PD-1
PD-1 (Programmed cell death protein-1) is a member of the B7/CD28 family that was proposed to serve as a first-tier co-inhibitory receptor together with CTLA-4, playing an important role in maintaining peripheral tolerance (Anderson et al., 2016; Fife and Bluestone, 2008). In contrast to CTLA-4 that inhibits the initial stage of activation of naïve T cells typically within lymphoid tissues, PD-1 down-regulates previously activated T cells during the effector phase of an immune response primarily in peripheral tissues, and may serve to maintain tolerance within locally infiltrated tissues, reducing organ damage (Keir et al., 2006). PD-1 signaling results in the inhibition of TCR-mediated responses including T cell proliferation, cytokine production, cytotoxicity, and reduced T cell survival (Buchbinder and Desai, 2016).
PD-1 is broadly expressed on T cells, B cells, and myeloid cells upon activation and its expression is maintained in settings of persistent antigenic stimulation (Liang et al., 2003; Yamazaki et al., 2002). The expression of its two ligands PD-L1 and PL-L2 is regulated by inflammatory stimuli, especially cytokines, with the expression of PD-L2 restricted to APCs, while PD-L1 is expressed by a wide range of cells including leukocytes, non-hematopoetic cells, and parenchymal cells (Chen, 2004). PD-L1 and PD-L2 are not expressed by resting human CNS cells, but inflammatory cytokines including IFN- and TNF- upregulate PD-L1 expression on human astrocytes and microglia in vitro (Pittet et al., 2011) and PD-L1 expression is detected on glial cells in MS lesions (Ortler et al., 2008; Pittet et al., 2011). PD-1 and PD-L1 are expressed on CNS infiltrating cells in EAE with a temporal pattern following the clinical course of the disease (Salama et al., 2003; Schreiner et al., 2008) and PD-L1 is also upregulated on vascular endothelial cells and microglia in mice with EAE (Liang et al., 2003; Magnus et al., 2005; Schreiner et al., 2008).
Several studies have shown that blockade of PD-L1 and/or PD-L2 increases the susceptibility to EAE and enhances disease severity (Salama et al., 2003; Zhu et al., 2006). While initial studies indicated that interactions between PD-1 and PD-L2 constitute the dominant pathway in regulating EAE, later studies showed that the relative importance of the two ligands is strain specific and dependent on the genetic background (Zhu et al., 2006). Experiments in knock-out (KO) mice showed that PD-1 and PD-L1, but not PD-L2, deficient mice developed more severe EAE (Carter et al., 2007; Latchman et al., 2004; Wang et al., 2010) and that genetic disruption of PD-L1 alone is sufficient to convert an EAE-resistant strain into a fully permissive one (Latchman et al., 2004). It is believed that increased CNS expression of PD-L1 during EAE serves as a negative feedback mechanism limiting encephalitogenic T cell responses and inflammatory tissue damage and it is likely that these effects are mediated at least in part by Tregs that constitutively express PD-L1 and upregulate PD-1 expression following TCR stimulation (Keir et al., 2007; Taylor et al., 2004). PD-1 deficient mice have reduced frequencies of Tregs in vivo, lowering the threshold for disease induction, and PD-1 expression contributes to the conversion of naïve myelin-specific CD4+ T cells into Tregs in vitro, correlating with their suppressive activity (Wang et al., 2010).
Single nucleotide polymorphisms (SNPs) in the Pdcd-1 gene that encodes PD-1, have been associated with several autoimmune diseases including MS (Kasagi et al., 2011). The only SNP associated with MS is the regulatory SNP PD1.3A that increases the probability of disease progression in MS (Kroner et al., 2005). This SNP is also associated with increased susceptibility to diseases such as systemic lupus erythematosus (SLE), diabetes, and rheumatoid arthritis (Kasagi et al., 2011). The PD1.3A polymorphism is not associated with a measurable change in expression of PD-1 on T cells, but functional experiments showed that PD-1-mediated inhibition of IFN- secretion is impaired in patients carrying the mutated allele, which could result in continuous activation of self-reactive T cells perpetuating CNS inflammation (Kroner et al., 2005).
A few studies have analyzed the expression of PD-1 and its ligands in patients with MS, suggesting that PD-1 and PD-L1 may be involved in controlling disease activity. Relative expression of PD-1 and PD-L1 was lower in total PBMCs from MS patients compared to healthy controls (Javan et al., 2016). The expression of PD-1 was significantly increased in MBP-stimulated T cells from MS patients with stable disease compared to patients with active disease, and the expression of PD-L1 was increased in B cells and monocytes from the same patients (Trabattoni et al., 2009). Interestingly, a large proportion of CD8+ T cells identified within MS lesions did not express PD-1 and are hence resistant to regulation through PD-L1 (Pittet et al., 2011). Treatment with IFN-β resulted in increased expression of PD-L1 and PD-L2 on monocytes as well as PD-L1 on CD4+ T cells both in vivo and in vitro (Schreiner et al., 2008; Wiesemann et al., 2008) and an upregulation of PD-L2 was associated with treatment response to IFN-β in a small cohort of MS patients (Wiesemann et al., 2008).
TIGIT/CD226
A more recently defined network of costimulatory molecules with similarities to the CD28/CTLA-4/CD80/CD86 family is the TIGIT (T cell immunoreceptor with Ig and ITIM domains)/CD226/CD112/CD155 network (Anderson et al., 2016). Both TIGIT and CD226 are expressed on NK cells, effector and memory T cells, and Tregs, and share the two ligands CD112 and CD155 that are expressed by APCs, T cells, and a variety of non-hematopoietic cells including tumor cells. In analogy with CTLA-4, TIGIT binds to its ligands with higher affinity than CD226 and provides an inhibitory signal, while CD226 acts as a costimulatory molecule for T cells.
Transgenic mice expressing TIGIT on T and B cells and mice treated with a TIGIT agonist are protected from EAE (Levin et al., 2011), while TIGIT deficient mice are susceptible to EAE and display higher frequencies of encephalitogenic T cells and higher levels of pro-inflammatory cytokines (Joller et al., 2011; Levin et al., 2011). TIGIT deficient mice developed spontaneous atypical EAE, characterized by neurological symptoms similar to those associated with Th17-driven EAE (Jäger et al., 2009) when crossed with MOG35-55-specific TCR transgenic 2D2 mice, suggesting that TIGIT is involved in regulating the threshold of T cell activation and is important in maintaining peripheral tolerance (Joller et al., 2011). In addition to a direct inhibitory effect on T cell activation and expansion (Joller et al., 2011; Levin et al., 2011; Lozano et al., 2012), TIGIT also inhibits immune responses by inducing tolerogenic DCs that secrete more IL-10 and less IL-12p40, by inhibiting cytotoxicity in NK cells, and by promoting Treg-mediated suppression through the induction of IL-10 and Fgl2 that selectively suppresses Th1 and Th17 responses (Anderson et al., 2016; Stanietsky et al., 2009).
Little is known about the role of TIGIT in MS, but preliminary data indicate that TIGIT is present in mononuclear cell infiltrates in glioblastoma multiforme while it’s nearly absent from inflammatory MS lesions (Lowther et al., 2015). In contrast, CD226 and CD155 were detected in glioblastoma infiltrates and as well as in MS lesions suggesting that TIGIT may be a checkpoint inhibitor for tumor evasion in the CNS, while lack of TIGIT may aggravate MS.
Genome-wide association studies have shown an association of the Gly307Ser SNP in the CD226 gene with MS susceptibility (Hafler et al., 2009). Using several large-scale expression quantitative trait loci (eQTL) datasets, a correlation between the MS risk haplotype rs763361T and reduced CD226 expression was observed in PBMCs and brain tissue (Liu et al., 2017a). Interestingly, more in-depth characterization of the CD226 genetic variant showed that healthy subjects carrying the MS risk haplotype had reduced CD226 expression on T cells, while MS patients had a CD226 expression level comparable to the risk haplotype and showed no haplotype-dependent differences (Piedavent-Salomon et al., 2015). CD226 promotes Th1 and Th17 responses and suppresses Th2 function, and blockade of CD226 delayed the onset and reduced the reduced severity of EAE (Dardalhon et al., 2005; Lozano et al., 2013; Zhang et al., 2016), raising the question how reduced expression of CD226 is associated with increased risk of MS. The answer to this apparent contradiction may be related to CD226 expression on Tregs, as CD226 is highly expressed in IL10-producing Tr1 cells as well as in classical Foxp3+ Tregs (Gagliani et al., 2013; Koyama et al., 2013). CD226 deficient Tregs showed reduced inhibitory activity, which was associated with an exacerbated course of EAE in CD226 deficient mice (Piedavent-Salomon et al., 2015). Similar to the situation with CTLA-4, treatment with anti-CD226 antibodies may predominantly affect effector T cells with high CD226 expression, while CD226 deficient mice may more accurately reflect the genetically encoded reduced CD226 expression observed in humans (Piedavent-Salomon et al., 2015).
CD226 and its two ligands also play an important role in NK cell mediated lysis of activated T cells (de Andrade et al., 2014). Antigen-activated T cells induce cytolytic activity in NK cells, a process which is dependent on CD155 (Gross et al., 2016). NK cells derived from MS patients exhibited reduced cytolytic activity in response to antigen-activated T cells due to reduced expression of CD226 on NK cells, which increase the threshold for NK cell activation, at the same time as their CD4+ T cells failed to upregulate CD155 upon antigen activation (Gross et al., 2016). Treatment with daclizumab, a humanized anti-CD25 antibody used for the treatment of MS, resulted in increased expression of CD155 on CD4+ T cells and restored cytolytic activity of NK cells (Gross et al., 2016).
LAG-3LAG-3 (lymphocyte activation gene-3) is a co-inhibitory receptor that is expressed by activated T cells, NK cells, B cells, and plasmacytoid DCs (Anderson et al., 2016; Sierro et al., 2011). LAG-3 negatively regulates proliferation and activation of T cells, while being crucial for Treg suppression. LAG-3 deficient mice do not spontaneously develop autoimmunity, but display augmented disease in susceptible strains. A recent study showed that CD4+ intraepithelial lymphocytes (IELs) isolated from the gut epithelium of MOG35-55 TCR transgenic mice can inhibit EAE on transfer (Kadowaki et al., 2016). These cells express immune regulatory molecules such as LAG-3, CTLA-4, and TGF-β and can inhibit T cell proliferation in vitro by a mechanism dependent on these molecules. LAG-3 blocking antibodies reduced the severity of EAE induced by MOG35-55 TCR transgenic IELs. Interestingly, these regulatory IELs were influenced by stimuli from the gut environment, such as the microbiota and aryl hydrocarbon receptor (AHR) ligands in the diet providing a mechanism for how the gut microbiome can control extra-intestinal autoimmune diseases such as MS.
Tim-3
Tim-3 (T cell immunoglobulin and mucin domain 3) is an inhibitory receptor that triggers apoptosis upon interaction with its ligand, galectin-9 (Gal-9) (Anderson et al., 2016). Tim-3 is highly expressed by terminally activated Th1 cells, expressed at lower levels on Th17 cells, and is absent on naïve T cells and Th2 cells (Hastings et al., 2009). In addition, Tim-3 is constitutively expressed by DCs and microglia (Anderson et al., 2007). Tim-3 is an inhibitory molecule regulating peripheral tolerance and the expansion of effector Th1 cells, preventing uncontrolled inflammation. It can be protective in autoimmunity, but has been more extensively studied in cancer and chronic viral infections, where it contributes to the dampening of protective immunity (Anderson et al., 2016).
Tim-3 positive cells accumulate in the CNS of mice with EAE and peak at the onset of clinical symptoms (Anderson et al., 2007; Monney et al., 2002). Administration of an anti-Tim-3 antibody during EAE induction resulted in hyperacute disease with high numbers of activated macrophages, increased numbers of inflammatory foci in the CNS, and increased mortality (Monney et al., 2002) and blockade of Tim-3 signaling resulted in hyperproliferation of Th1 cells, increased release of IFN- and IL-2, and lack of antigen-specific tolerance (Sabatos et al., 2003). Along the same line, silencing of Gal-9 with siRNA exacerbated EAE while in vivo administration of Gal-9 resulted in selective loss of IFN- producing cells, attenuated Th1 responses, and ameliorated EAE (Zhu et al., 2005). More recent data suggest that interfering with the Tim-3/Gal-9 axis in CD4+ T cells changes the pattern of inflammation in the brain and spinal cord due to differential effects on Th1 vs Th17 cells (Lee and Goverman, 2013). As Tim-3 signaling predominantly controls the expansion of pathogenic IFN- secreting Th1 cells, blockade of Tim-3 in Th1-mediated EAE exacerbates disease severity, while blockade in Th17-mediated EAE selectively increases the survival of Th1 cells, changing the ratio of Th1/Th17 cells and resulting in decreased inflammation in the brain, but not in the spinal cord.
Evidence suggests that CD4+ T cells from MS patients have a reduced ability to upregulate Tim-3 upon stimulation (Koguchi et al., 2006; Yang et al., 2008). CD4+ T cell clones generated from the CSF of MS patients expressed less Tim-3 than clones from controls and this was associated with higher levels of IFN- production (Koguchi et al., 2006). Th1 polarization increased IFN- secretion in clones from MS patients, but this was associated with a relative inability to upregulate Tim-3. Similarly, Tim-3 blockade during ex vivo activation of CD4+ T cells from peripheral blood of controls enhanced IFN- secretion, but had no effect on CD4+ T cells from untreated MS patients (Yang et al., 2008). Interestingly, treatment of MS patients with the disease modifying drugs IFN-β or glatiramer acetate restored the expression and function of Tim-3 (Yang et al., 2008).
MS patients with a benign course have increased expression of Tim-3 and Gal-9 in MBP-stimulated CD4+ T cells isolated from peripheral blood compared to other forms of MS, resulting in an augmented death rate of MBP-specific CD4+ T cells, while PPMS patients had reduced expression of the same two molecules (Saresella et al., 2014). Furthermore, patients with benign MS had decreased expression of Bat3 (HLA-B-associated transcript 3), a repressor of Tim-3 that protects Th1 cells from Gal-9 mediated cell death and facilitates the development of EAE (Rangachari et al., 2012), in MBP-stimulated CD4+ T cells, while Bat3 expression was increased in PPMS (Saresella et al., 2014).
High CSF levels of Gal-9 have been observed in two independent cohorts of SPMS patients when compared to RRMS patients (Burman and Svenningsson, 2016). Gal-9 levels did not correlate with markers of adaptive immunity and it was suggested that they were driven by cells of the innate immune system within the CNS such as activated astrocytes in chronically active white matter lesions (Anderson et al., 2007).
ICOS
ICOS (Inducible T-cell costimulator) is a costimulatory molecule structurally and genetically related to CD28 and CTLA-4 (Wikenheiser and Stumhofer, 2016). It is not constitutively expressed on resting T cells, but is rapidly induced on CD4+ and CD8+ T cells following TCR cross-linking or CD28 activation. ICOS binds to the ICOS-ligand (ICOSL) that is expressed on APCs and also on non-lymphoid cells such as fibroblasts, endothelial cells, and some epithelial cells. Like CD28, ICOS has positive costimulatory activities, including enhanced cytokine production, up-regulation of CD40L expression, and providing help for Ig isotype class switching by B cells.
The role for ICOS signaling in EAE is complex. ICOS deficient mice develop a severe form of EAE, characterized by massive CNS infiltrates even when induced on a genetically resistant background (Dong et al., 2001). Blockade of the ICOS/ICOSL interaction during the priming phase (1–10 days p.i.) exacerbated EAE, while blockade during the effector phase (9–20 days p.i.) attenuated the disease (Rottman et al., 2001). Effector and memory T cells have different costimulatory requirements, and blockade of ICOSL reduced memory cell-induced EAE but worsened effector cell-mediated disease (Elyaman et al., 2008). ICOS plays an important role in inducing the production of IL-10, a key immunoregulatory cytokine involved in the induction of regulatory T type 1 (Tr1) cells, and it is likely that the effects of ICOS blockade in EAE are related to impairment of the generation or function of IL-10 producing regulatory cells (Greve et al., 2004; Rojo et al., 2008; Sporici et al., 2001). ICOSL is upregulated in tissues in response to inflammatory cytokines and the ICOS/ICOSL system may dampen immune activation in EAE and contribute to terminating the inflammation. This was further supported in studies of mucosal tolerance in EAE, where ICOS deficient mice were resistant to the induction of mucosal tolerance to myelin peptides (Miyamoto et al., 2005), a process where IL-10-producing Tr1 cells are of crucial importance (Weiner et al., 2011).
It is also possible that other cell types are affected by ICOS/ICOSL signaling. Evidence suggest that helminth-infected MS patients have reduced disease activity (Correale and Farez, 2007), which has been linked to the induction of IL-10 producing regulatory B cells through a mechanism mediated at least in part by the ICOS/ICOSL pathway (Correale et al., 2008). T cells expressing the invariant V19i chain inhibited both the induction and progression of EAE through reduced production of inflammatory cytokines and increased secretion of IL-10 by B cells, a process that could be blocked by anti-ICOSL antibodies (Croxford et al., 2006).
Recent studies in patients with MS have focused on ICOS+ TFH cells. TFH cells play a fundamental role in humoral immunity by providing help for germinal center (GC) formation, B cell differentiation into plasma cells, and antibody production in secondary lymphoid tissues (Fan et al., 2015b). RR and SPMS patients have increased frequencies of activated ICOS+ TFH cells in peripheral blood compared to healthy controls (Romme Christensen et al., 2013) and the frequencies of ICOS+CCR7+ memory TFH cells were further increased in MS patients examined during exacerbation (Fan et al., 2015a). The frequency of ICOS+ TFH cells correlated with disease progression in progressive patients and with number of plasmablasts, suggesting that ICOS+ TFH cells may play a role in B cell activation (Romme Christensen et al., 2013).
CD40
CD40 is a member of the tumor necrosis factor (TNF) receptor family that is constitutively expressed by APCs, including DCs, B cells, and macrophages, and that can be expressed by endothelial cells, smooth muscle cells, fibroblasts, and epithelial cells (Chatzigeorgiou et al., 2009). CD40 binds to one ligand, CD40L (CD154) that is mainly expressed on activated T cells, but other cells such as endothelial cells, microglia, and astrocytes can be induced to express CD40L. CD40 was originally identified as a costimulatory molecule important in T cell-mediated B cell activation and differentiation, but signaling through CD40 also stimulates cells to produce and secrete cytokines and chemokines, express adhesion molecules, costimulatory molecules, and various enzymes, all of which are involved in the local inflammatory processes taking place during chronic inflammation (reviewed by (Chatzigeorgiou et al., 2009)).
Both CD40 and CD40L can be detected in the CNS of mice with EAE and their expression correlate with disease activity and production of IFN- (Issazadeh et al., 1998). Mice without functional CD40/CD40L signaling due to genetic deletion of CD40L or treatment with antagonistic anti-CD40L antibodies during the priming stage are resistant to EAE, while treatment during the peak of disease or during remission reduces the severity of EAE (Becher et al., 2001; Gerritse et al., 1996; Grewal et al., 1996; Howard et al., 1999; Howard and Miller, 2001). This was initially believed to be due to an inability of APCs to upregulate the expression of CD80/CD86 and produce IL-12, which is necessary for Th1 differentiation during the priming phase (Constantinescu et al., 1999; Grewal et al., 1996). Later studies revealed that even though encephalitogenic T cells were generated in CD40L treated mice and were able to enter the CNS, the expansion and retention of these cells were impaired (Howard and Miller, 2001). Using bone-marrow chimeras, it was demonstrated that mice lacking CD40 expression on CNS resident microglia developed less severe EAE with reduced number of inflammatory cell infiltrates in the CNS (Becher et al., 2001; Ponomarev et al., 2006). This was associated with an inability of encephalitogenic T cells to fully activate microglia, blocking their ability to upregulate CD80/CD86 expression and to present CNS antigens to T cells (Ponomarev et al., 2006). Interestingly, CD40 deficient microglia were able to achieve an intermediary stage of activation with increased MHC class II expression, but in the absence of CD80/CD86 costimulation, these cells may trigger a tolerogenic signal resulting in down-modulation of CNS infiltration. A more recent study showed that MHC class II+CD40dimCD86dimIL-10+ microglia are potent inducers of antigen-specific CD4+Foxp3+ Tregs in vitro and that blocking CD40/CD40L signaling resulted in reduced numbers of CD25+Foxp3- effector cells, but did not affect the number of CD25+Foxp3+ Tregs (Ebner et al., 2013).
Both CD40 and CD40L immunoreactivity has been detected in close proximity to each other in perivascular infiltrates in active MS lesions (Gerritse et al., 1996). MS patients have higher frequencies of CD40L+ T cells in blood compared to controls, with levels decreasing during IFN-β treatment (Filion et al., 2003; Jensen et al., 2001; Teleshova et al., 2000). MS patients also have an increased frequency of a subset of CD4+ T cells expressing CD40 compared to healthy subjects and controls with non-autoimmune diseases (Waid et al., 2014). These cells displayed a CD45ROhi phenotype, recognized myelin peptides, and produced both IFN- and IL-17 consistent with an autoaggressive effector phenotype. PBMCs from SPMS, but not RRMS, patients produced more IL-12 and IFN- when restimulated in vitro, a process which was dependent on CD40/CD40L signaling (Balashov et al., 1997).
Although CD40 is of vital importance for B cell development, little is known about effects of CD40/CD40L blockade on the B cell compartment in EAE. B10 cells constitute a small subset of regulatory B cells that inhibit antigen-specific inflammatory reactions through IL-10 secretion (Yanaba et al., 2008). These cells are dependent on IL-21 and CD40 for their development and expansion and can inhibit EAE when transferred into mice with established autoimmune disease (Yoshizaki et al., 2012). Memory B cells from MS patients respond with enhanced proliferation when stimulated with low dose CD40L compared to healthy subjects (Ireland et al., 2014). This hyper-responsiveness to CD40 signaling in B cells was associated with dysregulation of the canonical NF-kB pathway in MS patients (Chen et al., 2016) and was normalized in patients treated with glatiramer acetate (Ireland et al., 2014). One study reported increased frequencies of CD40+ B cells in RRMS and patients with a clinically isolated syndrome representing very early MS, especially around the time of a relapse (Mathias et al., 2017), but another study reported decreased frequencies of CD40+ B cells in RRMS patients (Field et al., 2015).
CD40 -1C>T (rs1883832) is a functional SNP, which has been studied extensively in different MS populations. A recent meta-analysis indicated that the rs1883832 SNP is associated with increased risk of MS (Qin et al., 2017). Intriguingly, the risk allele for MS at rs1883832 is associated with reduced expression of CD40 (Jacobson et al., 2005) suggesting a protective effect of CD40 in MS highlighting the complexity of costimulatory pathways that can affect regulatory as well as effector cells.
