Transient immunosuppression: a bridge between infection and the atypical autoimmunity of Guillain–Barré syndrome?

Abstract

Guillain–Barré syndrome (GBS) is an acute, usually monophasic, disorder of the peripheral nervous system that is assumed to be of immune-mediated pathogenesis. However, several clinical features and experimental findings of GBS are uncharacteristic for an immune-mediated disorder and set this condition apart from other disorders with a putative immune-mediated pathogenesis. These features include, among others, the monophasic nature of GBS, the lack of response to immunosuppressive (unlike immunomodulatory) therapy, the absence of a typical association with immunogenetic background and the inability to establish a valid and relevant animal model. We suggest a comprehensive hypothesis for the pathogenesis of GBS that is based on the assumption that the condition is due to a transient (or occasionally chronic) immune deficiency, as in most cases GBS follows an infection with pathogens known to induce immunosuppression. Such infections may be followed by breakdown of immune tolerance and induction of an immune attack on peripheral nerves. Mounting of the immune-mediated assault might be triggered either by the same infective pathogen or by secondary infection. Clearance of the infection and resumption of a normal immune response and tolerance eventually terminate the immune-mediated damage to the peripheral nerves and enable recovery. This hypothesis assumes that the entire sequence of events that culminates in GBS is due to transient exogenous factors and excludes a significant role for inherent host susceptibility, which explains the monophasic nature of the disorder.

Introduction

The term Guillain–Barré syndrome (GBS) encompasses several acute, paralytic disorders affecting the peripheral nervous system (PNS) that appear to share a similar aetiology and pathogenesis [1,2], and may even extend in some instances into the central nervous system (CNS), namely brainstem structures [3]. GBS manifests as an acute, monophasic disorder of variable severity, which is often triggered by an antecedent infection. The natural history, course and prognosis of GBS are well delineated and the currently recommended therapies seem to modify the morbidity associated with the disorder without a profound effect on the long-term outcome. GBS is currently considered a post-infectious, immune-mediated condition. This is based mainly on the pathology, the course, the response to immunomodulating therapies and the findings obtained in experimental animal models [1–5].

GBS, however, is distinct from most other systemic or neurological immune-mediated conditions. While the latter tend to be chronic with recurrent relapses (Table 1), in more than 95% of cases GBS is a single, once-in-a-lifetime episode. Considerable advance has been made over the past 15 years towards the recognition of the putative autoantigen and understanding of the pathogenetic mechanisms. As will be discussed later, other epidemiological, pathological and immunogenetic features of GBS are also either unusual for immune-mediated conditions or hard to explain. Importantly, an easily induced and consistently reproducible animal model mimicking the actual progression of the immune-mediated response in GBS is still lacking. This is in sharp contrast with other putative immune-mediated conditions, either those affecting the nervous system (such as myasthenia gravis or acute disseminated encephalomyelitis; ADEM) or disorders that involve other organs and have a systemic nature (e.g. systemic lupus erythematosus, SLE), where a reliable animal model is available. Indeed, in conditions where such a model is missing, such as multiple sclerosis (MS), the immune-mediated pathogenesis is disputed [6].

Table 1.

Course of putative neurological and systemic immune-mediated conditions.

Condition	Chronic	Relapsing	Monophasic
Neurological
MG	X	X
Multiple sclerosis	X	X
CIDP	X	X
ADEM		(X)	X
Paraneoplastic diseases	X
NMO		X
GBS		(X)	X
Systemic
DM	X
RA	X
SLE	X
Sjögren	X

ADEM, acute disseminated encephalomyelitis; CIDP, chronic inflammatory demyelinating polyneuropathy; DM, diabetes mellitus; GBS, Guillain–Barré syndrome; IDP, inflammatory demyelinating polyneuropathy; MG, myasthenia gravis; NMO, neuromyelitis optica; RA, rheumatoid arthritis; SLE, systemic lupus erythematosus.

The aim of this paper is to present the perplexing aspects of GBS and offer a modified explanation for the pathogenesis of this condition.

GBS features are atypical of most immune-mediated disorders, and are not explained fully by current knowledge of GBS pathophysiology

Patient population

Men are approximately 1·5 times more likely to be affected with GBS than women [7]. This is in contrast with most immune-mediated disorders that are inclined to afflict women more than men. Exceptions may include chronic inflammatory demyelinating neuropathy and multi-focal motor neuropathy. In Europe and North America, GBS incidence increases steadily with advancing age. This is also unusual, as conditions of putative immune pathogenesis tend to strike during the third to the fifth decades of life. Moreover, neonatal cases have been described and GBS should be included in the differential diagnosis of the ‘floppy infant’: approximately one-third of childhood cases occur before the age of 3 years. This is unusual for immune-mediated conditions that are likely to appear when the immune system matures and are therefore rare during infancy [8]. However, other conditions with putative immune-mediated pathogenesis such as myasthenia gravis and chronic inflammatory demyelinating polyneuropathy (CIDP) have been reported in childhood.

Clinical course

As noted, unlike most immune-mediated conditions (Table 1) GBS is a monophasic condition, in which recurrence is estimated to occur in fewer than 5% of patients [9,10]. Even then, recurrence is often associated with specific triggers such as vaccination or immune reconstitution. GBS follows an acute course, and may even present in a hyperacute or fulminant form [11], whereas most other human immune-mediated conditions are subacute or chronic. Nevertheless, acute forms of other conditions such as MS are recognized.

Most immune-mediated conditions are of heterogeneous clinical presentation and nature: the course of myasthenia gravis, neuromyelitis optica or MS as well as SLE or rheumatoid arthritis (RA) is often unpredictable and erratic. This is not the case with GBS, which has a rather limited set of core clinical features and a very similar and predictable course in the vast majority of patients [5]. This is especially striking considering that the neuropathy may be preceded by infections of varied aetiologies and that diverse immunopathogenetic mechanisms are implicated in the evolution of GBS.

Only two immunomodulatory therapies have been shown so far to be effective in GBS: plasma exchange and intravenous immunoglobulins (IVIg) [12]. Atypically for an immune-mediated disorder, GBS does not respond to immunosuppressive therapy and corticosteroids have been shown consistently to be ineffective in this disorder [13].

Of particular note for our discussion is the occurrence of GBS under conditions of immunological compromise [14–17]. These include not only patients who receive immunosuppressive agents for immune-mediated conditions [14], but also those under therapy following organ transplantation [15–17] and patients who are human immunodeficiency virus (HIV) seropositive or have acquired immune deficiency syndrome (AIDS) [18].

Pathology

In contrast to its clinical uniformity, pathological studies indicate a clear distinction between the demyelinating form (AIDP) and the axonal subtypes [acute motor axonal neuropathy (AMAN) and acute motor and sensory axonal neuropathy (AMSAN)] of GBS. In the uncommon event when pathological information is available, AMAN patients show an antibody mediated primary axonal degeneration whereas AIDP is characterized by a segmental demyelinating process affecting mainly nerve roots, and associated presumably with conduction blocks leading to muscle paralysis [19]. Three peculiar pathological observations should be noted, and these remain currently incompletely explained, as follows.

1. The brunt of the disease tends to fall on the proximal parts of the peripheral nerves. Indeed, the nerve roots demonstrate clinical (pain), electrophysiological and, when available, pathological evidence of involvement. Nevertheless, while the cerebrospinal fluid shows elevated protein levels due to alterations in the blood–nerve barrier enabling leakage of proteins, it does not reflect fully the inflammatory process present within the roots, i.e. there are no cells in the cerebrospinal fluid.

2. In many instances, clinical and electrophysiological evidence of demyelination with conduction block is not accompanied by pathological evidence of cellular infiltration, suggesting that demyelination could also be the result of antibody binding to myelin epitopes followed by complement activation [5]. This hypothesis is supported by experimental ex vivo model studies [20], and by findings of overlapping anti-ganglioside antibody profiles among the GBS subtypes [21], supporting the possibility that similar underlying mechanisms are associated with all GBS subtypes. Similarly, the overlapping antecedent infections and the occurrence of anti-GQ1b antibodies in Fisher syndrome and its CNS homologue Bickerstaff encephalitis, imply a common pathogenic mechanism for the entire spectrum of ‘Fisher–Bickerstaff’ and allow its inclusion in the broader spectrum of disorders classified as GBS [3].

3. In many cases of axonal GBS, recovery rates are rather fast and therefore inconsistent with actual axonal degeneration. In these cases a physiological block is postulated, but it is still unclear whether anti-ganglioside antibodies can block conduction by themselves, or whether additional soluble factors play a role in blocking normal conduction [5,22].

Immunogenetics

In many conditions with putative immune-mediated pathogenesis, there exists an immunogenetic background common to many of the patients, often involving such loci as the human leucocyte antigen (HLA) class I or class II ([23,24], Table 2). However, none of these loci has been associated with a predisposition or a risk for GBS and there are only rare reports of GBS occurring in more than one family member [25]. Most investigations have failed to reveal any association between HLA antigens and GBS, except in isolated instances: the HLA DQB1*03 in British Campylobacter jejuni-associated GBS [26] and the strong positive association with particular DQβ and DRβ positional residues in northern Chinese patients [27]. Despite the recent attention drawn to CD1 polymorphisms, it now appears that this T cell function-related molecule also fails to demonstrate consistent GBS-associated variations [28].

Table 2.

Some susceptibility loci associated with autoimmune and other immune-mediated diseases (modified from [14]).

Locus	Chromosomal location	Disorder
DQA1, DQB1, DRB1	6p21	Rheumatoid arthritis (RA), multiple sclerosis, IDDM
CTLA4	2q33.2	Hypothyroidism, IDDM
PTPN22	1p13	RA, Graves' disease, Hashimoto thyroiditis, SLE, myasthenia gravis, IDDM
FCRL3	1q21-q22	RA, autoimmune thyroid disease, SLE

IDDM, insulin-dependent diabetes mellitus; SLE, systemic lupus erythematosus.

Immunology and autoimmunity

There is abundant literature on the immunological abnormalities in GBS and there are several recent reviews that cover this subject [2,29,30]. Historically, the theory that GBS is an immune-mediated disorder was based on the histological findings of inflammatory infiltrates and macrophage-mediated myelin destruction, the presence of signs of immune system activation at disease onset and the similarity to the animal model experimental autoimmune neuritis (EAN) [31]. Despite these not unconvincing fragments of circumstantial evidence, none of the five criteria defining a disorder as autoimmune: identification of the autoantigen and the immunological mechanisms of tissue damage, passive transfer of the condition, an animal model generated by active immunization and response to immunotherapy [32] have been fulfilled unequivocally.

Current thought points to gangliosides as key autoantigens in GBS and attributes the immune-mediated attack to molecular mimicry between microbial antigens and nerve gangliosides. In brief, gangliosides, and GM1 in particular, are expressed in peripheral nerve tissue [33] and anti-gangliosides antibodies are present in 10–78% of patients [1,4,21,34,35]. There is molecular mimicry between gangliosides and lipo-oligosaccharides (LOS) of C. jejuni, the most common antecedent infection to trigger GBS, and C. jejuni-lipololysaccharide (LPS) can induce anti-ganglioside antibodies in experimental animals that are capable of causing nerve dysfunction [20,36–39]. Indeed, C. jejuni strains with no molecular semblance to gangliosides have not been associated with GBS cases [34]. Cell-mediated immunity may also have a role in induction of the demyelination, as T cells are sometimes present in peripheral nerves, and circulating T cells show activation against PNS myelin antigens such as P₀ and P₂[40].

None the less, the above observations of the involvement of humoral and cellular reactions in GBS still fail to explain key features of this disorder. This is particularly noticeable when applying these observations to alleged animal models of GBS, as follows.

1. Attempts to induce disease by passive transfer of serum and antibodies produced conflicting results [41,42]. Even high-affinity, high-titre anti-GM1 IgG antibodies transferred to wild-type mice caused an exceedingly mild neuropathy [43].

2. The classical animal model of GBS, experimental autoimmune neuritis (EAN), where cell invasion of peripheral nerves and their roots is associated with demyelination [44], results in axonal degeneration which appears to be more prominent than myelin damage [45], unlike in clinical GBS. Additionally, while in most GBS cases no consistent T cell reactivity to the same neural antigens has been demonstrated [40], the neuritis in EAN is mainly T cell-mediated. Importantly, the EAN model provides no insight into the mechanism(s) by which the autoimmune response is generated. Even the recently reported model of inducing motor weakness and axonal degeneration in rabbits by repeated immunization with GM1, mixed gangliosides or C. jejuni LOS in Complete Freund's adjuvant has major dissimilarities from clinical GBS [46,47]: the development of a chronic disease, unlike the monophasic GBS, and the need for repetitive immunizations with high doses of C. jejuni-LPS and highly potent adjuvants, of which the keyhole limpet haemocyanin may itself be pathogenic, as it cross-reacts with certain gangliosides [48].

3. Current experimental research does not explain why despite the strong antigenicity of the ganglioside-mimicking C. jejuni LOS [36], the vast majority of C. jejuni diarrhoea patients fail to develop ganglioside-reactive antibodies and subsequently GBS [4]. Of particular relevance to the present discussion is the gangliosides' peculiar ability to induce immunosuppression [49–51]. The role of ganglioside (or its mimicking LOS)-induced immunosuppression, which may be a crucial element in the induction of EAN, has never been investigated in animals. This issue will be discussed further below.

4. The current models do not explain why immunomodulation, but not immunosuppression, can shorten the incapacitation and the nadir of disability in patients during the acute attack (without an impact on prognosis) and why steroids have been shown repeatedly and consistently not to influence disease course and prognosis.

In summary, while ganglioside carbohydrate epitopes (especially GM1) might be plausible autoantigens in GBS, major questions concerning the mechanism of the immune-mediated response and nerve injury remain unanswered with our current knowledge. Relevant and valid animal models for GBS, which bear greater resemblance to the pathophysiology and clinical course encountered in patients, are still missing.

Hypothesis building blocks

A previously published review stated that ‘The occasional occurrence of Guillain–Barré syndrome in patients with suppressed immune function caused by AIDS or immune suppression after organ transplantation is difficult to explain’[1]. Is there a way to reconcile the features of GBS that are atypical of immune-mediated disorders with the evidence supporting a pivotal immune-mediated component in its pathogenesis?

We have speculated above on the data that are not explained satisfactorily by the current concepts (Table 3). Nevertheless, several undisputed observations may enable to bring together some of these discrepancies and apparent contradictions.

Table 3.

Summary of factors that are uncommon in immune-mediated conditions or do not favour an immune-mediated pathogensis.

Epidemiology and clinical features

Men afflicted more than women

The incidence increases with age

Cases reported in babies and infants

Monophasic condition

Homogeneous clinical entity

Non-response to corticosteroids and immunosuppressive therapy

Pathology

CSF does not show signs of inflammation

Lack of cellular infiltration in some nerves that show evidence of demyelination

Immunology

Lack of fulfillment of any of the five criteria required to define a condition as autoimmune, including identification of the culprit autoantigen

Lack of association with genetic loci that form the immunogenetic background of immune-mediated conditions

Lack of a valid and relevant animal model

CSF, cerebrospinal fluid.

Notably, infection within the preceding 6 weeks heralds GBS in about two-thirds of patients [52]. Often the pathogen is not identified, but four organisms have been recognized frequently and account for up to 60% of the preceding infections: C. jejuni, cytomegalovirus (CMV), Epstein–Barr virus (EBV) and Mycoplasma pneumoniae[53,54]. Another herpes virus irruption, varicella zoster virus (VZV), has also been recorded to precede GBS [52]. GBS has also been linked with HIV seropositivity and AIDS. GBS occurs most often at the time of seroconversion or in the early course of HIV infections, where the CD4 count is relatively unaffected, and then cellular infiltrates are present in the cerebrospinal fluid [55]. Occasionally, GBS is reported in advanced-stage AIDS, with CD4 count fewer than 50 cells/mm³ or during immune reconstitution [17,56].

Of great importance to our discussion is the ability of the five pathogens that top the list of GBS antecedent infections to either induce immunosuppression or to take place under defective immune surveillance. Thus, C. jejuni has antigens that cross-react with endogenous gangliosides and these antigens can induce immunosuppression [49–51]. CMV is not only a major cause of opportunistic infections in immunocompromised hosts such as AIDS [57,58] and organ transplant patients [59], but it can also manipulate the immune system in immunocompetent patients causing transient immunosuppression [59,60]. For example, monocytes infected in vitro with CMV undergo cellular redistribution of chemokine receptors CCR1, CCR2, CCR5 and CXCR4, such that these monocytes lose their sensitivity to signals initiating migration into sites of infection and for antigen presentation, which impairs their ability to recruit other immune cells [61]. Similarly, EBV, a lymphotropic herpes virus, is capable of inducing disorders, mainly malignancies, in immunocompromised individuals [62], but can also manipulate the immune system causing transient immune abnormalities [63]. Both varicella and herpes zoster due to VZV are conditions that affect the immunocompromised host [64]: the occurrence of shingles in a young adult should raise the possibility of some kind of immune deficiency. M. pneumoniae, while not recognized as a major cause of infection in immune deficient hosts [65], has been shown to be capable of manipulating the immune system [66]. Other infective agents associated with immunosuppression such as parvovirus B19 have also been associated with GBS [67].

All those pathogens have also been associated with immune-mediated conditions. While C. jejuni has been linked only with GBS, the three herpes viruses, CMV, EBV and VZV, have been implicated in other conditions of putative immune-mediated pathogenesis. For example, CMV has been connected with SLE [68], EBV with multiple immune-mediated conditions [69] and VZV has been proposed recently to cause multiple sclerosis [70]. Similarly, M. pneumoniae has been associated with a gamut of immune-mediated conditions [64,71].

The association between immunizations and GBS is problematic. A case–control study identified only a very small increase in the risk of GBS following influenza vaccination [72], but some vaccine batches associated with an increased risk for GBS (during the 1976 H1N1 epidemic) were found to be capable of inducing anti-ganglioside antibodies in mice [73]. With the older rabies vaccine versions, the increased risk of post-vaccination paralytic disease was attributed to the fact that the vaccine contained brain material, a problem that was eradicated with newer vaccine preparations [74,75]. Other immunizations have not been shown to carry an increased risk of GBS, except for a small increase of the risk associated inconsistently with the MCV4 meningococcal vaccine, and a number of case reports of GBS after hepatitis B vaccination, uncorroborated by the single epidemiological study that has been carried out [75]. Thus it seems that vaccines which do not contain nervous tissue cross-reactive antigens are associated with a very slight or negligible risk for GBS development. As vaccines are administered to large cohorts of healthy subjects, with intact immune functions, these findings would be consistent with our hypothesis as presented below.

A number of GBS cases were reported to occur in the wake of significant head trauma or surgical procedures such as brain or chest surgery (reviewed in [1–5]). After excluding the case of the heart transplant recipient in whom the immunosuppressed state is obvious, it is tempting to speculate on the role of physical and psychological stress as leading to altered immune function resulting in a susceptibility to develop GBS [76,77]. However, epidemiological data regarding the association of stressful, traumatic events with GBS are still unavailable. Future, prospective research on this subject might corroborate this point in support of our hypothesis.

Hypothesis

Only a minuscule proportion of patients infected with any of the aforementioned infectious agents go on to develop GBS. For example, whereas molecular mimicry is common to many clinical C. jejuni isolates, GBS develops in only 1/10⁴ of exposed individuals [4,78]. Clearly, more than simply antigenic mimicry is necessary to induce GBS, and most probably several factors need to co-exist and act in concert to induce the disorder.

Taken together, we propose the following: GBS is a disorder that develops in the context of immune deficiency or otherwise altered immune function. As failure or breakdown of immunological tolerance can result in autoimmunity and immune-mediated diseases [79–81], the immune-compromised state can be due to a transient factor such as an infection with a pathogen capable of manipulating the immune response. Recent data suggest that subsets of regulatory T cells, such as CD4⁺CD25⁺ or T helper type 1 (Th1) cells [82,83], are important in the pathogenesis of some organ-specific immune-mediated diseases such as type 1 diabetes mellitus [84]. Some of the pathogens associated with induction of GBS have indeed been shown to differentially affect these subpopulations of regulatory T lymphocytes. For example: an enhanced response to EBNA1 (an EBV nuclear antigen) was mediated by an expanded reservoir of EBNA1-specific central memory CD4⁽⁺⁾ Th1 precursors and Th1, but not Th17 cells in multiple sclerosis patients [85]. The subset of CD4⁺CD25 cells is also suppressed in EBV-associated lymphomas [86]. Similar findings were reported recently with CMV [87]. Indeed, two recent studies identified abnormality of circulating CD4⁽⁺⁾CD25⁽⁺⁾ regulatory T cells in GBS [88,89]. Additionally, a new subset of Th1 cells that predominantly produce interleukin (IL)-17 and induce autoimmunity has been discovered, and it is believed that this subset may be the major cell type involved in orchestrating tissue inflammation and autoimmunity (reviewed in [90]). In another, albeit chronic, presumably immune-mediated disease targeting peripheral nerve myelin – CIDP – an up-regulation of the Th1 cytokine IL-17 was detected in the CSF of non-treated patients [91].

Another physiological transient immune compromised condition is pregnancy [92]. Severe GBS cases were reported during pregnancy with rapid improvement when the pregnancy ended [93]. Alternatively, chronic states of immunosuppression, such that occur following organ transplantation with lifelong immunosuppressive therapy with ageing when cell-mediated immunity declines, or in AIDS patients, can provide a backdrop for GBS. This explains the occurrence of GBS under immune-compromised conditions, the increased incidence of GBS with age and the presence of this condition in infants lacking a mature immune system, as well as the unresponsiveness to corticosteroids. Is chronic immunodeficiency enough to set the stage for GBS? Not necessarily. Of note is the occurrence of GBS in immune compromised states only after additional immunosuppressive therapy is administered, not during the pre-existing chronic disease [94–96], again accounting for the transient course of GBS even in patients with chronic immunosuppressive conditions.

Theoretically, such a model of immune suppression might also explain the absence of cellular response in the cerebrospinal fluid and sometimes in the tissue. This is supported by the fact that during HIV seroconversion, when there is no immunosuppression, GBS is associated with CSF pleocytosis [97].

If the disorder is due to an aberrant immune state secondary to an infection, the transient and usually short time-span of most infections may account for the monophasic nature of GBS. Primary infection is followed by generation of lifelong immunity that prevents a second infection. The fact that GBS is a single episode argues against the possibility of an endogenous inherent susceptibility which is present in diseases such as SLE. Indeed, hardly any immunogenetic predisposition for GBS has been identified.

Is the immune-mediated tissue damage elicited by the same organism(s) that induce(s) the immunosuppression, or is that damage caused by other organisms or mechanisms? Considering that the same pathogens implicated in GBS have been also implicated in other immune-mediated conditions, it is tempting to assume that in GBS the same pathogen causes sequentially immunosuppression followed by immune-mediated tissue damage. Another piece of evidence in favour of this possibility is the fact that when immunosuppression is present, similar agents such as CMV are still associated with GBS occurrence [98,99]. However, the possibility that GBS is due to co-infection may also be postulated: the first pathogen may induce breakdown of immune tolerance while the other can elicit an immune-mediated attack on peripheral nerve antigens.

Thus, we hypothesize that transient immunosuppression, secondary to a specific infective pathogen or drug therapy and not due to host factors, sets the stage for an immune response against self-antigens. With so many factors required to act in concert, from the infection- or drug-induced immune suppression, through the second infection (if needed) to the specific alterations in the immune reaction and antigen-specific attack to the peripheral nerve damage, it is unlikely that GBS will be more than a single event in a lifetime, especially bearing in mind that most infections are transient and non-recurrent.

Disclosure

The authors have no conflict of interest to disclose.