Regulatory T cells in multiple sclerosis and myasthenia gravis

Tregs play a key role in maintaining self-tolerance, and their dysfunction is well documented in multiple autoimmune diseases including Type 1 diabetes, GBS, psoriasis, and others [1, 13‐17]. Tregs regulate immune response in the periphery predominantly by suppressing Teff cells. Although significant differences in the number of circulating Tregs in MS or MG patients relative to healthy controls are not frequently reported, Tregs from these patients are reported to have lower suppressive capabilities [1, 13, 60, 61]. This suggests that functional deficits in Tregs may contribute to the pathogenesis of MS and MG. For example, defects in Treg suppressor molecules have been linked to MS, such as reduced IL-10 production and genetic variations in CD25 [27, 55]. Likewise, MG patients have documented dysregulation in cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) expression, IL-2 sensitivity, and the levels of transforming growth factor beta (TGF-β) gene expression [25, 62, 63]. Mechanistically, lower Treg suppressive capabilities may lead to enhanced production of pro-inflammatory cytokines such as IL-6, IL-17, and IFN-γ, as well as activation of autoantibody producing B cells. Normally, Tregs suppress the production of these pro-inflammatory cytokines through contact-dependent and contact-independent suppression of Teff cells known to produce IL-6, IL-17, and IFN-γ [64, 65]. Functional Tregs are also believed to directly eliminate both antigen-presenting and autoantibody-producing B cells in a contact-dependent manner through secretion of perforin and granzyme as seen in lupus prone mice (NZB/W) [66]. This mechanism may be pivotal for Treg-based amelioration of MG and perhaps MS [2, 66, 67]. The autoreactive B cells can produce autoantibodies against acetylcholine receptor (AChR) or muscarinic receptor, thus causing membrane damage via activation of the complement cascade in MG; likewise, it has been suggested that autoantibodies may also be involved in the pathogenesis of MS [68, 69]. Furthermore, activated B cells are found to be upregulated in the neuroinflammatory sites of the CNS in MS patients and in the blood of MG patients [70, 71]. Although, the exact cause of autoreactive B cell dysregulation is still unknown, the reduction in Treg suppressive capabilities and the concomitant activation of autoreactive B cells may be linked [66, 72]. Finally, there is documented evidence of dysfunctional migratory receptors and cytokines in both MS and MG which are used by Tregs; therefore, repairing migratory defects may be another therapeutic strategy [73, 74].

There are multiple Treg subsets and a variety of suppressive mechanisms which they use. The two main Treg subtypes include natural Tregs (nTreg) and induced Tregs (iTreg), each of which express the FoxP3 transcription factor. The nTreg is generated in the thymus, while iTregs are generated in the periphery, further distinguished by their level of methylation at the FoxP3 promotor [75]. Tregs are considered essential for maintaining peripheral tolerance against self-antigens through a variety of soluble mediators including IL-10, IL-35, and TGF-β, and cell surface molecules such as CD25 and CTLA-4 (rapidly recycled to/from the cell surface) [76‐84]. IL-10 has been shown to inhibit T cells by preventing CD28 co-signaling that leads to tyrosine phosphorylation [85]. The predominant suppressive mechanism used by T regulatory type 1 cells (Tr1; CD4⁺IL-10⁺FoxP3⁻), found to be dysfunctional in MS patients, also involves IL-10 [27]. Another Treg subset, namely iTr35, has a suppression mechanism which is not fully understood [86]. However, IL-35 is associated with upregulation of inhibitory molecules programmed cell death protein 1, T cell immunoglobulin and mucin-domain containing-3, and lymphocyte activation protein 3 on T cells [87]. TGF-β has been found to direct antibody production toward non-inflammatory IgG4 and IgA isotypes [85]. Additionally, TGF-β allows for differentiation of iTregs from CD4⁺CD25⁻FoxP3⁻ Teff cells [88]. Notably, iTregs have been induced using other molecules such as retinoic acid [89]. The CTLA-4 and CD28 provide co-stimulatory signals when bound by their cognate ligands CD80 and CD86 expressed on antigen-presenting cells (APCs). T cell co-stimulation with CD28 leads to activation signals; in contrast, co-stimulation with CTLA-4 leads to inhibition signals [90]. CTLA-4 co-stimulation then leads to trans-endocytosis of its ligands into the T cell, thus depleting ligands for use in CD28 co-stimulation [72]. Although CD25 (IL-2Rα) expression is seen on both Tregs and Teff cells, it is constitutively expressed on Tregs while only transiently expressed on Teff cells upon activation [91]. IL-2Rα, when dimerized with the β-chain, gains high affinity for IL-2, thus it is thought to cause IL-2 depletion by Tregs which restricts IL-2-dependent activation of Teff [91]. Lastly, Tregs have been shown to suppress and destroy B cells using perforin which creates pores in the target cell membrane, as well as granzyme B, a serine protease which induces programed cell death via caspase activation [66, 92]. Perforin and granzyme B are important in regulatory T follicular cells (TFR; CD4⁺CXCR5⁺FoxP3⁺) which have been shown to play a key role in suppressing B cells in the GCs and are reduced in MG patients [93, 94].

MS is a neuroinflammatory disease characterized by demyelination and inflammatory infiltrates in the central nervous system (CNS) [95]. Immunologically, MS correlates with Treg dysfunction, enhanced Th1 and Th17 responses, and autoreactive B cell over activity [95, 96]. This immune disequilibrium may be caused by a loss of Treg suppression of Teff which leads to myelin destruction causing neuronal damage and neuroinflammation [96, 97]. Clinically, MS presents with multiple neurological deficits and is grouped into three categories: 85% of patients are initially diagnosed with a relapsing remitting disease, of which 50–60% progress to secondary progressive MS, having worsening deficits with or without relapse [98]. A small percentage of patients, about 10%, exhibit primary progressive MS, in which symptoms are present from onset and gradually worsen without notable remission [98]. One of the most well-characterized pre-clinical models for MS is the myelin oligodendrocyte glycoprotein (MOG)-induced EAE model [95, 99]. Other common self-antigen-specific EAE models tend to use myelin basic protein or proteolipid protein (PLP); however, MS patients have been suggested to have antibodies against multiple self-antigens, such as heat shock protein α B-crystallin or S100beta [95, 100‐103]. MOG-induced EAE is initiated by the injection of either full length recombinant MOG protein, or peptide fragments including a 21 amino acid peptide spanning MOG residues 35-55, together with complete Freund’s adjuvant followed by pertussis toxin injection [95, 99]. After immunization, APCs present MOG peptides to T cells leading to autoreactive T cell activation, migration into the CNS, and subsequent neuroinflammation.

A recent analysis of over 14,000 MS patients using ImmunoChip genotyping found abnormalities in genes involved in Treg IL-2 signaling, CD25, and CD127 [55]. While CD127 is expressed highly on Teff, but minimally on Tregs, molecules such as GITR, OX40, Helios, CD49b are also expressed on Tregs [22, 28, 47, 77]. Furthermore, many MS patients have decreased FoxP3 expression in Tregs, decreased Treg suppressive function, and decreased levels of Tr1 [1, 27]. Additionally, recent research suggests that Tregs are unable to properly infiltrate the CNS during the course of the disease [104‐106]. Brain biopsies from MS patients revealed that 30% of the lesions lacked FoxP3 expression [104]. Fas, a cellular apoptotic pathway receptor, is seen upregulated on Tregs in MS brain biopsies suggesting increased susceptibility to apoptosis [104]. Taken together, these findings suggest that Tregs might be restricted from migrating to neuroinflammatory sites or undergo apoptosis upon arrival. Treg dysfunction during the disease progression can also be noted through modulation of CD28 and CTLA-4 co-receptors mentioned earlier. Studies in EAE have shown that during relapse, there is increased expression of CD28, co-localizing with T cell receptor subunit CD3, in the CNS blood vessels and parenchyma [107]. During remission, CD28⁺ cells decrease while CTLA-4⁺ cells increase [107]. Considering Tregs constitutively express CTLA-4 while Teff do not, it is possible these cells are Tregs; however, the study did not look for co-expression of FoxP3 [107, 108]. Taken together, Treg abnormalities in MS appear to involve loss of suppressive capacity as well as defect in migration into the CNS.

Many treatments which increase Tregs in EAE have been found to be efficacious, while other treatments, although successful, showed no effect on Tregs [109]. Administration of an indoleamine 2,3-dioxygenase (IDO) metabolite which increases Treg number leads to significant amelioration of MOG-induced EAE [110]. Additionally, the effect of IL-10-based therapy has been evaluated in different EAE models; most resulted in reduced clinical scores, perhaps associated with an increase in Tr1 associated with IL-10 administration [111, 112]. Some of the most affirmative evidence advocating Treg augmenting treatments have utilized adoptive transfer techniques. Adoptive transfer of FoxP3⁻ Treg cells that ectopically expressed the transcription factor FoxA1 (suggested marker for a novel Treg) ameliorated EAE in IFN-β knockout (KO) mice and these cells were increased in response to treatment of MS patients with IFN-β, suggesting a role for this Treg subset [113]. Although adoptive transfer of CD4⁺CD25⁺FoxP3⁺ cells reduced EAE, they differed in their efficacy. Adoptive transfer of CNS-derived KLRG1⁻ and KLRG1⁺ cells modestly reduced the clinical score at the termination of the experiment without affecting peak response (KLRG1 is associated with natural killer cells but was tested due to its increased expression on Tregs when entering the CNS of EAE mice) [114]. Likewise, CNS-derived CD4⁺CD25⁺ cells reduced the peak clinical EAE score [115]. Finally, experiments knocking out known Treg proliferating molecules, such as IDO or IL-10, resulted in exacerbation of EAE [110, 116].

In contrast to the above findings, FoxP3gfp knock-in mice had an accumulation of Treg cells in the CNS, but it failed to control EAE [117]. Administration of cholesterol-reducing statins such as atorvastatin inhibited the development of EAE without increasing the frequency of FoxP3⁺ Treg cells or Th2 cells, yet there was a substantial increase in IL-10 expression [109]. Taken together, these data tend to suggest that modulation of FoxP3 Tregs may not always be necessary for the protective effect of certain treatment modalities for EAE. Inasmuch as administration of statins involve non-specific immunosuppression, it is hard to avoid undesirable consequences. Therefore, it appears that boosting the Treg cells may produce a desirable clinical outcome in MS.

Dysfunctional migration of Tregs to neuroinflammatory sites could have profound implications; for example, GCN2 (aids in CCL2-mediated migration) KO Tregs adoptively transferred into EAE mice were unable to enter the CNS and hence mice were unable to undergo remission [118]. Some chemoattractants (e.g., CCL17) or their receptors (e.g., CCR4 and CXCR3) appear to facilitate Treg homing to neuroinflammatory sites. Human CD4⁺CD25⁺ Treg cells express CCR4 which respond to mature dendritic cell-derived CCL17 and are thereby recruited to the site of inflammation. Butti et al. determined that CCR4 is indispensable for Treg recruitment into the cerebral spinal fluid (CSF) and that recruitment was necessary for EAE amelioration likely due to the increased CCL17 levels after administration of IL-4 gene therapy into the CSF [119]. Studies of malignant diseases also find a correlation between the levels of CCL17 and accumulation of FoxP3+ cells, implicating CCL17 as a Treg recruiter molecule [120, 121]. Therefore, CCL17 administration at specific sites may be utilized to recruit Tregs.

Molecules such as CXCR3 are important for Tr1 and Treg migration to the CNS [106, 122]. The chemokine receptor CXCR3 that promotes trafficking of activated T and natural killer cells in response to CXCL9, CXCL10, and CXCL11 is expressed heavily on Tr1 cells which can direct Tr1 to the CNS thereby implying a role in EAE [122]. Administration of CXCL10-Fc exacerbates EAE, whereas CXCL11-Fc leads to amelioration with subsequent polarization to Tr1 [122]. Since CXCR3 and CXCL10 are found elevated in neuroinflammatory sites in MS patients, over-recruitment of Teff may be occurring, thus use of CXCL11 could be a strategy to recruit additional Tr1 to these neuroinflammatory sites [123]. Additionally, CXCR3 KO mice have more severe EAE with reduced FoxP3⁺ cells in the CNS despite little change in total CD4⁺ infiltrates implying that CXCR3 may be crucial for Treg migration to the CNS [106]. Therefore, a strategy to increase CXCR3⁺ expression on FoxP3⁺ Tregs may be useful for recruitment to the CNS. Enhanced CXCR3 expression by Tregs has been seen with the use of IL-27, thus providing a treatment strategy to facilitate selective Treg migration [122, 124].

Many common MS treatments could result in adverse effects revealing a need for new therapeutic strategies [9, 125]. Additionally, current treatments such as IFN-β (Avonex, Rebif, and Betaseron), glatiramer acetate (GA; Copaxone), fingolimod (Gilenya), teriflunomide (Aubagio), dimethyl fumarate (Tecfidera), Natalizumab (Tysabri), and mitoxantrone, although beneficial, are not curative [9]. Corticosteroids, which have been in use for over 75 years, are still the treatment of choice for acute MS symptoms [9, 126]. However, they lack specificity and tend to cause severe adverse effects such as abnormal weight gain, behavioral changes, oral thrush, and others [126]. Over two decades ago, glatiramer acetate became one of the first treatments for reducing relapse rates in MS patients; however, it had only a modest efficacy in reducing disease activity [127]. Many newer treatments, such as the ones indicated above, have been implemented since then, yet nearly all can cause significant side effects [9, 128, 129]. Therefore, there is a need for novel treatments which reduce disadvantages of current treatments while providing greater efficacy.

Most of the clinical treatments do not target Tregs as specifically as new pre-clinical treatments can [130]. Despite different targets, some treatments have been shown to increase Treg numbers such as IFN-β, GA, fingolimod, and teriflunomide, while others, such as dimethyl fumarate, natalizumab, mitoxantrone, and corticosteroids, have resulted in reduced, unaffected, or contradictive effects on Treg numbers [131‐140]. Treatments such as the lipid-lowering drug atorvastatin mentioned earlier have no effect on the FoxP3 Treg population, indicating that increasing the Treg population, although may be desirable, may not always be necessary [109]. However, this review will discuss strategies to restore Treg abnormalities as therapeutic modalities that may be considered for further testing. These Treg abnormalities have been corrected using cytokines such as IL-10, IL-27, IL-35, bimolecular protein inhibitors (BPIs), indoleamine 2,3-dioxygenase (IDO), or the chemokine CXCL11 (Table 2; Fig. 1) [80, 106, 110, 111, 141]. Additionally, we propose modalities of treatment for further testing. These include Jagged-1/OX40L co-treatment and site-specific CCL17 administration for Treg recruitment (Table 2).

MG is a NMJ disorder characterized by worsening muscle weakness with sustained contraction which improves upon rest [2]. Damage to the NMJ ion channels is autoantibody mediated where autoantibodies bind and result in the formation of the membrane attack complex consisting of C5b, C6, C7, C8, and C9 leading to muscular membrane damage and muscle weakness [2, 173]. The most common autoantibodies involved are anti-AChR antibodies presenting in 85% of patients, anti-muscle-specific kinase presenting in 5–8%, and anti-low density lipoprotein receptor-related protein in 2–46% of patients [69, 174, 175]. Anti-AChR autoantibodies bind to the AChR at the NMJ which induces complement-mediated damage of the NMJ, cause production of inflammatory cytokines, and subsequent reduced muscle functionality. The autoreactive T cells, found both in MG and EAMG, provide the necessary help for the pathogenic autoantibody production [2]. Interestingly, 60% of MG patients develop thymic hyperplasia and 10% develop thymomas, suggesting T cell dysregulation as an etiology [176]. EAMG, the experimental model for MG, is typically induced upon immunization with Torpedo californica AChR emulsified in an adjuvant [177]. The EAMG model is reported to closely resemble the clinical, pharmacological, histological, electrophysical, immunological, and pathogenic mechanisms in MG, yet it does not show thymic hyperplasia characteristic of human MG [2, 178, 179]. Since its description in the early 1970s, the EAMG model has been indispensable in understanding MG [180].

Many studies have reported Treg dysfunction in MG [13, 28]. Data is conflicting whether reduced Treg numbers contribute to the disease pathogenesis [60, 181]. A recent genome-wide analysis of MG patients revealed CTLA-4 and RANK (receptor activator of nuclear factor-kappa B) gene loci to be associated with disease risk [56]. CTLA-4, implicated in Treg-mediated suppression as discussed earlier, is found downregulated in MG patients [182]. RANK ligand (RANKL), which is preferentially expressed on Tregs as compared to Teff [183], is thought to function by suppressing CD11c⁺ DC activation through the RANK-RANKL pathway [184]. Inhibition of RANKL hinders Treg expansion, whereas RANKL upregulation has been seen to induce Treg proliferation [184, 185]. Further, Tregs in MG patients have decreased expression of FoxP3 and IL-10 indicating a functional deficit, while they also have enhanced expression of Fas indicating increased susceptibility to Fas-mediated apoptosis [13, 28]. Immunologically, reduced Treg suppressive activity in MG patients is accompanied by elevated inflammatory cytokines (IL-6, IL-17, TNF-α, IL-12, and IL-1β), most of which are normally suppressed by Tregs [2, 28]. Decreased suppression from Tregs may also correlate with autoantibody production. We and others have shown that Tregs adoptively transferred into EAMG mice have significantly delayed disease progression, reduced numbers of autoreactive T cells, and lower levels of AChR antibodies [59, 67]. Adoptive transfer of Tregs from GM-CSF-treated EAMG mice also resulted in a reduction of pro-inflammatory cytokines IL-6, IL-17, and IFN-γ, and increase in FoxP3 and IL-10 expression [67]. Reduction in clinical scores and AChR antibodies after Treg adoptive transfer was likely due to reduced complement fixation, secondary to suppression of B cells by Tregs as was seen in a model of systemic lupus erythematosus (SLE) [92]. A possible mechanism by which Tregs might reduce autoantibody production and subsequent complement fixation is via destruction of autoreactive B cells by contact-dependent release of perforin and granzymes [66]. Taken together, addressing Treg abnormalities in MG patients may provide a strategy for therapeutic intervention.

Few studies have explored Treg migratory patterns in EAMG or MG, yet one chemokine receptor, CXCR5, involved in Treg migration has been implicated in MG [74, 186]. CXCR5 is used by follicular T helper cells (TFH; CD4⁺CXCR5⁺FoxP3⁻), TFR, and B cells to facilitate homing to the B cell zone in GCs [187]. The TFR can suppress TFH and B cell numbers in the GC; however, MG patients have a documented reduction in TFR cells along with a reduced TFR to TFH cell ratio [93, 188]. Furthermore, the ratio of TFR to TFH cells is found to be inversely correlated with clinical severity in patients treated with corticosteroids [93]. Thus, restoring the TFR to TFH imbalance would be useful for Treg-based therapeutics, and this could be achieved with the use of IL-2/mAb as seen in an experiment using EAMG [93, 189].

MG has a limited range of treatments; the clinical treatments usually include acetylcholinesterase inhibitors (pyridostigmine bromide), rituximab (anti-CD20), corticosteroids, immunosuppression (such as azathioprine), intravenous immunoglobulin (IVIG), plasmapheresis, and thymectomy, yet none focus on curing the underlying cause [11]. Most current treatments are unsuccessful in stopping disease reoccurrence after remission and may have unacceptable side effects after long-term use [11]. Although the mechanism is not understood, Tregs have been seen to modestly increase with the use of drugs not targeting Tregs, such as pyridostigmine, rituximab, azathioprine, and IVIG [190‐193]. Corticosteroids lack specificity and importantly cause severe adverse effects such as bruising, abnormal weight gain, behavior changes, oral thrush, and others [194]. Pyridostigmine bromide, the most commonly used acetylcholinesterase inhibitor, is the first line for symptomatic treatment [195]. Yet it provides only short-term relief, it is not a disease-modifying therapeutic (DMT), and it can cause a hypersensitivity rash [195]. Azathioprine, a commonly used DMT, causes side effects in roughly 20% of patients and failed to prevent relapse in roughly 33% of patients in a long-term study of 117 MG patients [196]. Thymectomy has proven to be a beneficial DMT; however, it does not cure the underlying disease, and patients still suffer side effects from other treatments because thymectomized patients still require additional interventions such as corticosteroids or azathioprine [197]. Taken together, there is a need for more efficacious DMTs focused on destroying autoreactive B cells to reduce autoantibody production and NMJ destruction. Although anti-B cell therapies are in practice like rituximab (anti-CD20 antibody), Tregs’ ability to destroy autoreactive B cells may provide relatively better specificity [68, 92]. Treg-augmenting therapies which might accomplish this task or treat MG Treg abnormalities include GM-CSF, IL-2/mAb complexes, OX40L-Jagged-1 co-treatment, and TGF-β administration.