Article Cite This: J. Am. Chem. Soc. 2017, 139, 14783-14791
pubs.acs.org/JACS
A Streptococcus pneumoniae Type 2 Oligosaccharide Glycoconjugate Elicits Opsonic Antibodies and Is Protective in an Animal Model of Invasive Pneumococcal Disease Madhu Emmadi,†,∥,⊥ Naeem Khan,†,∥ Lennart Lykke,†,∥ Katrin Reppe,§ Sharavathi G. Parameswarappa,†,⊥ Marilda P. Lisboa,†,⊥ Sandra-Maria Wienhold,§ Martin Witzenrath,§ Claney L. Pereira,*,†,⊥ and Peter H. Seeberger*,†,‡ †
Department of Biomolecular Systems, Max Planck Institute of Colloids and Interfaces, D-14424 Potsdam, Germany Department of Chemistry and Biochemistry, Freie Universität Berlin, Arnimallee 22, D-14195 Berlin, Germany § Department of Infectious Diseases and Pulmonary Medicine, Charité-Universitätsmedizin Berlin, Charitéplatz 1, 10117 Berlin, Germany ‡
S Supporting Information *
ABSTRACT: Invasive pneumococcal diseases (IPDs) remain the leading cause of vaccine-preventable childhood death, even though highly effective pneumococcal conjugate vaccines (PCVs) are used in national immunization programs in many developing countries. Licensed PCVs currently cover only 13 of the over 90 serotypes of Streptococcus pneumoniae (Sp), so nonvaccine serotypes are a major obstacle to the effective control of IPD. Sp serotype 2 (ST2) is such a nonvaccine serotype that is the main cause of IPD in many countries, including Nepal, Bangladesh, and Guatemala. Glycoconjugate vaccines based on synthetic oligosaccharides instead of isolated polysaccharides offer an attractive alternative to the traditional process for PCV development. To prevent the IPDs caused by ST2, we identified an effective ST2 neoglycoconjugate vaccine candidate that was identified using a medicinal chemistry approach. Glycan microarrays containing a series of synthetic glycans resembling portions of the ST2 capsular polysaccharide (CPS) repeating unit were used to screen human and rabbit sera and identify epitope hits. Synthetic hexasaccharide 2, resembling one repeating unit (RU) of ST2 CPS, emerged as a hit from the glycan array screens. Vaccination with neoglycoconjugates consisting of hexasaccharide 2 coupled to carrier protein CRM197 stimulates a T-cell-dependent B-cell response that induced CPS-specific opsonic antibodies in mice, resulting in killing of encapsulated bacteria by phagocytic activity. Subcutaneous immunization with neoglycoconjugate protected mice from transnasal challenge with the highly virulent ST2 strain NCTC 7466 by reducing the bacterial load in lung tissue and blood.
■
polysaccharides (CPS).8 However, the judicious selection of serotypes from among the more than 90 known serotypes9 for the formulation of PCVs is difficult. The prevailing capsular types vary with geography and age as well as over time, but fortunately, common serotypes are consistently identified around the world.10,11 Globally, about 20 serotypes are associated with >80% of IPDs occurring in all age groups, and the 13 most common serotypes cause approximately 75% of invasive disease in children.11 The licensed PCVs were designed to induce serotype-specific immunity for 65−90% of serotypes most frequently associated with IPDs in developed countries and are very effective in young children. The introduction of PCVs in developing countries12−15 has not
INTRODUCTION
Streptococcus pneumoniae (Sp), a Gram-positive encapsulated commensal bacterium, asymptomatically colonizes the human upper respiratory tract and is responsible for invasive pneumococcal diseases (IPDs) like pneumonia, septicemia, meningitis, and otitis media.1−3 Bacterial pneumonia in particular is globally the most common cause of vaccinepreventable deaths in younger children and the elderly,4 and IPDs in general are a major burden for the medical systems in both developed and developing countries. Pneumococcal diseases cause about 1.6 million deaths annually, mostly in developing countries.5 Pneumococcal conjugate vaccines (PCVs) saved the lives of millions of young children after the widespread emergence of antibiotic-resistant strains.6,7 Currently, licensed multivalent PCVs (Synflorix and Prevnar 13) are based on capsular © 2017 American Chemical Society
Received: August 21, 2017 Published: September 25, 2017 14783
DOI: 10.1021/jacs.7b07836 J. Am. Chem. Soc. 2017, 139, 14783−14791
Article
Journal of the American Chemical Society
new variant. Undoubtedly, inclusion of ST2 into marketed conjugate vaccines is essential to protect children from severe pneumococcal disease. To date, no efforts to create PCVs against ST2 using isolated CPS have been reported. Due to the problems associated with the isolation of natural CPS, synthetic oligosaccharides with precise length and structure have become an attractive alternative24,25 and have been shown to induce protective immune responses against Haemophilus influenzae type b, several pneumococcal serotypes, and Shigella.26−32 The repeating unit (RU), branched hexasaccharide 1 (Figure 1), of ST2 CPS33 was identified in 1988.34 To date, however, there has been no report concerning the synthesis of this glycan antigen and its immunological properties. Here, the first chemical synthesis of ST2 RU hexasaccharide 2 and its potential as a lead for the development of a semisynthetic conjugate vaccine against ST2 are described. A preparation containing the hexasaccharide CRM197−2 conjugate elicits a robust immune response in mice and promotes phagocytic killing of pneumococci. Active immunization and subsequent transnasal challenge with ST2 pneumococci revealed the potential of this neoglycoconjugate as a vaccine candidate.
Figure 1. Repeating unit of the S. pneumoniae serotype 2 CPS.
significantly reduced pneumococcal burden, since nonvaccine serotypes (NVS) are prevalent, and pneumococcal mortality and morbidity remain very high.16,17 Extending the serotype coverage of PCVs to include emerging and persistent NVS will help to reduce IPDs and save lives. Sp serotype 2 (ST2) is an emerging problem in many countries in Asia and Central America. IPD in children is caused by ST2 to a significant extent in low-income countries like Nepal (5%), Bangladesh (9%), and Guatemala (16%).18−21 Indeed, ST2 is the most prevalent serotype in Bangladesh, where it is responsible for 12% of meningitis cases.22 It is unclear whether the high meningitis rates are caused by the D39 strain isolated more than 90 years ago23 or an emerging
■
RESULTS AND DISCUSSION Oligosaccharide Assembly. In order to identify an oligosaccharide epitope that could serve as the basis for the development of a conjugate vaccine candidate,35,36 a series of oligosaccharides resembling the ST2 CPS was first synthesized. The β-rhamnoside, along with the 1,2-cis-glycosidic linkages and branching, renders hexasaccharide 1 (Figure 1), a
Figure 2. Retrosynthetic analysis of ST2 RU hexasaccharide 2. 14784
DOI: 10.1021/jacs.7b07836 J. Am. Chem. Soc. 2017, 139, 14783−14791
Article
Journal of the American Chemical Society Scheme 1. Synthesis of Disaccharides 4−6
challenging synthetic target. Retrosynthetic analysis of hexasaccharide repeating unit 2 provides guidance in accessing various related sequences for glycotope analyses using glycan microarrays (Figure 2). Target molecule 2 results from the global deprotection of fully protected hexasaccharide 3, which can be built up from disaccharide building blocks 4−6, via a [2 + 2 + 2] glycosylation strategy. These disaccharide units are in turn assembled from the monosaccharide building blocks Lrhamnose (7−9) and D-glucose (10−12) (see the Supporting Information). The synthesis of hexasaccharide 2 and related oligosaccharides commenced with the assembly of reducing end disaccharide acceptor 4 (Scheme 1a). Union of fully protected L-rhamnose thioglycoside 13 with acceptor 14 upon NIS/ TfOH activation was followed by cleavage of Fmoc by using triethylamine to afford disaccharide acceptor 8. Reducing end rhamnose disaccharide 4 was prepared by coupling rhamnose acceptor 8 with donor 7 and subsequent Fmoc cleavage (73% over two steps).
Figure 3. Identification of oligosaccharide epitopes by glycan microarray analysis of reference sera. (A) Synthetic glycans and native CPSs were immobilized on microarray slides, as described in the printing pattern and structure description on the right. Microarray slides were probed with (B) human reference sera (007sp) or (C) ST2 specific rabbit typing sera. To probe the binding specificity, both sera were incubated with either the negative control ST19F CPS (20 μg/ mL) or ST2 CPS (20 μg/mL) before glycan microarray analysis. Values of the mean fluorescence intensity in the presence or absence of CPS were compared. Data from triplicate determinations were plotted as mean ± SD minus background.
The challenging incorporation of β-rhamnosidic linkage37 required for the synthesis of disaccharide 5 (Figure 1) was
Scheme 2. Synthesis of Branched Hexasaccharide 2 RU of ST2
14785
DOI: 10.1021/jacs.7b07836 J. Am. Chem. Soc. 2017, 139, 14783−14791
Article
Journal of the American Chemical Society
Figure 4. Preparation and characterization of CRM197−2 neoglycoconjugate. (A) Schematic representation of CRM197−2 conjugate. Hexasaccharide 2 was covalently coupled with CRM197 using p-nitrophenyladipate ester as a coupling reagent. (B) MALDI-TOF analysis was used to determine the average molecular weight of the conjugate; CRM197 was used as a standard. (C) The CRM197−2 conjugate and CRM197 were resolved with 10% SDS−PAGE and stained with PageBlue protein staining solution. The molecular weight marker is indicated on the left.
Global deprotection of 3 delivered the conjugation-ready ST2 RU hexasaccharide 2. Using this synthetic strategy, various related oligosaccharide sequences were synthesized [Supporting Information(SI)]. Identification of an Oligosaccharide Vaccine Candidate by Glycan Microarray Experiments. Synthetic oligosaccharides resembling the ST2 RU (2), as well as related oligosaccharides (20−25), native ST2 CPS, ST19F CPS, cell wall polysaccharide (CWPS), and unrelated oligosaccharides as negative controls, were printed onto glycan microarray surfaces [Figures 3A and S1 (SI)] to screen for glycan-specific antibodies in the two commercially available reference sera, pooled human sera (007sp) and rabbit typing sera specific to serotype 2 CPS.41,42 The microarray data suggest that antibodies present in the human sera bind to hexasaccharide 2 and oligosaccharides 20−25 (Figure 3B), while those present in the rabbit typing sera bind hexasaccharide 2 and oligosaccharides 23 and 25 (Figure 3C), demonstrating that the α-D-GluA-(1→6)-α-D-Glc-(1→2) branch is an important substructure for strong antibody binding. Using an inhibition assay, the antibody binding specificity of the synthetic glycans on the array was confirmed by preincubating the sera with purified ST2 CPS (20 μg/mL). Indeed, binding of glycanspecific antibodies present in the sera declined significantly under these conditions, suggesting that these glycans mimic the epitope in native CPS; preincubation with unrelated ST19F CPS (20 μg/mL) failed to inhibit binding. The inhibition data suggest that antibodies binding to glycans 2, 23, and 25 that contain the disaccharide α-D-GluA-(1→6)-α-D-Glc-(1→2) branch feature were inhibited strongly compared to the oligosaccharides that lack this structure, thereby indicative of the importance of this substructure for antibody binding (Figure 3B,C). Taking into account that epitope length appears to play a role in antibody binding and that disaccharide 23 is an
addressed by installing a remote C3 picoloyl group for hydrogen-bond-mediated aglycon delivery in rhamnosyl thioglycoside 9 (Scheme 1 and the Supporting Information).38 Coupling of rhamnose 9 with glucose acceptor 10 afforded the desired β-linked disaccharide 15 (JC−H = 163 Hz) in 53% yield (Scheme 2b). The picoloyl group in 15 was removed by treatment with Cu(OAc)2, and the resulting C3′-hydroxyl group was acetylated to give disaccharide 16 (90% yield over two steps). Disaccharide 16 was converted into the corresponding glycosyl imidate over two steps by removal of the anomeric p-methoxyphenyl protecting group, followed by reaction with trichloroacetonitrile and DBU. The 1,2-cis linkage between glucose building blocks 1139 and 1240 was installed by in situ anomerization (Scheme 1c). Thioglycoside 11, by first converting to the corresponding glycosyl bromide, was then coupled with 12 in the presence of TBAI to afford the desired 1,2-cis-linked disaccharide 17 (JC−H = 174 Hz) in 52% yield. Under these conditions, elimination between two active methylene groups in the C6′-O-levulinoyl protecting group was observed. Removal of the ester protecting group from the C6′-O-position followed by oxidation of the primary hydroxyl group to the corresponding acid and esterification with methyl iodide afforded the desired disaccharide 6 (Scheme 1c). With disaccharides 4−6 in hand, the branched ST2 hexasaccharide repeating unit 2 was assembled (Scheme 2). Disaccharide imidate 5 served to glycosylate disaccharide 4 to give linear tetrasaccharide 18 in 68% yield (Scheme 2). Cleavage of the C2′-O-levulinoyl ester with hydrazine furnished tetrasaccharide acceptor 19, which was in turn coupled with thioglycoside 6 to afford fully protected hexasaccharide 3 (Scheme 2). The chelating solvent dioxane led to α-isomer exclusively, as confirmed by C−H coupled HSQC (JC−H = 170, 175, 175, 170, 158, and 166 Hz Supporting Information). 14786
DOI: 10.1021/jacs.7b07836 J. Am. Chem. Soc. 2017, 139, 14783−14791
Article
Journal of the American Chemical Society
Figure 5. Analysis of antibody titers in sera using glycan array and ELISA. (A) Immunization schedule with CRM197−2 conjugate. C57BL/6J mice (n = 5−8) were immunized subcutaneously with three doses of 2.2 μg hexasaccharide with or without alum. Control mice received either only PBS or alum in PBS. (B) The hyperimmune sera (day 35) were pooled and subjected to microarray analysis. (C) The end point titer of antibodies analyzed by ELISA. The CPS-specific (D) IgG, (E) IgM, (F) IgAI, and (G−J) IgG isotypes were also determined by ELISA. Antibody titers were determined in triplicate and plotted as mean ± SD.
humoral immune response was observed in the control groups. The glycan array data confirm that the CRM197−2 conjugate stimulates the production of ST2 CPS-specific antibodies. The CPS-specific antibody response raised against CRM197−2 conjugate was characterized in more detail by ELISA.46 Thus, 96-well microtiter plates were coated with ST2 CPS prior to incubation with mice sera raised against CRM197−2 conjugate with or without alum in serial dilutions (2-fold dilutions). PBS-immunized mice sera were used as a negative control. The ELISA data suggested that CRM197−2 conjugate induces very high titers of CPS-specific antibodies (Figure 5C). Quantitation of different classes of antibodies and IgG subtypes by end point ELISA revealed that mice immunized with the CRM197−2 conjugate formulated with alum showed class switching (Figure 5D−J), while those which received only conjugate stimulate primarily IgM antibodies (Figure 5E). None of the formulations induce the production of detectable amounts of IgA antibodies (Figure 5F). IgG1 and IgG3 contribute the bulk of the CPS-specific IgG titer, while a weak IgG2a and IgG2b response was observed only in mice vaccinated with CRM197−2 conjugate plus alum (Figure 5G− J). In contrast, mice vaccinated only with CRM197−2 conjugate fail to elicit CPS-specific IgG isotypes (Figure 5G− J). No antibodies were detected in the PBS control group sera, aside from some IgM background (Figure 5E). The observation that the CRM197−2 conjugate plus alum induces both CPS-
important substructure, we focused on RU hexasaccharide 2 as a synthetic antigen component of a vaccine candidate against ST2. Neoglycoconjugate CRM197−2 Induces a CPS-Specific Humoral Immune Response. Conjugation of oligosaccharide haptens to a carrier protein enhances glycan immunogenicity and provides T-cell help to induce somatic mutation and class switching of hapten-specific antibodies. Synthetic hexasaccharide 2 was conjugated to the nontoxic diphtheria toxin mutant CRM19743 that is commonly used as a carrier in licensed vaccines. Using p-nitrophenyladipate ester44 as a coupling agent (Figure 4A), 7.0 molecules of hexasaccharide 2 were covalently attached to each CRM197 molecule on average, as calculated by MALDI-TOF mass spectrometry (Figure 4B) and confirmed by SDS−PAGE (Figure 4C). The immunogenicity of CRM197−2 conjugate was assessed by immunizing female C57BL/6J mice (n = 5−8) with either the CRM197−2 conjugate mixed with alum or the conjugate alone in a prime boost regime (Figure 5A).45 The control groups received PBS or PBS plus alum. Post-immune sera (day 35) were analyzed using glycan arrays with fluorescently labeled goat anti-mouse secondary antibodies for detection [Figures 5B and S2 (SI)]. Mice immunized with CRM197−2 plus alum produce very high antibody titers, while mice vaccinated with conjugate exhibited only a weak antibody response. No 14787
DOI: 10.1021/jacs.7b07836 J. Am. Chem. Soc. 2017, 139, 14783−14791
Article
Journal of the American Chemical Society
Figure 6. Surface staining and the opsonophagocytic killing assay. (A) Anti-CRM197−2 antibodies bind to the surface of intact bacteria. The midlog-phase bacterial culture (OD600nm = 0.4−0.5) of serotype 2 (NCTC 7466) cells was stained with mouse polyclonal anti-CRM197−2 (black line histogram) or preimmune (dotted line histogram) sera. Alexa 635 conjugate goat anti-mouse IgG was used as secondary antibody (dashed histogram) and only bacterial cells (solid gray histogram) served as background controls. (B) Anti-CRM197−2 antibodies promote the phagocytosis of pneumococci. The model cell line (HL-60 cells) was incubated with NCTC 7466 cells preopsonized with hyperimmune sera raised against CRM197−2 conjugate with or without alum. Survival was assessed after 45 min incubation. Percent killing of pneumococci was calculated on the basis of viable pneumococcal colonies obtained relative to control sera. The background killing was analyzed by using PBS control sera. Data was represented as mean ± SD values in duplicates. (C) The antibody titer was calculated for 50% or more killing. Data were analyzed by the unpaired t test, and p values of
