1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54

YTUBE1322_proof ■ 22 May 2015 ■ 1/11

Tuberculosis xxx (2015) 1e11

Contents lists available at ScienceDirect

Tuberculosis journal homepage: http://intl.elsevierhealth.com/journals/tube

IMMUNOLOGICAL ASPECTS

Q2 Q1

Liposomal delivery of lipoarabinomannan triggers Mycobacterium tuberculosis specific T-cells Stephanie Kallert a, Sebastian Zenk a, Paul Walther b, Mark Grieshober a, Tanja Weil c, Steffen Stenger a, * a b c

Institute for Medical Microbiology and Hygiene, University Hospital Ulm, Albert Einstein Allee 11, 89081 Ulm, Germany Central Unit Electron Microscopy, Albert-Einstein-Allee 11, 89081 Ulm, Germany Institute for Organic Chemistry III/Macromolecular Chemistry, Albert-Einstein-Allee 11, 89081 Ulm, Germany

a r t i c l e i n f o

s u m m a r y

Article history: Received 22 January 2015 Received in revised form 2 April 2015 Accepted 8 April 2015

Lipoarabinomannan (LAM) is a major cell wall component of Mycobacterium tuberculosis (Mtb). LAM specific human T-lymphocytes release interferon-g (IFNg) and have antimicrobial activity against intracellular Mtb suggesting that they contribute to protection. Therefore the induction of LAM-specific memory T-cells is an attractive approach for the design of a new vaccine against tuberculosis. A prerequisite for the activation of LAM-specific T-cells is the efficient uptake and transport of the glycolipid antigen to the CD1 antigen presenting machinery. Based on the hydrophobicity of LAM we hypothesized that packaging of LAM into liposomes will support the activation of T-lymphocytes into antigen presenting cells. We prepared liposomes containing phosphatidylcholine, cholesterol, stearylated octaarginine and LAM via thin layer hydration method (LIPLAM). Flow cytometry analysis using fluorescently labelled LIPLAM showed an efficient uptake by antigen presenting cells. LAM delivered via liposomes was biologically active as demonstrated by the down-regulation of peroxisome proliferator activated receptor gamma (PPARg) protein expression. Importantly, LIPLAM induced higher IFNg production by primary human T-lymphocytes than purified LAM (2e16 times) or empty liposomes. These results suggest that the delivery of mycobacterial glycolipids via liposomes is a promising approach to promote the induction of M. tuberculosis specific T-cell responses. © 2015 Published by Elsevier Ltd.

Keywords: Mycobacterium tuberculosis Liposomes T-lymphocytes Vaccine

1. Introduction Mycobacterium tuberculosis (Mtb) is the etiologic agent of tuberculosis and remains to be a major health problem worldwide. The only approved vaccine is Bacillus Calmette Guerin (BCG) which was introduced in the first half of the 20th century [1,2]. Although BCG confers protection against severe disseminated forms of tuberculosis in childhood, the efficacy against pulmonary

Abbreviations: LAM, lipoarabinomannan; LIPLAM, liposomes containing LAM; LIP, liposomes without LAM; Mtb, Mycobacterium tuberculosis; APC, antigen presenting cell; MFI, mean fluorescent intensity; CM, complete medium; PPARg, peroxisome proliferator activated receptor gamma; GMM, glucose monomycolate; PI, propidium iodide; MR, mannose receptor. * Corresponding author. Tel.: þ49 731 50065300; fax: þ49 731 50065302. E-mail addresses: [email protected] (S. Kallert), sebastian.zenk@ uniklinik-ulm.de (S. Zenk), [email protected] (P. Walther), mark. [email protected] (M. Grieshober), [email protected] (T. Weil), [email protected] (S. Stenger).

tuberculosis in adults is limited. There are several approaches to develop new preventive vaccines and ten candidates have reached the level of clinical trials including live vaccines such as an improved BCG or genetically modified Mtb strains [3,4]. An alternative concept is to use immune-dominant antigens to booster a pre-existing BCG-induced memory. Booster antigens can be applied by using viral vectors [5e9] or protein adjuvant combinations [10e20]. In the past, the search for vaccine antigens has focused on proteins and peptides neglecting the immunogenic potential of glycolipids and lipids abundantly expressed in the mycobacterial cell wall. Mtb has an unusual complex cell wall which contains many lipid antigens, such as lipoarabinomannan (LAM), mycolic acid, glucose monomycolate (GMM) or mannosyl phosphodolichol [21]. Lipid antigens are presented by group 1 CD1 molecules which are nonpolymorphic cell-surface glycoproteins expressed mainly by dendritic cells. The loading of lipid antigens on CD1b molecules occurs in the same acidic late endosomes in which MHC class II molecules are loaded with peptide (MIICs) [22,23]. The acidic pH facilitates the

http://dx.doi.org/10.1016/j.tube.2015.04.001 1472-9792/© 2015 Published by Elsevier Ltd.

Please cite this article in press as: Kallert S, et al., Liposomal delivery of lipoarabinomannan triggers Mycobacterium tuberculosis specific T-cells, Tuberculosis (2015), http://dx.doi.org/10.1016/j.tube.2015.04.001

55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

YTUBE1322_proof ■ 22 May 2015 ■ 2/11

2

S. Kallert et al. / Tuberculosis xxx (2015) 1e11

loading of lipid antigens by promoting a conformational change in CD1b and thereby facilitating the access to the hydrophobic core of the protein [24]. LAM is an abundant glycolipid component of the mycobacterial cell wall. LAM comprises a group of heterogeneous molecules which share a basic tripartite structure, consisting of a diacyl glyceryl phosphatidyl-myo-inositol anchor, a carbohydrate backbone and capping motifs [25]. LAM structures differ in the glycosylation level as well as the amount of charged groups such as succinates and phosphates. LAM-specific T-cells are likely to contribute to protection against Mtb because they i) have cytolytic and bactericidal activity ii) induce pro-inflammatory cytokine release and iii) activate T-cells [26e29]. The protective efficacy of lipid specific T-cells in vivo was shown in guinea pigs where a CD1 restricted T-cell response as well as improved pulmonary pathology after Mtb challenge was observed following immunization with mycobacterial lipids [30]. A major challenge of inducing LAM specific memory T-cell responses is the delivery of the lipid antigen into antigen presenting cells (APC). Liposomes were first described in 1964 when spontaneous formation of lipid vesicles in a suspension was observed under the electron microscope [31]. The potential as transport vehicles was soon recognized and entrapment of therapeutic proteins in liposomes was established 1971 [32]. Since then liposomes have emerged as a versatile tool to encapsulate a wide range of therapeutic molecules and deliver them to defined compartments of the human body [33]. Several liposome based products are now approved for use in humans to treat diseases ranging from fungal infections to malignancies [33]. More recently liposomes have been employed to enhance the immunogenicity of protein-based subunit vaccines against tuberculosis. Liposomes offer the unique opportunity to combine activation of the innate immune system through adjuvant effects of the liposomal structure and the adaptive immune system through entrapped antigens. Here we investigate the possibility that liposomes promote the induction of primary human T-lymphocytes responding to the mycobacterial glycolipid antigen LAM. Our results demonstrate that purified LAM can be efficiently incorporated into liposomes (LIPLAM). LIPLAM retains biological activity, is efficiently taken up by APCs and improves T-cell activation as compared to purified LAM. These findings encourage the concept of delivering glycolipid antigens via liposomes for promoting protective T-cell responses and provide a platform to optimize the composition of liposomes in the context of vaccine development. 2. Materials and methods

of PPARg LAM H37Rv was used. Due to its robust and reproducible capacity to stimulate human CD1 restricted T-cells, T-cell stimulation was performed with LAM purified from Mycobacterium smegmatis. To screen for lipid-reactive blood donors, a total lipid extract from H37Rv was used. The following reagents were obtained through BEI Resources, NIAID, NIH: M. tuberculosis, Strain H37Rv, Purified Lipoarabinomannan (LAM), NR-14848, M. smegmatis, Purified Lipoarabinomannan (LAM), NR-14849, M. tuberculosis, Strain H37Rv, Total Lipids, NR-14837. The high purity of the preparations is documented by QC gel and Blot by BEI resources but not specifically quantified (http://www.beiresources.org/Catalog/antigen/ NR-14849.aspx). 2.2. Cell culture reagents Primary human blood cells were cultured in RPMI 1640 (Life Technologies) supplemented with 2 mM glutamine (Sigma), 10 mM HEPES, 13 mM NaHCO3, 100 mg/ml streptomycin, 60 mg/ml penicillin (all from Biochrom) and 5% heat-inactivated human AB serum (Sigma) (¼complete medium, CM). Human foreskin fibroblasts were isolated from the foreskin of infants by trypsin and used for experiments between passage 10 and 25. Fibroblasts were cultured in minimal essential medium þ GlutaMAX (Life Technologies), 5% foetal calf serum and 100 mg/ml gentamycin. HeLa 229 cells (American Type Culture Collection; CCL 2.1) were cultured in monolayers in RPMI 1640 þ GlutaMAX (Life technologies) and 10% foetal bovine serum (Life Technologies). 2.3. Liposome preparation Liposomes containing LAM were prepared by dissolving 100 mg LAM (BEI Resources, Manassas), egg phosphatidylcholine (Avanti polar lipids), cholesterol (Sigma) and stearylated octaarginine (Thermo Fisher Scientific) at a molar ratio of 7:3:0.5 in chloroform (VWR): methanol (Sigma) (1:1). The organic solvents were removed by rotary evaporation (Rotavapor RII, Büchi Labortechnik AG) and the dried lipid film was reconstituted by adding 1 ml PBS and sonication for 15 min. The liposome suspension was centrifuged (14,000 rpm, 30 min) and the liposome pellet was washed with PBS and afterwards applied to 10 freeze- and thaw-cycles in liquid nitrogen and 37  C (water bath) followed by mini-extrusion (11 times) through a polycarbonate membrane with a pore size of 100 nm (Mini Extruder, Avanti polar lipids). To compare different liposome preparations the OD was measured at 500 nm. Fluorescent liposomes were prepared by adding 1.10 -dioctadecyl-3.3.30 30 tetramethylindo-carbocyanine perchlorate (DIL-C, Sigma), a lipophilic dye at 0.2 mol% to the lipids in organic solvent.

2.1. Antibodies and reagents 2.4. Electron microscopy The following antibodies were used for flow cytometry or western blot: mouse anti-LAM (clone CS-35, provided by Ulrich Schaible, Research Center Borstel), PPARg rabbit mAb (clone C26H12), b-Actin rabbit mAb, (both Cell Signaling Technology), donkey anti mouse alkaline phosphatase conjugated antibody, donkey anti rabbit alkaline phosphatase conjugated antibody, goatanti-mouse-biotin (all Jackson ImmunoResearch Laboratories), MHC class II-biotin, MHC class II-PerCP, CD3-PerCP, CD1b-FITC, anti-mannose-receptor-PE, CD14-PerCP, CD56-PE (all BD Biosciences), Streptavidin-APC, CD86-PE, CD19-Biotin, CD54-PE, CD11c-APC, CD16-PE, CD40-APC, IgG1-APC, IgG2a-APC, IgG2a-PE, IgG1-FITC, IgG1-PerCP, IgG1-PE, IgG1-biotin, IgG2a-biotin, IgG2bFITC (all Invitrogen), CD1a-APC, CD80-FITC, CD3-FITC, IgG2a-FITC (both BioLegend), CD1c-APC, IFNg-PE, IgG1-PE (all Miltenyi), DCSIGN-FITC, CCR-6-FITC, CCR-7-FITC, IgG2b-FITC, IgG2a-FITC (all R&D Systems). For experiments addressing the protein expression

For freeze fracturing a 300 mesh copper grid was dipped into the LIPLAM suspension and positioned between two low mass copper platelets. This sandwich was fast frozen in liquid propane cooled by liquid nitrogen. The sample was then cryo-fractured in a BAF 300 freeze fracturing device at a temperature of 150  C in vacuum. The samples were coated by electron beam evaporation with 2 nm of platinum from an angle of 45 and with 10 nm of carbon perpendicular. The replica was cleaned with distilled water and imaged in a ZEISS 109 TEM at an accelerating voltage of 80 kV. 2.5. Preparation of CD1þ APCs and T-cell stimulation Buffy coats from healthy donors were provided by the German Red Cross located at the Institute of Transfusion Medicine (Ulm University, Ulm, Germany) or from healthy volunteers. PBMC were

Please cite this article in press as: Kallert S, et al., Liposomal delivery of lipoarabinomannan triggers Mycobacterium tuberculosis specific T-cells, Tuberculosis (2015), http://dx.doi.org/10.1016/j.tube.2015.04.001

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

YTUBE1322_proof ■ 22 May 2015 ■ 3/11

S. Kallert et al. / Tuberculosis xxx (2015) 1e11

3

isolated by density gradient centrifugation. CD1þ APC were differentiated from plastic adherent PBMC cultured in the presence of GM-CSF (10 ng/ml, Miltenyi) and IL-4 (10 ng/ml, Biolegend) for 4 d. CD1þ APC expressed MHC class II (95 ± 4%), CD1a (18 ± 11%), CD1b (57 ± 11%), CD1c (99 ± 1%), CD86 (95 ± 2%), CD40 (100 ± 0%), the mannose receptor (95 ± 3%), DC-SIGN (83 ± 11%), CD11c (100 ± 0%), CD54 (98 ± 2%), CD16 (19 ± 19%), as well as minor amounts of CD14 (3 ± 2%), CD80 (1 ± 0.9%), CCR-6 (1 ± 1%) and CCR-7 (3 ± 1%) (Suppl. Figure 1). Irradiated (30 Gy) CD1þ APCs were incubated with autologous non-adherent PBMCs at a ratio of 1:4 and stimulated overnight as indicated. IFNg release in the supernatant was determined by sandwich ELISA exactly as suggested by the manufacturer (Endogen). All studies were approved by the local ethical review committee (279/08).

pH 2.2) and re-probed using a b-Actin antibody (1:1000), followed by a donkey anti rabbit alkaline phosphatase (1:10,000) antibody (Jackson ImmunoResearch Laboratories). The densitometric evaluation of the western blots was performed with GelScan 5.1 (BioSciTec). For LAM detection, purified LAM (0.2 mg) or liposome preparations (2 ml) were boiled in Laemmli sample buffer and applied to an SDS PAGE followed by western blotting with a mouse anti LAM antibody (1:20).

2.6. Flow cytometry

3. Results

For cell surface labelling PBMCs were incubated for 30 min at 4  C with the respective antibodies. For intracellular IFNg staining, CD1þ APCs and non-adherent PBMCs were stimulated overnight and Brefeldin A (10 mg/ml, Sigma) was added for 4 h at 37  C. Afterwards the cells were fixed for 20 min with 4% paraformaldehyde (Sigma) on ice, permeabilized using PermWash (BD Biosciences) and stained with anti IFNg for 30 min at 4  C. Annexin V/propidium iodide staining was performed using the “FITC Annexin V Apoptosis Detection Kit I” from BD Biosciences following the manufacturer's protocol. Data was recorded using a FACSCalibur™, BD Biosciences and for the analysis of flow cytometry data the program FlowJo, Version 7.6.5 (Tree Star Inc.) was used. Detection of LAM on liposomes was performed by labelling rhodamine-LIPLAM and rhodamine-LIP with a monoclonal mouse anti LAM antibody (1:20). Staining was detected by a goat-antimouse biotinylated antibody (Jackson ImmunoResearch Laboratories) and Streptavidin-APC (Invitrogen).

3.1. Preparation of liposomes containing lipoarabinomannan

2.9. Statistical analysis The results are presented as mean ± SD. Paired student's t test was used to determine statistical significance between different conditions. Differences were considered significant if p < 0.05.

0.1  106 CD1þ APCs were plated in 200 ml CM per chamber of a 8 chamber slide (Nunc), and incubated overnight with liposomes. Slides were fixed with 4% paraformaldehyde and mounted with aquatex (Merck). Pictures were taken using a Axio Imager M2 microscope (Carl Zeiss, Jena).

LAM-containing liposomes (LIPLAM) were prepared by mixing phosphatidylcholine, cholesterol and octaarginine conjugated to stearin to increase internalization (Suppl. Figure 2) [34e36]. Transmission electron microscopy (TEM) of cryo fractured vesicles showed a single particle suspension of liposomes. The fracture plane in the higher magnification depicts a representative area with three differential fractured unilamellar liposomes (Figure 1A). Liposomes consisting of the framework lipids only (LIP) had comparable size and phenotype (data not shown). To ascertain that LAM was included in LIPLAM, liposome preparations were applied to an SDS gel, followed by western blotting and detection with a specific LAM antibody. LIPLAM showed a band at the same size as purified LAM (Figure 1B). The band was absent in the lane with empty liposomes confirming the integration of LAM into the liposomes. To estimate the amount of LAM in LIPLAM we performed western blot analysis using purified LAM with defined concentration as a standard. Comparison of signal intensities between LIPLAM and purified LAM revealed that 2 ml LIPLAM contain approximately 0.1 mg of LAM (data not shown). To ascertain that LAM is exposed on the liposomal surface, we stained LAM entrapped in liposomes using a monoclonal antibody. Flow cytometry analysis revealed the LAM was expressed on the surface of LIPLAM but not empty liposomes (Figure 1C).

2.8. Western blotting

3.2. LIPLAM does not affect the viability of PBMC or CD1þ APC

CD1þ APCs were incubated with LIP, LIPLAM or LAM overnight. Cells were lysed using modified RIPA buffer (1 M PBS (Life technologies); 1% NP-40 (Fluka); 0.5% sodium deoxycholate (Fluka); 0.1% SDS (Roth)) and protease inhibitor cocktail tablets as well as phosphatase-inhibitor cocktail tablets (both from Roche). Cell debris was removed by centrifugation (14,000 rpm, 10 min) and protein content in the supernatant was determined (BCA protein assay, Pierce). 25 mg of protein was boiled for 10 min in Laemmli sample buffer (2 mM SDS (Roth); 0.3% Glycerol (Sigma); 63 mM TriseHCl (Sigma), pH 6; 0.03% bromophenole blue (Biomol); 0.15% mercaptoethanol (Sigma)) and analyzed by SDSePage (12%) and western blot. The membranes were incubated overnight with an antibody against peroxisome proliferator activated receptor gamma (PPARg) (1:1000) and afterwards with a donkey anti rabbit alkaline phosphatase (1:10,000) antibody (Jackson ImmunoResearch Laboratories). Proteins were detected by chemiluminescence (CDP Star, Roche) following the manufacturer's protocol. The same membrane was stripped (0.2 M Glycine (AppliChem GmbH); 3.5 mM SDS (Roth); 0.1% Tween 20 (Sigma),

To assess whether LIPLAM affects the viability of primary human cells, freshly isolated PBMC were incubated overnight with different dilutions of LIPLAM and the frequency of apoptotic- and necrotic cells was determined by annexin V and propidium iodide staining. The frequency of apoptotic or necrotic cells was 20.3 ± 9.9 for untreated cells and was not significantly affected by LIPLAM (1:10) resulting in 18.8 ± 9.4 (Figure 2A, B). Similarly, the viability of CD1þ APCs, the target cell population of liposomes in our model, was not influenced (Suppl. Figure 3).

2.7. Fluorescence microscopy

3.3. Uptake of LIPLAM by CD1þ APCs To monitor the uptake of liposomes by CD1þ APC, LIPLAM was fluorescently labelled with a lipophilic carbocyanine dye which intercalates into the lipid bilayer and stains LIPLAM bright red (rhodamine-LIPLAM). CD1þ APCs were incubated overnight with rhodamine-LIPLAM and the uptake was analyzed by fluorescence microscopy. Z-layer images showed bright-red staining of the majority of the cells. The punctate staining pattern suggested that

Please cite this article in press as: Kallert S, et al., Liposomal delivery of lipoarabinomannan triggers Mycobacterium tuberculosis specific T-cells, Tuberculosis (2015), http://dx.doi.org/10.1016/j.tube.2015.04.001

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

YTUBE1322_proof ■ 22 May 2015 ■ 4/11

4

S. Kallert et al. / Tuberculosis xxx (2015) 1e11

Figure 1. Preparation of liposomes containing Lipoarabinomannan. (A) A copper grid was immersed into the LIPLAM suspension and sandwiched between two copper platelets followed by fast freezing, cryo-fracture and coating with platinum and carbon. Representative areas of fractured unilamellar liposomes are shown in two magnifications. (B) LIPLAM (2 ml), LIP (2 ml) and purified LAM (0.2 mg) were analyzed by western blotting with a monoclonal antibody directed against LAM (1:20). The panel shows a representative result of four western blots with different liposome preparations. (C) Rhodamine-LIPLAM (red line) and rhodamine-LIP (blue line) were stained with a monoclonal mouse antibody directed against LAM (1:20). Binding was detected by using a goat-anti-mouse-biotinylated antibody and strepatavidin-APC and analyzed by flow cytometry. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

the liposomes accumulate in small intracellular vesicles (Figure 3A). To ascertain that LIPLAM is taken up by APC and not solely attached to the cell surface, we co-stained cell cultures with markers specific for T-cells, B-cells and macrophages and incubated the cultures overnight with fluorescent LIPLAM. The macrophage and lymphocyte gate was set according to the forward sideward scatter characteristics and the three subpopulations were independently analyzed via flow cytometry. MHCIIþ cells efficiently phagocytosed fluorescent LIPLAM, whereas T- and B-cells remained unlabelled (Figure 3B). Uptake of liposomes by MHCIIþ cells was dose-dependent with the highest concentration (1:100) leading to 83 ± 5.7% positive cells (Figure 3C). The number of positive cells (Figure 3C) as well as the mean fluorescent intensity of single cells increased significantly from 2.8 ± 2.8 (no LIPLAM) to 124 ± 34 (LIPLAM 1:100) (Figure 3D), suggesting that increasing concentrations of LIPLAM increase the frequency of positive cells as well as the amount of LIPLAM taken up per cell.

A likely candidate for mediating the interaction between the mannose residues of LAM and macrophages is the macrophage mannose receptor (MR) [37] which is expressed in CD1þ APCs but not by T- or B-cells (data not shown). To investigate whether the selective uptake of LIPLAM by CD1þ APCs and T- and B-cells is dependent on the MR, we pre-incubated CD1þ APCs with blocking antibodies to the MR but observed no effect on the uptake of LIPLAM (data not shown). To predict the bioavailability of LIPLAM when administered in vivo, we investigated whether endothelial cells (HeLa) or fibroblasts could be loaded with LIPLAM. HeLa cells and primary foreskin fibroblasts were incubated overnight with rhodamine-LIPLAM and the uptake was evaluated by flow cytometry and fluorescence microscopy. Endothelial cells and fibroblasts stained positively for rhodamine-LIPLAM indicating efficient uptake of the liposomes to the same magnitude as CD1þ APCs (Figure 3E). Since neither primary foreskin fibroblasts nor HeLa cells express the MR (data not shown), the MR is unlikely to be critically involved in the uptake of LIPLAM by human cells.

Please cite this article in press as: Kallert S, et al., Liposomal delivery of lipoarabinomannan triggers Mycobacterium tuberculosis specific T-cells, Tuberculosis (2015), http://dx.doi.org/10.1016/j.tube.2015.04.001

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

YTUBE1322_proof ■ 22 May 2015 ■ 5/11

S. Kallert et al. / Tuberculosis xxx (2015) 1e11

5

Figure 2. LIPLAM does not affect the viability of PBMC. (A) 0.5  106 freshly isolated PBMCs were incubated with medium alone or LIPLAM (1:100). After overnight incubation apoptotic and necrotic cells were detected by annexin V-FITC and PI-staining. The panel shows a representative result of one donor out of five. (B) 0.5  106 freshly isolated PBMCs were incubated with medium alone, LPS (100 ng/ml) or LIPLAM and annexin V PI-staining was performed. The figures give the average percentage of annexin V and PI positive cells of five independent donors þ SD.

3.4. LIPLAM down regulates PPARg protein expression To assess the biological activity of LIPLAM, we measured the release of pro-inflammatory cytokines by CD1þ APCs. However purified LAM and LIPLAM in low concentrations failed to stimulate the release of TNFa or IL-12 by CD1þ APC (data not shown). In searching for an alternative approach to quantify the biological activity of LIPLAM, we measured the protein expression of PPARg in CD1þ APCs. PPARg belongs to the superfamily of lipid-activator nuclear receptors and the expression in macrophages is modulated by purified LAM [38]. Untreated CD1þ APCs expressed basal protein levels of PPARg (Figure 4A) as has been reported for IL-4/IL13-mediated alternatively activated macrophages [39]. Purified

LAM significantly reduced PPARg protein levels in a dose dependent manner as determined by western blotting (Figure 4A). Incubation of CD1þ APCs with purified LAM or LIPLAM induced a comparable dose-dependent inhibition of PPARg protein expression (LAM: 21 ± 10%; LIPLAM 27 ± 11%) (Figure 4B, C) while empty liposomes did not affect the expression of PPARg (Suppl. Figure 4). Therefore the biological activity of LAM is retained in LIPLAM. 3.5. LIPLAM induces CD1b-restricted IFNg release in human Tlymphocytes After we had demonstrated that LIPLAM is efficiently taken up by CD1þ APC and exerts biological activity on the host cell, we next

Please cite this article in press as: Kallert S, et al., Liposomal delivery of lipoarabinomannan triggers Mycobacterium tuberculosis specific T-cells, Tuberculosis (2015), http://dx.doi.org/10.1016/j.tube.2015.04.001

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

YTUBE1322_proof ■ 22 May 2015 ■ 6/11

6

S. Kallert et al. / Tuberculosis xxx (2015) 1e11

Figure 3. Uptake of LIPLAM by CD1þ APCs. (A) 0.1  106 CD1þ APCs were incubated overnight with unstained- or rhodamine-LIPLAM (1:100) in an 8 chamber slide and fixed with PFA. The pictures show Z-Layer images of representative areas of three independent experiments. (B) 0.2  106 CD1þ APCs and 0.2  106 non-adherent PBMCs and PBMCs were incubated overnight with medium alone or rhodamine-LIPLAM (1:100). After 16 h cells were harvested and stained for MHC class II-APC, CD3-PerCP or CD19. The results of one out of three independent experiments are shown. (C) 0.2  106 CD1þ APCs and 0.2  106 non-adherent PBMCs were incubated overnight with medium alone or increasing dilutions of LIPLAM and stained for MHCII- and CD3-expression after 16 h. The average percentage þ SD of MHCIIþ/rhodamine positive cells from three different donors is shown and in (D) the average mean fluorescence intensity (MFI) is shown. Asterisk indicates significant differences as compared to the untreated controls (*p < 0.05; **p < 0.005). (E) 0.1  105 human fibroblasts or HeLa cells were incubated with medium alone or rhodamine-LIPLAM for 16 h and analyzed by fluorescence microscopy as well as flow cytometry.

Please cite this article in press as: Kallert S, et al., Liposomal delivery of lipoarabinomannan triggers Mycobacterium tuberculosis specific T-cells, Tuberculosis (2015), http://dx.doi.org/10.1016/j.tube.2015.04.001

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

YTUBE1322_proof ■ 22 May 2015 ■ 7/11

S. Kallert et al. / Tuberculosis xxx (2015) 1e11

7

Figure 3. (continued).

tested whether LIPLAM activates antigen-specific T-cells. We screened 30 healthy students from regions in which tuberculosis is endemic (Eastern Europe, India, China) for the presence of T-cells reactive to a crude mixture of mycobacterial lipids [40]. Seven (23%) students showed a robust, reproducible T-cell activation as measured by IFNg release in response to mycobacterial lipids and were further evaluated for LAM-reactivity. Of these seven, five responded to LAM (71%) and were recruited for the experiments involving LIPLAM induced T-cell activation. CD1þ APCs and autologous non-adherent PBMCs (ratio of 1:4) were incubated with increasing concentrations of LIPLAM and IFNg release was measured after overnight incubation. LIPLAM induced IFNg-release in a dose-dependent manner in the donors tested (Figure 5A). The percentage of IFNg producing cells also increased on average from 0.06 ± 0.03% in unstimulated cells to 0.74 ± 0.14% after stimulation with LIPLAM (Figure 5B). This effect was also dose-dependent as determined by intracellular cytokine staining (Figure 5C). IFNg release was LAM-specific because neutralizing antibodies to CD1b significantly reduced the amount of IFNg (Figure 5D). To provide formal evidence that T-cells are the source for IFNg, we recruited additional LIPLAM responsive donors and combined the IFNg staining of LIPLAM-stimulated cells with antibodies directed against T- and Natural Killer cells, the two candidate populations for IFNg release in non-adherent PBMCs. In both donors IFNg was expressed by CD3þ T-cells and NK cells contributed to a variable extent (Figure 5E). Together with our findings that CD1b neutralizing antibodies reduce LIPLAM mediated IFNg release, this provides formal evidence that LIPLAM triggers antigen specific IFNg expression in CD3þ T-cells. 3.6. Liposomal delivery promotes LAM-specific T-cell activation To determine whether liposomal delivery improves LAMspecific activation of T-cells, we compared the IFNg induction of LIPLAM, purified LAM and LIP. LIPLAM was more effective in inducing IFNg release in primary human cells as compared to purified LAM or LIP (Figure 6A). In parallel the percentage of IFNg

producing lymphocytes was tenfold increased by stimulation with LIPLAM as compared to purified LAM (Figure 6B). Taken together our results demonstrate that purified LAM can be efficiently delivered into macrophages via liposomes providing an attractive strategy to boost glycolipid-specific T-cell responses. 4. Discussion Efficient delivery of vaccine antigens into macrophages is a crucial step for the induction of protective memory T-cell responses. Here we introduce a liposome-based delivery system for glycolipid antigens that induces robust Th1-biased T-cell responses in primary human cells. This system provides a preclinical platform for the optimization of liposome/glycolipid conjugates and will facilitate the selection of vaccine candidates to be pushed forward to sophisticated immunogenicity- and challenge experiments in animals. We selected LAM as a model antigen for our study because glycolipid specific CD1 restricted T-cells are associated with protective immunity in human tuberculosis, albeit direct evidence is still lacking [41]. LAM specific T-cells exert various effector functions for controlling mycobacterial infections including the secretion of Th1-cytokines, lysing infected host cells and directly killing the intracellular pathogen [27e29,42]. One evasion mechanism of Mtb which highlights the importance of lipid specific T-cell responses is the down-regulation of group 1 CD1 expression in infected macrophages thereby suppressing T-cell activation [43]. In the context of vaccination it is noteworthy that natural Mtb-infection induces long-lasting memory T-cells which can be detected more than 6 years after infection and are absent in Mtb-naïve individuals (unpublished observation). Besides the well-known adjuvant effects of liposomes on innate immunity [44] this study raises possibilities to improve antigen presentation and T-cell activation. First, the framework structure of the liposomes provides a biochemical matrix that can be modified to improve the targeting, uptake and shuttling of the immunogenic compound into the antigen presenting machinery. The liposomes

Please cite this article in press as: Kallert S, et al., Liposomal delivery of lipoarabinomannan triggers Mycobacterium tuberculosis specific T-cells, Tuberculosis (2015), http://dx.doi.org/10.1016/j.tube.2015.04.001

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

YTUBE1322_proof ■ 22 May 2015 ■ 8/11

8

S. Kallert et al. / Tuberculosis xxx (2015) 1e11

Figure 4. LIPLAM down regulates PPARg protein expression. (A) 1.5  106 CD1þ APCs were treated with medium or LAM for 48 h and PPARg expression was analyzed by western blot using a monoclonal PPARg-antibody (1:1000). The blot was stripped and re-probed using a b-Actin antibody (1:1000) as a loading control. One representative blot out of five independent experiments with different donors is shown. (B) 1.5  106 CD1þ APCs were treated with LAM or LIPLAM for 48 h and analyzed for PPARg protein expression. The panel shows one representative western blot out of four. (C) The western blots were analyzed by densitometry based on the b-Actin protein expression. The value of the untreated control was set to 100% (0% inhibition) and the inhibition of the PPARg protein expression was calculated based on this value. The average result of the inhibition of PPARg protein expression by LAM and LIPLAM þ SD from four independent experiments is shown.

used in this study consisted of egg phosphatidylcholine, cholesterol and stearylated octaarginine. Octaarginine is a cell-penetrating peptide [45] consisting of eight positively charged arginines that support the delivery of a spectrum of molecules ranging from DNA over peptides to nanoparticles and liposomes [34e36,46]. Here, we demonstrate that stearylated octaarginine improves the uptake of a glycolipid antigen into CD1þ APCs (Suppl. Figure 2). In vivo administration of octaarginine-containing liposomes induced a delayed type hypersensitivity reaction in the skin of BCG primed guinea pigs [47] and a Th1-skewed immune response as well as upregulation of the antimicrobial peptide granulysin in rhesus macaques [48,49] using glucose monomycolate (GMM) as a vaccine antigen. Side by side comparison of glycolipid containing liposomes ± octaarginine is required to determine the direct contribution of this compound to immunogenicity in vivo.

Secondly, liposomes can be designed to direct vaccine antigens to specific cell types. In a very recent study GMM-containing liposomes were decorated with Siglec-7, a ligand specific for dendritic cells. This elegant approach enhanced the GMM specific response of a human T-cell clone [50]. Our study indicates that this approach can be extended to primary human T-cells and encourages the implementation of this strategy to additional glycolipid antigens such as LAM. Finally, liposomes improve the activation of classical cytotoxic Tcells (CTL) since the antigens are more efficiently shuttled into the MHC class I pathway [51e54]. Cytotoxic T-lymphocytes are increasingly acknowledged as an important component of protective immunity against intracellular bacteria including Mtb. Major functions include the release of Th1-cytokines, lysis of Mtb-infected host cells and direct antimicrobial activity [55]. Glycolipids,

Please cite this article in press as: Kallert S, et al., Liposomal delivery of lipoarabinomannan triggers Mycobacterium tuberculosis specific T-cells, Tuberculosis (2015), http://dx.doi.org/10.1016/j.tube.2015.04.001

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

YTUBE1322_proof ■ 22 May 2015 ■ 9/11

S. Kallert et al. / Tuberculosis xxx (2015) 1e11

9

Figure 5. LIPLAM induces CD1b-restricted IFNg release in human T-Lymphocytes. (A) 0.4  106 non-adherent PBMCs and 0.1  106 autologous CD1þ APCs were incubated overnight with increasing dilutions of LIPLAM and the concentration of IFNg in the supernatant was measured by ELISA. LIPLAM induced IFNg release of two individual donors is shown. (B) The graph gives the percentage of IFNgþ lymphocytes after stimulation with LIPLAM (1:10) of three individual donors. (C) PBMC and autologous CD1þ APC were stimulated overnight with increasing dilutions of LIPLAM. The percentage of IFNg positive cells within the lymphocyte gate was determined by intracellular cytokine staining. One representative donor out of three with comparable results is shown. (D) CD1þ APCs were pre-incubated for 1 h with CD1b-neutralizing antibodies (10 mg/ml) before autologous nonadherent PBMCs at a ratio of 1:4 and LIPLAM (1:10) were added overnight. The effect of the CD1b-neutralizing antibody on the LIPLAM induced IFNg release was calculated by setting the IFNg release without CD1b neutralizing antibody as 100%. The figure shows the average result ± SD of three independent experiments. (E) PBMC and autologous CD1þ APC were stimulated overnight with LIPLAM and stained for CD3, CD56 and IFNg. The graph gives one representative result out of two.

Please cite this article in press as: Kallert S, et al., Liposomal delivery of lipoarabinomannan triggers Mycobacterium tuberculosis specific T-cells, Tuberculosis (2015), http://dx.doi.org/10.1016/j.tube.2015.04.001

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

YTUBE1322_proof ■ 22 May 2015 ■ 10/11

10

S. Kallert et al. / Tuberculosis xxx (2015) 1e11

Figure 6. Liposomal delivery of purified LAM promotes IFNg release. (A) CD1þ APCs and autologous non-adherent PBMCs were incubated at a ratio of 1:4 overnight with LAM (10 mg/ml), LIPLAM (1:10) or LIP (1:10) and the concentration of IFNg in the supernatant was measured by ELISA. One representative result out of five with comparable results is shown. (B) CD1þ APCs and autologous non-adherent PBMCs were stimulated overnight and intracellular IFNg expression was determined by intracellular staining. One representative donor out of three with comparable results is shown.

including LAM, also induce CTL with antimicrobial activity [28,42] offering the intriguing possibility to combine protein and glycolipid antigens in liposomes to achieve synergistic MHC class I- and CD1b-triggered CTL activity. Ongoing experiments are designed to characterise the phenotype and function of LIPLAM-responsive Tcells, a task which is complicated by the limited access to LAMreactive donors and the low frequency of antigen-specific Tlymphocytes. In summary our study demonstrates that LAM-mediated activation of primary human T-cells is improved by liposomal delivery. The simple composition of the liposomes used here provides a starting basis for future experiments to optimize the uptake, intracellular transport and presentation of glycolipid antigens with the ultimate vision of inducing protective and long lasting memory T-cell responses in humans. Conflicts of interest None. Acknowledgements We are grateful for the technical support and advice of Yvonne Perrie, Goutam Pramanik, Benjamin Hagemann and Daniel Mayer

as well as to Christian Sinzger (University Hospital, Ulm) for supplying human foreskin fibroblasts. This work was funded by the 7th frame work project (NewTBVAC) and the Horizon 2020 program (TBVAC2020) of the European Union. S.K. was supported by a doctoral fellowship by the Land Baden-Württemberg (“Kooperative Promotionskolleg Pharmazeutische Biotechnologie der Hochschule Biberach und der Universit€ at Ulm”) and the International Graduate School in Molecular Medicine Ulm. Funding:

None.

Competing interests: Ethical approval:

None declared. Not required.

Appendix A. Supplementary material Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.tube.2015.04.001. References [1] Crispen R. History of BCG and its substrains. Prog Clin Biol Res 1989;310: 35e50.

Please cite this article in press as: Kallert S, et al., Liposomal delivery of lipoarabinomannan triggers Mycobacterium tuberculosis specific T-cells, Tuberculosis (2015), http://dx.doi.org/10.1016/j.tube.2015.04.001

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58

YTUBE1322_proof ■ 22 May 2015 ■ 11/11

S. Kallert et al. / Tuberculosis xxx (2015) 1e11 [2] Moore DF. The history and development of the BCG. Practitioner 1983;227(1376):ieiii. [3] Andersen P, Kaufmann SH. Novel vaccination strategies against tuberculosis. Cold Spring Harb Perspect Med 2014;4(6). [4] Montagnani C, et al. Vaccine against tuberculosis: what's new? BMC Infect Dis 2014;14(Suppl. 1):S2. [5] McShane H, et al. Recombinant modified vaccinia virus Ankara expressing antigen 85A boosts BCG-primed and naturally acquired antimycobacterial immunity in humans. Nat Med 2004;10(11):1240e4. [6] Beveridge NE, et al. Immunisation with BCG and recombinant MVA85A induces long-lasting, polyfunctional Mycobacterium tuberculosis-specific CD4þ memory T lymphocyte populations. Eur J Immunol 2007;37(11):3089e100. [7] Verreck FA, et al. MVA.85A boosting of BCG and an attenuated, phoP deficient M. tuberculosis vaccine both show protective efficacy against tuberculosis in rhesus macaques. PLoS ONE 2009;4(4):e5264. [8] Tameris MD, et al. Safety and efficacy of MVA85A, a new tuberculosis vaccine, in infants previously vaccinated with BCG: a randomised, placebo-controlled phase 2b trial. Lancet 2013;381(9871):1021e8. [9] White AD, et al. Evaluation of the safety and immunogenicity of a candidate tuberculosis vaccine, MVA85A, delivered by aerosol to the lungs of macaques. Clin Vaccine Immunol 2013;20(5):663e72. [10] Dietrich J, et al. Exchanging ESAT6 with TB10.4 in an Ag85B fusion moleculebased tuberculosis subunit vaccine: efficient protection and ESAT6-based sensitive monitoring of vaccine efficacy. J Immunol 2005;174(10):6332e9. [11] Aagaard C, et al. Protection and polyfunctional T cells induced by Ag85BTB10.4/IC31 against Mycobacterium tuberculosis is highly dependent on the antigen dose. PLoS ONE 2009;4(6):e5930. [12] Billeskov R, et al. Comparing adjuvanted H28 and modified vaccinia virus ankara expressing H28 in a mouse and a non-human primate tuberculosis model. PLoS ONE 2013;8(8):e72185. [13] Reed SG, et al. Defined tuberculosis vaccine, Mtb72F/AS02A, evidence of protection in cynomolgus monkeys. Proc Natl Acad Sci U S A 2009;106(7): 2301e6. [14] Day CL, et al. Induction and regulation of T-cell immunity by the novel tuberculosis vaccine M72/AS01 in South African adults. Am J Respir Crit Care Med 2013;188(4):492e502. [15] Leroux-Roels I, et al. Improved CD4(þ) T cell responses to Mycobacterium tuberculosis in PPD-negative adults by M72/AS01 as compared to the M72/ AS02 and Mtb72F/AS02 tuberculosis candidate vaccine formulations: a randomized trial. Vaccine 2013;31(17):2196e206. [16] Bertholet S, et al. A defined tuberculosis vaccine candidate boosts BCG and protects against multidrug-resistant Mycobacterium tuberculosis. Sci Transl Med 2010;2(53):53e74. [17] Coler RN, et al. Therapeutic immunization against Mycobacterium tuberculosis is an effective adjunct to antibiotic treatment. J Infect Dis 2013;207(8): 1242e52. [18] Weinrich Olsen A, et al. Protection of mice with a tuberculosis subunit vaccine based on a fusion protein of antigen 85b and esat-6. Infect Immun 2001;69(5): 2773e8. [19] van Dissel JT, et al. Ag85B-ESAT-6 adjuvanted with IC31 promotes strong and long-lived Mycobacterium tuberculosis specific T cell responses in naive human volunteers. Vaccine 2010;28(20):3571e81. [20] Aagaard C, et al. A multistage tuberculosis vaccine that confers efficient protection before and after exposure. Nat Med 2011;17(2):189e94. [21] Brennan PJ, Nikaido H. The envelope of mycobacteria. Annu Rev Biochem 1995;64:29e63. [22] Prigozy TI, et al. The mannose receptor delivers lipoglycan antigens to endosomes for presentation to T cells by CD1b molecules. Immunity 1997;6(2):187e97. [23] Sugita M, et al. Cytoplasmic tail-dependent localization of CD1b antigenpresenting molecules to MIICs. Science 1996;273(5273):349e52. [24] Ernst WA, et al. Molecular interaction of CD1b with lipoglycan antigens. Immunity 1998;8(3):331e40. [25] Mishra AK, et al. Lipoarabinomannan and related glycoconjugates: structure, biogenesis and role in Mycobacterium tuberculosis physiology and hostepathogen interaction. FEMS Microbiol Rev 2011;35(6):1126e57. [26] Moody DB, et al. Structural requirements for glycolipid antigen recognition by CD1b-restricted T cells. Science 1997;278(5336):283e6. [27] Porcelli S, Morita CT, Brenner MB. CD1b restricts the response of human CD48-T lymphocytes to a microbial antigen. Nature 1992;360(6404):593e7. [28] Stenger S, et al. Differential effects of cytolytic T cell subsets on intracellular infection. Science 1997;276(5319):1684e7.

11

[29] Gilleron M, et al. Diacylated sulfoglycolipids are novel mycobacterial antigens stimulating CD1-restricted T cells during infection with Mycobacterium tuberculosis. J Exp Med 2004;199(5):649e59. [30] Hiromatsu K, et al. Induction of CD1-restricted immune responses in guinea pigs by immunization with mycobacterial lipid antigens. J Immunol 2002;169(1):330e9. [31] Bangham AD, Horne RW. Negative staining of phospholipids and their structural modification by surface-active agents as observed in the electron microscope. J Mol Biol 1964;8:660e8. [32] Gregoriadis G, Leathwood PD, Ryman BE. Enzyme entrapment in liposomes. FEBS Lett 1971;14(2):95e9. [33] Madni A, et al. Liposomal drug delivery: a versatile platform for challenging clinical applications. J Pharm Pharm Sci 2014;17(3):401e26. [34] Nakamura T, et al. Efficient MHC class I presentation by controlled intracellular trafficking of antigens in octaarginine-modified liposomes. Mol Ther 2008;16(8):1507e14. [35] Hattori Y, et al. Glycerol monomycolate, a latent tuberculosis-associated mycobacterial lipid, induces eosinophilic hypersensitivity responses in guinea pigs. Biochem Biophys Res Commun 2011;409(2):304e7. [36] Khalil IA, et al. High density of octaarginine stimulates macropinocytosis leading to efficient intracellular trafficking for gene expression. J Biol Chem 2006;281(6):3544e51. [37] Schlesinger LS, Hull SR, Kaufman TM. Binding of the terminal mannosyl units of lipoarabinomannan from a virulent strain of Mycobacterium tuberculosis to human macrophages. J Immunol 1994;152(8):4070e9. [38] Almeida PE, et al. PPARgamma expression and function in mycobacterial infection: roles in lipid metabolism, immunity, and bacterial killing. PPAR Res 2012;2012:383829. [39] Gordon S. Alternative activation of macrophages. Nat Rev Immunol 2003;3(1): 23e35. [40] Bastian M, et al. Mycobacterial lipopeptides elicit CD4þ CTLs in Mycobacterium tuberculosis-infected humans. J Immunol 2008;180(5):3436e46. [41] Van Rhijn I, Ly D, Moody DB. CD1a, CD1b, and CD1c in immunity against mycobacteria. Adv Exp Med Biol 2013;783:181e97. [42] Stenger S, et al. An antimicrobial activity of cytolytic T cells mediated by granulysin. Science 1998;282(5386):121e5. [43] Stenger S, Niazi KR, Modlin RL. Down-regulation of CD1 on antigenpresenting cells by infection with Mycobacterium tuberculosis. J Immunol 1998;161(7):3582e8. [44] Christensen D, et al. Liposome-based cationic adjuvant formulations (CAF): past, present, and future. J Liposome Res 2009;19(1):2e11. [45] Futaki S, et al. Arginine-rich peptides. An abundant source of membranepermeable peptides having potential as carriers for intracellular protein delivery. J Biol Chem 2001;276(8):5836e40. [46] Gupta B, Levchenko TS, Torchilin VP. Intracellular delivery of large molecules and small particles by cell-penetrating proteins and peptides. Adv Drug Deliv Rev 2005;57(4):637e51. [47] Komori T, et al. A microbial glycolipid functions as a new class of target antigen for delayed-type hypersensitivity. J Biol Chem 2011;286(19):16800e6. [48] Morita D, et al. Major T cell response to a mycolyl glycolipid is mediated by CD1c molecules in rhesus macaques. Infect Immun 2013;81(1):311e6. [49] Morita D, et al. Th1-skewed tissue responses to a mycolyl glycolipid in mycobacteria-infected rhesus macaques. Biochem Biophys Res Commun 2013;441(1):108e13. [50] Kawasaki N, et al. Targeted delivery of mycobacterial antigens to human dendritic cells via Siglec-7 induces robust T cell activation. J Immunol 2014;193(4):1560e6. [51] Alving CR, Wassef NM. Cytotoxic T lymphocytes induced by liposomal antigens: mechanisms of immunological presentation. AIDS Res Hum Retroviruses 1994;10(Suppl. 2):S91e4. [52] Rao M, et al. Visualization of peptides derived from liposome-encapsulated proteins in the trans-Golgi area of macrophages. Immunol Lett 1997;59(2): 99e105. [53] Richards RL, et al. Liposomes containing lipid A serve as an adjuvant for induction of antibody and cytotoxic T-cell responses against RTS, S malaria antigen. Infect Immun 1998;66(6):2859e65. [54] Korsholm KS, et al. Induction of CD8þ T-cell responses against subunit antigens by the novel cationic liposomal CAF09 adjuvant. Vaccine 2014;32(31): 3927e35. [55] Cooper AM. T cells in mycobacterial infection and disease. Curr Opin Immunol 2009;21(4):378e84.

Please cite this article in press as: Kallert S, et al., Liposomal delivery of lipoarabinomannan triggers Mycobacterium tuberculosis specific T-cells, Tuberculosis (2015), http://dx.doi.org/10.1016/j.tube.2015.04.001

59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116