Activation of dendritic cells by liposomes prepared from phosphatidylinositol mannosides from Mycobacterium bovis bacillus Calmette-Guerin and adjuvant activity in vivo

Abstract

Liposome vesicles could be formed at 65 degrees C from the chloroform-soluble, total polar lipids (TPL) extracted from Mycobacterium bovis bacillus Calmette-Guérin (BCG). Mice immunized with ovalbumin (OVA) entrapped in TPL liposomes produced both anti-OVA antibody and cytotoxic T lymphocyte responses. Murine bone marrow-derived dendritic cells were activated to secrete interleukin-6 (IL-6), IL-12, and tumor necrosis factor upon exposure to antigen-free TPL liposomes. Three phosphoglycolipids and three phospholipids comprising 96% of TPL were identified as phosphatidylinositol dimannoside, palmitoyl-phosphatidylinositol dimannoside, dipalmitoyl-phosphatidylinositol dimannoside, phosphatidylinositol, phosphatidylethanolamine, and cardiolipin. The activation of dendritic cells by liposomes prepared from each purified lipid component of TPL was evaluated in vitro. A basal activity of phosphatidylinositol liposomes to activate proinflammatory cytokine production appeared to be attributable to the tuberculosteric fatty acyl 19:0 chain characteristic of mycobacterial glycerolipids, as similar lipids lacking tuberculosteric chains showed little activity. Phosphatidylinositol dimannoside was identified as the primary lipid that activated dendritic cells to produce amounts of proinflammatory cytokines several times higher than the basal level, indicating the importance of mannose residues. Although the activity of phosphatidylinositol dimannoside was little influenced by palmitoylation of mannose at C-6, a further palmitoylation at inositol C-3 diminished the induction levels of IL-6 and IL-12. Further, OVA entrapped in palmitoyl-phosphatidylinositol dimannoside liposomes was delivered to dendritic cells for major histocompatibility complex class I presentation more effectively than TPL OVA-liposomes. BCG liposomes containing mannose lipids caused up-regulation of costimulatory molecules and CD40. Thus, the inclusion of pure phosphatidylinositol mannosides of BCG in lipid vesicle vaccines represents a simple and efficient option for targeting antigen delivery and providing immune stimulation.

FIG. 1.

Ability of BCG TPL liposomes to have adjuvant effects on a humoral response in mice. C57BL/6 mice received subcutaneous injections at 0 and 21 days of 15 μg of OVA (naïve, or no adjuvant) or OVA admixed with alum or entrapped in liposomes. Liposomes were prepared from BCG TPL (100%) or diluted with commercial PC-PG-cholesterol (1.8:0.2:1.5, mol%) to achieve 50, 10, and 0% (wt/wt) of BCG lipids. BCG+OVA indicates mice injected with 15 μg of OVA admixed just prior to injections with an equivalent amount (0.5 mg) of antigen-free TPL liposomes. Anti-OVA antibody titrations were performed on blood samples taken on day 31. Similar data not shown were obtained for day 41. Data are the average ± standard error of the means of results for four mice per group (six mice per group for 100% BCG lipids). None of the means among 100% BCG-OVA, 50% BCG-OVA, or alum+OVA were statistically different (95% confidence) by an unpaired t test. Paired groups indicated by symbols (+, Ο, and *) were significantly different, with P values of 0.0097, 0.0013, and 0.0024, respectively.

FIG. 2.

Ability of BCG liposomes to have an adjuvant effect on the CD8⁺-T-cell response in mice. C57BL/6 mice received subcutaneous injections at 0 and 21 days of 15 μg of OVA homogenized in CFA followed by OVA in incomplete Freund's or of OVA entrapped in BCG TPL liposomes (BCG-OVA). Also shown are the results from a single injection of 10⁶ live recombinant BCG cells expressing the main CTL epitope of OVA. Splenic cells were harvested 10 weeks after first injections for CTL and ELISPOT assays. CTL results are shown in panel A for various effector:target (E:T) cell ratios, where EL.4 cells represent nonspecific targets and EG.7 cells are specific targets expressing OVA_257-264 peptide. IFN-γ-specific ELISPOT precursor frequencies are shown in panel B. ELISPOT frequencies (with peptide) for BCG-OVA, live BCG, and CFA+OVA are significantly different from results for the näive case (two-tailed t test, P values 0.0001, 0.0001, and 0.0303, respectively). The mean frequency of BCG-OVA compared to that of CFA+OVA (with peptide) was not significant (P value of 0.1306).

FIG. 3.

Effect of antigen loading in BCG liposomes on induction of an immune response in mice. At 0 and 21 days C57BL/6 mice received subcutaneous injections consisting of 15 μg of OVA entrapped at various loadings in BCG TPL liposomes. Splenic cells from duplicate mice were harvested and pooled 6 weeks after the first injections for CTL and ELISPOT assays as described in the legend of Fig. 2. ELISPOT precursor frequencies are shown in panel B. Data are the means ± standard deviations of assays performed in triplicate.

FIG. 4.

Thin-layer chromatogram showing the separation of BCG TPL into six major and one minor lipid fractions. Fractions are numbered 1 to 7, from most polar to least polar. Lipids (applied at the bottom of the plate) are BCG TPL and reference standards PI, l-α-phosphatidyl-l-serine, PG, and PE. An acidic solvent was used to develop the plate. Lipids were located by spraying with phosphate detection reagent.

FIG. 5.

Structures of lipids PI, PIM₂, palm₁-PIM₂, and palm₂-PIM₂ present in the TPL of BCG. Structural linkage details have been described previously (4, 21). The position sn-1 versus sn-2 of tuberculosteric acid may be reversed.

FIG. 6.

Activation of bone marrow dendritic cells by BCG liposomes composed of palm₁-PIM₂ and PI. The average diameters of antigen-free liposomes were 93 ± 41 nm for TPL, 160 ± 95 nm for palm₁-PIM₂, and 162 ± 84 nm for PI. LPS from E. coli is shown as a positive control for activation. Sensitivity of the TNF bioassay was <0.1 pg/ml; values below the detection limit are graphed as zero. These experiments were repeated three times with the same results. O.D., optical density.

FIG. 7.

IL-12 and IL-6 secretions from bone marrow dendritic cells induced by liposomes composed of BCG lipids compared to secretions induced by liposomes prepared from commercial lipids. Commercial lipids were PI and cardiolipin (CL) from soybean (soy), DMPC, DPPE, and DMPG. Liposomes and LPS were added to bone marrow dendritic cell cultures at the indicated concentrations of dry weight per milliliter and incubated 72 h before assay. Panels a and b represent separate experiments performed with different primary cultures of dendritic cells. These experiments were repeated three times with the same results.

FIG. 8.

Up-regulation of costimulatory molecule CD80 on bone marrow dendritic cells exposed to antigen-free BCG liposomes. Dendritic cells were stimulated with 25 μg of the various liposomes or LPS per ml for 24 h, and cell surface expression of CD80 was assessed by flow cytometry. The circular gates indicate the percentages of CD80⁺ cells.