Critical role for Invariant chain in CD1d-mediated selection and maturation of Vα14-invariant NKT cells

Abstract

The development and maturation of Vα14 invariant (i)NKT cells in mice requires CD1d-mediated lipid antigen presentation in the thymus and the periphery. Cortical thymocytes mediate positive selection, while professional APCs are involved in thymic negative selection and in terminal maturation of iNKT cells in the periphery. CD1d requires entry in the endosomal pathway to allow antigen acquisition for assembly as lipid/CD1d complexes for display to iNKT cells. This process involves tyrosine-based sorting motifs in the CD1d cytoplasmic tail and invariant chain (Ii) that CD1d associates with in the endoplasmic reticulum. The function of Ii in iNKT cell thymic development and peripheral maturation had not been fully understood. Using mice deficient in Ii and the Ii-processing enzyme cathepsin S (catS), we addressed this question. Ii^−/− mice but not catS^−/− mice developed significantly fewer iNKT cells in thymus, that were less mature as measured by CD44 and NK1.1 expression. Ii^−/− mice but not catS^−/− mice developed fewer Vβ7⁺ cells in their iNKT TCR repertoire than WT counterparts, indicative of a change in endogenous glycolipid antigen/CD1d-mediated iNKT cell selection. Finally, using a Mycobacterium tuberculosis infection model in macrophages, we show that iNKT developed in Ii^−/− but not catS^−/− mice have defective effector function. Our data support a role for professional APCs expressing Ii, but no role for catS in the thymic development and peripheral terminal maturation of iNKT cells.

Introduction

The development of iNKT cells is a strictly thymus-dependent process that involves the presentation of antigenic glycolipid/CD1d complexes to developing T cells. The antigen presenting cells (APCs) that mediate the positive selection of iNKT cells are cortical thymocytes, while professional APCs (i.e., dendritic cells, DCs) are involved in negative selection and terminal maturation [1,2]. CD1d-mediated lipid antigen presentation by APCs is probably an important contributor to host defense against pathogens [3], and may have equally important roles in triggering iNKT cells under conditions of non-infectious cellular stress, such as in type I diabetes [4-6].

CD1d/lipid complexes are assembled in the endoplasmic reticulum (ER), composed of CD1d heavy chain, beta2-microglobulin (β2m) and endogenous lipids. After assembly, stable CD1d/lipid complexes are allowed to transit to the cell surface, followed by the rapid, within 30 minutes, introduction into the endocytic pathway, mediated by a tyrosine-based motif present in the CD1 cytosolic tail or through interaction with invariant chain (Ii)/Class II MHC complexes [7,8]. Most mouse CD1d protein is found in endosomal compartments under steady-state conditions with low levels of CD1d present at the cell surface. Lysosomal enzymes are involved in the processing and degradation of glycolipids for exchange and binding to CD1d [8-12], after which CD1d/lipid complexes traffic back to the plasma membrane for presentation. Cathepsin S (catS) is a lysosomal enzyme that cleaves the N-terminal segment of Ii. In absence of catS, the Class II-associated fragment Ii-p10 that contains the Class II MHC endosomal sorting signals is retained, resulting in the accumulation of Class II MHC complexes in endosomal compartments [13-15]. As a consequence, build up of Ii/Class II MHC complexes causes the enlargement of endosomal compartments, possibly affecting the loading of CD1d with antigenic lipid cargo.

The involvement of the Class II MHC chaperone Ii in CD1d sorting had been suggested to allow for iNKT cell development, selection and maturation, but is not entirely clear. Mice that lack catS appeared to have decreased numbers of thymic iNKT cells [16], a finding that was not supported by others [9,17] and was not investigated further. In professional APCs, CD1d transport to the endosomal pathway requires association with Ii [4,7,8,18], supporting the possibility that Ii mediates CD1d endosomal transport if professional APCs play a role in thymic development of iNKT cells. Similarly, there is data to suggest a role for Ii in peripheral maturation of iNKT cells [4,7,8,18], but the effector function of iNKT cells matured in the absence of Ii has not been evaluated.

We set out to resolve the role of Ii and catS in CD1d-mediated selection and development of iNKT cells. By using Ii^−/− and catS^−/− mice, we show a role for Ii but not catS in the thymic selection and peripheral terminal differentiation of CD1d-restricted iNKT cells in spleen. Ii^−/− but not catS^−/− mice exhibit a Vβ-bias in the TCR repertoire of thymic iNKT cells. As readout for iNKT cell effector function, we exploited a model of Mycobacterium tuberculosis (Mtb) infected macrophages (Mϕ), for which we earlier showed that iNKT cells are both necessary and sufficient to suppress Mtb replication [19]. Addition of catS^−/− or WT splenocytes to Mtb-infected Mϕ resulted in a reduction in bacterial CFU. Splenocytes from Ii^−/− mice, however, did not suppress intracellular Mtb replication in infected Mϕ. Only WT splenocytes were able to produce interferon-γ (IFN-γ) and tumor necrosis factor (TNF) upon stimulation by Mtb-infected WT Mϕ. Furthermore, we tested iNKT cell effector function by stimulating iNKT cells with αGalactosylceramide (αGalCer), a potent stimulant of CD1d-restricted iNKT cells [20,21]. Both catS^−/− and Ii^−/− iNKT cells were able to produce IFN-γ but produced less TNF. It is well established that cortical thymocytes mediate iNKT cell positive selection [2,22-24], while professional APCs mediate thymic negative selection [1] and peripheral terminal maturation of iNKT cells [2]. We show here that Ii, but not catS, expression in APCs facilitates the generation of functional CD1d-restricted iNKT cells.

1. Materials and Methods

1.1. Mice

We used six- to eight- week old mice, all on C57Bl/6 background. WT mice were acquired from The Jackson Laboratory (Bar Harbor, ME, USA). Cathepsin S knockout (catS^−/−) mice [25], CD1d-knockout (CD1d^−/−) mice [26] and Ii-knockout (Ii^−/−) mice [27] were described previously. Mice were bred and maintained in a barrier facility and studies were performed according to institutional guidelines for animal use and care.

1.2. Antibodies and reagents

The National Institute of Allergy and Infectious Disease MHC Tetramer Core facility (Atlanta) kindly provided the CD1d:PBS-57 (αgalactosylceramide, αGalCer analog) tetramers. Dr G. Besra generated αGalCer for culture experiments. The following antibodies were used in this study: rat anti-mouse CD1d (1B1, Biolegend), rat anti-mouse CD4 (RM4-5, BD Pharmingen), rat anti-mouse CD8 (53-6.7, BD Pharmingen), Armenian Hamster anti-mouse CD11c (HL3, BD Pharmingen), rat anti-mouse CD44 (IM7, eBioscience), Armenian hamster anti-mouse CD69 (H1.2F3, Biolegend), rat anti-mouse B220 (RA3-6B2, Biolegend), rat anti-mouse F4/80, NK1.1 (PK136, Biolegend), rat anti-mouse Vβ7 (TR310, BD Pharmingen), rat anti-mouse Vβ8 (KJ16, Serotec) and Armenian Hamster anti-mouse TCRβ (H57-597, Biolegend).

1.3. Vα14 iNKT cell enrichment from the liver

Mouse livers were perfused with PBS and homogenized trough a stainless steel mesh. The liver cells were spun for 5 min at 2000 rpm and taken up in 9 ml of PBS, mixed with 5ml of 80% Percoll. This cell/PBS/Percoll suspension was layered over 3 ml 80% Percoll and the gradient was spun for 20 min at 2000 rpm. Liver lymphocytes accumulating at the interface were collected and washed with PBS after which red blood cells were lysed using ammonium chloride.

1.4. Flow cytometry

Single-cell suspensions from thymus, spleen and liver were resuspended in FACS buffer (2% FCS, 2 mM sodium azide and anti-FcγRII/III Ab (2.4G2) in PBS) and incubated for 20 min with the relevant conjugated mAb on ice. Mtb-infected cells were fixed with 1% paraformaldehyde overnight before analysis. At least 100,000 cells per sample were analyzed on a FACScanto (Becton Dickinson) using FlowJo software.

1.5. In vitro culture of peritoneal Mϕ

Mice received intraperitoneal injections with sterile 3% thioglycolate medium. After 4 days elicited peritoneal cells were harvested by intraperitoneal lavage as described earlier [19]. Mϕ, purified by positive selection with CD11b-microbeads (Miltenyi Biotech) were >95% pure (F4/80⁺ CD11b⁺), as determined by flow cytometry (data not shown). 5×10⁵ purified Mϕ were cultured in 24-well flat-bottom plates (or 1×10⁵ in 96-well plates) in complete RPMI 1640 medium (Invitrogen Life Technologies) supplemented with 10% fetal calf serum (HyClone), penicillin/streptomycin, L-glutamine, sodium-pyruvate, 2-ME, nonessential amino acids, essential amino acids, and HEPES buffer (all from Gibco).

1.6. In vitro bacteria infections

Virulent Mtb (H37Rv) were grown to mid-log phase in Middlebrook 7H9 medium containing 10% albumin/dextrose/catalase enrichment (BD Biosciences) under shaking conditions, 37°C / 5% CO₂. After opsonization for 5 min using RPMI 1640 medium with 2% human serum (Gemini Bio-Products), 10% FBS, and 0.05% Tween 80, mycobacteria were washed twice with complete medium without antibiotics. After passage through a 5-μm syringe filter (Millipore), mycobacteria were counted in a Petroff-Hausser chamber and added to purified Mϕ at a multiplicity of infection (MOI) of 10:1. After 2hrs, unbound Mtb were removed by extensive washing with complete RPMI medium without antibiotics. Infected Mϕ were cultured overnight before the addition of splenocytes (see below) and the next day all wells were washed twice with RPMI medium without antibiotics.

1.7. Splenocytes

Spleens were aseptically removed and mechanically homogenized with a 3-ml syringe plunger. Erythrocytes were lysed with RBC lysis buffer (1 mM KHCO3, 0.15 M NaCl and 0.1 mM sodium-EDTA, pH 7.3). After washing, cell viability was determined using trypan blue. Splenocytes, as a source of iNKT cells, were resuspended in complete RPMI 1640 medium without antibiotics and 2.5×10⁶ splenocytes/well were added to cultures of Mtb-infected Mϕ in 24-well plates (5:1, or 5×10⁵ in 96-well plates).

1.8. Determination of colony forming units

Mycobacterial growth was quantified 24 hrs after Mϕ infection (day 1) and 72 hr after co-culture with or without naïve splenocytes (day 4). Cells were lysed by removing the culture supernatant and adding distilled water containing 1% Triton X-100 for 5 min. 10-fold dilutions, made in PBS containing 0.02% Tween-80, were plated on Middlebrook 7H11 agar plates. Three weeks after incubation at 37°C in a humidified CO₂ atmosphere, the numbers of colonies on plates were counted.

1.9. iNKT stimulation assays

CD11b⁺ MACS-purified thioglycolate-induced Mϕ (1×10⁵ in 96-well plates) were stimulated for 4 h with 3-30 ng/ml αGalCer. After two washes, 1×10⁵ freshly isolated and enriched (B220⁻, CD11b⁻ and CD11c⁻) spleen iNKT cells were added to the culture for 48 h. Enrichment was performed by negative depletion of B220⁺, CD11b⁺ and CD11c⁺ cells with magnetic beads using autoMACs. After 48 h supernatants were harvested for cytokine detection.

1.10. Cytokine detection

Culture supernatants were assayed for IFN-γ and TNF by standard sandwich ELISA. ELISAs were done in accordance with the manufacturer’s instructions (eBiosciences). Samples were read at 405 nm on SoftMax Pro ELISA analysis software (Molecular Devices).

1.11. Statistical analysis

Data are shown as mean ± standard error of the mean (SEM). Unpaired two-tailed t-test was used to compare two groups. To compare multiple different groups to their control, we used one-way ANOVA. A p-value of at least 0.05 was considered statistically significant. Analysis was performed using Prism 4.0 for Mac software (GraphPad Software, Inc).

2. Results

2.1. Reduction in thymic development of iNKT cells in Ii^−/− mice

Mice deficient in Ii and catS exhibit modifications in the endosomal pathway in professional APCs: Ii^−/− cells exhibit defects in endosomal fusion and resulting in small Class II MHC-positive endosomal compartments, while catS^−/− cells are defective in endosomal degradation of Ii remnants resulting in enlarged endosomal compartments that lack multivesicular structures [14,15]. As CD1d localization to specialized endosomal compartments is required for the loading of antigenic lipids onto CD1d, the absence of Ii or catS may affect CD1d-driven iNKT cell development. Multiple studies had attempted to address a role for Ii or catS in iNKT cell biology, mostly by staining for iNKT cell numbers or use of iNKT cell hybridomas in culture experiments [7,9,16,17]. Our own approach was to analyze the phenotype and function of primary iNKT cells developed in the complete absence of Ii or catS. We first confirmed the number of iNKT cells in WT, Ii^−/− and catS^−/− mice using αGalCer-loaded CD1d-tetramers. CD1d^−/− mice served as controls that do not support the development of CD1d-restricted iNKT cells. We found that lack of Ii or catS does not significantly alter the percentage or absolute number of iNKT cells in spleen and liver (Fig. 1). The percentage of thymic iNKT cells in Ii^−/− mice was reduced compared to WT (* p<0.05). However, this trend was not recapitulated in absolute iNKT cell numbers, perhaps as a consequence of the biased thymocyte repertoire in Ii^−/− mice, lacking most Class II MHC-restricted CD4 T cells [27,29].

Fig. 1. Defective iNKT cell development in Ii^−/− mice.

Representative flow cytometry dot-plots of thymus, spleen and liver are shown for different mouse genotypes. iNKT cells are detected by using anti-TCRβ antibody and a-galactosylceramide (αGalCer)-loaded CD1d-tetramers. Absolute numbers of iNKT cells in thymus, spleen and liver were calculated and displayed in scatter plots. While thymic iNKT cell percentages in Ii^−/− mice are slightly decreased, catS^−/− mice exhibit normal iNKT cell percentages in thymus. Average percentages ± SEM are summarized in each dot-plot (n=6-7). Shown is one representative experiment out of three. Statistical testing was done using a one-way ANOVA with Dunnett’s post-test comparing each experimental group to WT.

2.2. Reduction in thymic and peripheral maturation of iNKT cells in Ii^−/− mice

While Ii-deficient mice develop slightly fewer iNKT cells, there is still a sizeable population present in thymus. In Ii^−/− spleen and liver, normal iNKT cell percentages and numbers are present, probably because of homeostatic CD1d-independent expansion [28] and supported by the lymphopenic CD4 T cell-deficient environment in Ii^−/− mice [27,29]. We asked whether CD1d-restricted iNKT cells that develop in the absence of Ii or catS have a distinction in maturation phenotype. To this end, we stained cells from thymus, spleen and liver with αGalCer-loaded CD1d-tetramers and antibodies against TCRβ and the maturation markers CD44, CD69 and NK1.1. Thymic iNKT cell development can be divided into three successive stages: stage 1: CD44-NK1.1-, stage 2: CD44⁺NK1.1-, and mature stage 3: CD44⁺NK1.1⁺CD69⁺, as described in previous studies [4,30-34]. The transition in development from immature NK1.1⁻ to mature NK1.1⁺ Vα14 iNKT cells is an important step during the terminal differentiation of iNKT cells, which continues outside the thymus and requires the presence of CD1d [2,35,36]. There are slightly lower percentages of iNKT cells expressing CD69 in spleen from catS^−/− mice (Fig. 2A) However, the maturation of thymic and liver iNKT cells from catS^−/− mice, based on expression of CD69 is similar to WT (Fig. 2A). The maturation of iNKT cells from catS^−/− mice, based on expression of NK1.1 is similar to WT (Fig. 2B). The iNKT cells from Ii^−/− mice are delayed in maturation with significantly lower percentages expressing CD69 and NK1.1 during stage 3 of maturation in the thymus (** p<0.01). Correspondingly, we found elevated percentages of iNKT cells in stage 2 (CD44⁺ NK1.1⁻) in Ii^−/− thymus compared to WT (** p<0.01). However, peripheral maturation, measured in the spleen and liver, is only marginally, if at all decreased (* p<0.05 in spleen). We observed that in Ii^−/− liver, more iNKT cells than in WT liver express a CD44⁺NK1.1⁺ phenotype. This possibly reflects local activation of liver-resident iNKT cells, rather than increased migration of mature iNKT cells, since absolute iNKT cell numbers are similar in Ii^−/− and WT livers (Fig. 1). Thus, although Ii may influence iNKT cell differentiation from stage 2 (CD44⁺NK1.1⁻) to the mature stage 3 (CD44⁺NK1.1⁺CD69⁺) in the thymus, this delay in iNKT cell maturation is partially overcome in the periphery. We have previously reported similar kinetics for iNKT cell development in mice in which thymic CD1d endosomal trafficking was delayed due to presence of a cytosolic tag [4].

Fig. 2. Defective iNKT cell maturation in Ii^−/− mice.

Cells from thymus, spleen and liver from different mouse genotypes were stained with αGalCer-loaded CD1d-tetramers and antibodies against TCRβ, CD44, CD69 (A) and NK1.1 (B) and analyzed by flow cytometry. Dots in the scatter-plots represent individual mice. Representative flow cytometry dot-plots displaying iNKT cell maturation based on CD44 and NK1.1 expression are shown for the different mouse genotypes. iNKT cells from Ii^−/− mice are delayed in maturation with significantly lower percentages expressing CD69 and NK1.1 in the thymus and spleen (** p<0.01 and * p<0.05, respectively), but not in liver. iNKT cells from catS^−/− mice display lower percentages expressing CD69 in the spleen (* p<0.05), but not in thymus or liver, nor for NK1.1. (C) Skewing of TCR Vβ repertoire in Ii^−/− mice. Cells from thymus, spleen and liver, from different mice were stained with αGalCer-loaded CD1d-tetramers in combination with antibodies against TCRβ, Vβ7 and Vβ8, and analyzed by flow cytometry. Bars represent mean percentages ± standard deviation of Vβ7 (black, left axis) and Vβ8 (white, right axis) positive iNKT cells (5≤n≤7). Ii^−/− (** p<0.01) mice display significantly less Vβ7 positive iNKT cells in the thymus, but not in the periphery compared to catS^−/− and WT mice. Shown are results from one representative experiment out of three. Statistical testing was done using a one-way ANOVA with Dunnett’s post-test comparing each experimental group to WT.

2.3. Bias in TCR Vβ chain usage of iNKT cells in Ii^−/− mice

We considered that a change in intracellular sorting could expose CD1d to endogenous glycolipid antigens that were not normally encountered. This raised the possibility that such a change in binding of endogenous glycolipid antigens by CD1d would alter iNKT cell selection. The selection of Vβ TCR domains assembled with the invariant Vα14 chain shapes the avidity of TCRs on iNKT cells for endogenous ligand/CD1d complexes, to which developing iNKT cells are exposed during thymic selection [28]. A change in Vβ usage would suggest a role for Ii and catS-expressing professional APCs in thymic selection of iNKT cells. We found that Ii^−/− but not catS^−/− mice display significantly less Vβ7 positive iNKT cells in the thymus compared to catS^−/− and WT mice (p<0.01), but normal contribution of Vβ7 iNKT cells in spleen and liver (Fig. 2C). Thus, thymic selection of iNKT cells by lipid/CD1d complexes appears influenced by Ii, but not catS-expressing professional APCs.

2.4. No suppression of M. tuberculosis replication by iNKT cells from Ii^−/− mice

We earlier showed that CD1d-restricted iNKT cells are able to suppress intracellular Mtb replication [19]. We asked whether iNKT cells that developed in the complete absence of Ii or catS could suppress the replication of Mtb in Mϕ. Confirming previous findings [19], we found that splenocytes from CD1d^−/− mice, which lack iNKT cells, are unable to restrict intracellular Mtb replication compared to WT splenocytes (Fig. 3A). Although to a lesser extent the addition of catS^−/− splenocytes to Mtb-infected WT Mϕ still led to a significant reduction in bacterial CFU after 4 days (*** p<0.001), compared to the addition of WT splenocytes (*** p<0.001) (Fig. 3A). However, Ii^−/− splenocytes did not suppress intracellular Mtb replication in infected WT Mϕ, and the growth of Mtb was not statistically different compared to Mϕ that were cultured in the absence of splenocytes (Fig. 3A).

Fig. 3. Naïve splenocytes from Ii^−/− mice are functionally impaired and are unable to suppress M. tuberculosis replication in WT macrophages.

Splenocytes (spl) from CD1d^−/− or invariant chain knockout (Ii^−/−) mice are unable to limit M. tuberculosis (Mtb) replication (A) and fail to produce cytokines upon stimulation (B). Mtb-infected WT Mϕ were cultured alone or in the presence of splenocytes from uninfected WT, CD1d^−/−, Ii^−/− or catS^−/− mice. Splenocytes were added to the Mtb-infected WT Mϕ 1 day post-infection. (A) Growth of Mtb in the infected Mϕ was assessed by determining CFU on day 1 and day 4 post-infection. Addition of WT or catS^−/− splenocytes (*** p<0.001) but not CD1d^−/− or Ii^−/− splenocytes (NS) led to a significant reduction in bacterial CFU after 4 days. Statistical testing was done using a one-way ANOVA with Dunnett’s post-test comparing each experimental group to WT Mtb-infected Mϕ cultured alone for 4 days (d4). Shown is one representative experiment out of three. Dots in the scatter-plots represent individual CFU counts of replicate cultures (n = 3-6, *** p<0.001). (B) Day 4 IFN-γ and TNF production by spleen iNKT cells stimulated by Mtb-infected WT Mϕ was assessed by ELISA. Ii^−/− and catS^−/− iNKT cells fail to produce IFN-γ and TNF upon stimulation with Mtb-infected WT Mϕ. Statistical testing was done using one-way ANOVA with Dunnett’s post-test comparing each experimental group to WT per condition. WT Mϕ + mutant splenocytes and mutant splenocytes alone function as negative controls. (C) iNKT cells from Ii^−/− and catS^−/− mice display altered cytokine production upon stimulation with αGalCer by Mϕ. WT peritoneal Mϕ were cocultured with purified spleen iNKT cells from WT, CD1d^−/−, Ii^−/− or catS^−/− mice, in the presence of different amounts of αGalCer. WT Mϕ + mutant iNKT and mutant iNKT alone were included as negative controls. Results displayed in bar graphs are representative of two experiments. Shown are the mean of triplicate values with SEM. Ii^−/− and catS^−/− iNKT cells were capable of producing normal levels of IFN-γ, but produced significantly lower amounts of TNF upon activation with αGalCer (*** p<0.001 and ** p<0.01). Statistical testing was done using one-way ANOVA with Dunnett’s post-test comparing each experimental group to WT per condition.

2.5. Impaired cytokine production by iNKT cells from catS^−/− and Ii^−/− mice

Mϕ activation is an important part of the host defense against Mtb mediated by IFN-γ [37] and TNF [38], which are both readily produced by activated iNKT cells. We determined cytokine production by spleen iNKT cells upon stimulation by Mtb-infected Mϕ in vitro. The addition of WT splenocytes, but not catS^−/− or Ii^−/− splenocytes, resulted in significant production of TNF and IFN-γ (Fig. 3B). We did not detect any significant levels of IL-4 in the in vitro Mtb infection assays (data not shown). We additionally tested iNKT cell activation by stimulating purified spleen iNKT cells with αGalCer presented by WT Mϕ. We did not find any significant difference in IFNγ production by Ii^−/− or catS^−/− iNKT cells compared to WT iNKT cells. However, we observed significantly lower levels of TNF produced by iNKT cells from mice deficient for Ii or catS (Fig. 3C). Given the bias in immature iNKT cells (CD44⁺NK1.1⁻) in Ii−/− thymus and spleen and the demonstrated bias of immature iNKT cells towards T-helper 2 (T_H2) phenotype [32], an increase in IL4 production upon activation was expected. However, we did not observe any significant IL4 production in these assays (data not shown).

2.6. Inability of naïve purified iNKT cells from Ii^−/− mice to suppress M. tuberculosis replication in WT macrophages

Finally, we addressed directly whether purified iNKT cells from WT and Ii−/− spleens can suppress Mtb replication in infected WT Mϕ. Mtb-infected WT Mϕ were cultured alone or in the presence of purified iNKT cells from uninfected WT or Ii^−/− mice. We added purified iNKT cells to the Mtb-infected WT Mϕ at 1 day post-infection and determined growth of Mtb in infected Mϕ by determining CFU on day 1 and day 4 post-infection. Addition of WT iNKT cells (*** p<0.001) but not Ii^−/− splenocytes (NS) led to a significant reduction in bacterial CFU after 4 days. Thus, iNKT cells developed in the complete absence of Ii are not only deficient in cytokine production, but are defective at inhibition of Mtb replication.

Discussion

Invariant (i)NKT cells are a unique subset of T lymphocytes that use a conserved TCR that recognizes both self and foreign glycolipids when presented in complex with the antigen presenting molecule CD1d. It is now well established that endosomal localization of CD1 molecules including mouse CD1d, is important for the presentation of glycolipid antigen/CD1 complexes at the cell surface [4,39-41]. From the endoplasmic reticulum, where CD1d heavy chain is assembled into stable complexes with beta-2-microglobulin and endogenous lipid, CD1d complexes are transported via the trans-Golgi network to the cell surface. CD1d complexes are rapidly thereafter introduced into endosomal compartments through association with the Class II MHC-associated invariant chain (Ii), and by use of a tyrosine-based endosomal sorting motif present in the cytoplasmic tail of CD1d [4,7,8]. Exchange of self-lipids for antigenic lipid in CD1d complexes occurs in specialized compartments to which endocytosed CD1d has access, and that contain necessary co-factors including lysosomal lipid-transfer proteins such as saposins and microsomal triglyceride transfer protein [12,42,43]. CD1d, through its ability to bind selective glycolipid antigens from endosomal compartments, controls the development, thymic selection and maturation as well as the peripheral maturation and function of iNKT cells.

The development of iNKT cell precursors separates from mainstream thymocytes at the CD4⁺CD8⁺ thymocyte stage, and involves the random rearrangement and selection of the canonical Vα14 TCR. Expression of this Vα14-invariant TCR allows for the positive selection of these CD4⁺CD8⁺ thymocytes on neighboring CD1d-expressing thymocyte counterparts [2,22-24], cells that do not express Ii. The periphery is seeded primarily with immature iNKT cells that undergo further maturation in the periphery [30,34]. The terminal differentiation of iNKT cells involves antigen presentation by professional APCs, but the necessity for Ii in these cells to develop fully functional iNKT cells, able to fight pathogen infection was not yet known.

iNKT cells themselves do not express Ii or the Ii proteolytic enzyme catS, but rely on CD1d-mediated lipid antigen presentation, which may be controlled by function of Ii or catS. Multiple lines of research suggest a role for Ii in CD1d endosomal transport. First, Ii controls endosomal architecture and transport of endosomal content [14,44,45]. Second, Ii can promote the adherence between Ii molecules embedded in distinct vesicular membranes and thereby promoting endosomal docking and fusion [46,47]. Third, Ii alone or Ii/Class II MHC complexes can mediate entry of CD1d in late endosomal compartments: mouse CD1d can be co-immunoprecipitated with Ii [7] and human CD1d associates with Ii/Class II MHC complexes in the endoplasmic reticulum and late endosomal compartments [18]. Finally, concerning CD1d specifically, Ii can introduce CD1d into endosomal compartments [4,7]. CatS has also been implicated in CD1d endosomal processes. CatS proteolysis is critical for endosomal antigen loading of Class II MHC complexes by cleavage of Class II-associated Ii-p10 to CLIP, and was implicated in CD1d-mediated antigen presentation [16]. However, a study from the same year did not confirm this data [17]. Endosomal architecture in catS^−/− APCs is changed dramatically, causing enlarged compartments devoid of multivesicular structures [14]. Either morphological changes or a direct proteolytic defect in catS^−/− APCs may conceivably cause aberrant glycolipid loading of CD1d. A study carefully delineating CD1d-mediated iNKT cell thymic and peripheral development and function in Ii^−/− or catS^−/− mice has not been reported.

In this study, we combined detailed phenotype analysis of iNKT cells with study of effector function with standard iNKT cell activation assays involving presentation of αGalCer as well as in an established Mycobacterium tuberculosis macrophage infection model [19]. iNKT cells are frequently activated early during infection and have an immunomodulatory role that can either exacerbate or ameliorate the outcome of infection (reviewed in [48]; [49]). While the absence of CD1d-restricted T cells does not worsen the outcome to tuberculosis infection in the mouse model, specific activation of iNKT cells with αGalCer, significantly prolongs the survival of infected mice [21,50]. Other data suggests that iNKT cells make a contribution to anti-mycobacterial immunity early during infection [19,51]. One possible explanation for why CD1d^−/− mice, which lack the iNKT cells, are not more susceptible to Mtb than WT mice is that the contribution of iNKT cells to anti-mycobacterial immunity is redundant. However, iNKT cells obtained from uninfected mice are unique in their ability to recognize Mtb-infected macrophages in vitro and suppress mycobacterial replication [19]. The capacity of iNKT cells to inhibit Mtb growth requires CD1d expression by the infected macrophage, although whether a self/endogenous or microbial antigen is recognized, is unknown. We have used this well characterized model to determine whether the alteration of iNKT cell numbers in Ii-deficient mice results in a functional alteration [9,16].

Our work complements earlier studies using Ii^−/− or catS^−/− APCs co-cultures involving iNKT cell hybridomas and use of model antigens [7,9,16]. Our data shows that Ii but not catS may be involved in thymic development of fully functional iNKT cells. iNKT cells developed in absence of Ii were slightly decreased in numbers and exhibited a bias in TCR Vβ repertoire coupled to the TCR Vα14, suggesting a change in endogenous glycolipid/CD1d-mediated selection. Ii^−/− mice moreover displayed increased percentages of iNKT cells in the immature stage 2 (CD44⁺NK1.1⁻) at the expense of mature 3 (CD44⁺NK1.1⁺) iNKT cells. It has been demonstrated that the Tec family tyrosine kinase Itk is necessary for optimal NK1.1 up-regulation and hence final maturation of NKT cells [32]. Future experiments may verify any convergence of Tec family kinase Itk and Ii⁺ APCs in directing iNKT cell maturation. iNKT cells from Ii^−/− mice were unable to restrict Mycobacterium tuberculosis replication in infected macrophages. In addition, they failed to produce IFN-γ or TNF, cytokines which have been shown to play a role in the host response against Mtb [37,38]. Furthermore, despite increased percentages of immature (CD44⁻NK1.1⁻) iNKT cells in the spleen, we did not observe significant IL4 production by Ii^−/− iNKT cells. CatS^−/− mice, however, developed iNKT cells in normal ratios and were still able to suppress Mtb replication despite impaired production of IFN-γ and TNF, suggesting a non-critical role for catS in iNKT cell development and effector function. We conclude that Ii-expressing professional APCs support thymic development of iNKT cells. It is well established that iNKT cells positive selection occurs by CD4⁺CD8⁺ double positive cortical thymocytes [2,22,23]. Targeted expression of CD1d limited to professional APCs did not support the development of iNKT cells [24]. However, CD1d-mediated presentation by professional APCs is important for iNKT cell development. Experiments in which human CD1d was expressed on thymic CD11c⁺ APCs suggested a role for professional APCs in the negative selection of thymic iNKT cells [1]. In the periphery, professional APCs mediate the terminal maturation of newly formed iNKT cells, for acquisition of full effector function [2]. Our experiments described here establish a role for Ii during thymic development of iNKT cells, which is independent of catS. Moreover, we show an essential requirement for Ii in the periphery to iNKT cell acquisition of full effector function, in the ability to fight Mtb infection.

Fig. 4. Naïve purified iNKT cells from Ii^−/− mice are unable to suppress M. tuberculosis replication in WT macrophages.

Purified iNKT cells from invariant chain knockout (Ii^−/−) mice are unable to limit M. tuberculosis (Mtb) replication. Mtb-infected WT Mϕ were cultured alone or in the presence of purified iNKT cells from uninfected WT or Ii^−/− mice. Enrichment of iNKT cells was performed by negative depletion of B220⁺, CD11b⁺ and CD11c⁺ cells with magnetic beads using autoMACs. Purified iNKT cells were added to the Mtb-infected WT Mϕ 1 day post-infection. Growth of Mtb in the infected Mϕ was assessed by determining CFU on day 1 and day 4 post-infection. Addition of WT iNKT cells (*** p<0.001) but not Ii^−/− splenocytes (NS) led to a significant reduction in bacterial CFU after 4 days. Statistical testing was done using unpaired two-tailed t-test comparing the Ii^−/− group to WT Mtb-infected Mϕ cultured alone for 4 days (d4). Dots in the scatter-plots represent individual CFU counts of replicate cultures (n = 3, *** p<0.001).

Vitae

Fenna C. M. Sillé received her combined B.A. and M.Sc. in biology in 2004 from the University of Groningen, The Netherlands. She performed her PhD thesis research in the Brigham and Women’s Hospital in Boston under supervision of Marianne Boes. In 2010 she received her Ph.D. in immunology from the Utrecht University, The Netherlands.

Constance Martin received her B.S. in Biology from Case Western Reserve University in Cleveland, OH in 2006. She is currently a Ph.D. student in the Department of Immunology & Infectious Diseases at the Harvard School of Public Health, Boston MA.

Pushpa Jayaraman received her Ph. D. in Pathobiology in 2006 from the University of Washington. She is currently a post-doctoral fellow working with Sam Behar, Brigham and Women’s Hospital, Boston on the role of Tim3 mediated immunoregulation of innate and adaptive immunity towards Mycobacterium tuberculosis. She is a recipient of post-doctoral research training fellowship from American Lung Association (RT-123085-N).

Alissa Rothchild received her B.S. in biology in 2005 from Brown University and is currently a PhD graduate student in the immunology program in the Division of Medical Science at Harvard University.

Samuel M. Behar is an Associate Professor in Rheumatology, Immunology, and Allergy at Brigham and Women’s Hospital, Harvard Medical School, and in the Department of Immunology and Infectious Diseases at the Harvard School of Public Health.

Marianne Boes was Assistant Professor in Dermatology at Brigham and Women’s Hospital during part of these studies, and is currently Associate Professor in Immunology at the University Medical Centre in Utrecht, the Netherlands.