actions. Besides both having circadian rhythms with similar nocturnal peaks, TNF-a and IL-1P seem to act cooperatively in animals to prolong NREM sleep. IL-1 consists of a family of peptides the biological activity of which depends on the net balance of the activity of agonists such as IL-1P, and several natural counter regulatory factors particularly IL-1 receptor antagonist. Blocking TNF-a and IL-1 by a variety of means alters sleep regulation. The mechanism is unknown, but increases in either of these cytokines lead to enhanced production of the other via induction of transcription factor NF kappa B (Krueger et al. 1998). Experimental data from animals indicate that these cytokines increase in the brain during sleep deprivation (Krueger, Obal, Fang, Kubota, and Taishi 2000). Other cytokines that are reported to increase NREM sleep include IL-2, IL-6, IL-8, and IL-18 (Kapsimalis, Richardson, Opp, and Kryger 2005).
Cytokines that disrupt slow wave sleep (stages 3+4) have been less well examined, but the best-established are IL-4 and IL-10. These cytokines function by inhibiting the production of IL-1 and TNF-a, probably via inhibition of NF kappa B activation (Krueger et al. 2000). Intracerebral injections in rabbits of a cell permeable inhibitor peptide of the transcription factor, NF kappa B inhibit both spontaneous and IL-1P induced sleep (Kubota et al. 2000). Also intracerebral injections of TNF-a and IL-1 inhibitors in rabbits significantly reduced spontaneous NREM sleep, whereas pretreatment with inhibitors of these cytokines significantly attenuated sleep rebound after sleep deprivation. IL-4 has been shown in animals to reduce slow wave sleep (Kushikata, Fang, Wang, and Krueger 1998). IL-10 is spontaneously produced by both lymphocytes and monocytes, inhibits the production of TNF-a, and inhibits slow wave sleep in rabbits (Kushikata, Fang, and Krueger 1999). In contrast, IL-10 knockout mice have increased slow wave sleep (Toth and Opp 2001). Other cytokines reported to have sleep inhibiting effects are IL-13 and TGF-p.
The evidence suggestive of abnormalities in immune regulation in CFS stimulated our group to do an intensive study of immunological cell populations and cytokine gene message in a carefully delineated group of CFS patients matched with sedentary healthy controls. We believe that one of the reasons for inconsistent data on the immune system in CFS lies in the fact that extremely sedentary CFS patients are compared to active healthy controls, and level of fitness can affect immunological activity. Looking at the data using a number of classical statistical approaches, we could find no consistent differences between the groups (Natelson et al. 1998; Zhang et al. 1999). Moreover, we recently completed a comprehensive review of all controlled studies of immunity in CFS and were unable to find uniformly consistent results (Natelson, Haghighi, and Ponzio 2002). However, using a new neural nets methodology designed specifically to identify small differences between groups, we did find preliminary evidence that IL-4 has a role in CFS (Hanson, Gause, and Natelson 2001). The higher IL-4 levels in CFS suggest a Type 2 shift in cytokines. It may be that these subtle differences in IL-4 would be magnified if blood were samples collected in sleeping CFS patients.
We have, however, found clear cytokine abnormalities in veterans of the first Gulf War with CFS: sick veterans had higher levels of gene expression for IL-2, IL-10, INF-y, and TNF-a than healthy veteran controls, and factor analysis showed that
Th-2-related cytokines mediated the cognitive dysfunction reported by Gulf veterans with CFS (Brimacombe, Zhang, Lange, and Natelson 2002b). We do not interpret the results of these positive studies to mean that CFS in Gulf veterans is necessarily different from that in nonveterans with CFS. Instead, we believe these differences emerged because of our finding that Gulf veterans are more immunologically homogeneous than nonveterans (Zhang et al. 1999). Such homogeneity would allow subtle but important differences to be detected.
Robust circadian rhythms are known to exist for the cytokines. In fact, IL-1, and TNF-a, cytokines which are known to enhance sleep, peak during nighttime sleep and are low during the day (Krueger et al. 1998; Moldofsky, Lue, Eisen, Keystone, and Gorczynski 1986). Thus, sampling these cytokines in daytime could, in and of itself, be one reason why more striking cytokine abnormalities have not been found in CFS.
It may be that sleep-disrupting cytokines are relatively up-regulated and sleep-producing cytokines relatively down-regulated in some patients with CFS, and that these changes in cytokines lead to sleep disturbance throughout the night. Hence there might not be abnormalities in the absolute levels of either sleep promoting or sleep inhibiting cytokines (both of which might be in the normal range) but rather in their ratio or difference. We believe this postulated cytokine imbalance is exacerbated by exertion—leading to further disruption of sleep and increased CFS symptoms.
Exercise and level of fitness are certainly known to influence cytokines. In our study of cytokine gene expression 10 min after an acute stress test in seven sedentary men, we found levels of IL-1P, IL-4, IL-6, and IL-10 did not change while levels of TNF-a decreased (Natelson et al. 1996). Our data for IL-1P, IL-4, IL-6, and IL-10 are consistent with the literature (except for IL-1P which increases in plasma in trained athletes doing high intensity exercise) (Moldoveanu, Shephard, and Shek 2001). For TNF-a, most studies were of highly trained athletes doing long duration exercise, and these reported increases in plasma levels with no change in gene expression (Moldoveanu, Shephard, and Shek 2000).
There are a few studies on CFS patients after exercise. We followed cytokine gene expression before and after a standard maximal exertion test in CFS and sedentary matched healthy controls (LaManca et al. 1999). We found no effect of exercise on IL-4 or IL-10 for either group. Exercise produced a decrease in TNF-a in both groups, but the CFS group did not differ from the controls (LaManca et al. 1999). Three studies besides our own also evaluated the effect of exercise on cytokines, including IL-1P, INF-a, TNF-a, and TGF-P (Cannon et al. 1997; Cannon, Angel, Ball, Abad, Fagioli, and Komaroff 1999; Lloyd, Gandevia, Brockman, Hales, and Wakefield 1994; Peterson et al. 1994); again none showed that CFS patients responded differently than controls. However, there are two major limitations to these studies. They only assessed cytokines shortly after exercise and never during the night, and none examined plasma cytokine levels, gene message and cytokine secretion, concurrently.
Also the complexity of the immune network and its positive and negative interactions with the HPA and with peptides involved in energy homeostasis have thus far not been taken into account in evaluating neuroimmune modulation of sleep in CFS. For example, growth hormone is sleep promoting while corticotrophin releasing hormone enhances wakefulness. Levels of both are affected by the release of cytokines such as IL-6, which in turn helps regulate their level. For example, IL-6 stimulates the HP A axis and increases levels of cortisol, but cortisol inhibits IL-6 release (Path et al. 2000). Ghrelin, an endogenous ligand of growth hormone, produced in the stomach, which stimulates both appetite and promotes sleep and in excess can give rise to obesity, has not been studied in CFS. Obesity in turn is now recognized as an important cause of fatigue and to be characteristic of some types of CFS particularly those associated with disturbed breathing during sleep (Vollmer-Conna, Aslakson, and White 2006). The role of these multiple factors on sleep is unlikely to be equal, suggesting that solving the calculus of relating sleep to neuropeptides and cytokines is likely to be quite challenging.
Another issue is methodological and has to do with the assessment of cytokines themselves. There simply is no "gold standard" assay that assures that one is getting a good representation of the biological activity of any one cytokine. The problems here are many. First cytokines are often bound to carrier proteins in the circulation that can interfere with measurement and may influence biological activity; these could consist of multifunctional liver-derived binding proteins, cytokine-specific soluble receptors or cytokine-specific soluble receptor antagonists (Cannon 2000). These issues suggest there is a problem in looking only at plasma cytokines. Another problem is that plasma cytokines usually circulate in picomolar concentrations. Thus, what one finds in the plasma may not be representative of what is happening in the immunological microenvironment. One can get a sense of this microenvironment by turning to cytokine gene expression. Some cytokines are produced mainly by cells of a single lineage or phenotype, whereas others can be produced by many. The selection of a specific cell type for study can affect yield and introduce bias. The cells directly involved in pathogenesis may be sequestered in tissues, and may be of unknown phenotype. In this situation, biologically relevant data can often be obtained following stimulation of peripheral cells in vitro with an appropriate stimulus. Such a functional assessment permits accurate measurement of cytokine production, consumption and neutralization. To deal with the lack of a "gold standard" to quantify cytokines, it may be necessary to use multiple approaches including quantifying serum levels of cytokines, measurements of RNA message, and an assessment of the frequencies of cytokine-producing cells.
Was this article helpful?