Targeted delivery of lipid antigen to macrophages via the CD169/sialoadhesin endocytic pathway induces robust invariant natural killer T cell activation

Abstract

Invariant natural killer T (iNKT) cells induce a protective immune response triggered by foreign glycolipid antigens bound to CD1d on antigen-presenting cells (APCs). A limitation of using glycolipid antigens to stimulate immune responses in human patients has been the inability to target them to the most effective APCs. Recent studies have implicated phagocytic CD169⁺ macrophages as major APCs in lymph nodes for priming iNKT cells in mice immunized with glycolipid antigen in particulate form. CD169 is known as sialoadhesin (Sn), a macrophage-specific adhesion and endocytic receptor of the siglec family that recognizes sialic acid containing glycans as ligands. We have recently developed liposomes decorated with glycan ligands for CD169/Sn suitable for targeted delivery to macrophages via CD169/Sn-mediated endocytosis. Here we show that targeted delivery of a lipid antigen to CD169⁺ macrophages in vivo results in robust iNKT cell activation in liver and spleen using nanogram amounts of antigen. Activation of iNKT cells is abrogated in Cd169^−/− mice and is macrophage-dependent, demonstrating that targeting CD169⁺ macrophages is sufficient for systemic activation of iNKT cells. When pulsed with targeted liposomes, human monocyte–derived dendritic cells expressing CD169/Sn activated human iNKT cells, demonstrating the conservation of the CD169/Sn endocytic pathway capable of presenting lipid antigens to iNKT cells.

Invariant natural killer T (iNKT) cells play an important role in the innate arm of the immune system to modulate the subsequent acquired immune responses. They are activated by antigen-presenting cells (APCs) displaying glycolipid antigens loaded on the antigen-presenting molecule CD1d (1). Glycolipid antigens recognized by iNKT cells can be captured by phagocytic APCs and ultimately loaded onto CD1d in lysosomes (2). Although all APCs expressing CD1d have the potential to present glycolipid antigens, studies involving depletion of specific types of APCs, and/or cell type–specific deletion of the Cd1d allele, have revealed that APC subtypes have differing abilities to prime iNKT cells with lipid antigens (3, 4).

Recent studies have documented CD169⁺ macrophages as a versatile APC for T cells (5, 6). CD169⁺ macrophages are found in various tissues including spleen, liver, and lymph nodes (7–9). This subset has been implicated in the activation of iNKT cells in lymph nodes of mice injected with glycolipid antigen alpha-galactosylceramide (α-GalCer)-coated particles (6). However, although iNKT cells are also robustly activated in liver and spleen, the specific role of CD169⁺ macrophages in the activation of iNKT cells in these tissues has not been documented because antigen-coated particles in these tissues are taken up by CD169 negative phagocytic APCs (10).

In addition to being a marker of a subset of macrophages, CD169 is also a member of the siglec family called sialoadhesin (Sn), an adhesion and endocytic receptor that recognizes sialic acid containing glycans as ligands (11, 12). CD169/Sn binds and internalizes the sialylated virus and bacteria, suggesting that CD169/Sn may serve as a receptor for sialylated pathogens (12, 13). CD169/Sn follows the clathrin-mediated endocytosis and it constitutively recycles between the cell surface and endosomes (14, 15). We and others have investigated the potential for targeting CD169⁺ macrophages using glycan ligand–decorated liposomes or antibodies as targeting agents to deliver cargo specifically into these cells via the endocytosis of the siglec receptor (14–16). Thus, CD169/Sn is capable of carrying cargo into CD169⁺ macrophages by an endocytic mechanism distinct from the phagocytic pathway used for the uptake of lipid-coated particles studied previously.

In this report, we investigated the ability of CD169⁺ macrophages to induce systemic activation of iNKT cells. We used a high-affinity glycan ligand–bearing liposomes to selectively deliver lipid antigens via the CD169/Sn endocytic pathway. We found that ligand-targeted liposomes are captured by CD169⁺ macrophages and potently prime iNKT cells in liver and spleen. These effects occur in a CD169/Sn-dependent manner, because no activation is seen with the targeted liposomes in CD169-deficient mice. Thus, we conclude that iNKT cells can be efficiently activated by targeting macrophages via the CD169/Sn endocytic pathway.

Results

Generation of CD169/Sn-Specific Liposomes That Deliver α-GalCer to CD169⁺ Macrophages.

To assess the involvement of CD169⁺ macrophages in the presentation of lipid antigens to iNKT cells, we formulated α-GalCer into CD169/Sn-targeted liposomes that display a high-affinity glycan ligand of CD169/Sn (Fig. 1A). The two ligands shown, 9-N-biphenylcarbonyl ^(BPC)NeuAc and 9-N-(4H-thieno[3,2-c]chromene-2-carbonyl) ^(TCC)NeuAc, differ in their substituent at the nine-position of sialic acid, with the latter exhibiting the highest selectivity for CD169/Sn over other mouse and human siglecs (15, 16). Although both ligands were used interchangeably in in vitro experiments, only ^TCCNeuAc was used for in vivo experiments. Alexa647-labeled ^TCCNeuAc liposomes specifically targeted CD169⁺ macrophages in the spleen upon i.v. injection into WT mice (Fig. 1B). Using Abs on various cell surface markers, we identified CD169⁺ macrophages in the spleen by flow cytometry (Fig. S1 A and B). In contrast to previous reports using lipid-coated particles (10), there was no significant uptake of the ^TCCNeuAc liposomes by other phagocytic APCs including CD169⁻ macrophages and dendritic cells (DCs). Nontargeted (naked) liposomes did not bind to any cells. In vitro analysis of the uptake of Alexa647-labeled ^TCCNeuAc liposomes by bone marrow–derived macrophages (BMMs) showed that the targeted, but not naked, liposomes were internalized by BMMs in a CD169/Sn-dependent manner and delivered to the lysosomes (Fig. S2), the compartment documented to load lipid antigens onto CD1d (2). Furthermore, CD169⁺ macrophages express CD1d, suggesting that they may present lipid antigens to iNKT cells in the spleen (Fig. S1C). We also found that most of CD169⁺ macrophages in spleen are CD11c⁺ (Fig. S1C), which could explain the unexpected deletion of these macrophages in the CD11c-diphtheria toxin receptor transgenic mice upon the toxin injection (17).

Fig. 1.

Targeting CD169⁺ macrophages with glycan ligand decorated liposomes. (A) Targeted liposomes display glycan ligands of CD169/Sn (^BPCNeuAc or ^TCCNeuAc) linked to PEGylated lipids. Targeted liposomes with α-GalCer activate iNKT cells via CD169/Sn-mediated endocytosis. (B) B6 mice were injected i.v. with Alexa647-labeled liposomes or buffer alone. After 1 h, splenocytes were analyzed by flow cytometry. Each cell population defined by the gating strategy in Fig. S1A was assessed for Alexa 647 staining: targeted liposomes (^TCCNeuAc; gray, filled), nontargeted liposomes (naked; broken line), or buffer control (solid line). Data are representative of three independent experiments with similar results.

CD169/Sn-Targeted Liposomes with Lipid Antigen Robustly Activate iNKT Cells.

To test if lipid antigen delivery to macrophages through CD169/Sn-mediated endocytosis leads to activation of iNKT cells, we i.v. injected the ^TCCNeuAc liposomes containing 2 ng α-GalCer into mice and measured the cytokine production in iNKT cells in the liver and spleen. As shown in Fig. 2 A and B, after 1.5 h, robust IFN-γ and IL-4 production by iNKT cells was observed in the WT mice injected with ^TCCNeuAc liposomal α-GalCer. We observed that iNKT cells in the liver are more activated than those in the spleen. This may be simply because liver has the largest number of CD169 expressing macrophages (9, 18), rather than the difference in accessibility of iNKT cells to the APCs in these organs, because both in the liver and spleen, iNKT cells have favorable access to APCs that filtrate lipid antigens from blood stream (10, 19). Activation was dependent on CD169/Sn because iNKT cells were not primed in Cd169^−/− mice by ^TCCNeuAc liposomal α-GalCer. Also, there was little or no activation of iNKT cells with naked liposomes loaded with α-GalCer or free α-GalCer when the same amount of glycolipid was used as in the liposomes (2 ng per mouse) (Fig. 2 B and C). These data demonstrate the highly efficient activation of iNKT cells by delivering α-GalCer to CD169⁺ macrophages. The lack of response in Cd169^−/− mice was not the result of a diminished capability to prime iNKT cells because Cd169^−/− mice showed equivalent CD1d expression on macrophages (Fig. S1D) and produced equivalent cytokine responses to that of WT mice in response to naked liposomes with a much higher dose of α-GalCer (50 ng) (Fig. 2D). As further controls, there was no iNKT activation with ^TCCNeuAc liposomes without α-GalCer (Fig. 2E), and anti-CD1d blocking antibody inhibited the activation (Fig. S3), demonstrating that iNKT cell antigen–dependent activation requires delivery of lipid antigen through the CD169/Sn endocytic pathway.

Fig. 2.

CD169/Sn-targeted liposomes with α-GalCer activate iNKT cells in vivo. (A) WT and Cd169^−/− mice were injected i.v. with indicated liposomes (2 ng α-GalCer per mouse) or buffer only (control). After 1.5 h, liver lymphocytes and splenocytes were analyzed by flow cytometry for intracellular production of IFN-γ and IL-4 in iNKT cells. Intracellular staining of iNKT cells (B220⁻TCRβ⁺CD1dtetramer⁺) in the liver and spleen is shown. (B) Quantification of iNKT cell activation. The percentage of IFN-γ⁺IL-4⁺iNKT cells in the liver and spleen of three mice is shown. (C) Targeting of CD169/Sn endocytic pathway with ^TCCNeuAc liposomal α-GalCer enhances iNKT cell activation by α-GalCer. To compare the efficiency of NKT cell activation between liposomal and free α-GalCer, WT mice were injected i.v. with the 2 ng α-GalCer formulated in naked and ^TCCNeuAc liposomes or suspended in buffer, or buffer only (control). After 1.5 h, the intracellular IFN-γ and IL-4 from liver iNKT cells were analyzed as in A. (D) Cd169^−/− mice exhibit equivalent cytokine responses to that of WT mice in response to naked liposomes with a high dose of α-GalCer. WT and Cd169^−/− mice were injected i.v. with naked liposomes with 50 ng of α-GalCer. iNKT cell activation in the liver and spleen were analyzed as in A. (E) iNKT cell activation by ^TCCNeuAc liposomes requires α-GalCer. WT mice were injected i.v. with ^TCCNeuAc liposomes with or without α-GalCer, or buffer only (control). iNKT cell activation in the liver and spleen were analyzed as in A. Each symbol represents individual animals (n = 3). Error bars indicate SD. Data are representative of at least two independent experiments with similar results. Statistical analyses were performed by Student t test. *P < 0.05; **P < 0.01; ***P < 0.001; N.S., statistically not significant.

CD169/Sn is expressed on a subset of macrophages in various tissues (7, 8). To confirm that iNKT cell activation is mediated by CD169⁺ macrophages, we initially injected clodronate liposomes (CLL) to specifically deplete macrophages (20). Splenic macrophages were significantly depleted 24 h after CLL injection, whereas iNKT cells were unaffected (Fig. 3 A and B). CLL-treated mice exhibited minimal iNKT activation in response to ^TCCNeuAc liposomes with α-GalCer in the spleen (Fig. 3C), consistent with CD169/Sn macrophages being responsible for the iNKT cell activation by ^TCCNeuAc liposomes. CLL treatment also abrogated liver iNKT cell activation by ^TCCNeuAc liposomes (Fig. S4). It is notable that CLL treatment depletes all CD169⁺ macrophages in the spleen (21) and CLL-injected animals did not respond to ^TCCNeuAc liposomal α-GalCer. Although CLL did not clear all CD11b⁺CD11c⁻ cells in the spleen, the residual cells appear not to be involved in the ^TCCNeuAc liposome–mediated iNKT cell activation.

Fig. 3.

The CD169⁺ macrophages present lipid antigens to iNKT cells in the spleen. (A) CLL depletes macrophages, but not iNKT cells, in the spleen. Mice were injected i.v. with CLL or buffer alone (control). After 24 h, splenocytes were analyzed by flow cytometry. Gated population shows macrophages (left, CD11b⁺CD11c^dim-neg) and iNKT cells (right, TCRβ⁺NK1.1⁺), respectively. (B) Assessment of CLL treatment on a number of macrophages and iNKT cells is shown for three mice. (C) Macrophage-depleted mice do not respond to ^TCCNeuAc liposomal α-GalCer. WT mice treated with CLL or buffer only (control) were injected i.v. with ^TCCNeuAc liposomal α-GalCer. After 1.5 h, splenic iNKT cell activation was analyzed as in Fig. 2A. (D) Sorted CD169⁺ macrophages activate iNKT cell hybridoma in vitro. Mice were injected i.v. with ^TCCNeuAc liposomal α-GalCer (20 ng per mouse). After 0.5 h, indicated APCs in the spleen were sorted based on the sorting strategies in Fig. S5. The sorted cells were cultured with iNKT cell hybridoma 1.2. After 20 h, IL-2 in the culture supernatant was measured by ELISA. Data are representative of three independent experiments with similar results.

To further confirm the role of the CD169⁺ macrophage subset, we injected mice with ^TCCNeuAc liposomal α-GalCer and then isolated CD169⁺, CD169⁻ macrophages, and DCs (Fig. S5). These cell populations, exposed in vivo to α-GalCer, were then tested for their ability to activate iNKT cell lines in vitro. Of the three cell populations, only the CD169⁺ macrophages activated iNKT cells (Fig. 3D), demonstrating that the CD169⁺ macrophages, and not other phagocytic APC subsets, were responsible for the iNKT cell priming by ^TCCNeuAc liposomes. The specificity of stimulation by the CD169⁺ macrophages was observed reproducibly, and although the amount of IL-2 released was low, this is likely a reflection of low amount of antigen administered (20 ng per mouse).

CD169/Sn-Endocytic Pathway Leads to Lipid Antigen Presentation to iNKT Cells.

To further investigate the CD169/Sn-endocytic pathway in presentation of lipid antigen to iNKT cells, we used BMMs pulsed with CD169/Sn-targeted liposomes in vitro, followed by adoptive transfer into naive WT mice. LPS-treated BMMs bound to the targeted liposomes (^BPCNeuAc) in a CD169/Sn-dependent manner (15). BMMs pulsed with the targeted liposomes containing α-GalCer induced IFN-γ and IL-4 production in iNKT cells in liver and spleen of recipient mice, whereas BMMs pulsed with naked liposomes with α-GalCer did not (Fig. 4 A and B). The percentage of iNKT cells producing IFN-γ was greater than the percentage producing IL-4, which also has been observed when DCs pulsed with larger amounts of α-GalCer were injected into mice (22). Using Cd169^−/− BMMs, no cytokine production on iNKT cells was detected in the recipient mice, despite similar levels of CD1d expression by the CD169-deficient BMMs (Fig. S1E). Furthermore, BMMs pulsed with ^BPCNeuAc liposomes in the absence of α-GalCer did not induce IFN-γ and IL-4 production in iNKT cells (Fig. 4C). Taken together, these data suggest that CD169/Sn-mediated endocytosis in macrophages is linked to lipid antigen presentation to iNKT cells.

Fig. 4.

Delivery of α-GalCer via the CD169/Sn-endocytic pathway results in lipid antigen presentation to iNKT cells. (A) BMMs pulsed with ^BPCNeuAc liposomal α-GalCer activate iNKT cells in a CD169/Sn-dependent manner. WT and Cd169^−/− BMMs were incubated with the indicated liposomes with α-GalCer (0.4 ng) or buffer only (control). Cells were washed and transferred to naive mice. After 9 h, intracellular IFN-γ and IL-4 in iNKT cells in the liver and spleen were analyzed as in Fig. 2A. (B) Quantification of iNKT cell activation in A. The percentage of IFNγ⁺ iNKT cells in the liver and spleen is shown. (C) iNKT cell activation by BMMs pulsed with ^BPCNeuAc liposomes requires α-GalCer. BMMs were pulsed with ^BPCNeuAc liposomes with or without α-GalCer or buffer only (control) and transferred to naïve animals. iNKT cell activation was analyzed as described. Data are representative of at least two independent experiments with similar results.

Human Mo-DCs Pulsed with CD169/Sn-Targeted Liposomes Containing α-GalCer Activate Human iNKT Cells.

Because CD169/Sn is conserved among the mammalian species (12), we were interested in determining if CD169/Sn bearing human phagocytes could be targeted with CD169/Sn-targeted liposomes to activate human iNKT cells. In humans, CD169/Sn is expressed on macrophages in both lymphoid and nonlymphoid organs (23). Other immune cell lineages such as monocyte-derived DCs (Mo-DCs) and monocytes have been shown to express CD169/Sn upon type I IFN stimulation in vitro (24), coincident with the findings of CD169⁺ monocytes in patients suffering from systemic lupus erythematosus and HIV infection (12, 25, 26). Consistent with these reports, we found that human Mo-DCs treated with IFN-α expressed CD169/Sn and CD1d (Fig. S6). These cells bound specifically to ^TCCNeuAc liposomes, and binding was completely inhibited by the addition of the anti-human CD169/Sn Ab (Fig. 5 A and B).

Fig. 5.

Human CD169/Sn endocytic pathway is linked to lipid antigen presentation to iNKT cells. (A) Alexa647-labeled ^TCCNeuAc liposomes bind to Mo-DCs. Mo-DCs treated with IFN-α were incubated with 5 μM of ^TCCNeuAc liposomes (gray, filled), Naked liposomes (broken line), buffer only (solid line). Cells were washed and analyzed by flow cytometry. (B) The binding of ^TCCNeuAc liposomes to the Mo-DCs is CD169/Sn-dependent. Mo-DCs were first incubated with 10 μg/mL of indicated Abs or buffer only for 30 min at 4 °C, and then incubated with the liposomes (5 μM) or buffer only (control). The cells were washed and analyzed by flow cytometry. The binding of liposomes is shown as mean fluorescence intensity (MFI). (C) Mo-DCs pulsed with ^TCCNeuAc liposomal α-GalCer activate human iNKT cells. Mo-DCs were incubated with different amount of α-GalCer formulated in naked and ^TCCNeuAc liposomes or suspended in buffer, or buffer only (control). The cells were washed and cultured with human iNKT cell lines. After 20 h, IFN-γ in the culture supernatants was measured by ELISA. (D) The iNKT cell activation by Mo-DCs is CD169/Sn-dependent. Mo-DCs were incubated first with 10 μg/mL of indicated Abs, followed by the addition of ^TCCNeuAc liposomes. Cells were washed and used as in C. Data are representative of at least two independent experiments with similar results.

To assess the potential for the CD169/Sn pathway in lipid antigen presentation, we cultured human iNKT cell lines with Mo-DCs pulsed with the targeted liposomes. As shown in Fig. 5C, Mo-DCs pulsed with ^TCCNeuAc liposomal α-GalCer induced significantly more IFNγ from human iNKT cells than naked liposomes with α-GalCer. ^TCCNeuAc liposomal α-GalCer was 100-fold more effective than α-GalCer alone. Importantly, this activation only occurred if the α-GalCer was present in the liposomes and was inhibited by the treatment of Mo-DCs with anti-CD169/Sn Ab before the addition of liposomes (Fig. 5D). These data demonstrate that the link between CD169/Sn-mediated endocytosis and lipid antigen presentation is conserved among the species.

Discussion

It is generally accepted that various APCs can present lipid antigens to iNKT cells (2). Several studies have shown that DCs are the main APCs for presenting free α-GalCer in vivo (3, 4, 27), but in one study, macrophages were most important in the liver (27). However, because of the importance of iNKT cells in modulating the immune response, understanding the capabilities of individual APC subsets provides the opportunity to tailor the immune response. With a view toward immune therapy for cancer using glycolipid antigens, pulsing the iNKT cell glycolipid on APCs is more effective than injecting a free glycolipid antigen, both in mice (22, 28) and in clinical trials (29). Cell-based therapies are complex and expensive, however; therefore, it is important to develop means of targeting glycolipid antigens to the appropriate APC in vivo.

Several reports have investigated APC presentation of lipid antigen to iNKT cells when delivered using lipid-coated silica particles (6, 10, 30). In lymph nodes, α-GalCer particles injected intraperitoneally were preferentially delivered to subcapsular sinus macrophages expressing CD169, presumably because these macrophages were directly exposed to the lymphatic flow and capable of capturing the particles by phagocytic activity, independent of CD169/Sn function (6, 7). However, in spleen, other phagocytic APCs were found to actively take up and present α-GalCer–coated particles to iNKT cells (10). In contrast, when lipid particles were formulated with hen egg lysozyme (HEL), they were actively taken up by HEL specific B cells, resulting in B-cell presentation of lipid antigen to iNKT cells, which in turn provided help from iNKT cells to the B cells for differentiation and antibody production (30). In this case, HEL plays two roles—to actively target the α-GalCer to a B-cell subset and to activate the cell for its efficient processing and presentation via CD1d to iNKT cells.

Here we have used a glycan ligand of CD169/Sn to specifically target the CD169⁺ macrophage via CD169/Sn-mediated endocytosis. CD169⁺ macrophages and iNKT cells are abundant in spleen and liver (9, 19). Although the ability of CD169⁺ macrophages to process and present antigen to iNKT cells has been documented in lymph nodes (6), it has not been possible until now to assess the ability of CD169⁺ macrophages to induce systemic activation of the more numerous iNKT cells outside lymph nodes. Likewise, it has not been possible to assess the role of the CD169 endocytic pathway for lipid antigen delivery and presentation. We find that targeting even small amounts of α-GalCer specifically to CD169⁺ macrophages is sufficient to cause systemic activation of iNKT cells. Because the enhanced iNKT cell activation is dependent on CD169/Sn, our data clearly define CD169/Sn as an endocytic receptor capable of delivering lipid antigen for presentation to iNKT cells. It is also notable that antigen presentation to iNKT cells occurs in the absence of macrophage activation including costimulatory molecule induction and IL-12 production (Fig. S7), confirming that the increased iNKT cell activation is due to the enhanced uptake of α-GalCer by CD169⁺ macrophages, not to bystander effects such as IL-12–mediated stimulation of iNKT cells (31).

iNKT cell activation by α-GalCer leads to rapid IL-4 and IFN-γ production. Induction of these cytokines by α-GalCer is known to have a modulatory effect on various disease models in mice (22, 32). Although we found that ^TCCNeuAc liposomal α-GalCer induced acute induction of IL-4 at 100-fold lower doses than free α-GalCer (Fig. S8), the ^TCCNeuAc liposomes resulted in only a modest induction of IFN-γ in the serum at 24 h, which is the peak time following injection of naked or free α-GalCer. Most of this late production of IFN-γ is due to the synthesis by NK cells that are activated following stimulation of iNKT cells (33). Although DCs that have interacted with iNKT cells are efficient at this secondary activation of NK cells (3), the results suggest that the targeted CD169⁺ macrophages are much less capable of NK cell activation.

A recent report has shown that CD169/Sn mediates robust immune responses against selected sialylated pathogens in vivo (13). Sialylated Campylobacter jejuni injected i.v. is captured by CD169/Sn via the sialic acid recognition, resulting in rapid secretion of inflammatory cytokines; this effect is not seen in CD169/Sn knockout mice. The results are consistent with sialylated C. jejuni being endocytosed into the compartments where Toll-like receptors are localized, rather than direct signaling by CD169/Sn from the cell surface. Indeed, we and others have shown that cross-linking of CD169/Sn does not induce the costimulatory molecule and inflammatory IL-12 from macrophages (Fig. S7) (34). Together with our data, we suggest that the CD169/Sn-endocytic pathway may orchestrate systemic immune responses including cytokine production and lipid antigen presentation to iNKT cells for sialylated pathogens such as Group B Streptococcus and C. jejuni (13, 35, 36).

Materials and Methods

Mice and Human Subjects.

C57BL/6J mice were maintained in a specific pathogen-free condition in the animal facility at The Scripps Research Institute and used in the accordance with the guidelines of the Institutional Animal Care Committee at the National Institutes of Health. Cd169^−/− mice were obtained from A. Varki, University of California, La Jolla, CA) with the permission from P. R. Crocker (University of Dundee, United Kingdom). Peripheral blood from healthy donors was obtained after informed consent from normal blood donor service at The Scripps Research Institute and used in the accordance to Institutional Review Board at The Scripps Research Institute.

Reagents.

Antibodies were from Biolegend unless otherwise indicated: anti-mouse CD1d (1B1), B220 (RA3-6B2), CD169/Sn (3D6.112), IL-4 (BVD6-24G2, eBioscience), T cell receptor β (TCRβ) (H57-597), CD11b (M1/70), CD11c (N418), Gr-1, (RB6-8C5), NK1.1 (PK136), CD68 (FA-11), CD80 (16-10A1), CD86 (GL-1), H-2K^b (28-8-6), I-A^b (AF6-120.1), anti-human CD1d (51.1), and CD169/Sn (7-239). α-GalCer was a gift from K. L. Matta (Roswell Park Cancer Institute, Buffalo, NY). α-GalCer or PBS57/mCD1d tetramer (National Institutes of Health tetramer core facility) was used to identify iNKT cells. CLLs were obtained from Nico van Rooijen (Vrije Universiteit, Amsterdam, The Netherlands). Ficoll-Plaque plus and Percoll plus (GE Healthcare) were used for the density gradient centrifugation to separate peripheral blood mononuclear cells and hepatic lymphocytes respectively. ELISA kits (Biolegend) were used to measure mouse IFN-γ, IL-4, IL-12p40, and human IFN-γ.

Liposome Preparation.

Liposomes used in this study were formulated as described (37). Briefly, 5 mol% ^BPCNeuAc liposomes were composed of distearoyl phosphatidylcholine (Avanti Polar Lipids): cholesterol (Chol, Sigma-Aldrich): PEG-distearoyl phosphoethanolamine (DSPE, Nippon Oil & Fats): ^BPCNeuAc-α2,3-Gal-β1,4-GlcNAc-PEG-DSPE: α-GalCer in a 57:37:0:5:1 molar ratio. Naked liposomes were composed of 5 mol% PEG-DSPE. ^TCCNeuAc liposomes were composed of 0.3 mol% ^TCCNeuAc-α2,3-Gal-β1,4-GlcNAc-PEG-DSPE instead of ^BPCNeuAc trisaccharide-PEG-DSPE. Fluorescent liposomes contained 0.2 mol% of Alexa Fluor 647-PEG-DSPE. All components were mixed and lyophilized. The lyophilized lipids were hydrated with 1 mL PBS, sonicated, then extruded until the size became around 100 nm measured by Zetasizer (Malvern).

Flow Cytometry.

Cells were washed with HBSS containing 0.1% BSA, 2 mM EDTA, and 0.1% NaN₃ (FACS buffer), blocked with anti-mouse CD16/32 (2.4G2, BD Biosciences) or anti-human CD32 (3D3, BD Biosciences), and stained with indicated antibodies. Stained cells were washed with FACS buffer and analyzed by FACS Caliber or LSR II (BD Biosciences). Acquired data were analyzed with Flowjo (Tree Star). A total of 1 μg/mL propidium iodide was used to exclude the dead cells. For the staining of intracellular cytokines, cells were first stained for cell surface markers, washed with FACS buffer, then fixed and permeabilized with cytofix/cytoperm solution (BD Biosciences). The fixed cells were further stained with the indicated Abs for cytokines, then washed with perm/wash buffer (eBioscience) and analyzed. For the liposome-binding analysis, cells were first incubated with indicated liposomes at 37 °C for 30 min, then washed and stained with indicated Abs.

Preparation of BMMs.

BMMs were prepared from mouse bone marrow cell culture with the RPMI medium (Invitrogen) supplemented with 10% (vol/vol) heat-inactivated FCS, 2 mM glutamine, 100 U/mL penicillin, 100 μg/mL streptomycin, 1 mM nonessential amino acids, 1 mM sodium pyruvate, 50 μM 2-melcaptoethanol, and 10% (vol/vol) L929 cell conditioned medium. At day 6, 100 ng/mL of LPS (Invivogen) was added to the culture to induce CD169/Sn expression. At day 8 adherent cells were used as BMMs.

Generation of Human Mo-DCs.

Human monocytes were separated with a CD14-positive selection kit (Miltenyi) from buffy coats of peripheral blood obtained from healthy donors. Isolated monocytes were cultured with the same medium used for BMM culture with 20 ng/mL of human IL-4 and GM-CSF (Peprotech) instead of the L929 conditioned medium. At day 6, the floating cells were harvested and cultured with the same medium, with 500 U/mL of human IFN-α (R&D systems) to induce human CD169/Sn expression. At day 8, the cells were used as Mo-DCs expressing CD169/Sn.

Analysis of iNKT Cell Activation.

For the lipid antigen presentation by BMMs and Mo-DCs, those cells were pulsed with indicated reagents for 30 min at 37 °C in the sterile FACS buffer and washed. To check mouse iNKT cell activation by BMMs, 5 × 10⁵ of BMMs was i.v. transferred to WT mice. After 9 h, intracellular cytokines as well as cell surface markers of iNKT cells in the liver and spleen of recipient animals were analyzed by flow cytometry. To assess the activation of human iNKT cells by Mo-DCs, 3 × 10⁴ of Mo-DCs was cocultured with 5 × 10⁴ of human iNKT cells prepared as described (38). After 20 h, the culture supernatants were collected and human IFN-γ was measured by ELISA. Alternatively mice were injected with indicated reagents through the tail vein. After 1.5 h, cytokine production by iNKT cells was measured by intracellular flow cytometry. To analyze the effect of macrophage depletion on iNKT cell activation by liposomes, first we injected 1:1 diluted CLL i.v. to deplete macrophages in vivo, After 24 h, we injected ^TCCNeuAc liposomes i.v. and analyzed iNKT cell activation as described previously. To determine if isolated CD169/Sn macrophages activate the mouse iNKT cell hybridoma 1.2, ^TCCNeuAc liposome with α-GalCer was injected i.v. into WT mice. After 0.5 h, splenocytes were harvested, stained with antibodies, and subjected to cell sorting by FACS. 2.5 × 10⁴ of sorted cells were cocultured with 2.5 × 10⁴ of the hybridoma 1.2 for 20 h. The amount of IL-2 in the culture supernatants was measured by ELISA.

Microscopy.

Mouse liver perfused with PBS was harvested and incubated with 4% (wt/wt) paraformaldehyde in PBS for 24 h at 4 °C. Fixed liver was frozen in Tissue-Tek O.C.T. Compound (Sakura) and sectioned (8 μm thick). Liver sections were incubated with RPMI medium containing 10% (vol/vol) FCS and 1 μg/mL 2.4G2 for 1 h at 25 °C. Then the sections were incubated with 1 μg/mL of Alexa 488–labled anti-CD68 (FA-11, Biolegend) and 1 μg/mL DAPI for 30 min at 25 °C. The sections were washed with PBS and mounted with anti-fade mounting solution (Life Technologies). Images were obtained on a Bio-Rad (Zeiss) Radiance 2100 Rainbow laser scanning confocal microscope. To see the colocalization of internalized Alexa647-labeled ^TCCNeuAc liposomes with lysosomal marker LAMP-1, BMMs on the cover slide were incubated with ^TCCNeuAc liposomes in the PBS containing 0.01% BSA for 30 min at 37 °C. The cells were washed, fixed, and permeabilized with PBS containing 0.01% Saponin and 0.01% BSA for 3 min. The cells were then stained with anti-LAMP1 (ab24170, Abcam) and DAPI for 30 min at 25 °C. The cells were washed, mounted, and analyzed as described.

Statistical Analysis.

Student t test was used for statistical analysis on Prism software (GraphPad). P < 0.05 was considered statistically significant.