Development of carbohydrate based next-generation anti-pertussis vaccines

Abstract

Pertussis is a highly contagious respiratory disease caused by the Gram-negative bacterial pathogen, Bordetella pertussis. Despite high global vaccination rates, pertussis is resurging worldwide. Here we discuss the development of current pertussis vaccines and their limitations, which highlight the need for new vaccines that can protect against the disease and prevent development of the carrier state, thereby reducing transmission. The lipo-oligosaccharide of Bp is an attractive antigen for vaccine development as the anti-glycan antibodies could have bactericidal activities. The structure of the lipo-oligosaccharide has been determined and its immunological properties analyzed. Strategies enabling the expression, isolation, and bioconjugation have been presented. However, obtaining the saccharide on a large scale with high purity remains one of the main obstacles. Chemical synthesis provides a complementary approach to accessing the carbohydrate epitopes in a pure and structurally well-defined form. The first total synthesis of the non-reducing end pertussis pentasaccharide is discussed. The conjugate of the synthetic glycan with a powerful immunogenic carrier, bacteriophage Qβ, results in high levels and long-lasting anti-glycan IgG antibodies, paving the way for the development of a new generation of anti-pertussis vaccines with high bactericidal activities and biocompatibilities.

1. Introduction

The obligate human pathogen Bordetella pertussis (Bp) is the etiological agent of whooping cough also known as pertussis.^1-2 Despite sustained efforts and high global vaccine coverage (85–99 %), this disease remains endemic worldwide, and has not been eradicated from any country.³ In this review, we will discuss pertussis vaccines with a focus on the efforts to develop newer carbohydrate-based vaccines. For descriptions of the microbiology, immunology, infection and disease characteristics, and pathogenesis of Bp, we refer readers to some excellent recent reviews.^4-8

A Bavarian physician, Adalbert Friedrich Marcus, cataloged more than 40 European names for pertussis.⁹ However, pertussis is commonly known as whooping cough or 100-day cough. The outbreaks of pertussis were first documented by de Baillou in the 16th century.¹⁰ Bp is a gramnegative, aerobic encapsulated coccobacillus of the genus Bordetella. After inhalation, Bp specifically adheres to the ciliated epithelium of the respiratory tract, replicates and produces an arsenal of virulence factors. Some of the well-studied virulence factors include pertussis toxin, adenylate cyclase toxin, filamentous hemagglutinin, pertactin, lipo-oligosaccharide (LOS), and Bordetella polysaccharide (Bps).^4,11-13

Pertussis usually starts with cold-like symptoms with occasional mild cough or fever. One to two weeks post-infection, well-known pertussis symptoms appear. These include paroxysms of multiple rapid coughs, appearance of the high-pitched “whoop” sound, vomiting and exhaustion. Pertussis can also cause extra-respiratory complications, which may include cardiovascular ailments, skin problems and impact on sleep patterns of both children and adults. Neonates, infants, and younger children are at extremely high risk of developing pertussis and exhibiting severe symptoms.^14,2 In contrast, Bp infections cause mild symptoms or are asymptomatic in vaccinated older children and adults.^15-16 In general, pertussis is not life-threatening in adults but can be fatal in infants.^2,17 Older children and adults are important drivers of pertussis outbreaks since they serve as carriers and sources of bacterial transmission.¹⁸

Before the introduction of Bp vaccines, pertussis was one of the major causes of childhood mortality in the United States and more than 260,000 deaths were attributed to pertussis annually. According to the Centers for Disease Control and Prevention (CDC), the highest pertussis incidence has been reported in babies less than one year old for the last 20 years in the United States with an incidence rate of about 60 per 100,000 persons in 2019.¹⁹ However, since many of the symptoms of pertussis are similar to other respiratory infections, the global incidence of pertussis is severely underreported. An estimated 24.1 million cases and 160,000 deaths due to pertussis occur annually.²⁰

2. Brief history of pertussis vaccine development and associated challenges

The best way to prevent the disease pertussis is vaccination, which is recommended for routine use by the CDC and the World Health Organization. Madsen prepared the first pertussis vaccine back in 1929, which was composed of a suspension of Bp bacteria in rabbit blood.²¹ During two years of observation, a lower death rate among vaccine recipients than that of the non-vaccinees demonstrated the vaccine’s effectiveness. The more well-known tri-functional vaccine Diphtheria-Tetanus-Pertussis (DTP), which blended whole-cell pertussis (wP) vaccines with diphtheria and tetanus toxoids with alum as the adjuvant, was developed and administered widely in the United States. Since wP vaccines were developed and standardized, the number of Bp related deaths decreased²² by more than 75% compared with the pre-vaccine era.²³ However, the wP vaccine was associated with some safety issues, as it can occasionally have serious adverse effects such as local reactions, fever, and seizures.²⁴

With the increasing public scrutiny on vaccine safety, a safer acellular vaccine (aP) composed of pertussis toxin and filamentous hemagglutinin was introduced in Japan in 1981. Subsequently, similar aPs consisting of one to five pertussis components have now been adopted by most developed countries.^25-26 Although mass vaccination has greatly reduced morbidity and mortality, the world has been experiencing a resurgence of pertussis in recent decades, especially after the introduction of the aP vaccine. Even in affluent nations with high vaccination rates, ^27,28 the number of pertussis cases reached a 60-year high in places like the US and Australia.²⁹ The infection rate of pertussis was reported to be 1–6 %,^30-31 making pertussis the most prevalent vaccine-preventable disease in developed countries.

3. Advantages and limitations of aP

The histamine-sensitizing, lymphocyte-leukocyte-promoting and islet-activating (HSF-LPF-IAP) antigen that caused major harmful effects of pertussis was named by Pittman as “pertussis toxin”.³² They also suggested that antitoxin against pertussis toxin would be an effective defense against the disease. The initial aP vaccine composed of pertussis toxin and filamentous hemagglutinin showed excellent efficacy as well as much lower reactogenicity compared to the wP vaccine.³³ The current commercial aP vaccine may contain up to five virulence factors from pertussis, including pertussis toxin, filamentous hemagglutinin, pertactin, and fimbriae 2&3. As discussed below, although the number of antigens and their amounts in different aP can vary, there is general agreement that pertussis toxin is an essential component in the aP vaccines.³⁴

Early clinical trials of aP vaccines demonstrated comparable efficacy with wP in preventing severe diseases.^35-36 However, epidemiological data show that aP vaccines generated shorter immunity compared to wP vaccines. The anti-pertussis immune responses from aP have reduced efficacy after 3–5 years and only 10% of the children vaccinated with aP are still protected at the recommended time for an adolescent booster.³⁷ To understand the differential protection provided by wP and aP vaccines, it is essential to understand the phenotype of the immune responses generated. Similar to natural infection, wPs mainly induce Th1-dependent immune response, and elicit IgG2 antibody subclasses as well as IgG1 and IgG3. On the contrary, aPs induce Th2/Th1 mixed or more Th2-skewed responses.^38-40 Although Th2 cells were thought to be important for promoting antibody response against extracellular pathogens such as bacteria, Mills and coworkers⁴¹ reported that the Th2 type response was actually dispensable for protection against Bp. Knockout of interleukin (IL) 4 gene that was important for inducing differentiation of naïve T helper cells (Th) to Th2 cells did not affect the clearance of bacteria. Bacterial clearance was mediated by CD4⁺IL-17⁺ (Th17) T cells that primarily recruited neutrophils to the lungs promoting the killing of Bp. IL-17 knockout mice challenged with Bp had much slower clearance of bacteria, confirming the importance of Th17 T cells in protection. However, aP adjuvanted with alum does not elicit Th17 responses. This dichotomy in immune polarization between wP and aP is a second major explanation for the resurgence of pertussis in vaccinated populations. The redundant Th2 component in immune responses for aP vaccines may have even caused the rare type hypersensitivity reactions seen in children after a fourth or fifth injection.^42-43 Protection by wP vaccines mainly depends on the production of interferon–γ (IFN-γ) by Th1 cells. Although the role of Th17 response upon immunization with wP vaccines is not well understood yet, the induction of Th17 cells via IL-1 might explain the better protection provided by wP vaccines. Recent efforts in the field are focused on developing next-generation vaccines that elicit Th1/17 polarized immune responses. These include substituting alum with Th1/17 eliciting adjuvant such as CpG and STING agonists,⁴⁴ novel lipopeptide TLR2 agonists identified in Bp⁴⁵ or adding adjuvants such as BcfA which attenuates the Th2 responses primed by alum.⁴⁶

Another limitation of aP is the inability to prevent nasal colonization of Bp. aP immunized individuals harbor a reservoir of bacteria in the nose, and thus can transmit the infection to vulnerable populations. This was directly demonstrated in studies conducted by the Merkel lab using a baboon model of Bp infection.⁴⁷ The baboons were immunized with wP or aP with the same schedule as human infants before being challenged with Bp. Although post-vaccination serum analysis indicated comparable levels of antibodies against the four main virulence factors and aP also successfully prevented leukocytosis, analysis of nasopharyngeal washes showed persistent colonization of bacteria in aP vaccinated baboons. To mimic the transmission through cough illness in human disease, they either co-housed unimmunized challenged animals with aP vaccinated animals or naïve animals with aP immunized challenged animals. Surprisingly, the transmission of bacteria was observed in both cases despite vaccination. Their results showed that aP failed to prevent colonization of Bp and transmission to healthy individuals. These data suggest that transmission of the infection by vaccinated people with a nasal reservoir is another mechanism leading to the resurgence of pertussis.

It has been found that immunization with aP vaccines did not improve bactericidal activities.⁴⁸ In a survey of 34 pairs of pre- and post-immunization serum samples from adults done by Weiss and coworkers, no significant increases in bactericidal activity at high serum concentrations were observed after vaccination in spite of elevated titers of IgG against all three exotoxins (pertussis toxin, filamentous hemagglutinin, and pertactin) in the vaccine.⁴⁹ Those results suggested that antibodies induced by aP vaccines might neutralize the exotoxins and reduce related symptoms, but do not directly kill the pathogens. Thus, new vaccines that can not only reduce disease symptoms, but also enhance bactericidal activities would be highly desired.

4. Pertussis LOS as the antigen for next generation of pertussis vaccines

New pertussis vaccines need to be developed due to the inadequacy of current aP vaccines. Returning to wP vaccinations, however, is not a viable option given the potential adverse effects. The development of next generation aP that elicits T1/17 immune responses and provides long lived protection is a critical goal for the field. It is especially important to include novel antigens that may generate stronger immunity and bactericidal antibodies. Below, we discuss the structure and function of one such antigen, LOS and its role as a component of next generation pertussis vaccines.

Polysaccharides are major components of the microbial cell walls, which can contain relatively high amounts and varying types of capsular polysaccharides (CPS) or lipopolysaccharides (LPS).⁵⁰ Gram negative bacteria typically express LPS on the surface, which are generally consisted of three parts: lipid A, core oligosaccharide, and O-specific antigen (Figure 1). LPS and CPS are important virulence factors because they can stimulate many biological processes such as bacterial colonization, blocking phagocytosis, and interfering with leukocyte migration and adhesion. These CPS and LPS can be recognized by receptors of the innate immune system leading to the production of cytokines, chemokines, etc.^51-52

Figure 1.

Common structures of LPS of Gram-negative bacteria.

The cell surface lipidated glycans of Bp lacks the O-specific chain seen in the homopolymer of 2,3-dideoxy-2,3-diacetamidogalactosaminuronic acid found in B. bronchiseptica and B. parapertussis, two other bacteria of the same genus,⁵³ and are thus known as LOS rather than LPS. Eldering isolated the LOS of Bp for the first time using an acid method in 1941, and later MacLennan reported a hot phenol-water method as a more commonly used method of Bp LOS extraction.⁵⁴ Sodium dodecyl sulfate (SDS)-PAGE analysis of Bp LOS showed two separate bands. The more abundant band, which is referred to as band A migrating more slowly on the gel, represents a LOS containing a dodecasaccharide core, while the small band, the faster migrating band B, lacks a distal trisaccharide.

The biological function of Bp LOS is similar to those of the LPSs isolated from other Gram-negative bacteria. It was found that LOS worked synergistically with tracheal cytotoxin inducing epithelial NO production exclusively in non-ciliated cells, which eventually caused the disruption of ciliated cells in the respiratory mucosa.⁵⁵ LOS also protects Bp from the innate immunity in the respiratory tract mediated by surfactant proteins A and D, which bind to the lipid A region and core saccharide region respectively, inducing aggregation and acting as opsonins for macrophages and neutrophils.^56-57 The binding with proteins A and D could be shielded mainly by the distal trisaccharide in LOS, as Bp mutants lacking the trisaccharide can be aggregated and permeabilized by the innate defense mechanism.

Besides LOS, other carbohydrates have been found to be produced on the Bp surface, which include Bps, a member of the poly-β-1,6-N-acetyl-d-glucosamine (PNAG) family of polysaccharides,¹³ and the CPS.⁵⁸ Bps can promote respiratory tract colonization of Bp by functioning as an adhesin, resisting killing by serum, complement and antimicrobial peptides.,^13,59,60 However, while Bps has been hypothesized as a potential vaccine antigen, to date it has not been experimentally shown to function as a protective antigen. For Bp CPS, to the best of our knowledge, its structures have not been determined and little is known about its biological functions.⁶¹

5. Structure elucidation of pertussis LOS

Detailed elucidation of the structure of pertussis LOS was pioneered by Caroff and coworkers.^62-64 They extracted the LOS from Bp strain 1414 following the hot phenol-water method and cleaved lipid A from LOS under mild SDS conditions. Selective hydrolysis of fatty acids from the major component of lipid A with hydroxylamine or sodium hydroxide pinpointed the linkage of fatty acids, while the distribution of fatty acids on the β-(1 → 6)-linked glucosamine disaccharide backbone was resolved by analyzing the fragmentation pattern in fast atom bombardment mass spectrometry data.⁶² The ensemble of those results defined the structure for lipid A, as shown in Figure 2. A minor species was less in molecular weight by 226 Da, presumably from losing one hydroxytetradecanoic acid.

Figure 2.

Structure of the major molecular species present in Bp lipid A.

Structural characterization of the heptasaccharide adjacent to lipid A was first attempted by Chaby and coworkers.^65-68 Following hydrolysis with hydrochloric acid of various concentrations, di- or tri-saccharide subunits at the non-reducing end of the heptasaccharide were harvested albeit in low yields. These di- or tri-saccharides were further digested into monosaccharides, which were then subjected to chemical modifications such as reduction or acetylation and compared with standard samples in HPLC. The structure of the heptose was proven to be l-glycero-d-mannoheptose by chemical degradation. Enzymolysis by stereochemistry-specific glycosidases revealed the correct configuration for glycosidic linkage. Eventually, the substitution positions of glucosamine and glucuronic acid on heptose were determined to be 7 and 2 respectively by analyzing fragments from the reduction with NaB³H₄ and cleavage with NaIO₄.^65-66 Similarly, the structure of another trisaccharide purified from the acidic hydrolysis was characterized as the 4-O-(2-amino-2-deoxy-α-d-glucopyranosyl)-6-O-(2-amino-2-deoxy-α-d-galactopyranuronyl)-d-glycopyranose configuration.⁶⁷ Due to the lability of 3-deoxy-d-manno-2-octulosonic acid (Kdo) in strong acid, they applied a nitrous acid-cleavage protocol and successfully separated a Kdo-containing oligosaccharide that was claimed to be a tetrasaccharide with the configuration of d-glucopyranosyl-β-(1 → 3)-d-glycopyranuronyl-β-(1 → 2)-l-glycero-d-mannoheptopyranosyl-α-(1 → 5)-3-deoxy-d-manno-2-octulosonic acid. By overlapping the shared monosaccharide units in all fragments, they proposed the structure of the reducing end heptasaccharide as in Figure 3.

Figure 3.

Proposed structure of the heptasaccharide present at the reducing end of LOS-1 isolated from Bp.

However, the determination of the precise structure of the tetrasaccharide highly relied on the colorimetric estimation of monosaccharides. The ratio of hexoses to heptoses was erroneously determined to be 1:1 rather than the actual value of 1:2, and therefore the wrong structure was proposed. After sequential dephosphorylation with hydrofluoric acid, deamination with nitrous acid and hydrolysis promoted by SDS, Szabó and coworkers isolated a hexasaccharide from the Bp endotoxin.⁶³ Methylation analysis revealed another mannoheptose in the structure, which was 3, 4-disubstituted and not detected in the previous research. Although the anomeric configurations were not easy to assign in the hexasaccharide due to overlapped signals in nuclear magnetic resonance (NMR) spectroscopy, exhaustive Smith degradation with NaIO₄ truncated all but the non-reducing end disaccharide and the glycosidic linkage was determined to be α by NMR analysis. The anomeric stereochemistry of the glucuronic acid was also determined to be α by NMR, which was incorrectly assigned to be β by the cleavage reaction with a commercial β-d-glucuronidase.

Nitrous acid deamination afforded the distal pentasaccharide from the LOS of Bp stain 1414. Relative structure of the five subunits in the pentasaccharide was resolved by NMR analysis, which suggested the existence of α-2-acetamidoglucose (α-GlcNAc), β-2-acetamido-4-N-methyl-2,4,6-trideoxy-galactose (β-Fuc2NAc4NMe), β-2,3-diacetamido-2,3-dideoxy-mannuronic acid (β-Man2NAc3NAcA), α-mannoheptose (α-ManHep) and anhydromannitol. By hydrolyzing the pentasaccharide with hydrochloric acid, the α-GlcNAc and α-Hep subunits were isolated and the absolute configurations were determined to be D and L, D by gas chromatography (GC)-MS.⁶⁴ Potential energy calculation in combination with nuclear Overhauser effect (NOE) measurements helped determining the structures of the other three monosaccharides as shown in Figure 4. Based on all data obtained, the complete structure of Bp LOS was proposed as in Figure 5.

Figure 4.

Structure of the trisaccharide present in the distal pentasaccharide from Bp. Some short inter-proton distances that could be observed in the NOE experiments and used to determine the absolute configuration of the sugars are shown.

Figure 5.

Complete structure of the LOS of Bp strain 1414 corresponding to band A. The lipid A structure is highlighted in blue in the rectangle box. Compared to band A, the structure of band B lacks the terminal trisaccharide GlcNAc-Man2NAc3NAcA-Fuc2NAc4NMe highlighted in red in the oval.

It is interesting to note that the four species of Bordetella, Bp, B. parapertussis, B. bronchiseptica, and B. holmesii were found to express similar core oligosaccharides.⁶⁹ The reducing end hexasaccharide (Figure 6) is conserved among all the investigated oligosaccharides of Bordetella. This core oligosaccharide may be useful in developing a cross-protective vaccine against infections with various bacteria from the Bordetella genus.

Figure 6.

Structure of the core hexasaccharide shared by the four species of Bordetella, Bp, B. parapertussis, B. bronchiseptica, and B. holmesii.

6. Immunological analysis and epitope profiling of pertussis LOS

For next generation vaccine development, it is important to identify the key factors in the protective immune responses to pertussis. Dolby and coworkers characterized fractions of antibodies in the Bp antisera and identified a complement-mediated bactericidal antibody that is capable of killing Bp.^70-71 Interestingly, the bactericidal activity pattern did not correlate well with any of the antibody titers for agglutinin, antihaemagglutinin or anti-histamine sensitizing factors, suggesting that the antibody was stimulated by a complete different antigen, which was referred to as “the bactericidal antigen”. Extraction of the endotoxin LOS from six strains of Bp and immunization at multiple doses of the LOS coupled with a carrier protein elicited bactericidal antibodies, which suggested that the antigen was actually LOS.⁷¹

Much research has been performed to further identify antigenic determinants on the LOS of Bp. Brodeur and coworkers described the preparation of both monoclonal antibodies specific for band A or band B by the hybridoma technology.^72-73 The LOS band A-specific antibodies were found to react well with strains of LOS AB phenotype but not the atypical strain 134 of LOS B phenotype. At least two antigenic epitopes were discovered with five out of the seven LOS band A-specific antibodies found to recognize the same epitope.⁷² Although band B-specific antibody BL-8 bound to strain 134 and led to moderate lytic activity, it failed to affect the predominant strains that express both LOS band A and B, presumably due to the limited expression and accessibility of epitopes on the cell membrane.⁷³

Structure elucidation of the pertussis LOS greatly expedited epitope mapping for antigen design. Chaby and coworkers immunized mice with strain 1414, which carried a majority of band A LOS.⁷⁴ Three monoclonal antibodies were tested against LOS from Bp1414, BpA100, B. bronchiseptica, B. parapertussis and different subparts of band A LOS. All three monoclonal antibodies bound to the terminal pentasaccharide, which was located far from lipid A, with recognition of GlcNAc-Man2NAc3NAcA, Fuc2NAc4NMe-GlcN and Hep-GlcN respectively. The carbohydrate region proximal to lipid A was considered poorly immunogenic because of the failure in generating an antibody that could bind to strain bpA100 lacking the GlcNAc-Man2NAc3NAcA-Fuc2NAc4NMe trisaccharide. Niedziela and coworkers conjugated the pentasaccharide cleaved from band A LOS with nitrous acid (Figure 7) to tetanus toxoid for mice immunization to generate polyclonal antibodies.⁷⁵ Those antibodies failed to bind with strain 606 in western blot, which carries only band B LOS. It implies that the immunogenic epitopes were mainly located on the distal trisaccharide GlcNAc-Man2NAc3NAcA-Fuc2NAc4NMe. Saturation transfer difference NMR experiments (STD-NMR) were also employed to investigate the structural elements on the pentasaccharide that contributed to the binding epitope (Figure 7). Although the information provided by STD-NMR supported that the major components of antigenic epitopes were on the trisaccharide, weak signals on the ManHep suggested that the heptose unit may also play a role in the immunogenicity of LOS. It was further confirmed by Robbins and coworkers, who reported that repetitive expression of the distal trisaccharide from a B. bronchiseptica mutant led to decreased binding affinity with anti-LOS antibodies.⁷⁶

Figure 7.

Structure of the pentasaccharide cleaved from band A LOS used for conjugate vaccine study. The complex of this pentasaccharide with an anti-LOS mAb was also studied by STD-NMR. The protons in close contact with antibodies as identified by STD-NMR are indicated by asterisks.

7. Pertussis LOS-based glycoconjugate vaccines

Despite the antibodies elicited from LOS in wP vaccines, it is not feasible to directly use Bp LOS as a vaccine. Many LPSs from Gram-negative bacteria have been reported to be highly immunogenic, but Bp LOS was poorly immunogenic when injected alone, probably due to its low molecular weight. Unlike proteins, the LOS is a T-independent antigen and triggers B cells to secret predominantly IgM with low binding affinity. Moreover, LOS was suspected to be the main culprit for the side effects of wP vaccines because of its endotoxin activity.⁷⁷ Therefore, it is necessary to remove the endotoxin determinant lipid A from LOS. Covalent conjugation of the immunogenic carbohydrate region with a carrier protein results in uptake by B cells and antigen presentation through MHC II to CD4⁺ T cells. Cytokines secreted by the activated T cells will stimulate the maturation of B cells and cause the class-switching from IgM to high-affinity IgG, which helps achieve higher-level and longer-lasting immune responses. Although the relative abundance of two bands of oligosaccharides may vary across different strains, such as band B being the major component in Bp strain A100, the structure of Bp core saccharide remains relatively conserved, which is different from virulence factors included in the current aP vaccines. This conclusion was supported by an investigation of the structure of Bp LOS from pre- and post- vaccination era,⁷⁸ suggesting that those antigens may be promising epitopes for vaccines.

Niedziela and coworkers conjugated the pentasaccharide from Bp LOS with tetanus toxoid (TT) through reductive amination with the free aldehyde group on the reducing-end anhydromannose.⁷⁵ They found that the mouse polyclonal antibodies against such a conjugate bound strongly with the whole wild-type LOS of Bp as well as the live Bp bacteria. It was reported previously that LOS worked synergistically with tracheal cytotoxin and induced the release of NO from secretory epithelial cells.⁵⁵ Such effect was significantly inhibited by the polyclonal antibodies against the pentasaccharide-TT conjugate along with lower production of IL6 and TNF-α, which are both pro-inflammatory cytokines involved in the inflammatory process stimulated by LOS.

Instead of choosing the most immunogenic distal pentasaccharide as the antigen, Robbins and coworkers cleaved the whole dodecasaccharide from LOS with 1 % acetic acid (Figure 8).⁷⁶ In their study, a mutant of B. bronchiseptica lacking the complete wbm locus and did not express the O-specific chain was used as an alternative source of LOS, due to the ease in its culture as compared to Bp. The purified dodecasaccharide shared the same structure as that in Bp, except that ~50% of the non-reducing end GlcNAc was replaced by GalNAc. The glycan obtained contained a ketone moiety at the reducing end due to Kdo cleavage. Similarly, LOS from Bp Tohama I was isolated. As a protein carrier, bovine serum albumin (BSA), was modified with an oxoamine (ONH₂) linker (Figure 7), to which the ketone bearing oligosaccharide was conjugated via oxime formation. Mice were immunized with the resulting BSA/OS conjugate, which successfully produced antisera that were bactericidal against Bp Tohama I strain. The bactericidal activity of the post-immune sera against Bp correlated with the levels of anti-glycan antibodies determined by the enzyme linked immunosorbent assay (ELISA).

Figure 8.

Conjugation of core oligosaccharide (OS) from mutant B. bronchiseptica or Bp with BSA for vaccine studies.

8. First chemical synthesis of the antigenic determinant Bp pentasaccharide

For vaccine studies, the availability of sufficient quantities of pure LOS in a conjugatable form is critical. While traditional LOS based vaccine studies focused on antigens isolated from bacteria, this approach has limitations due to the heterogeneity of glycans expressed on bacterial surfaces and the safety concerns due to the need for culturing large quantities of pathogenic bacteria. Chemical synthesis of the LOS can be a complementary strategy for vaccine development.

Based on the understanding of the importance of the distal pentasaccharide of pertussis LOS as a key epitope for the generation of protective antibodies, Huang, Deora, and coworkers developed the first total synthesis of the unique non-reducing end pentasaccharide from Bp LPS.⁷⁹ As this epitope contains three rare sugar units (2-acetamido-2,4,6-trideoxy-4-methylamino l-galactose, l-glycero-d-manno-heptose, and 2,3-diacetamido-2,3-dideoxy d-mannuronic acid) and multiple 1,2-cis glycosyl linkages, it has presented significant challenges in building block preparation as well as stereochemical control in glycosylations.

Several factors influenced the synthetic design of the target pentasaccharide 1 (Scheme 1). E unit was installed onto D before C to form the CDE branched trisaccharide because the 4-OH of D may be too hindered if the 6-OH of D were glycosylated with C first. Trichloroethoxy carbonyl (Troc), a known participating neighboring group, was used as the N-protective group for 2-amine of C unit fucosamine (building block 4) to facilitate the formation of 1,2-trans linkage between C and D. Interestingly, it was found that the protective group on the 4-N of the fucosamine building block significantly impacted the stereoselectivity in glycosylation. While building block 5 led to high 1,2-trans selectivity presumably aided by the Troc group, the corresponding donor with an additional benzyloxycarbonyl (Cbz) on the 4-N produced the undesired 1,2-cis glycoside as the major product. The 2-amine moiety in the D unit was acetylated rather than keeping it as a free amine for better synthetic feasibility. Acetylation was also necessary since Niedziela and coworkers recently found that free 2-NH₂ in the LOS may lead to glycan epitope cleavage from the carrier protein during antigen processing in cells, resulting in much lower immune response.⁸⁰ For the linkage between amino-mannuronic acid B and fucosamine C, the direct formation of the 1,2-cis glycosylation bond was difficult. Instead, an indirect route involving the 3-amino glucose derivative 5, whose 2-O stereochemistry could be stereospecifically inverted after glycosylation, was utilized. Thus, sequential reactions of 2, 3, 4, 5 and 6 with protective group or functional group manipulation in between glycosylation reactions successfully produced the target pertussis pentasaccharide 1 for the first time.

Scheme 1.

Summary of the first chemical synthesis of pertussis associated pentasaccharide 1.

9. Evaluation of synthetic pertussis-like pentasaccharide as a conjugate vaccine candidate

For potential vaccine development, the synthetic pertussis-like pentasaccharide 1 was covalently conjugated with a powerful carrier system, the bacteriophage Qβ^81-82 through an amide bond. Bacteriophage Qβ can typically display several hundred copies of antigen on each particle in an organized manner. In addition, it can provide the requisite Th help both in mice and in humans to potentiate the B cell responses. To benchmark the Qβ conjugate, in a similar manner, pentasaccharide 1 was conjugated with a gold standard carrier, keyhole limpet hemocyanin (KLH), to establish the effect of carrier protein in eliciting anti-glycan antibodies. Mice were immunized with the Qβ-1 and KLH-1 conjugates respectively. As shown in Figure 9, mice immunized with Qβ-1 produced significantly higher anti-glycan IgG titers compared to those immunized with Qβ only or KLH-1 conjugate on day 35 after the initial immunization. Furthermore, the antibodies induced by Qβ-1 were long-lasting, which remained at a high level even on 607 days following the initial immunization. Furthermore, the antibodies could be boosted suggesting the induction of memory B cells through vaccination with Qβ-1. The post-immune sera also recognized well multiple laboratory as well as clinical strains of Bp suggesting the potential of Qβ-1 as a new anti-pertussis vaccine. The in vivo protective effects of Qβ-1 in suitable animal models are yet to be established.

Figure 9.

a) Comparison of serum anti-pertussis glycan IgG titers from mice immunized at different doses of Qβ-glycan vs Qβ only or the comparison group KLH-glycan. Qβ-glycan conjugate induced significantly higher anti-glycan IgG antibody responses; b) The IgG antibody responses induced by the Qβ-glycan conjugate was long lasting; c) The antibodies elicited by Qβ-glycan recognized multiple laboratories and clinical strains of Bp. ns: non-significant; *** p < 0.001; **** p < 0.0001.

10. Future outlook

While vaccines have drastically changed the landscape of the prevention of pertussis diseases, there are significant drawbacks to available vaccines. The wP vaccines can lead to serious side effects. Although the aP vaccines are more biocompatible, they fail to induce durable protection. Furthermore, the antibodies produced by aP vaccines can effectively reduce the disease symptoms, but do not kill the bacteria. Thus, new vaccines that can elicit long-term robust bactericidal immune responses are highly attractive.

Carbohydrates present on the surface of bacteria have been used as immunogens to develop carbohydrate-based vaccines against many pathogens such as Streptococcus pneumoniae, Neisseria meningitides and Haemophilus influenza.⁸³ Conjugates of Bp LOS with a carrier protein elicit bactericidal antibodies, which were reported to overcome the BrkA protein-induced resistance against complement-dependent bactericidal pathway.^34,84 This along with the highly conserved structure of LOS across strains renders LOS an appealing target for vaccine development. However, isolating LOS in abundance from the highly aerosol transmissible pathogen remains a big obstacle, and the alternative way of expressing LOS in mutant B. bronchiseptica results in heterogeneity in the desired structure. The obtainable sequences of oligosaccharides limited by LOS processing methods also impose restrictions on screening for the optimal immunogenic epitope and understanding the adaptive immunity against LOS. Total synthesis of pertussis associated glycan is an important approach to producing structurally well-defined epitopes for vaccine studies. Recently, the non-reducing end pentasaccharide of pertussis LOS has been successfully prepared, laying the groundwork for the preparation of the full LOS sequence.

Preliminary analysis of the conjugate of the synthetic pentasaccharide with the Qβ carrier showed that the pentasaccharide is immunogenic in mice. The conjugate elicited strong and long-lasting anti-glycan IgG responses, which recognized multiple strains of the bacteria, highlighting the potential of synthetic carbohydrate-based vaccines. Optimization of the carbohydrate epitope may help the development of a more effective vaccine against Bp, which is why, novel strategies remain to be developed to facilitate the design of a convergent synthetic route for the pentasaccharide, the reducing end-core of dodecasaccharide and/or the whole dodecasaccharide. That synthetic carbohydrate chemistry provides flexibility to probe epitope specificity and can add to the arsenal for the development of effective anti-pertussis vaccines.

Data availability

As a review article, no original data is reported in this work.