Synthesis, Liposomal Formulation, and Immunological Evaluation of a

Article Cite This: J. Med. Chem. 2018, 61, 4918−4927

pubs.acs.org/jmc

Synthesis, Liposomal Formulation, and Immunological Evaluation of a Minimalistic Carbohydrate-α-GalCer Vaccine Candidate Felix Broecker,†,‡,§,▽ Sebastian Götze,†,‡,▽ Jonathan Hudon,†,▽,¶ Dominea C. K. Rathwell,†,∥ Claney L. Pereira,†,⊥ Pierre Stallforth,†,# Chakkumkal Anish,†,○ and Peter H. Seeberger*,†,‡ †

Department of Biomolecular Systems, Max Planck Institute of Colloids and Interfaces, Am Mühlenberg 1, 14424 Potsdam, Germany Institute of Chemistry and Biochemistry, Freie Universität Berlin, Arnimallee 22, 14195 Berlin, Germany

‡

J. Med. Chem. 2018.61:4918-4927. Downloaded from pubs.acs.org by UNIV OF TOLEDO on 06/17/18. For personal use only.

S Supporting Information *

ABSTRACT: Fully synthetic glycan-based vaccines hold great potential as preventive and therapeutic vaccines against infectious diseases as well as cancer. Here, we present a two-component platform based on the facile conjugation of carbohydrate antigens to α-galactosylceramide (α-GalCer) to yield fully synthetic vaccine candidates. Formulation of the cancer-associated Tn antigen glycolipid model vaccine candidate into liposomes of different sizes and subsequent immunization of mice generated specific, high-affinity antibodies against the carbohydrate antigen with characteristics of T cell-dependent immunity. Liposome formulation elicited more reproducible glycan immunity than a conventional glycoconjugate vaccine bearing the same glycan antigen did. Further evaluation of the immune response revealed that the size of the liposomes influenced the glycan antibody responses toward either a cellular (Th1) or a humoral (Th2) immune phenotype. The glycolipid vaccine platform affords strong and robust antiglycan antibody responses in vivo without the need for an external adjuvant.

■

INTRODUCTION Novel preventive and therapeutic vaccines are urgently needed to fight antibiotic-resistant bacteria1−3 as well as different forms of cancer.4 Surface-exposed carbohydrates on pathogens and tumor cells are particularly promising antigens for vaccines against infectious diseases or cancer, respectively. Carbohydrates are usually weak, T cell independent antigens that require presention on immunogenic carrier molecules to stimulate T cell help and induce long-lasting antibody responses.5 Conventionally, isolated carbohydrate antigens are coupled to immunogenic protein carriers to furnish glycoconjugates. These are the constituents of licensed vaccines against various bacterial pathogens including Streptococcus pneumoniae, Neisseria meningitidis, and Haemophilus inf luenzae type B.6 Glycoconjugate vaccines, however, require external adjuvants to efficiently confer immunity7 and constant cooling to prevent protein degradation or aggregation.8 Moreover, glycoconjugates prepared with small oligosaccharides or selfantigenic tumor-associated carbohydrate antigens (TACA) frequently fail to induce glycan-specific antibodies due to the immunodominant carrier protein.9 Oligosaccharides derived from microorganisms10 or tumor tissue11 have been shown to elicit potent anti-infective and tumor-killing antibody responses, respectively, both in model animals and humans. Yet, in many cases carbohydrate antigens cannot be accessed by isolation in sufficient purity for clinical use.12,13 In efforts to overcome © 2018 American Chemical Society

these limitations, various fully synthetic vaccine candidates based on microbial or tumor-associated carbohydrate antigens have been reported. For example, three-component vaccine candidates composed of a carbohydrate antigen B cell epitope, a peptidic T cell epitope, and a Toll-like receptor (TLR) agonist that serves as internal adjuvant have been shown to promote strong antiglycan antibody responses in mice.14,15 The immunogenicity of small glycans also benefits from their multivalent presentation on peptide carriers,16,17 virus particles,18,19 or self-assembling structures.20 Another method to boost the immunological activity of oligosaccharides is direct conjugation with a potent immune stimulant such as monophosphoryl lipid A (MPLA)21,22 or α-galactosylceramide (αGalCer),23 which furnishes peptide-free conjugate vaccine candidates. Synthetic vaccine candidates have the advantage of being single chemical entities that can be fully characterized with standard methods such as nuclear magnetic resonance (NMR) and high-resolution mass spectrometry (HRMS). Thus, they can be treated as a small molecule drug, enable rational structure−activity relationship (SAR) studies, and may simplify the vaccine application process of regulatory agencies. On the other hand, elaborate multicomponent vaccine candidates Received: February 23, 2018 Published: May 9, 2018 4918

DOI: 10.1021/acs.jmedchem.8b00312 J. Med. Chem. 2018, 61, 4918−4927

Journal of Medicinal Chemistry

Article

Scheme 1. Synthesis of Carbohydrate Conjugates 4, 5, and 6a

a

Reagents and conditions: (a) bis(4-nitrophenyl) carbonate, trimethylamine, pyridine, 79%; (b) 3, pyridine, trimethylamine, 42%; (c) pyridine, stearoyl chloride, 46%; (d) triethylamine, di-N-succinimidyl adipate dissolved in DMSO; (e) 7, 100 mM sodium phosphate buffer, pH 7.4. A characterization of 6 by SDS−PAGE and MALDI-TOF MS analysis is shown in Figure S1.

(∼120 nm) favored the production of IgG1 associated with Th2-mediated immunity. Thus, modulating the size of liposomes allows for directing antiglycan antibody responses toward either cellular (Th1) or humoral (Th2) immunity. Overall, this study establishes a vaccine platform based on fully synthetic saccharide-α-GalCer conjugates for the robust generation of high-affinity IgG even against the weakly immunogenic carbohydrate antigen Tn.

require multistep synthetic routes and a considerable amount of resources. Therefore, a simple platform that allows for easy access to fully synthetic vaccine candidates is highly desirable. Here, we report on the in vivo efficacy of a fully synthetic saccharide-glycolipid conjugate vaccine platform that allows for the multivalent display of glycan antigens with minimal synthetic effort. The weakly immunogenic cancer-associated Tn (α-N-acetylgalactosamine linked to serine or threonine)24,25 served as a model antigen that was efficiently coupled to the glycosphingolipid α-GalCer. This internal immune stimulatory moiety has recently been shown to promote strong antibody responses when coformulated26 or covalently coupled with polysaccharides23 or peptides.27,28 The liposome-formulated self-adjuvanting Tn-α-GalCer conjugate consistently induced strong anti-Tn antibody responses characterized by an immunoglobulin switch to IgG and affinity maturation in all immunized mice. Control liposomes prepared with stearoylated Tn lacking the α-GalCer moiety were also immunogenic but elicited lower antibody levels. In contrast, a semisynthetic glycoconjugate composed of the Tn antigen and the carrier protein CRM197 induced antiglycan IgG only in a subset of immunized mice. Larger liposomes (∼400 nm diameter) promoted Th1-type IgG2a antibodies, while smaller ones

■

RESULTS AND DISCUSSION

Chemical Synthesis of Carbohydrate−Lipid Conjugates. α-GalCer is a CD1d ligand and a potent stimulant of invariant natural killer T (iNKT) cells that enhances interactions with antigen presenting cells (APC) via CD40 signaling.29 This glycolipid has been extensively investigated in SAR studies as an adjuvant and as a component of synthetic vaccines.27,28,30,31 We reasoned that conjugation of α-GalCer to a defined synthetic carbohydrate antigen will furnish easily accessible but effective carbohydrate−lipid vaccine candidate conjugates to circumvent the elaborate synthetic effort required for other fully synthetic carbohydrate vaccines. Ideally, the αGalCer conjugate should be able to induce high antibody titers even against weakly immunogenic carbohydrate antigens of 4919

DOI: 10.1021/acs.jmedchem.8b00312 J. Med. Chem. 2018, 61, 4918−4927

Journal of Medicinal Chemistry

Article

Figure 1. Immunogenicity of liposomes and a glycoconjugate presenting the Tn antigen. (a) Immunization regime. Groups of six mice received three injections comprising 1.7 μg of Tn antigen at the indicated time points. Serum at week 5 was subjected to microarray-assisted serum IgG1 analysis. (b) Spotting pattern of microarray slides shown in panels c and d. Glycans were spotted at 0.25 and 0.5 mM (small and large circles, respectively) and CRM197 at 1 μM, all in quadruplicates. (c) Representative microarray scan showing the binding pattern of GS-I lectin to the spotted compounds. Blue signals represent fluorescence excited at 488 nm. (d) Exemplary microarray scans of mice immunized with the indicated antigen formulations at week 5. Each panel represents the IgG1 response of one individual mouse. Yellow signals represent fluorescence excited at 594 nm.

lesterol at 20.3:13.6:1 (DSPC/cholesterol/4 or 5) or 20.3:13.6:1:1 (DSPC/cholesterol: 5:1) molar ratios, respectively. The combination of DSPC and cholesterol for liposomal vaccines has been shown to induce strong antibody responses in small animals before.40 Liposomes with mean diameters (Zaverage values) of about 120 and 400 nm, as determined by dynamic light scattering (DLS), were obtained using filters with pore sizes of 100 or 400 nm (Figure S2). There was no difference in size distribution between the different antigens. Controls included micelles of 4 prepared by ultrasonication of a lipid suspension, the CRM197 glycoconjugate 6 formulated with aluminum hydroxide (Alum) adjuvant, soluble Tn 3 coformulated with 1 by ultrasonication, and soluble 3 alone. To assess the primary antibody response to the abovementioned antigen formulations, groups of six mice were immunized via the subcutaneous (s.c.) route three times in twoweek intervals, each dose containing 1.7 μg of Tn (Figure 1a). The serum IgG responses were followed by microarray analysis with slides presenting Tn antigen 3 and the disaccharide substructure of S. pneumoniae serotype 4 capsular polysaccharide 8 (α-GalNAc-(1 → 4)-α-Gal)41 as a control antigen (Figure 1b). Furthermore, CRM197 was spotted on the microarray slides to detect IgG to the carrier protein in mice immunized with 6. Successful spotting of 3 and 8 was verified with the Bandeiraea (Grif fonia) simplicifolia lectin-I (GS-I)42 that bound to both oligosaccharides (Figure 1c). All mice immunized with liposomes of 4 and 5 mounted consistent serum IgG responses to Tn antigen 3 without detectable cross-reactivity to 8 (Figure 1d). In contrast, only three of the six mice immunized with Alum-adjuvanted glycoconjugate 6 elicited IgG to 3, and the magnitudes of responses were more variable. The other three mice showed weak IgG binding to 8 that shares the terminal αGalNAc residue with the Tn antigen. Thus, liposomes induced more consistent and specific anti-3 IgG responses than did glycoconjugate 6. Differential IgG responses over time were quantified by microarray-inferred mean fluorescence intensity (MFI) signals of 3. A fourth immunization dose given after 21 weeks that contained 0.3 μg of Tn aimed at detecting memory responses.

synthetic origin. Hence, we selected the TACA Tn antigen as a model compound to validate the immunogenic properties of our α-GalCer based vaccine platform. It is known that the Tn antigen is only weakly immunogenic, especially in a monomeric form,32 and is usually coupled to immunogenic carrier proteins,33 peptidic T-helper cell epitopes,34 or TLR ligands35 to afford strong antibody responses in mammals. In order to create a molecule to be easily conjugated with synthetic carbohydrates that have been equipped with an amino linker for conjugation,36 linker-equipped α-GalCer 123 was reacted with bis(4-nitrophenyl) carbonate to yield activated αGalCer 2 in 79% yield (Scheme 1). The activated carbamate represents an ideal functional group for conjugation since carbohydrate antigens usually do not contain amino groups or other nucleophilic moieties.25,33,37,38 Thus, the subsequent attack of the amino linker equipped synthetic carbohydrate is selective. In addition, carbamate 2 can be purified by simple silica chromatography and stored at −20 °C for several days without decomposition. In the final step, synthetic Tn antigen 314 was coupled to carbamate 2 in a standard substitution reaction to form urea linkage and glycolipid 4 in 42% yield. The resulting glycolipid represents a fully synthetic, peptide-free, homogeneous vaccine candidate. Glycolipid 5 served as the control and was prepared by an amide bond forming reaction between Tn antigen 3 and stearoyl chloride in 46% yield. It was reasoned that comparing the immunogenicity of liposomes containing either 4 or 5 should reveal the effect of α-GalCer on total anti-Tn antibody levels, affinity maturation, and Th-skewing. Finally, a conventional glycoconjugate 6 that consists of the carrier protein CRM197 covalently linked to 3 via di-N-succinimidyl adipate cross-linker chemistry was prepared to enable comparison of the immune responses.39 Liposomes Multivalently Displaying the Tn Cancer Antigen Promote Long-Lasting IgG Responses in Vivo. For multivalent display of the Tn antigen, size-defined liposomes were prepared by lipid extrusion that contained 4, 5, or equimolar amounts of 5 and 1. The liposomes contained 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC) and cho4920

DOI: 10.1021/acs.jmedchem.8b00312 J. Med. Chem. 2018, 61, 4918−4927

Journal of Medicinal Chemistry

Article

Figure 2. Serum anti-Tn antibody responses in immunized mice. (a) Microarray-inferred serum IgG expressed as mean fluorescence intensity (MFI) values are shown. Each data point represents a mean ± SEM of six mice. Statistical significance was inferred by comparing IgG levels of mice immunized with 4-containing liposomes to those immunized with 5-containing liposomes. Note that week 21 and 22 IgG levels in response to micelles of 4 represent pooled serum (n = 6) instead of the mean of individual sera. (b) Comparison of IgG levels at week 5. Horizontal bars indicate mean values. (c) SPR-inferred antibody responses using pooled serum diluted 1:100. Levels of significance in panels a and b were inferred by twosided unpaired Student’s t-tests with *p ≤ 0.05 and **p ≤ 0.01.

Tn but an increase in anti-CRM197 IgG1 and IgG2a (Figure S3). The latter finding indicated that 6 elicited immunological memory to the carrier protein but not to the glycan antigen. The other two control groups immunized with coformulated 3 and 1 or soluble 3 did not mount any detectable anti-Tn IgG responses. Liposomes of 4 and 5 but not those of 5 coformulated with 1 proved to be suitable for robust induction of anti-Tn antibodies with hallmarks of T cell-dependency such as isotype switch to IgG and the formation of immunological memory. Therefore, we focused antibody analyses on these four groups of mice (liposomes of 4 or 5, each 120 or 400 nm-sized). To determine the effect of liposome size and the adjuvant effect of α-GalCer, we compared the magnitudes of primary IgG responses at week 5, 1 week after the third immunization, elicited by liposomes consisting of either 4 or 5 (Figure 2b). Induction of IgG1 did not benefit from the internal adjuvant, but IgG1 levels were higher for the smaller liposomes of both 4 and 5, albeit only with statistical significance in the case of 5. IgG2a responses were higher for the larger liposomes of 4 but not significantly increased when compared to liposomes of 5. In contrast, IgG3 levels were significantly higher in mice immunized with liposomes of 4 compared to 5. Liposomes with a diameter of 400 nm prepared with 4 induced the highest levels of IgG3 among all groups.

All mice immunized with liposomes containing 4 or 5 mounted primary anti-Tn IgG1, IgG2, and IgG3 responses (Figure 2a). Liposomes of 4 induced detectable IgG levels, especially IgG2a and IgG3, already 1 week after the first immunization, whereas with 5, antibodies appeared only after the second or third injection. While T-cell-dependent IgG responses are generally not expected before 2 weeks after initial immunization, it is known that murine Ig2a and IgG3 but not IgG1 can be produced without T cell help.43,44 Thus, it appeared that liposomes of 4 elicited both T-cell-dependent and T-cellindependent IgGs, whereas liposomes of 5 raised T-celldependent antibodies only. Compared to 5, 120 nm liposomes of 4 elicited higher IgG3 levels. At 400 nm size, 4 elicited higher primary anti-Tn IgG responses for all three subtypes, especially IgG3. IgG of all subtypes remained detectable for 21 weeks and could then be boosted with liposomes bearing 4 or 5, indicating the presence of memory B cells. In contrast, liposomes coformulated with 5 and 1 did not induce any detectable primary IgG response but a weak secondary response after immunization at week 21. Micelles of 4 elicited IgGs in all immunized mice with comparable kinetics and isotype distribution to the liposomes prepared with 4, and elicited a strong secondary response. The Alum-adjuvanted glycoconjugate 6 elicited anti-Tn IgG in three of the six immunized mice only, and there was no detectable secondary IgG response to 4921

DOI: 10.1021/acs.jmedchem.8b00312 J. Med. Chem. 2018, 61, 4918−4927

Journal of Medicinal Chemistry

Article

micelles of 4 and the Alum-adjuvanted glycoconjugate 6 in this analysis. The latter adjuvant is known to primarily induce Th2type IgG.46 Responses elicited by 400 nm liposomes showed higher IgG2a to IgG1 ratios than those induced by the smaller liposomes but without difference between 4 and 5. Antibodies raised with micelles of 4 ranged in-between. The lowest ratio was detected for 6 that elicited primarily IgG1 likely through the action of the Alum adjuvant. Finally, we investigated whether liposome size or the presence of α-GalCer influenced IgG affinity maturation. High-affinity antibodies are usually resistant to urea-mediated dissociation from the antigen. Therefore, we adapted an ELISA procedure to microarray analysis that includes a 10 min incubation step of bound serum IgG with 7 M urea to dissociate low-affinity antibodies.47 The ratio of measured binding signals to nonurea-treated serum samples serves as a measure of overall IgG affinity (Figure 3b). Affinity of IgG increased from week 3 (after second immunization) to week 5 (after third immunization) in pooled serum obtained from mice immunized with all four liposome preparations, whereby larger liposomes elicited IgG of higher affinity. The α-GalCer internal adjuvant did not appear to influence affinity maturation. The size-dependent effect on affinity maturation was verified by SPR analysis of pooled week 5 serum (Figure 3c). Binding affinity is affected by both association (ka) and dissociation (kd) rate constants, but determination of ka requires knowledge of the unknown serum antibody concentration. In contrast, kd is concentration-independent and recognized as the major parameter influencing antibody affinity.47,48 We therefore estimated the kd values by SPR through fitting the sensorgram curves during dissociation state. As expected, the larger liposomes elicited higher affinity antibodies (smaller kd values) when compared to the smaller liposomes.

Next, we compared the total primary antibody responses by using surface plasmon resonance (SPR) (Figure 2c). Pooled serum obtained at week 5 was passed through a sensor chip functionalized with Tn antigen 3. Preimmune serum served as control and did not show any binding signals to 3. Anti-Tn antibodies were detected in week 5 serum of all four groups. The magnitude of the binding signal expressed as response units (RUs) allowed for ranking the primary total antibody responses. The highest signals were detected for the 400 nm liposomes of 4 (∼800 RUs) followed by liposomes of 5 with the same size (∼400 RUs). The smaller liposomes led to weaker signals irrespective of the antigen (∼200 RUs for both 4 and 5). The SPR studies confirmed that liposomes of 4 with a diameter of 400 nm were the most potent inducer of anti-Tn antibodies. This was attributed mainly to strong IgG3 responses, as indicated by the microarray studies. Furthermore, serum IgA and/or IgM might contribute to total antibody responses as well. Size of Liposomes Influences Th Phenotype and Affinity Maturation. Next, we sought to evaluate the quality of IgG responses elicited by the different liposomes. One factor that influences the activity of IgG is the subclass distribution. In mice, IgG1 is primarily induced by Th2-type responses, whereas Th1 favors the generation of IgG2 and IgG3.45 Therefore, we studied the IgG2a to IgG1 ratio of IgG elicited by the liposomes after the third immunization (week 5) by dividing the respective fluorescent signal intensities (Figure 3a). We also included serum obtained from mice immunized with

■

CONCLUSIONS We disclose a fully synthetic vaccine platform based on readily accessible carbohydrate-α-GalCer conjugates that are multivalently displayed on liposomes. Using the weakly immunogenic Tn as the model antigen, the liposomes elicited strong and consistent antiglycan IgG responses in mice with hallmarks of T cell-dependent immunity such as isotype switching to IgG1 and affinity maturation, presumably mediated by iNKT cells. While glycoconjugate6 and three-component14,15 vaccines rely on classical T cells activated by MHC-presented peptides to generate B cell help, iNKT cells can help B cells to mount IgG responses with affinity maturation and isotype switch in the absence of peptidic T cell epitopes.26,43 Instead, the invariant T cell receptor of iNKT cells is activated by (glyco)lipids presented on the nonpolymorphic MHC class I molecule CD1d.43 Thus, B cell help was likely generated via activation of iNKT cells because the liposomes used in this study contain only glycolipids and no peptides. The antibody responses induced by liposomes were highly specific and more reproducible than those elicited by a conventional Tn-glycoconjugate vaccine candidate with the CRM197 carrier protein. Although generating a relevant antitumor cell response would require the presentation of Tn antigen within mucin peptide sequences,14,20,25,34 we demonstrate that our vaccine platform is suitable to generate robust IgG responses to small and weakly immunogenic carbohydrate antigens. Detailed investigation of the immune responses also revealed that the size of the liposomes influences antibody

Figure 3. IgG2a to IgG1 ratios and affinity maturation in immunized mice. (a) IgG2a to IgG1 ratios were calculated by dividing microarrayinferred binding signals of six mice per group except for 6 where the ratio was determined from three mice that raised anti-Tn IgG. Bars show the mean ± SEM. (b) Pooled serum was used to calculate the percentage of urea-resistant IgG to Tn antigen. Bars represent the mean ± SD of two experiments. (c) SPR-inferred dissociation rate constants (kd values). From top to bottom, pooled serum diluted 1:100, 1:200, 1:400, 1:800, and 1:1600 at week 5 was passed through the SPR sensor chip. Red lines indicate fitting of the dissociation stage that was used to determine the indicated kd values. 4922

DOI: 10.1021/acs.jmedchem.8b00312 J. Med. Chem. 2018, 61, 4918−4927

Journal of Medicinal Chemistry

Article

raphy). 1H- and 13C NMR were recorded either on a Varian 400 (400 MHz) or Varian 600 (600 MHz) in CDCl3, pyridine-d5, and CD3OD with chemical shifts referenced to internal standards CDCl3 (7.26 ppm 1 H, 77.1 ppm 13C), pyridine-d5 (8.74 ppm 1H, 150.4 ppm 13C), and CD3OD (3.31 ppm 1H, 49.0 ppm 13C) unless otherwise stated. Coupling constants are reported in Hertz (Hz). Splitting patterns are indicated as s, singlet; d, doublet; and t, triplet for 1H NMR data. Signals were assigned by means of 1H−1H COSY and 1H−13C HSQC spectra. ESI mass spectral analyses were performed by the MS service at the Institute of Chemistry and Biochemistry at the Free University of Berlin using a modified MAT 711 spectrometer (Varian MAT). All biologically tested compounds were highly pure as judged by 1H NMR spectroscopy. We did not assess their purity by qNMR; however, we here investigated vaccine candidates that act by stimulating the immune system, a mechanism of action not expected to be significantly affected by any possible minor impurities. Furthermore, the control in vivo experiments using 5 + 1 as separate compounds did not elicit any detectable immune response. This confirmed that the observed immune responses in the other groups could have only resulted from the glycolipids interacting with the immune system and were not due to nonspecific effects. Materials. The α-GalCer derivative 1 was prepared as reported by Cavallari et al.23 Tn antigen 3 was prepared according to procedures reported by Buskas et al.15 Disaccharide 8 was synthesized according to procedures reported by Geissner et al.41 (2S,3S,4R)-1-(6-(4-Nitrophenyl-6′-hexylcarbamate)-α-D-galactopyranosyl)-2-hexacosanoylaminooctadecane-3,4-diol (activated α-GalCer, 2). α-GalCer derivative 1 (3.9 mg, 4.1 μmol) was dissolved in pyridine (0.5 mL). Bis(4-nitrophenyl) carbonate (6.1 mg, 20 μmol) was added to this solution followed by triethylamine (25 μL, 0.179 mmol). The resulting yellow solution was stirred at room temperature for 12 h before it was evaporated to dryness. The crude product was purified by column chromatography (SiO2 column packed, pre-washed with 100% MeOH then washed with 100% CH2H2, product loaded, then eluting gradient of CH2H2/MeOH 0% → 20% MeOH) to yield glycolipid 2 (3.6 mg, 3.2 μmol, 79% yield) as a pale yellow oil. 1H NMR (400 MHz, pyridine-d5) δ 9.02 (t, J = 5.6 Hz, 1H, NH of carbamate), 8.49 (d, J = 8.7 Hz, 1H, NH of amide), 8.26 (d, J = 9.2 Hz, 2H), 7.54 (d, J = 9.2 Hz, 2H), 5.56 (d, J = 3.8 Hz, 1H, Gal-1), 5.30−5.24 (m, 2H), 4.73−4.61 (m, 2H), 4.51 (t, J = 6.0 Hz, 1H), 4.45−4.38 (m, 3H), 4.37−4.31 (m, 2H), 4.10 (dd, J = 9.8, 5.7 Hz, 1H), 4.02 (dd, J = 9.8, 6.5 Hz, 1H), 3.52 (td, J = 9.2, 2.7 Hz, 2H), 3.45 (dd, J = 13.0, 6.9 Hz, 2H), 2.46 (t, J = 7.2 Hz, 2H), 2.36− 2.23 (m, 1H), 1.97−1.79 (m, 4H), 1.74−1.65 (m, 3H), 1.64−1.54 (m, 2H), 1.48−1.18 (m, 74H), 0.87 (t, J = 6.7 Hz, 6H, Me of lipids); 13C NMR (101 MHz, pyridine-d5) δ 173.6 (amide), 157.8 (carbamate), 154.6, 145.2, 125.9, 123.0, 102.0 (Gal-1), 77.1, 73.0, 72.0, 71.9, 71.6, 71.3, 71.3, 70.7, 69.2, 51.8, 42.0, 37.3, 34.8, 32.6, 32.6, 30.9, 30.7, 30.6, 30.5, 30.5, 30.5, 30.4, 30.42, 30.40, 30.35, 30.28, 30.13, 30.12, 27.5, 27.0, 26.9, 26.7, 23.5, 14.8 (Me of lipids). m/z (ESI) found: [M + Na]+, 1144.8373 C63H115N3O13 requires [M + Na]+, 1144.8323. Tn Antigen-α-GalCer Conjugate (4). Activated α-GalCer 2 (9.1 mg, 8.11 μmol) was dissolved in pyridine (0.8 mL). Tn antigen 3 (4.8 mg, 11 μmol) was added to this solution. Afterward, triethylamine (2.5 μL, 18 μmol) was added to the reaction mixture that was stirred for 18 h. The reaction mixture was evaporated to dryness and redissolved in MeOH/CH2Cl2 (1 mL, 1:1). DOWEX 50WX-8 acidic resin (25 mg) was added, and the slurry was stirred for 15 min to remove any excess amine. The filtrate was evaporated to dryness, and the residue was purified by column chromatography (SiO2, CH2H2/MeOH 0% → 20% MeOH) to yield glycolipid 4 (4.8 mg, 3.4 μmol, 42% yield) as a white powder. 1H NMR (600 MHz, CDCl3/CD3OD, 1:1) δ 5.12 (d, J = 3.8 Hz, 1H, GalNAc-1), 5.09 (d, J = 3.8 Hz, 1H, Gal-1), 4.71 (d, J = 2.5 Hz, 1H, α-Thr), 4.52−4.43 (m, 2H, Gal-2, β-Thr), 4.41 (dd, J = 9.9, 4.2 Hz, 1H), 4.18−4.08 (m, 4H), 4.04 (dd, J = 10.0, 3.8 Hz, 1H, GalNAc-2), 4.02−3.92 (m, 4H), 3.88 (dd, J = 10.2, 5.3 Hz, 1H), 3.86− 3.81 (m, 2H), 3.81−3.77 (m, 1H), 3.75−3.70 (m, 2H), 3.47−3.28 (m, 6H, 3× C(O)NH-CH2), 2.44 (t, J = 7.6 Hz, 2H, HNC(O)-CH2 of fatty acid), 2.35 (s, 3H, Me of Ac), 2.29 (s, 3H, Me of Ac), 1.89− 1.78 (m, 6H), 1.75−1.69 (m, 2H) 1.64−1.44 (m, 79H), 1.11 (t, J = 7.0

titers, affinity maturation, as well as Th-skewing. Larger liposomes with a diameter of 400 nm elicited higher affinity Th1-type antibodies. The latter aspect is particularly intriguing as modifying the size of liposomes would allow for directing immune responses to the desired activity, for instance, to generate cytotoxic Th1 immunity against tumor cells or intracellular pathogens. On the other hand, smaller liposomes may be suitable to induce Th2 type immunity against extracellular pathogens. For protein antigens, it has been shown before that smaller liposomes (225 nm) Th1 responses in mice.49,50 This observation has been attributed to the ability of larger particles to promote secretion of the Th1-directing cytokine interleukin-12 in macrophages.49 Furthermore, larger particles (>225 nm) are subject to phagocytosis, mainly by macrophages and perhaps dendritic cell progenitors, whereas smaller ones (