A neuro-immune model of Myalgic Encephalomyelitis/Chronic fatigue syndrome
- First Online:
- Cite this article as:
- Morris, G. & Maes, M. Metab Brain Dis (2013) 28: 523. doi:10.1007/s11011-012-9324-8
- 1.1k Views
This paper proposes a neuro-immune model for Myalgic Encephalomyelitis/Chronic fatigue syndrome (ME/CFS). A wide range of immunological and neurological abnormalities have been reported in people suffering from ME/CFS. They include abnormalities in proinflammatory cytokines, raised production of nuclear factor-κB, mitochondrial dysfunctions, autoimmune responses, autonomic disturbances and brain pathology. Raised levels of oxidative and nitrosative stress (O&NS), together with reduced levels of antioxidants are indicative of an immuno-inflammatory pathology. A number of different pathogens have been reported either as triggering or maintaining factors. Our model proposes that initial infection and immune activation caused by a number of possible pathogens leads to a state of chronic peripheral immune activation driven by activated O&NS pathways that lead to progressive damage of self epitopes even when the initial infection has been cleared. Subsequent activation of autoreactive T cells conspiring with O&NS pathways cause further damage and provoke chronic activation of immuno-inflammatory pathways. The subsequent upregulation of proinflammatory compounds may activate microglia via the vagus nerve. Elevated proinflammatory cytokines together with raised O&NS conspire to produce mitochondrial damage. The subsequent ATP deficit together with inflammation and O&NS are responsible for the landmark symptoms of ME/CFS, including post-exertional malaise. Raised levels of O&NS subsequently cause progressive elevation of autoimmune activity facilitated by molecular mimicry, bystander activation or epitope spreading. These processes provoke central nervous system (CNS) activation in an attempt to restore immune homeostatsis. This model proposes that the antagonistic activities of the CNS response to peripheral inflammation, O&NS and chronic immune activation are responsible for the remitting-relapsing nature of ME/CFS. Leads for future research are suggested based on this neuro-immune model.
KeywordsChronic fatigue syndromeInflammationCytokinesDepressionTryptophanOxidative and nitrosative stressMitochondriaAutoimmune
World Health Organization
Chronic fatigue syndrome
Tumor necrosis factor
Oxidative and nitrosative stress
Nuclear factor κB
2-5A-dependent ribonuclease L
Protein kinase R
Transforming growth factor
Blood brain barrier
Single-photon emission computed tomography
Positron emission tomography
- T MRI
Tesla magnetic-resonance imaging
Nicotinamide adenine dinucleotide phosphate
Signal Transducer and Activator of Transcription 3
Forkhead box P3
Peripheral blood mononuclear cells
Single nucleotide polymorphisms
Myalgic Encephalomyelitis (ME) has been described in the medical literature since the ‘30s. Characteristics of ME are post-exertional malaise, chronic fatigue, autonomic symptoms, hyperalgesia, malaise, neurocognitive dysfunctions, and a chronic and relapsing-remitting course. The World Health Organization (WHO) acknowledges ME as a disease of the nervous system (World Health Organization 1992). Much later, in the ‘80s and ‘90s, the diagnosis of chronic fatigue syndrome (CFS) was introduced. The definition of CFS is primarily based upon the presence of chronic fatigue lasting more than 6 months. According to a commonly used diagnostic classification of CFS (Fukuda et al. 1994), chronic fatigue must be accompanied by four or more of the following symptoms: substantial impairment in short-term memory or concentration; sore throat; tender lymph nodes; muscle pain; multi-joint pain without swelling or redness; headache of a new type, pattern or severity; unrefreshing sleep; and post-exertional malaise lasting more than 24 h. ME and CFS are therefore different diagnostic concepts (Maes et al. 2012a; Twisk and Maes 2009).
In an attempt to define a more homogenous patient population, a panel of ME/CFS experts has published consensus criteria for ME (Carruthers et al. 2011). Using multivariate statistical analyses, Maes et al. (2012a) showed that the diagnosis of CFS according to Fukuda’s criteria (Fukuda et al. 1994) defines a heterogeneous population of individuals with chronic fatigue, of whom individuals with ME are a subset (Maes et al. 2012a). Moreover, the data show that post-exertional malaise is a key component of ME and differentiates ME from CFS (Maes et al. 2012a). Thus, the diagnosis of ME, CFS and ME/CFS refer to different concepts, ranging from chronic fatigue together with physio-somatic symptoms (Fukuda et al. 1994), to a symptom complex where post-exertional malaise is a key factor (Maes et al. 2012a). Patients with ME display the unique characteristic of producing a range of objectively measurable abnormal physiological responses to exercise and a global worsening of symptoms following even trivial increases in cognitive or physical activity (De Becker et al. 2000; Nijs et al. 2005; Lane et al. 1998). Obviously these differences in diagnostic classifications and criteria have blurred the results of scientific research. In this paper we will use the term ME/CFS, since most research until now has been performed on ME/CFS and not on ME as defined recently (Maes et al. 2012a; Carruthers et al. 2011).
Recently an explanatory model has been developed in support of the WHO’s classification that considers ME to be a neuro-immune disorder (Maes 2009; Maes and Twisk 2010; Maes et al. 2012b). This model provides considerable evidence that the pathophysiology of ME/CFS is associated with immuno-inflammatory pathways, which lead to neurological aberrations, including behavioral responses, and neuro-endocrine, autonomic and brain dysfunctions.
The presence of raised levels of the pro-inflammatory cytokines (PICs), tumor necrosis factor (TNF)α, interleukin-6 (IL-6) and IL-1β, is a consistent finding in patients with ME/CFS who show signs of a chronically activated and highly disregulated immune system. These include the presence of autoimmune responses, loss of natural killer (NK) function and raised levels of the anti-inflammatory cytokines. Studies investigating the pathophysiology of ME/CFS have consistently revealed a wide range of aberrations in oxidative and nitrosative stress (O&NS) pathways and lowered antioxidant defenses. Perhaps unsurprisingly then evidence of mitochondrial damage has been reported in patients with ME/CFS. Studies have reported activation of intracellular signaling molecules, e.g. nuclear factor κB (NF-κB) and the 2′-5′-oligoadenylate (2-5A)/2-5A-dependent ribonuclease L (RNaseL) and the protein kinase R (PKR) pathways. A new immuno-inflammatory pathway in ME/CFS that may cause neuro-immune dysfunctions and even neurological disease is bacterial translocation.
Here we propose a key role for changes in immuno-inflammatory pathways and resultant neuro-inflammation, in conjunction with aberrations in O&NS pathways, in producing the wide range of symptoms reported and abnormalities identified in people with ME/CFS. This paper proposes that an interaction between an activated immune system, O&NS, and mitochondrial damage, explains the symptoms of ME/CFS. This is an area of active research and much needs to be confirmed by new results. As such there are gaps, which we have filled with established knowledge on immune and neuro-immune mechanisms and directions for future research.
Pathophysiological findings in ME/CFS
Immuno-inflammatory pathways in ME/CFS
Activation of immuno-inflammatory pathways is a consistent finding in ME/CFS (Moss et al. 1999; Cannon et al. 1997). Increased levels of PICs, such as IL-1β, TNFα and IL-6 have been reported in people with ME/CFS (Maes et al. 2012a; Broderick et al. 2010). The levels of IL-1 and TNFα show a positive significant correlation with fatigue, autonomic symptoms and flu like symptoms (Maes et al. 2012b). Other markers of inflammation, e.g. lysozyme and polymorph nuclear neutrophil elastase, are also elevated (Maes et al. 2012b). The levels of the latter correlate significantly and positively with levels of memory and cognitive impairment and phenomenological experience of infection. Increased levels of acute phase reactants, e.g. C-reactive protein and the α2 protein fraction obtained during protein electrophoresis, further underscore the presence of an inflammatory response (Buchwald et al. 1997; Maes et al. 2005a; Spence et al. 2008). Also the production of cyclo-oxygenase 2 (COX-2) is significantly elevated in patients suffering from ME/CFS (Maes et al. 2007b). The elevated levels of COX-2 are additionally related to key symptoms of ME/CFS, such as a subjective experience of infection, hyperalgesia, fatigue, and neurocognitive disorders (Maes et al. 2007b).
Neopterin, an indicant of cell-mediated immune activation is significantly increased in ME/CFS patients (Maes et al. 2012b). The same study showed that severity of fatigue, autonomic and flu like symptoms also correlate positively with serum neopterin concentrations. Some but not all studies found increased production of interferon (IFN)γ in ME/CFS (review: Maes and Twisk 2010). Increased levels of IL-12 were detected by Fletcher et al. (2009), suggesting a Thelper (Th)1-like response in some patients. No significant differences in the Th17 cytokines, IL-17 and IL-23, were observed in patients with ME/CFS as compared with controls (Fletcher et al. 2009). Nevertheless, associations between fatigue scores and IL-17 levels were established in patients with primary Sjögren's syndrome (Haldorsen et al. 2011). This finding together with increased levels of PICs, including IL-1, IL-6 and TNFα, further suggest that a Th17 response may occur in some patients with chronic fatigue. In addition, the lowered expression of CD69 activation markers on CD4+ T cells in ME/CFS (Mihaylova et al. 2007) may further skew towards Th17 responses. Indeed, CD69 promotes some pathways that inhibit Th17 cell differentiation (Martín and Sánchez-Madrid 2011).
A state of immune activation is further evidenced by elevated expression of surface markers CD38+, HLA-DR on CD8+T cells (Landay et al. 1991), and higher percentages of CD8+T cells and higher numbers of CD8+T cell class 11 activation markers (Klimas et al. 1990). We have reviewed that a Th2 response with increased levels of IL-4 or IL-5 may occur in part of the patients (Maes et al. 2012b; Fletcher et al. 2009). Other studies show signs of immunosuppression, including lowered NK cell activity and raised levels of immunoregulatory cytokines, IL-10 and transforming growth factor (TGF)-β1 (Chao et al. 1991; Ogawa et al. 1998; Fletcher et al. 2010; Bennett et al. 1997; Nakamura et al. 2010). All in all, while a chronic, low-grade inflammatory response accompanies ME/CFS, some patient groups feature elevations of Th1, Th2, Treg and maybe Th17 cytokines. Some ME/CFS patients show a mixed response with activation of Th1 and Th2 cytokines. Th1 and Th2 subsets are indeed not mutually exclusive, as Th17 and Treg subsets, but together facilitate recovery from infection (Bot et al. 2004; Afzali et al. 2007). For example, IL-4 and Th2 immune responses monitor Th1 responses and maintain homeostasis (Bot et al. 2004). It may be that the different cytokine profiles in ME/CFS depend on subgroups or staging of the disease, e.g. acute flares versus partial remission. In ‘Mechanistic explanations for the chronic and relapsing–remitting nature of ME/CFS’ we will discuss how Th1 versus Th2 dominance and Treg functions may play a role in the relapsing–remitting nature of ME/CFS.
A number of factors predispose to chronic inflammation. IgG1 deficiencies often result in a decreased level of total IgG (hypogammaglobulinemia). Along with IgG1, the IgG3 subclass is most frequently present in the antibody response to protein antigens. Decreased IgG3 levels are frequently associated with IgG1 deficiency (Herrod 1993). IgG1 and IgG3 deficiency has been associated with a history of recurrent infectious. IgG1 and IgG3 levels are reduced in ME/CFS (Bassi et al. 2008; Hilgers and Frank 1994). Another feature which predisposes to chronic inflammation is the decreased ratio of ω3/ω6 polyunsaturated fatty acids (PUFAs) in patients with ME/CFS (Maes et al. 2005b). Omega-3 fatty acids have anti-inflammatory properties, whereas omega-6 fatty acids have pro-inflammatory effects (Simopoulos 2002). A genetic predisposition is repeatedly observed in patients with ME/CFS. For example, a significant increase in TNF–857 TT and CT genotypes, likely to be associated with increased production of TNFα and IFNγ; human leucocyte antigen class II allele HLA-DQA1*01; C/T polymorphism leading to His161Arg substitution in IL-17F; and IFN-γ +874T/A and IL-10 -592C/A polymorphisms (review: Maes and Twisk 2010).
O&NS pathways in ME/CFS
ME/CFS is accompanied by increased reactive oxygen and nitrogen species (ROS/RNS) and activated O&NS pathways (Jammes et al. 2005; Meeus et al. 2009; Maes et al. 2006b, 2009b). Frequently reported abnormalities include O&NS damage to proteins, lipids and DNA (Maes et al. 2009b; Maes and Twisk 2010). Patients with ME/CFS additionally show increased IgM-mediated responses to different NO-adducts, indicating chronically increased nitric oxide (NO) production (Maes et al. 2006b), and increased production of inducible NO synthase (iNOS) (Maes et al. 2007b). Increased IgM-mediated responses directed to palmitic and myristic acid and the NO-adducts show that O&NS have modified the structure of fatty acids and proteins (Maes et al. 2006b). Reduced levels of the antioxidants zinc, coenzyme Q10, dehydroepiandrosterone sulfate and glutathione have also been reported (Manuel-y-Keenoy et al. 2000, 2001; Kennedy et al. 2003, 2005; Maes et al. 2005a, 2006a, 2009a; Maes 2011a). Low grade inflammation, activated O&NS pathways and impaired oxidative defenses likely interact to increase the magnitude of abnormality in each leading to a vicious cycle (Maes and Twisk 2010). The symptoms of fatigue, malaise and pain, positively and significantly correlate with markers of lipid peroxidation and protein oxidation/nitrosylation (Kennedy et al. 2003, 2005; Maes et al. 2006b; Maes and Twisk 2010; Maes 2011a). Studies have demonstrated increased ventricular lactate in ME/CFS using 3 Tesla proton magnetic-resonance spectroscopy (Murrough et al. 2010; Mathew et al. 2009; Shungu et al. 2012). Ventricular CSF lactate was significantly increased in ME/CFS as compared to normal volunteers (Murrough et al. 2010). During exercise lactate is associated with the production of radical oxygen species and acidosis (Groussard et al. 2000). Lactate is an antioxidant that scavenges free radicals and attenuates lipid peroxidation, whereas acidosis is a potent oxidative condition (Groussard et al. 2000).
The presence of O&NS is known to lead to mitochondrial damage in neurological diseases (Lin and Beal 2006; Trushina and McMurray 2007) and is likely the source of mitochondrial dysfunction seen in people with ME/CFS (Maes and Twisk 2010). Specific and highly sensitive alterations in the availability of adenosine triphosphate (ATP), inadequate supplies of ATP, and oxidative phosphorylation have been reported in neutrophil mitochondria of individuals with ME/CFS (Myhill et al. 2009; Vermeulen et al. 2010). Another study reported that all individuals with physio-somatic symptoms in depression had a significantly lower ATP production in muscle biopsies compared to controls (Gardner and Boles 2008).
Intracellular signaling networks
A number of different signaling networks are upregulated in ME/CFS. Studies have demonstrated higher production of unstimulated, and TNFα- and PMA-stimulated NF-κB, cyclo-oxygenase-2 and inducible iNOS (Maes et al. 2007a, b). NF-κβ is a major upstream molecule, which regulates immuno-inflammatory and O&NS responses. Upon activation by viral products, lipopolysaccharide (LPS), a constituent of the outer membrane of gram negative bacteria, PICs and free radicals, NF-κβ is translocated from the cytoplasma to the nucleus where it binds DNA promoter sequences and induces transcriptional activation of PICs, such as IL-1, IL-6, and TNFα, COX-2 and inducible iNOS (Brasier 2006). Upon activation, the inhibitory IκB proteins, that hold NF-κB in an inactive state, are tagged from proteosomal degradation following phosphorylation by IκB kinases. In many inflammatory disorders, NF-κB in chronically upregulated. In ME/CFS, the increased production of NF-κB is associated with fatigue and a subjective feeling of infection (Maes et al. 2007a). Moreover, there were significant and positive intercorrelations between the production of NF-κB, COX-2, and inducible iNOS, suggesting that the increased production of COX-2 and iNOS is driven by the transcription factor NF-κB. Thus, increased production of NF-κB is response to viral and bacterial infections may be one key mechanism modulating the immuno-inflammatory and O&NS pathways in ME/CFS.
The 2-5A/RNaseL and PKR pathways are also upregulated in individuals with ME/CFS (Vojdani and Lapp 1999). Importantly, the latter pathways have antiproliferative and antiviral effects and are induced by interferons in response to viral infections. Moreover, the enzyme that synthesizes 2-5A, i.e. 2′-5′-oligoadenylate synthetase, is activated during bacterial infections. PKR is also activated by bacterial LPS and PICs, including IL-1 and TNFα.
The presence of intestinal permeability and resultant bacterial translocation into the lamina propina, the lymph nodes and possibly the systemic circulation, is evidenced by the presence of IgM and IgA mediated immune responses against LPS of different gram negative bacteria in up to 40–60 % of people with ME/CFS (Maes et al. 2012c). This may be due to the presence of systemic inflammation, which can reduce the integrity of epithelial tight junctions. The IgA responses positively and significantly correlate with the levels of PICs, such as IL-1β and TNFα, and neopterin (Maes et al. 2012c). LPS from translocated commensal bacteria may provoke a peripheral immuno-inflammatory response by binding to the Toll-like receptor-4 (TLR4) complex, which in turn activates intracellular signaling pathways, such as NF-κB, which subsequently will induce genes of PICs, iNOS and COX-2 (Maes et al. 2012c). Increased LPS may also explain other immuno-inflammatory signs in ME/CFS, such as increased neopterin and elastase levels. Indeed, LPS may stimulate the production of neopterin by monocytes/macrophages and elastase by neutrophils and monocytes (Maes et al. 2012c).
Increased levels of B cells such as CD19+, CD 20+ and CD21+ have been reported (Tirelli et al. 1994). IgM mediated autoimmune responses towards self antigens whose structure has been damaged by O&NS also occurs in ME/CFS (Maes et al. 2006b). These involve autoantibodies to anchorage molecules, palmitic and myristinic acid, which are essential for normal cellular function (Maes et al. 2006b). Other autoantibodies towards Sa antigen, cell nucleus, and cardiolopin and serotonin, gangliosides, vasoactive neuropeptides, and cholinergic receptors have all been reported in large groups of ME/CFS patients (review: Maes. 2011b; Nishikai et al. 2001; Ortega-Hernandez and Shoenfeld 2009). In the next sections we will further explain the mechanisms that may cause autoimmune responses in ME/CFS.
Role of pathogens in ME/CFS
Viral and bacterial infections probably play an important role in ME/CFS, either as trigger or maintaining factors (Maes and Twisk 2010; Maes 2011a; Komaroff and Cho 2011; Broderick et al. 2010). The overall picture of activated immuno-inflammatory pathways in ME/CFS, e.g. increased NF-κB, 2-5A/RNaseL and PKR pathways, PICs and bacterial translocation, support the hypothesis that viral and bacterial infections play a major role in the onset of ME/CFS (Maes and Twisk 2010). Not only the immunological anomalies but also the distinctive gene expression are in accordance with viral or bacterial infections (Maes and Twisk 2010; Broderick et al. 2010). ME/CFS often begins with a severe flu-like infection (Naess et al. 2010) and epidemics of ME/CFS have repeatedly been described in the literature (Strickland et al. 2001). A number of pathogens have been associated with ME/CFS such as Borrelia burgdorferi, P. aeruginosa, Mycoplasma, Chlamydia, Cytomegalovirus, HHV-6, and EBV, Coxsackie and Aspergillus (review: Maes and Twisk 2010).
A wide range of pathogens commonly found infecting people with ME/CFS are capable of establishing prolonged infections as a result of developing sophisticated adaptations to avoid or mitigate the effects of the host immune response (Hilgers and Frank 1994; Zhang et al. 2010; Stephensen et al. 2005). Borrelia burgdorferi, for example, persists in the host and has strategies for immune evasion that include immunogenic surface-exposed proteins as potential virulence determinants (Radolf 1994). Borrelia burgdorferi undergoes antigenic variation, so that the configuration of the exposed proteins also changes in the host over time (Bankhead and Chaconas 2007). P. aeruginosa evades human complement attack by binding human regulators to its surface (Kunert et al. 2007). Mycoplasma can evade the host response by generating variable surface antigen proteins, which continually change their antigenicity (Simmons and Dybvig 2007; Le Grand et al. 1996). Chlamydiae hijack signaling pathways that prevent the host cells from undergoing apoptosis and protect the infected cells from attack by host defenses (Zhong 2009). Each human herpes virus studied has evolved cell biological solutions to avoid or mitigate problems posed by the human immune response (Gewurz et al. 2007). Herpes virus evasion of T-cell immunity includes latency, restriction of viral gene expression to immuno-privileged sites (CNS), interference with complement, cytokines, apoptosis and NK cell activity, and expression of immuno-evasins, i.e. molecules that disrupt normal immune physiology. Coxsackie B3 employs a virally-encoded protein, i.e. 3Cpro, to degrade molecules, which function downstream of pattern recognition receptor signaling (Mukherjee et al. 2011; Wessels et al. 2006). Aspergillus species employ the following strategies: binding to complement regulators, which avoids the host complement attack (Behnsen et al. 2008; Radolf 1994); shielding of stimulatory surface recognition molecules; induction of anti-inflammatory signals; complement evasion; and shedding of decoy components (Chai et al. 2009).
Transition to a chronic inflammatory state
Not only persistent infections can create a chronic inflammatory state, but also acute microbial infections can lead to chronic systemic inflammation, whether the infection becomes chronic or is cleared by the immune system (Pincus 2005). Processes that may explain chronic systemic inflammation are, for example, damage by O&NS and autoimmune responses generated by infections or mounted against self-antigens that have been modified by O&NS (Maes et al. 2006b). Thus, activation of O&NS pathways may cause O&NS damage depending on the state of the antioxidant defenses, and lead to elevations of NF-κB levels. O&NS damage to lipids and proteins can change their structure so that they are regarded as 'non self', leading to the generation of auto-immune responses and autoreactive T cells (Maes et al. 2006a, b).
Different infections can initiate or exacerbate a chronic inflammatory condition and cause autoimmune responses, using processes described as molecular mimicry, bystander damage and epitope spreading (Fujinami et al. 2006). Autoreactive T cells can be formed when the pathogen contains epitopes, which are cross reactive to host molecules, leading to autoimmune targeting of both pathogen and self epitopes. Epitope spreading and bystander damage are other sources of chronic immune activation, inflammation and autoimmune responses. Bystander cell death, which is caused by pathogen specific T cells, may increase the release of self-epitopes that induce de novo T cell responses. Pathogen-specific T cells may initiate bystander activation by migrating to areas of infection, where they encounter pathogen-infected cells that present pathogen peptides in the context of MHC human leukocyte antigen class I molecules to T cells. The CD8+ T cells recognize these infected cells and will release cytotoxic granules which may result in the killing of the infected cell. Consequently, the dying cells, the CD8+ T cells and macrophages within the inflammatory focus, release PICs and ROS/RNS, which may cause bystander damage of uninfected but neighbouring cells (Fujinami et al. 2006). This may result in additional immunopathology (Duke 1989; Smyth and Sedgwick 1998), because self proteins may be released and engulfed by dendritic cells, macrophages and antigen presenting cells (Fujinami et al. 2006). The latter cells are subsequently lysed by cytotoxic CD4+T cells (Hahn et al. 1995).
It is also possible that the initial infection may create autoreactive T cells, but insufficient inflammation to induce overt pathology (Hahn et al. 1995). Nevertheless, autoreactive T cells can be activated by subsequent infections with a broad range of pathogens leading to overt inflammatory damage (Hahn et al. 1995). This fertile field, induced by the interaction with the primary infecting agent, can change with time (Fujinami et al. 2006). The initial priming infection and subsequent challenges that produce symptoms may be separated by many years and sometimes more than one challenge is needed to induce symptoms. A fertile field can also be generated when an infection with a pathogen, which shows molecular mimicry to self CNS molecules, primes autoreactive T cells without initiating autoimmune inflammatory disease. Later challenges may then trigger these autoimmune cells to cause overt symptoms (Fujinami et al. 2006).
From peripheral inflammation to neuro-inflammation
It is well known that peripheral inflammation may induce a behavioral complex characterized by fever, anorexia, weight loss, and behavior inhibition, including reduced locomotor activity and listlessness (Goehler et al. 2005). Circulating PICs and other inflammatory and O&NS molecules may affect brain function through several routes, mainly humoral and neural pathways (Ferrari and Tarelli 2011). The blood brain barrier (BBB) controls the passage of inflammatory compounds from the plasma to the brain. There are several ways by which these substances can cross the BBB (Perry et al. 2010). Firstly, these substances can enter the brain through the areas that lack a BBB, e.g. the circumventricular organs (Whitton 2007). Secondly, cytokines may cross the BBB using specific transporters (Kim and de Vellis 2005). Thirdly, BBB permeability may be increased as a consequence of immuno-inflammatory stimuli. For example, Th17 T cells can infiltrate the brain through the BBB via loosening of the endothelial tight junctions by increased secretion of IL-17 and IL-22 (Jadidi-Niaragh and Mirshafiey 2011). Fourthly, peripheral stimuli may activate endothelial cells which in turn may induce immune signals in the CNS (Ransohoff and Perry 2009). For example, Parkinson’s Disease is accompanied by breakdown of the BBB which allows noxious blood-borne substances to penetrate in the CNS (Kortekaas et al. 2005). Likewise, disruption of the BBB allows the extravasation of PICs and immune cells, which can activate microglia and, therefore, induce neuroprogression (Leonard and Maes 2012). Neuroprogression is a progressive process, which is induced by inflammatory and O&NS processes, and that is characterized by neurodegeneration, reduced neurogenesis and neuronal plasticity, and/or neuronal apoptosis (Berk et al. 2011).
The second route of immune-to-brain communication is the neural pathway, which transmits peripheral inflammatory signals to the CNS via the autonomic nervous system. The most important afferent pathway that signals peripheral inflammation to the CNS is the vagus nerve. These neural pathways are stimulated by inflammatory signals that increase the levels of brain cytokines (Perry et al. 2003, 2010; Tracey 2002). The role of the vagus nerve is crucial in inducing neuro-inflammation and activation of microglia (Goehler et al. 1997, 2005; Herkenham et al. 1998; Ericsson et al. 1997; Benarroch 2011; Blatteis and Sehic 1997; Ek et al. 1998; Saper 1995; Gallaher et al. 2012; Riazi et al. 2008; Engler et al. 2011). Not only PICs, but also prostaglandins and complement factors may activate the vagal nerve pathways, via chemoreceptors situated on glomus cells in the vagal paraganglia (Goehler et al. 1997). The latter serve as a monitoring system that detects immuno-inflammatory stimuli in the lymph and spleen. Vagal afferents terminate in the dorsal vagal complex of the caudal medulla, the area postrema, the nucleus of the solitary tract and the dorsal motor nucleus of the vagus (Saper 1995). Ascending immuno-inflammatory signals activate microglia in the nucleus of the solitary tract and the doral motor nucleus of the vagus (Gallaher et al. 2012) and the hippocampus (Riazi et al. 2008). Subsequently, these peripheral immuno-inflammatory signals activate the production of PICs in the amygadala, which acts to coordinate behavioral, CNS and autonomic responses (Engler et al. 2011). These bottom-up pathways may explain that systemic inflammation is accompanied by neuro-inflammation and behavioral changes. For example, systemic LPS administration may elicit neuro-inflammation with increased brain TNFα, which remains elevated for months and is associated with behavior symptoms, including fatigue, malaise, and hyperalagesia (Qin et al. 2007). Moreover, once the immuno-inflammatory signals have arrived in the CNS via the vagal afferents, a counter-regulatory anti-inflammatory reflex response is mounted that tends to downregulate the peripheral immuno-inflammatory response (Perry et al. 2007; Tracey 2002). These counter-regulatory responses will be discussed in ‘Mechanistic explanations for the chronic and relapsing–remitting nature of ME/CFS’.
Neurological dysfunctions in ME/CFS
Single-photon emission computed tomography (SPECT) scans may reveal significantly lower cortical/cerebellar regional cerebral blood flow frequently in the frontal, parietal, temporal, occipital, brain stem and throughout the cerebral cortex (Schwartz et al. 1994; Biswal et al. 2011; Patrick et al. 2008). Positron emission tomography (PET) scans may reveal decreased density of 5-HT transporter in the brain (Yamamoto et al. 2004). Proton magnetic resonance spectroscopy has revealed various abnormalities in metabolic pathways in the brain (Chaudhuri et al. 2003; Chaudhuri and Behan 2004). Barnden et al. (2011) detected that white matter volume decreased with fatigue duration and they found signs of impaired cerebrovascular autoregulation. The findings suggest that fatigue onset is accompanied by an insult that suppresses cognitive activity and modulates the autonomic system (Barnden et al. 2011). Elevated numbers of punctuate lesions have been detected in ME/CFS without psychiatric comorbidity in the frontal lobes as compared to controls (Lange et al. 1999). 3 Tesla magnetic-resonance imaging (3 T MRI) scans have revealed significant neuroanatomical changes, including reduced white matter volume (Puri et al. 2011). A 1.5 T MRI, on the other hand, showed no abnormalities suggesting brain atrophy or white matter lesions in patients with ME/CFS (Perrin et al. 2010). Greco et al. (1997) found no specific white matter abnormalities in ME/CFS. On SPECT imaging, ME/CFS shares some similarities with AIDS Dementia Complex: acute changes in radionuclide uptake in the younger population may be caused by inflammatory processes at the cellular or micro-vascular level (Schwartz et al. 1994). The findings in ME/CFS are consistent with the hypothesis that this disease results from a viral infection of neurons, glia or vasculature (Schwartz et al. 1994). Thus, viral infection can provoke neurological dysfunction by interfering with intra-cellular mechanisms or membrane transport systems or cerebral hypoperfusion due to vasculitis. All in all, these brain studies could indicate metabolic dysfunctions, neuroinflammation or maybe neuroprogression in patients with ME/CFS.
Given the different findings on activated immuno-inflammatory pathways in ME/CFS, it comes as no surprise therefore that autonomic (Newton et al. 2007; Winkler et al. 2004) and endocrine abnormalities (Van Den Eede et al. 2007; Demitrack et al. 1991; Demitrack and Crofford 1998) accompany ME/CFS. Recently we have reviewed the autonomic dysfunctions and orthostatic abnormalities in ME/CFS (Maes and Twisk 2009). Autonomic symptoms are associated with cardiovascular abnormalities and may be explained by activated immuno-inflammatory, intracellular and O&NS pathways, including increased TNFα, NO-related mechanisms and NF-κB (Maes et al. 2011; Maes and Twisk 2009). Hypothalamic-pituitary-adrenal (HPA) axis hypofunction has been established in part of individuals with ME/CFS (Scott et al. 1998). This entails decreased production of cortisol, blunted diurnal variation of cortisol, and blunted responses to a variety of provocation tests, including psychological and physical stressors (under review). Also these alterations in HPA-axis function may be explained by immuno-inflammatory and O&NS effects, including increased NO production attenuating HPA-axis functions (under review).
Mechanistic explanations for the onset of ME/CFS symptoms including post exertional malaise
This section will propose an immuno-inflammatory explanation for the hallmark symptoms of the disease, including fatigue, malaise, neurocognitive symptoms and a range of abnormal responses to exertion that may be delayed by 24 or even 48 h, which is often labeled as post exertional malaise (VanNess et al. 2010; Van Oosterwijck et al. 2010; Carruthers et al. 2011). The term refers to abnormal responses to even a trivial increase in normal levels of physical or neurocognitive activity. Post activity relapse may therefore be a better term. This phenomenon is often reported by ME patients to be similar to the acute phase of influenza, with symptoms indicating infection/inflammation, e.g. sore throat, lymph tenderness or swelling, malaise, hyperalgesia and brain fog (Twisk and Maes 2009). Insignificant increases in physical activity or minor cognitive tasks may exacerbate immune dysfunction, inflammation and O&NS thereby producing further signs of the disease (Twisk and Maes 2009). The effect may be delayed, but is predictable and accumulative, varying in duration dependant on disease severity and accumulative activity levels (VanNess et al. 2010; Van Oosterwijck et al. 2010; Carruthers et al. 2011).
In ME/CFS there are significant correlations between typical symptoms, such as fatigue, malaise, a subjective feeling of infection, sleep disorders, autonomic symptoms and post exertional malaise and activated O&NS (see above) and immuno-inflammatory pathways, including increased neopterin and TNFα concentrations (Maes et al. 2012a). This indicates that these characteristic symptoms of ME/CFS are in part mediated through effects of immuno-inflammatory and O&NS pathways. We have previously reviewed that PICs, e.g. IL-1 and TNFα, and Th1-like cytokines, such as IFNγ, activated O&NS pathways and lowered antioxidant levels, including coenzyme Q10 and zinc, may elicit the characteristic symptoms of ME/CFS, such as fatigue, malaise, and neurocognitive and autonomic symptoms (Maes and Twisk 2010; Maes et al. 2012b; Anderson et al. 2012). For example, PICs are heavily implicated in the cause of fatigue in multiple sclerosis (MS) (Heesen et al. 2006). Fatigue is also one of the most disabling symptoms in MS, with up to two thirds of patients describing fatigue as their main complaint (Schwid et al. 2002). TNFα, as a principal proinflammatory mediator, is associated with MS-related fatigue. This is in support of a pathogenic role of the MS-related inflammatory process in the development of fatigue (Flachenecker et al. 2004). Also, the production of Th1 cytokines, e.g. IFNγ, in MS patients is associated with the extent of fatigue (Pokryszko-Dragan et al. 2012). Other studies have reported that TNFα, IL-1β and IL-6 could be involved in the pathophysiology of mental fatigue through their ability to attenuate the astroglial clearance of extracellular glutamate (Rönnbäck and Elisabeth Hansson 2004). An adverse effect caused by specific PICs, such as IL-1, on learning and memory capacities, has been observed in several experimental systems. IL-6 facilitates lipopolysaccharide-induced disruption in working memory and expression of other PICs (Sparkman et al. 2006). TFNα and IL-1β regulate sleep architecture together with IL-6 (Krueger and Majade 1987; Cavadini et al. 2007; Obal and Krueger 2003) and can lead to sleep wake reversal and even complete loss of circadian control (Obal and Krueger 2003).
As explained above, bacterial translocation through binding to the TLR4 complex induces NF-κB and immuno-inflammatory and O&NS pathways and thus may cause behavioral and neuro-immune changes, including autoimmune responses directed against gangliosides (Maes et al. 2012c). Raised levels of IFNγ, IL-1, IL-2 and TNFα activate the tryptophan-degrading enzyme, indoleamine 2,3-dioxygenase (IDO), which causes alterations in the tryptophan catabolite (TRYCAT) pathway (Anderson et al. 2012). Increased IDO activation lowers plasma tryptophan and enhances the production of neurotoxic catabolites, e.g. quinolinic acid, a strong agonist of the glutamatergic N-methyl-D-aspartate (NMDA) receptor. Decreased tryptophan levels may cause physio-somatic symptoms, such as fatigue, autonomic symptoms, hyperalgesia and somatic presentations (Anderson et al. 2012). Disorders in the TRYCAT pathway are detected in post-viral fatigue syndrome (Anderson et al. 2012).
In the presence of increased levels of IL-1 and TFNα, ATP consumption is greatly accelerated, and levels of anaerobic glycolysis and glucose consumption are increased (Berg et al. 2003). The combined effects of these actions cause a greater metabolic demand even in the resting state (Berg et al. 2003). TNFα increases ROS production in the mitochondrial electron transport chain. NO is known to inhibit mitochondrial respiration, due to reversible binding to cytochrome C oxidase thus competing with oxygen and inhibiting energy production (Brown and Bal-Price 2003). Moreover, NO can be converted to other reactive species, including peroxynitrite, NO2, N2O3, and S-nitrosothiols, that may inhibit mitochondrial respiration and activate mitochondrial permeability transition, which eventually may trigger neuronal apoptosis or necrosis (Brown and Bal-Price 2003). PICs can directly disrupt glucose homeostasis and the availability of ATP. TNFα, IL-1β and IFNγ may significantly inhibit insulin secretion by beta cells in the pancreas (Kiely et al. 2007). This increases cellular glucose consumption and lactate and glutamate levels (Kiely et al. 2007). The resultant decrease in ATP may shift cellular metabolism towards a catabolic state, showing how mitochondrial dysfunction can severely impair glucose metabolism. PICs in general lead to the upregulation of nicotinamide adenine dinucleotide phosphate (NADP) oxidase and resultant increased production of ROS (Morgan et al. 2007). This process may generate superoxide anion dependent damage and formation of peroxinitrite which subsequently may contribute to permanent damage to mitochondria (Mohankumar et al. 1991; Shintani et al. 1993; Linthorst et al. 1995). IL-1 not only causes hypoglycemia, but also resets glucose homeostasis in the brain; both phenomena together favor an increased uptake of glucose by immune cells during inflammation (Del Rey et al. 2006). All in all, glucose usage can easily exceed supply and the increased demand for ATP cannot be met since mitochondrial functions are compromised by TNFα and NO. Cytokines can also lead to cerebral hypoperfusion via prolonged exposure to IL-1β (Maher et al. 2003). The ensuing lack of glucose would potentially explain the brain fog that so many people with this illness complain of.
TNFα reduces mitochondrial respiration via effects at the level of complex III of the electron transport chain and by ROS production (Samavati et al. 2008; Gudz et al. 1997). The latter may induce opening of mitochondrial permeability transition pores, which may cause uncoupling of oxidative phosphorylation (Chernyak 1997). The condition of increased ATP consumption and mitochondrial damage means that exceeding a threshold of physical and mental exertion will lead to all available ATP becoming exhausted and thus a crash may ensue. It takes damaged mitochondria time to replenish said ATP, leading to a profound post-exertional disability. Anything that increases TNFα activity during this “recharge” period, such as stress, will greatly delay recharge and possibly create longer term disability by further damaging the inner membrane. Chronically elevated cytokines act on the brain and hypothalamus to create a new homeostatis of glucose metabolism, and the autonomic nervous system controlling heart rate, respiratory rate, blood pressure, vascular tone, gastric peristalsis, etc. (Del Rey et al. 2006). The autonomic symptoms observed in ME/CFS could be obvious consequences of such a shift.
Increased TNFα and O&NS, iNOS, low ATP and high TGFβ1 levels deplete glutathione (Castagna et al. 1995; Phelps et al. 1995). A depletion in glutathione levels is often accompanied by a partial blockade in the methylation cycle. Moreover, nitric and nitrous acid inhibit methionine synthase (Layser 1978; Danishpajooh et al. 2001; Kerkeni et al. 2006; Culley et al. 2007). A number of studies mentioned earlier have demonstrated reduced glutathione in people with ME/CFS. A recent study confirmed these earlier findings and found that low glutathione levels and high lactate levels correlated with measurements of physical health and disability (Shungu et al. 2012).
Mechanistic explanations for the chronic and relapsing–remitting nature of ME/CFS
ME/CFS is characterized by highly increased IgM-related autoimmune response directed against disrupted lipid membrane components, by-products of lipid peroxidation, anchorage molecules, and NO-modified amino-acids as compared with normal controls (Maes et al. 2006b). These results show that self-epitopes, which are normally not detected by the immune system, have become immunogenic following damage by O&NS (Maes et al. 2006b). A distinctive pattern of proteins modified by oxidative stress in people with ME/CFS was also reported in a recent study involving proteomic analysis (Schutzer et al. 2011).
Such antibodies directed against lipids have also been detected in remitting-relapsing MS (Geffard et al. 2002; Boullerne et al. 1996). However the latter study reported that there was no significant level of IgM against azelaic acid detected, whereas Maes et al. (2006b) found increased IgM responses directed against azelaic acid in ME/CFS patients. The circulating IgM antibody titers in MS appear to be related to the presence of inflammation, because they increase during relapses and decrease during remissions (Bodet et al. 2004). This suggests that the level of immune activation, O&NS and magnitude of autoimmunity may in fact be greater in ME/CFS than in MS.
Increased levels of peroxinitrite and nitrosothiols and lowered levels of reduced glutathione have been found in the CSF of patients with MS, suggesting a role for nitrosative stress in MS (Calabrese et al. 2003). This situation is also found in ME/CFS, as discussed earlier, and therefore following factors could contribute to relapses in MS and ME/CFS. Innate pattern recognition receptors recognize O&NS damaged self epitopes, and the various components of the innate immune system, including innate IgM antibodies, collaborate to mediate their removal (Weismann and Binder 2012). In MS, IL-6 and TNFα levels are significantly increased during relapses as compared to remission (Hautecoeur et al. 1997), as are IL-17 and IL-23 (Wang et al. 2011). Plasmacytoid dendritic cells, which are elevated in the CNS during relapses, may be explained by either virus infections or down regulatory processes (Longhini et al. 2011). This dendritic subset is important in stimulating or downregulating effector T cells in MS (Geffard et al. 2002; Bayley-Bucktrout et al. 2008). By stimulating the maturation of dendritic cells, and upregulating the membrane expression of CD80, CD86, CD83, and HLA-DR (Csillag et al. 2010), O&NS induces maturation of dendritic cells (Kantengwa et al. 2003). Moreover, NF-κB is activated by threshold levels of oxidative stress (Khalaf et al. 2010) leading to the activation of autoreactive T cells. Thus chronic inflammation, due to the causes mentioned earlier, leads to a cycle where there is increasing damage to neoepitopes, ultimately leading to a surge of PICs due to immune activation and a cascade of further inflammatory damage and dysfunction of mitochondria, constituting a flare up of active disease or relapse.
We tentatively propose that (partial) remission is induced by the response of the CNS to immune activation. The CNS modulates immune responses through neuronal and hormonal effects. Thus, the sympathetic and parasympathetic system and neuroendocrine axes attenuate immune responses, whereas the peripheral nervous systems reinforce local immune responses (Sternberg 2006). In the first stage, the peripheral immune and nervous systems work together to activate local immune responses and to eliminate invading pathogens. In a later stage, however, counter-regulatory neuronal and neuroendocrine mechanisms are mounted that tend to terminate inflammation and thus restore host homeostasis (Sternberg 2006). Activation of the vagus nerve induces the release of transmitters that facilitate Th2 responses (Liang et al. 2011). Activation of the vagus nerve decreases the expression of M1 monocytes, which produce TNFα, IL-6, and IL-1β. Together, this indicates that activation of the vagus neural circuit produces an anti-inflammatory cellular milieu (Valdes-Ferrer et al. 2010).
The role of the peripheral nervous system in modulating the initial effects of immune activation and tissue inflammation is complex and beyond the scope of this paper. The antagonistic actions of the factors amplifying immuno-inflammatory pathways and the CNS would potentially account for a relapsing-remitting pattern. The findings that the T cell pool of remitted MS patients is in a state of anergy (Fransson et al. 2009) is of particular interest. Thus, active MS is associated with activated Th1/Th17 responses as discussed earlier, while it appears that people in remission have their immune system biased towards Th2 or Treg phenotypes. The immune abnormalities reported in people with ME/CFS vary between Th1 and maybe Th17, Th2 and Th2/Treg responses. This is the pattern one would expect if the relapsing-remitting nature of ME/CFS is produced by the interactions described above. A Th17 profile is induced by prolonged elevation of IL-6 and IL-23, which are in themselves the product of elevated O&NS pathways and chronic inflammation. IL-6 inhibits forkhead box P3 (scurfin, Foxp3) by activating the transcription of Signal Transducer and Activator of Transcription 3 (STAT-3) and hence attenuates the functions of Treg cells (Samanta et al. 2008; Korn et al. 2009). Also, IL-23 inhibits Foxp3 and thus promotes Th17 differentiation (Zhu et al. 2010; Lal et al. 2011; Mus et al. 2010).
Another plausible explanation for the chronic relapsing remitting nature of the disease revolves around tumor suppressor p53. P53 is activated by factors that cause DNA damage, including free radicals. P53 arrests the cell cycle, activates DNA repair and initiates apoptosis. P53 is also crucial in matching increased mitochondrial function to increases in demand for ATP (Niu et al. 2005; Matoba et al. 2006). P53 regulates energy metabolism by balancing glycolysis and oxidative phosphorylation. In conditions of oxidative stress, p53 activation results in decreased glycolysis. In such an environment, p53 expression acts to upregulate glutathione. The ultimate aim of both functions is the reduction of oxidative stress. p53 also regulates mitochondrial oxygen production via the synthesis of cytochrome C oxidase. In normal circumstances activation of p53 leads to an upregulation of mitochondrial respiration. NF-κB and p53 have antagonistic strategies in the cell and therefore cannot function simultaneously in the same cell (Ak and Levine 2010). Thus, upon activation of one of these two transcription factors the other is downregulated. Moreover, STAT-3 activation leads to the downregulation of p53 expression (Niu et al. 2005), while the latter is required for Foxp3 expression (Jung et al. 2010).
Therefore, it may be hypothesized that ME/CFS may be accompanied by lowered p53 expression. Lower transcription of p53 causes a shift from ATP production by oxidative phosphorylation to glycolysis. This can lead to a severe energy deficit and high levels of lactate production (Xiang et al. 2010; Holley and St Clair 2009). Thus, increased levels of NF-κB as observed in ME/CFS may cause a gradual downregulation of p53 (Holley and St Clair 2009), which may lead to a gradual worsening of symptoms. Increasing levels of mitochondrial damage however act to “switch on” the transcription of p53. This can have the effects of raising levels of Foxp3 and thus a reduction in symptoms. P53 activation can also induce cell-cycle arrest and allow mitochondrial healing. A drop in free radical production acts to suspend the activation of p53. This allows the levels of NF-κB and oxidative stress to downregulate p53 and Foxp3, and remission is terminated. A number of such cycles are possible with a worsening of active disease over time. The presence of raised anticardiolipin antibodies in up to 95 % of patients with ME/CFS (Klein and Berg 1995) is consistent with the presence of peroxidative mitochondrial damage and loss of cytochrome C oxidase function (Paradies et al. 1998).
Future research should further examine the different inflammatory and O&NS pathways, the transition to autoimmune responses and damage to mitochondria, the intracellular signaling networks involved, neuro-inflammation and brain metabolic changes in relation to the new diagnostic criteria (Maes et al. 2012a). We suggest that future research should be based on cohort selection using PET scans to detect activated microglia (Gerhard et al. 2006). We would suggest objective testing for the existence of BBB permeability using dynamic Gadolinium-DTPA MRI scanning (Larsson and Tofts 1992). The initial challenge would be to isolate people with ME from those with CFS and CF (Maes et al. 2012a) via the use of PET, SPECT scans, 3 T MRI or Gadolinium-DTPA MRI.
Mass spectroscopy and proteomic approaches for detecting and characterizing nitrosatively and oxidatively modified lipids, proteins and DNA produce specific sensitive and reproducible results (Aulak et al. 2001). Luminex assays for the determination of cytokine levels in supernatants from mitogen stimulated cultured peripheral blood mononuclear cells (PMBCs) are regarded as producing reliable and reproducible results while minimizing intra and inter laboratory variation (Griffiths et al. 2002; Fulton et al. 1997; Fulya et al. 2006; de Jager et al. 2009). Real time reverse transcriptase PCR provides a sensitive and quantitative measure of cytokine mRNA levels in PBMCs. STAT gene expression, p53 expression and Foxp3 transcription should be measured.
Mitochondrial dysfunction and ATP levels should be assessed using a combination of nuclear magnetic resonance spectroscopy, (31)P magnetic-resonance spectroscopy and blood tests measuring ATP synthesis in PMBCs (Dykens et al. 2008; Chaumeil et al. 2009). Exercise stress testing has the potential to do harm to people with ME/CFS and thus the non-ischemic repeat handgrip test should be considered as a simple non-invasive test for mitochondrial dysfunction (Meulemans et al. 2007). The aerobic forearm test provides the basis of a simple measure of mitochondrial function and could be adapted to perform measurements over a 3 day period (Garrabou et al. 2006).
Genetic polymorphisms in the DNA repair mechanisms and O&NS pathways, single nucleotide polymorphisms (SNPs) in the NF-κB and p53 pathways may further reveal genetic vulnerabilities to develop ME/CFS. Next generation sequencing offers a rapid and sensitive method for detecting the wide range of SNPs in the human genome and offers promise for investigating SNPs in patients with ME and CFS now that scientific diagnostic criteria have been established (Maes et al. 2012a; Goya et al. 2010; Karchin 2009).
The authors would like to thank Victoria Storey and Jane Clout for secretarial services. There was no specific financial support for this study. The authors declare that they do not have a conflict of interest.