SEARCH

SEARCH BY CITATION

Keywords:

  • autoimmunity;
  • CD1;
  • MAIT;
  • NK T cells;
  • therapy

Summary

  1. Top of page
  2. Summary
  3. Natural killer T cells
  4. NK T cells in autoimmunity
  5. NK T cells as therapeutic agents in autoimmunity
  6. Conclusion and perspectives
  7. Acknowledgements
  8. Disclosure
  9. References

T cells reactive to lipids and restricted by major histocompatibility complex (MHC) class I-like molecules represent more than 15% of all lymphocytes in human blood. This heterogeneous population of innate cells includes the invariant natural killer T cells (iNK T), type II NK T cells, CD1a,b,c-restricted T cells and mucosal-associated invariant T (MAIT) cells. These populations are implicated in cancer, infection and autoimmunity. In this review, we focus on the role of these cells in autoimmunity. We summarize data obtained in humans and preclinical models of autoimmune diseases such as primary biliary cirrhosis, type 1 diabetes, multiple sclerosis, systemic lupus erythematosus, rheumatoid arthritis, psoriasis and atherosclerosis. We also discuss the promise of NK T cell manipulations: restoration of function, specific activation, depletion and the relevance of these treatments to human autoimmune diseases.


Natural killer T cells

  1. Top of page
  2. Summary
  3. Natural killer T cells
  4. NK T cells in autoimmunity
  5. NK T cells as therapeutic agents in autoimmunity
  6. Conclusion and perspectives
  7. Acknowledgements
  8. Disclosure
  9. References

Natural killer T (NK T) cells were first described in the 1990s. These cells were characterized as a subset of T cells that share some characteristics with innate NK cells. NK T cells are present in mice, humans and other mammalian species [1]. Classically, NK T cells are divided into three subsets: type I, or invariant NK (iNK) T, type II NK T and NK T-like cells (Fig. 1). NK T cells represent a heterogeneous class of cells restricted by major histocompatibility complex (MHC) class I-like molecules such as CD1a,b,c,d, and MR1. These non-polymorphic molecules present non-protein antigens such as glycolipids and induce NK T cell activation [2]. NK T cells modulate immune responses by producing large amounts of cytokines and by the expression of various surface molecules. NK T cells influence the development of innate and adaptive immune responses. It is essential to understand more clearly the role of each NK T cell subset in the protection or exacerbation of various pathologies, and to determine if they can be manipulated therapeutically in autoimmune diseases.

image

Figure 1. Natural killer (NK T) cell populations. NK T cells can be divided into three groups: invariant NK (iNK) T (blue background), type II NK T (yellow background) and NK T-like cells (pink background). Each group is composed of distinct subsets.

Download figure to PowerPoint

Type I NK T or iNK T cells

Type I NK T cells, or iNK T cells, express an invariant T cell receptor α chain (TCR-α), Vα14-Jα18 in mice and Vα24-Jα18 in humans, and are associated with a limited set of TCR-β chains (Vβ2, 7 or 8·3 in mice and Vβ11 in humans). This T cell subset recognizes glycolipids presented by the MHC class I-like molecule, CD1d. iNK T cells specifically recognize the glycolipid α-galactosylceramide (α-GalCer) presented by CD1d [3,4]. As shown in Fig. 1, iNK T cells can be divided into distinct CD4+, CD4CD8 double-negative (DN) or CD8+ (in humans only) subsets [5]. Not all NK T cells express the NK1·1 (CD161 in humans) marker [1]. In humans, CD4+ iNK T cells produce T helper type 1 (Th1) and Th2 cytokines and CD4 iNK T cells produce primarily Th1 cytokines. This dichotomy is not observed in mice [6,7]; however, functional subsets have been identified: iNK  T NK1·1[8], iNK T interleukin (IL)-17+ (iNK T17) [9,10] and iNK T IL-17RB+[11]. CD4, usually considered as a co-receptor for binding to MHC class II, is thought to interact with CD1d, thereby potentiating iNK T cell activation [12]. A new subset of iNK T cells has been described recently. These cells are reactive to α-GalCer, express the TCR-α chain Vα10-Jα50, the NK1·1 marker and secrete interleukin (IL)-4, IL-10, IL-13, IL-17 and interferon (IFN)-γ after TCR activation [13].

Type II NK T cells

Type II NK T cells express a more diverse TCR-α chain repertoire (such as Vα3·2-Jα7/9 or Vα1-Jα7/9 in mice), a limited TCR-β chain (such as Vβ8 in mice) [14], and are present in humans [15]. Like type I NK T cells, these cells are CD1d-restricted. However, they do not recognize α-GalCer, but instead recognize other antigens such as sulphatide [16], lysophosphatidylcholine [17] or non-lipid small molecules [18]. A subset of γδ T cells, expressing TCR Vγ4 in mice, is restricted to CD1d, but their antigen specificity has not been identified [19].

NK T-like cells

Mucosal-associated invariant T (MAIT) cells express an invariant TCR-α chain (Vα19-Jα33 in mice and Vα7·2-Jα33 in humans) and are restricted to the non-polymorphic MHC class I-like MR1 molecule [20]. In humans, a monoclonal antibody allows the specific detection of MAIT cells, which are primarily CD8+ (but may be DN), express CD161 and secrete tumour necrosis factor (TNF)-α, IFN-γ and IL-17 [21]. The nature of the antigen(s) presented by MR1 remains to be determined. One study suggested that synthetic α-mannosyl ceramide derivatives activate MAIT cells [22]. However, a subsequent study did not confirm this original observation [23].

CD1a, CD1b and CD1c MHC class I-like molecules present lipid antigens [24]. These molecules, well defined in humans, are absent in mice. In human blood, 10% of T cells are restricted to these molecules (2% are reactive to CD1a, 1% to CD1b and 7% to CD1c) [25–27]. These T cells express αβ TCR [27] or γδ TCR [28], but their role in autoimmunity remains unknown, as no cell-type specific markers were available until recently [27,29].

NK T cells in autoimmunity

  1. Top of page
  2. Summary
  3. Natural killer T cells
  4. NK T cells in autoimmunity
  5. NK T cells as therapeutic agents in autoimmunity
  6. Conclusion and perspectives
  7. Acknowledgements
  8. Disclosure
  9. References

There are two primary phases in the development of autoimmune disease: the initiation phase and the chronic phase (Fig. 2). Because of the difficulty in determining the cause of tolerance breakdown in these pathologies, studies on the role of NK T cells in autoimmune disease initiation are limited. One interesting study suggests that iNK T cells are key players in the initiation of primary biliary cirrhosis (PBC). PBC is a chronic lethal autoimmune disease characterized by the destruction of small intrahepatic bile ducts by autoreactive T cells. In a PBC mouse model, iNK T activation upon infection by Novosphingobium aromaticivorans initiates liver injury [30]. Unfortunately, it will be difficult to interfere in established PBC by modulating iNK T cell function because, at the time of diagnosis, iNK T cells are no longer required. For other autoimmune diseases, the contribution of NK T cells may be due to defective immunoregulation by NK T cells or inappropriate NK T cell activation (Fig. 2; Table 1).

image

Figure 2. Implication of natural killer (NK) T cells in human autoimmune diseases. The development of autoimmune diseases (blue arrow) can be divided into an initial and chronic phase. In primary biliary cirrhosis, invariant NK (iNK) T cells play a key role in the initial phase, whereas in other autoimmune diseases NK T cells can be involved at different phases of pathogenesis (e.g. psoriasis or multiple sclerosis). While some autoimmune diseases are associated with a defective pool of NK T cells (e.g. multiple sclerosis, rheumatoid arthritis, systemic lupus erythematosus or type 1 diabetes), others are associated with inappropriate activation (e.g. psoriasis, atherosclerosis).

Download figure to PowerPoint

Table 1. Role of invariant natural killer (iNK) T, type II NK T and NK T-like cells in autoimmune diseases.Thumbnail image of

Diseases linked to a defective pool of NK T cells

A functionally defective pool of NK T cells has been described in several autoimmune diseases, such as multiple sclerosis (MS), systemic lupus erythematosus (SLE), rheumatoid arthritis (RA), type 1 diabetes (T1D), Crohn's disease, Graves' disease and Sjögren syndrome [31,32].

MS.  MS is characterized by neurological symptoms, including muscle spasms, muscle weakness and difficulty of movement. In MS, autoreactive T cells induce damage in the myelin sheath around the axons of the brain and spinal cord. In experimental autoimmune encephalomyelitis (EAE), a mouse model of MS, iNK T cells infiltrate the central nervous system (CNS). Mice devoid of iNK T cells (Jα18-deficient mice) develop a more severe EAE than control mice [33]. We have shown that increasing the number of iNK T cells protects mice from EAE by inhibiting Th1 and Th17 autoimmune responses [34,35]. This protection is independent of CD1d [35]. Recently, another group showed that iNK T cells, producing IL-4 or IL-10, inhibit Th1 responses and reduce EAE severity [33]. In the blood of MS patients, total iNK T cell frequency is decreased [31,36]. Under remission, CD4+ iNK T cells secrete large amounts of IL-4 that could favour a Th2 bias, suggesting a beneficial role of this subset [36]. In contrast to mouse models, iNK T cells have not been detected in human CNS lesions [37].

An increased number of type II NK T cells are observed in the CNS during EAE, and treatment of mice with sulphatide prevents development of the disease [16]. Increasing the number of MAIT cells (Vα19 TCR transgenic mice) protects mice against the induction and progression of EAE. Mice devoid of MAIT cells (MR1-deficient mice) present an exacerbated form of EAE. In Vα19 transgenic mice, as well as in wild-type mice subjected to adoptive transfer with MAIT cells, these cells modulate EAE severity by reducing the production of inflammatory cytokines and enhancing B cell IL-10 secretion in an inducible T cell co-stimulatory (ICOS)-B7RP-1 manner [38]. Polymerase chain reaction (PCR) analysis suggests that MAIT cells accumulate in human CNS [39]. More recently, flow cytometry analysis shows that MS patients harbour a lower frequency of MAIT cells in blood compared to healthy controls. The authors observed a positive correlation between clinical recovery and increase in MAIT cell frequency and that MAIT cells suppress IFN-γ production by T cells in vitro in a contact-dependent manner [40].

CD1b-reactive T cells are more frequent in the blood of MS patients than in healthy individuals. These cells respond to several glycolipids from the CNS and release IFN-γ and TNF-α[41]. Their role, as well as the role of CNS self-lipids (e.g. ganglioside, sulphatide) in NK T cell activation, remains to be investigated [42].

SLE.  SLE is characterized by a range of symptoms, including arthritis, facial rash, pleuritis, pericarditis and photosensitivity. Inappropriate activation of autoreactive T cells and autoantibody production cause acute and chronic inflammation of various tissues such as skin, kidney, joints and the nervous system. Two SLE mouse models (MRL-lpr and SLE pristane-induced) exhibit a reduced number of iNK T cells at disease onset in secondary lymphoid organs [43,44]. However, New Zealand black/white (NZB/W) F1 mice do not have a defect in NK T cell frequency and iNK T cells are hyperactive, as indicated by cytokine production (IFN-γ and IL-4) [45]. Treatment of 3-month-old (NZB/W) F1 mice with anti-CD1d blocking antibodies decreases disease severity, wherein iNK T cells interact with B cells to promote production of autoantibodies [45,46]. Paradoxically, CD1d-deficient (NZB/W) F1 mice develop an exacerbated disease [47], similar to CD1d-deficient MRL-lpr[48]. The regulatory role of iNK T cells on B cell activation has also been described in another SLE mouse model. Injection of apoptotic cells induces autoreactive B cell activation and production of anti-DNA immunoglobulin (Ig)G in C57BL/6 mice. Autoimmune responses are increased in CD1d- and Jα18-deficient mice, which present immune complex deposition in the kidneys. CD1d expression on B cells is required for their suppression by iNK T cells [49]. These observations suggest that in the early phase of SLE development iNK T cells are protective, but promote autoantibody production later.

iNK T cell numbers decrease in the blood of SLE patients compared to healthy controls [31,50]. The reduced numbers affect DN, CD4+ and CD8+ subsets [50]. In addition, iNK T cells from SLE patients are functionally defective [50,51]. iNK T cell default is associated with a defect of lipid antigen presentation by immature B cells from SLE patients [50].

CD1c-restricted T cell lines derived from SLE patients are more activated than cells from healthy individuals. These cells provide help to B cells in secreting pathogenic IgG antibodies [52], suggesting a pathogenic role.

RA.  RA is characterized by joint deformity and loss of movement. RA autoantibodies and autoreactive T cells induce chronic inflammation in the synovial membrane of the joint. There is no evidence of a decrease in iNK T cell numbers in a collagen-induced arthritis (CIA) RA mouse model [53]. On the contrary, mice devoid of iNK T cells (Jα18-deficient mice) present an attenuated form of RA [54,55]. Recently, it has been shown that iNK T cells are activated in early-stage CIA and anti-CD1d blocking antibody treatment improves the clinical signs of arthritis [56]. The pathogenic mechanism of iNK T cells is unclear. One report demonstrated that antibodies activate iNK T cells directly through FcγRIII in an antibody-induced arthritis mouse model [55]. In that model, iNK T cells inhibit TGF-β production and promote arthritis by producing IL-4 and IFN-γ[57]. In contrast to mouse models, low numbers of circulating iNK T cells (particularly the DN subset) have been described in RA patients [51,58–60]. iNK T cells were detected in the synovium of patients and are biased towards Th0-like cytokine profiles upon α-GalCer activation [61]. Interestingly, in RA patients treated with anti-CD20, iNK T cell numbers increased, suggesting a beneficial role for these cells [60].

Another recent study showed that, in mice, an immunodominant peptide of mouse collagen presented by CD1d activates type II NK T cells, which inhibits the development of CIA [62]. However, additional studies are still needed to characterize NK T cells more clearly in mice and humans.

Mice devoid of MAIT cells (MR1-deficient mice) develop a milder disease than control mice, suggesting that MAIT cells promote inflammation and exacerbate RA. However, the mechanism remains unknown [63].

T1D.  T1D is characterized by hyperglycaemia, polyuria, polydipsia, polyphagia and weight loss, and is lethal in the absence of insulin treatment. T1D is a chronic autoimmune disease in which insulin-secreting pancreatic β cells are destroyed selectively. It is thought to be a Th1-mediated disease with involvement of CD8+ T cells and macrophages. Several mouse model studies provide a converging picture of a protective role for iNK T cells in T1D [3]. iNK T cell numbers are reduced in young non-obese diabetic (NOD) mice [64,65], and increasing their number by adoptive transfer [66,67] or via the introduction of a Vα14-Jα18 transgene inhibits development of T1D [66]. Moreover, CD1d deficiency exacerbates diabetes in NOD mice [68]. Early reports suggest that iNK T cell protection is associated with induction of a Th2 response to islet autoantigens [69–72]. However, experiments based on the transfer of monoclonal anti-islet T cells showed that iNK T cells inhibit differentiation of autoreactive T cells into effector cells during their priming in pancreatic lymph nodes [73,74]. Defective priming of autoreactive T cells could reflect the ability of iNK T cells to promote recruitment of tolerogenic dendritic cells [74,75]. We described recently the functional dichotomy between CD4+ and DN iNK T cell subsets in the regulation of T1D. While CD4+ iNK T cells strongly protect NOD mice against diabetes, DN iNK T cells (containing iNK T17 cells) increase diabetes incidence. Importantly, exacerbation of diabetes by DN iNK T cells is abrogated by treating with an anti-IL-17 blocking antibody [76]. Interestingly, NOD mice contain a higher frequency of iNK T17 cells and fewer CD4+ iNK T cells compared to non-autoimmune C57BL/6 mice [76,77].

Contrary to autoimmune diseases cited previously, there is no clear evidence of a role for iNK T cells in T1D aetiology. PCR analysis found a lower frequency of DN iNK T cells in diabetic blood compared to discordant diabetic twins without the disease [78]. However, flow cytometry analysis showed similar Vα24+ CD1d-tetramer+ iNK T cell frequency in discordant monozygotic twins [79]. Other studies reporting either low or high iNK T cell numbers in the blood of diabetic patients have been published [80,81]. Of note, analysis of several mouse strains showed that iNK T cell frequency in blood is not correlated with their frequency in lymphoid tissues [82]. Functional studies show that iNK T clones from pancreatic lymph nodes of diabetic patients exhibit defective IL-4 production [83].

Type II NK T cells inhibit diabetes progression in NOD mice. Diabetes protection was observed in Vα3·2-Vβ9 TCR transgenic NOD mice harbouring elevated numbers of type II NK T cells as well as by adoptive transfer [84]. These type II NK T cells dampen the diabetogenic T cell response through regulatory mechanisms involving programmed cell death ligand 1 (PD-L1) and ICOS molecules [85].

Similarly, increasing the number of MAIT cells via the introduction of a Vα19-Jα33 TCR transgene in NOD mice reduces T1D onset significantly [86]. However, the mechanism by which MAIT cells prevent diabetes, as well as the role of these cells in T1D patients, remains to be elucidated.

Diseases linked to an inappropriate activation of NK T cells

Psoriasis.  Psoriasis is characterized by the presence of red dry plaques on the skin. In psoriasis, innate and autoreactive T cells induce inflammation through TNF-α production, leading to abnormal proliferation of skin cells. In mouse models, NK T cells infiltrate the psoriatic plaques [87,88]. Similarly, human studies revealed an increased iNK T cell number, particularly of the CD4+ subset, in psoriatic lesions compared to healthy skin [87,89]. Furthermore, CD1d expression is higher in keratinocytes from psoriasis patients, and NK T cells co-cultured with keratinocytes from psoriasis patients produce IFN-γ[87]. Together, these mouse and human data suggest the involvement of NK T cells in psoriatic skin lesions.

Atherosclerosis.  Atherosclerosis is involved in the development of cardiovascular diseases and exhibits aspects of autoimmune disease, including the presence of autoantibodies and autoreactive T cells against heat shock protein 60 (HSP60) [90]. In atherosclerosis, accumulation of immune cells and lipid particles in blood vessels leads to narrowing of the arterial lumen and causes thrombosis. Mouse models have shown the pro-atherogenic effect of iNK T cells [91–93]. ApoE-CD1d double-deficient mice exhibit a 25% decrease in lesion size [93]. CD4+ iNK T cells appear to be responsible for the proatherogenic activity of iNK T cells due to production of more proinflammatory cytokines (IL-2, TNF-α, IFN-γ) than DN iNK T cells [94]. In humans, iNK T cells are present and CD1d expression is enhanced in atherosclerotic plaques [95–97]. As observed in the mouse model, CD4+ iNK T cells infiltrated human atherosclerotic lesions. Infiltrating iNK T cells secrete large amounts of IL-8, a chemoattractant for immune cells [97]. Furthermore, enhanced CD1a,b,c expression in macrophages was observed in atherosclerotic plaques compared to healthy controls [98]. These observations suggest a role for NK T-like cells in atherosclerosis.

Human NK T cell deficiency and autoimmunity

Genetic defects affecting lymphocyte signalling pathways (e.g. ITK, XIAP, SH2D1A), lipid transfer and processing proteins (e.g. MTP, NCP2) are associated with dysfunction of iNK T cells (reduced/absence of function and number) in humans. Patients with these disorders seem more susceptible to selective viral infections (e.g. Epstein–Barr virus), but do not present with autoimmune disorders. It is possible that development of autoimmune disorders is hampered by the fact that patients affected by these mutations experience a shortened life expectancy, wherein treatments such as stem cell transplantation may be performed to ameliorate symptoms [99].

NK T cells as therapeutic agents in autoimmunity

  1. Top of page
  2. Summary
  3. Natural killer T cells
  4. NK T cells in autoimmunity
  5. NK T cells as therapeutic agents in autoimmunity
  6. Conclusion and perspectives
  7. Acknowledgements
  8. Disclosure
  9. References

Harnessing of iNK T cell using specific ligands

Several autoimmune diseases exhibit a defective pool or function of NK T cells. During the past 10 years, many molecules have been tested for their ability to activate iNK T cells.

α-GalCer treatment.  The glycolipid α-GalCer stimulates iNK T cells in mice and humans. Recognition of the CD1d–αGalCer complex by the semi-invariant TCR of iNK T cells results in rapid production of cytokines. Single or repeated injections of α-GalCer in mice give different outcomes. A single injection of α-GalCer induces IL-12 production and CD40 up-regulation by dendritic cells (DC) [100] and CD40L up-regulation on iNK T cells. The interaction between these two cell types induces a strong secretion of IFN-γ and IL-4 by iNK T cells and DC maturation [101]. This cross-talk leads to activation of NK cells (through IFN-γ produced by iNK T cells) and conventional CD4 and CD8 T cells (through mature DCs) [102]. On the contrary, repeated α-GalCer injections biased DC maturation towards a tolerogenic phenotype in an IL-10 dependent manner [103]. Furthermore, iNK T cells become unable to produce IFN-γ and IL-17 but their IL-4 production, although weaker, persists [76]. Both mechanisms probably contribute to inhibition of pathogenic autoreactive T cell responses. Therefore, repeated α-GalCer treatments may represent an attractive strategy for preventing autoimmune diseases as treatment in mice is protective against EAE [104,105], SLE [106], RA [53,107] and T1D [70–72,108]. However, depending on the timing and frequency of injections, age and sex of the mice, α-GalCer may exacerbate some autoimmune diseases [57,105,109,110]. α-GalCer could also be deleterious in the context of atherosclerosis [92,93], allergic reaction [111] and asthma [112]. IL-4 secretion by iNK T cells during repeated α-GalCer treatment could promote the development of asthma through IgE induction and eosinophil recruitment [111–113], although the precise role of iNK T cells in asthma remains controversial [114]. Together, these data suggest that α-GalCer treatment might not be the most appropriate to prevent autoimmune diseases. In this regard, other iNK T cell agonists have been generated and tested in mouse models.

α-GalCer analogues: a perspective.  Structural modifications of α-GalCer influence the iNK T cytokine secretion profile towards Th1 or Th2 [115]. The analogue OCH skews T cell responses towards Th2 through the production of IL-4 by iNK T cells, and a single OCH injection inhibits EAE [116]. This protective effect has been confirmed in other autoimmune diseases such as CIA [117], T1D [108] and colitis [118] in mice. Another Th2-biased analogue, C20:2, protects NOD mice against diabetes. The C20:2 molecule seems to favour the generation of tolerogenic DCs and inhibits IL-12 production by DCs [70,119]. However, as OCH and C20:2 skew T cell responses towards a Th2 profile, these molecules could promote the development of asthma. A new analogue, C16:0, that induces only moderate IFN-γ and IL-4 production by iNK T cells, is more efficient than α-GalCer in preventing T1D in NOD mice [120]. Because C16:0 induces very little IL-4 production, it may be a good candidate for a T1D clinical trials. It would be interesting to evaluate further the efficacy of C16:0 in other T1D models, such as virus-induced diabetes and other autoimmune diseases. Moreover, it would be important to determine the ability of C16:0 to reverse an established disease.

NK T cell agonists in clinical trials.  Phase I cancer clinical trials revealed that soluble α-GalCer treatment is safe, but exerts moderate immunostimulatory effects [91–93]. This difference between humans versus mice might reflect the lower frequency of iNK T cells in humans. iNK T cells represent 0·2–0·5% of blood lymphocytes in mice versus 0·01–1% in humans [94] and 30% of liver lymphocytes in mouse versus 1% in humans. This lower frequency in humans suggests that α-GalCer analogue therapy might be less efficient in humans than in mice. Because iNK T cell numbers are quite variable in humans, individuals with a higher iNK T cell number should be favoured for iNK T cell-specific therapy. However, it will be important to investigate the expansion ability of iNK T cells from individuals exhibiting different iNK T cell frequencies. Moreover, it will be important to analyse iNK T cell subsets in patients before and during iNK T cell therapy to determine the effect of iNK T cell analogues on different subsets. Further investigation is required on type II NK T cells and MAIT cells before using them for therapeutic purposes. Furthermore, the interplay between type I, type II and NK  T-like cells during glycolipid treatment remains poorly characterized. Interestingly, researchers have noted the activation of type II NK T cells by sulphatide-induced anergy in type I NK T cells in a mouse model of inflammatory liver disease [121].

Restoration of iNK T cell numbers

In vitro iNK T cell expansion.  Increasing the number of iNK T cells by adoptive transfer reduces significantly the progression of autoimmune diseases in mouse models [66,67]. In humans, a Phase I clinical trial showed that injection of in vitro expanded iNK T cells is safe and well tolerated [122]. This strategy could have the advantage of expanding and selecting defined subsets of iNK T cells (e.g. CD4+ iNK T cells in MS).

Enhanced self-ligand presentation.  The role of self-antigen(s) and the mechanisms triggering NK T cell activation in autoimmune diseases remain unknown. Mouse NK T cell clones can be activated by endogenous tumour cell ligands [123]. Microbial infections enhance the expression of glucosylceramide synthase, leading to the synthesis of β-glucosylceramide (β-GlcCer). This self-glycolipid presented by CD1d activates iNK T cells and induces their proliferation [124]. iNK T cell function may be promoted by enhancing the expression of glucosylceramide synthase (or other enzymes) that increases presentation of self-glycolipids capable of activating iNK T cells.

Interestingly, IFN-β treatment ameliorates the disease in MS patients [125]. This treatment increases the frequency and enhances the function of iNK T cells (IL-4, IL-5 and IFN-γ production) in the blood of MS patients. This iNK T cell modulation is mediated by DCs that up-regulate CD1d and CD40 expression [126].

Therapeutic approach for autoimmune diseases associated with inappropriate NK T cell activation

Diseases with inappropriate NK T cell activation (e.g. psoriasis or atherosclerosis) are characterized by elevated CD1d expression in lesions. Antibodies against CD1d have been developed, and could be used to block NK T cell activation. However, anti-CD1d antibodies added to human PBMC cultures induce IL-12 production by DC [127]. Therefore, such antibodies might not be effective for inhibiting the development of autoimmune diseases. Another approach could be the depletion of NK T cells by using specific antibodies. Recently, a monoclonal antibody recognizing human iNK T cells has been generated [128], and could be modified to induce depletion, rather than activation.

Conclusion and perspectives

  1. Top of page
  2. Summary
  3. Natural killer T cells
  4. NK T cells in autoimmunity
  5. NK T cells as therapeutic agents in autoimmunity
  6. Conclusion and perspectives
  7. Acknowledgements
  8. Disclosure
  9. References

Studies in patients and animal models of autoimmune diseases describe different roles for NK T and iNK T cell subsets. For example, CD4+ iNK T cells prevent T1D in NOD mice, whereas iNK T17 cells aggravate the disease. In MS patients under remission, CD4+ iNK T cells secrete large amounts of IL-4, suggesting a beneficial role of these cells. In contrast, CD4+ iNK T cells infiltrate lesions in psoriasis and atherosclerosis, and might be pathogenic. All these data suggest that protection or exacerbation of autoimmune diseases by iNK T cells may be due to disequilibrium between the different subsets (Fig. 3). As highlighted in previous reviews [129,130], most studies used methods that do not identify iNK T cells clearly (e.g. Vα24 PCR, TCR+ NK1·1+ or CD56+ CD3+ staining). CD8+ iNK T cells, representing 20% of human iNK T cells in blood, were rarely analysed and most of the studies focused on CD4+ or DN cells. Similarly, the only cytokines produced by NK T cells were IL-4 and IFN-γ, and only a few studies explored the secretion of IL-2, IL-5, IL-13, IL-17, granulocyte–macrophage colony-stimulating factor (GM-CSF), TGF-β or chemokines. It is important to note that iNK T cell number is higher in mice than in humans, whereas type II NK T [131] and MAIT cells [132] are more abundant in humans than in mice. CD1a,b,c-restricted T cells are present in humans, but not mice, due to deletion of CD1 genes in mice, which suggests that CD1d-restricted NK T cells might compensate for these cell populations. Humanized mice expressing these molecules have been generated [133], and it would be interesting to cross them into genetic backgrounds susceptible to autoimmune disease. Future human studies should focus on less characterized innate T cells. Similarly, the role of self-ligands, cytokine environment and accessory molecules (e.g. NKG2D) required for NK T cell activation should be investigated further [130,134,135]. More extensive research must be performed on specific tissues. For example, the characterization of NK T cells in the pancreas of type 1 diabetic patients remains to be investigated. New mouse models of autoimmune diseases would be useful, in particular, to understand apparent discrepancies in the role of NK T cells between mouse models and human diseases [136]. For example, in mouse models of RA, iNK T cells have a deleterious role, whereas in human RA, iNK T cells seem to exhibit a protective role. It would be interesting to analyse the behaviour of human iNK T cells in humanized mice reconstituted with human stem cells.

image

Figure 3. Hypothesis regarding the effect of various natural killer (NK) T subsets in autoimmunity. NK T cell subsets exert different roles in autoimmune diseases. For example, in non-obese diabetic (NOD) mice developing type 1 diabetes CD4+ invariant NK (iNK) T cells are protective, whereas iNK T17 cells enhance disease incidence. Therefore, protection or exacerbation of autoimmune diseases by NK T cells could be due to disequilibrium between different cell subsets.

Download figure to PowerPoint

Acknowledgements

  1. Top of page
  2. Summary
  3. Natural killer T cells
  4. NK T cells in autoimmunity
  5. NK T cells as therapeutic agents in autoimmunity
  6. Conclusion and perspectives
  7. Acknowledgements
  8. Disclosure
  9. References

Y.S and L.G. are supported by doctoral fellowships from the Region Ile-de-France (CODDIM) and from Paris 5 University, respectively. A.L. is supported by funds from INSERM, CNRS, ANR-09-GENO-023, ANR-10-MIDI-010 and Laboratoire d'Excellence INFLAMEX. A. Lehuen is recipient of an APHP-CNRS Contrat Hospitalier de Recherche Translationelle.

References

  1. Top of page
  2. Summary
  3. Natural killer T cells
  4. NK T cells in autoimmunity
  5. NK T cells as therapeutic agents in autoimmunity
  6. Conclusion and perspectives
  7. Acknowledgements
  8. Disclosure
  9. References