Chinese Chemical Letters  2018, Vol. 29 Issue (1): 19-26   PDF    
Recent progress of fully synthetic carbohydrate-based vaccine using TLR agonist as build-in adjuvant
Zhifang Zhou, Han Lin, Chen Li, Zhimeng Wu    
Key Laboratory of Carbohydrate Chemistry & Biotechnology, Ministry of Education, School of Biotechnology, Jiangnan University, Wuxi 214122, China
Abstract: Fully synthetic vaccine, in which one or multi-molecular antigens are conjugated to a synthetic carrier with well-defined chemical structure, is a new direction to develop carbohydrate-based vaccine against cancer and pathogens. Toll like receptor (TLR) agonists with the ability to stimulate immune response have been widely investigated and been applied as build-in adjuvants to construct fully synthetic vaccines. In particular, remarkable progress has been achieved in recent years in the development of vaccines constructed with the agonists of TLR1/2, TLR2/6 and TLR4 and tumor-associated carbohydrate antigens (TACAs). These di-, tri-or multi-component vaccine candidates showed attractive immunological properties. This review highlights recent advances in developing full synthetic carbohydrate antigen based vaccines, with an emphasis on the structure-activity relationships that provide a primary basis for future vaccine design and immunotherapy developing.
Key words: Fully synthetic vaccine     TLR agonist     Build-in adjuvant     Carbohydrate-based vaccine     TACAs    
1. Introduction

Since Edward Jenner discovered that cowpox could prevent human from the threat of smallpox infection, vaccine has become one of the most important and successful strategy to protect against infectious diseases [1]. Currently, the licensed vaccines in clinical are made from attenuated or killed pathogens, toxoids, proteins or polysaccharide antigens which are isolated from pathogens, and so on [2]. Despite the great achievement in vaccines, there are still some concerns and issues about them, such as limited immunological efficacy in certain populations, safety issues and lacking vaccines for some serious diseases. Among the potential immune targets, carbohydrate antigens, usually abundantly exposed on the cell surface of pathogens with highly conserved and specific chemical structures, are attractive and important targets for the development of vaccines and immunotherapy [3].

Typically, carbohydrate antigens alone only induce short-term and T cell independent immunity, especially in infants and children [3]. To overcome this problem, the conventional approach is coupling carbohydrate antigens to carrier proteins as conjugate vaccines. In this way, the immune responses induced against carbohydrate antigens can be switched to T cell dependent immunity that is more potent, functional and with immune memory [4]. This strategy is quite straightforward and has been widely applied in licensed vaccines, such as the vaccines against Haemophilus influenzae type b (Hib) and Neisseria meningitidis A, C, Y and W-135. However, the glycoprotein-based vaccine strategy still has drawbacks. One of them is that the strong immunities against the carrier proteins may suppress the immunities against carbohydrate antigen moieties and lead to poor vaccine efficacy [5]. Meanwhile, the coupling sites on the carrier protein are various and the loadings of the carbohydrate antigens are difficult to control [6]. Additionally, using the naturally derived polysaccharides, which are heterogeneous and complex antigens, to produce protein conjugate vaccines meets the challenges in quality control and safety standards during the purification procedures [7]. Furthermore, an external adjuvant is widely used and necessary in the vaccine composing carrier protein, which may cause serious side effects [8]. As a result, new vaccine techniques and strategies are still imperatively needed.

To avoid the problems described above and improve the vaccine efficacy, a number of small molecules with immunostimulatory abilities, such as the agonists of toll like receptor (TLR), were explored to build in the novel vaccine constructs, and showed promising potential in enhancing the immunogenicity of carbohydrate-based vaccines. In this review, we are discussing these fully synthetic carbohydrate-based vaccines developed in recent years using TLR agonists as build-in adjuvant, including lipopeptides (recognized by TLR1/2 heterodimers and TLR2/6 heterodimers), lipid A (recognized by TLR4) and C-phosphate-G DNA (CpG-DNA, recognized by TLR9). The design of the carbohydrate-based vaccine with build-in adjuvant and immunological studies are covered in this review.

2. Vaccine development using TLR ligands as build-in adjuvant 2.1. Introduction of TLRs and their ligands

Recent technological advances and deep studies in immunology are paving the way for digging more and more immunological stimulators that can be used in vaccine design as delivery vehicles, carriers, and build-in adjuvants. Most of these immunological stimulators are pathogen-associated molecular patterns (PAMPs) or damage associated molecular patterns (DAMPs) that can be recognized by distinct pattern-recognition receptors (PRPs), and activate specific signaling pathways leading to distinct immunological responses [9]. TLRs, which are type Ⅰ integral membrane glycoproteins with the extracellular domains containing varying numbers of leucine-rich-repeat motifs and a cytoplasmic signaling domain homologous to that of the interleukin 1 receptor, as one kind of PRPs, have been identified to play a critical role in the innate immune recognition [9]. Each subfamily of TLRs can recognize specific PAMPs: The subfamily of TLR1, TLR2 and TLR6 recognized lipids derived from bacteria or mycoplasma, the TLR7, TLR8 and TLR9 recognize nucleic acids of bacteria or mycobacteria, while the TLR4 subfamily recognizes specific ligands with very different structures, such as lipopolysaccharide (LPS) derived from gramnegative bacteria, the plant diterpene paclitaxel, and some fusion proteins from virus or heat-shock proteins (Fig. 1). The general signaling pathways induced by different TLRs share common mechanisms. As shown in Fig. 1, after the microbial components binding, TLRs dimerize and change conformation to recruit Tollinterleukin 1 (IL-1) receptor (TIR) domain-containing adaptors TIRAP and MyD88, which further induces downstream signal transduction and activates NF-κB to produce inflammatory cytokines. The signal transduction of TLR4 has some differences in that LPS bind to TLR4 associated with MD2 to form a complex, TLR4-MD2-LPS. This complex is then internalized to endosome and recruits TRAM and TRIF to induce a late-phase NF-kB activation [9].

Download:
Fig. 1. Simple description about PAMP recognition by cell surface receptors TLR1/2, TLR2/6, TLR4, and intracellular receptor TLR9.

2.2. TLR1/2 and TLR2/6 ligands

In 1983, Hopp's group firstly used the dipalmityl-lysyl as carrier to construct lipid-conjugated vaccine bearing hepatitis B virus surface antigen (HBsAg) peptide and found that this novel conjugate elicited significant higher HBsAg-specific antibody titer than the traditional protein conjugate using KLH as carrier [10]. Since then, the lipid moieties were employed to conjugate with peptide or carbohydrate antigens to enhance the immunogenicity of vaccines. Mono-, di- and tri- palmitoylated peptides have been demonstrated to activate dendritic cells and induce robust cellular immunity. For example, mono-palmitoylated (Pam) peptide conjugate, either using PamLys or PamCys lipid moiety, has been demonstrated to provoke strong antigen-specific cytotoxic T lymphocytes (CTL) activity against peptide-pulsed target cells in the absence of external adjuvant [11, 12]. Tri-palmitoylated cysteine (Pam3C) firstly described by Jung and Bessler [13], has been developed into the most popular lipid motif that is widely used in recent anti-cancer vaccine design. Toyokuni et al. initially proved that carbohydrate antigen coupled with Pam3C could elicit robust immune responses without external adjuvant [14]. Danishefsky and Lloyd et al. developed a fully synthetic antitumor vaccine using a cluster of three tumor-associated carbohydrate antigens (TACAs) with a build-in adjuvant Pam3C. These total synthetic vaccine constructs can be effective immunogens and have the potential to be produced in large scale with simplified regulatory approval [15]. However, those fully synthetic vaccines can only elicit IgM antibodies, which indicated that a T-helper cell (Th) epitope was needed in the vaccine constructs.

To result in a more efficient T cell immune response and class switch to IgG antibodies, Boons et al. extensively investigated the synthesis and immunological studies of three-component vaccines containing lipopeptide Pam3C, Th epitope peptide and the tumor-associated carbohydrate antigen [6]. Based on this concept, they have constructed many fully synthetic vaccines with tumor-associated carbohydrate antigens, especially the glycosylated mucin 1 (MUC1) related peptides. Tumor-associated mucin MUC1 peptide motifs, with truncated glycosylation compared to normal mucins, are ranked second attractive targets for immunotherapy development and have been widely used in anticancer vaccines [16, 17]. In Boons' group, the Tn glycosylated MUC1 epitope (TSAPDT (α-GalNAc) RPAP) was conjugated with Pam3CSK4 (conjugate 1) or Pam2CSK4 (conjugate 3) and a Th epitope by a combination of solid-phase peptide synthesis (SPPS) and liposome-mediated native chemical ligation (NCL) [18, 19](Fig. 2). Besides the excellent antigenicity of the three-component vaccines, they also found that the conjugate composing Pam3CSK4 provoked higher titers of MUC1 epitope-specific antibody than that composing Pam2CSK4 [18, 19]. Additionally, several factors were essential for the efficacy of these three-component vaccines, including the tetra-lysine moiety (KKKK, K4), the covalent attachment of the three components, and the liposome formulation [19]. Compared with the vaccine containing MUC1 epitope without glycosylation, the one with glycosylation led to more tumor size reduction in mouse mammary model [19]. The antibodies induced by the vaccine with glycosylated MUC1 epitope could induce efficient antibody-dependent cytotoxicity (ADCC) towards MUC1-expressing cancer cells. The activated CTLs and CD8+ T cells isolated from the lymph nodes of the mice immunized with glycosylated conjugates could recognize a wider range of antigen structure including glycosylated and unglycosylated MUC1 epitope. In contrast, the T cells induced by unglycosylated vaccines could not recognize the glycosylated conjugates [19]. Both in vitro and in vivo evaluation revealed that the glycosylation of the MUC1 peptide was critical for eliciting proper immune response.

Download:
Fig. 2. Chemical structure of Pam3CSK4 conjugated with T-helper epitope and MUC1 epitopes with Tn or sTn glycosylation.

In addition to the NCL strategy, Boons' group designed and synthesized another three-component cancer vaccine candidate 2 with natural linkers by a linear solid phase peptide synthesis strategy, where MUC1 epitopes bearing sTn glycosylation which could not be achieved by NCL strategy (Fig. 2). Immunization studies demonstrated that conjugate 2 elicit both humoral and cellular immune responses [20]. This result further showed the vaccine candidate with sTn antigen could break immune tolerance and induce the immune response against tumor cells. The conjugates 1 and 2 elicited robust IgG isotypes indicating a mixed Th1/Th2 response was involved. The antibody can further induce ADCC toward to MUC1-expressing mammary cancer cells. Moreover, the enzyme-linked immune spot (ELISPOT) assay revealed that the conjugate 2 with sTn glycosylation could provoke somewhat stronger CD8+ T cell response than that with Tn glycosylation [20].

Li and Kunz et al. developed two-component and threecomponent strategy to design and synthesize fully synthetic antitumor vaccines by employing full length MUC1 variable number tandem repeat (VNTR, HGVTSAPDTRPAPGSTAPPA, 20 amino acids) with Tn, T or sTn antigens at threonine-9 and/or serine-15 sites [21, 22] (Fig. 3). The two-component vaccines 4 were constructed by coupling MUC1 glycopeptides to Pam3CSK4 through thioether ligation with a spacer group directly. The threecomponent vaccines 5, 6 and 7 were synthesized by adding the P4 and P2 Th epitopes in the middle, respectively. At first, they investigated the vaccine formulation to improve the immunological activities of these synthesized vaccine candidates. It was found that conjugate in PBS buffer could provoke the highest antibody titers compared to liposome formulation or with Freund's complete adjuvant (CFA). The antibody titers induced by these conjugates followed the order: 7 >6 > 5 > 4. Therefore, threecomponent vaccines 7 with P2 epitope and two-Tn glycosylation sites showed the best ability to increase antibody titers. The cross reaction assay revealed that the antibodies elicited by these conjugates were MUC1-Tn specific but not against Tn antigen itself. The in vitro binding assay of tumor cells and complementdependent cytotoxicity (CDC) experiments further confirmed the antibodies could recognize the tumor cells and induce strong immunological cytotoxicity.

Download:
Fig. 3. Chemical structure of Pam3CSK4 with different antigens containing full length glycosylated MUC1 tandem repeat.

To further enhance the immunogenicity of MUC1-based vaccines, they developed the cluster strategy to construct vaccines. A branched oligo-lysine core was applied to cluster di- or tetra-MUC1 peptides in one molecule by click chemistry [23, 24] (Fig. 4). Interestingly, conjugate 11 elicited the highest titer of MUC1-specific antibodies, while the antisera elicited by conjugate 13 exhibited the strongest binding ability and induced the most efficient CDC to MCF-7 cells. The IgG2a/IgG1 ratio induced by 13 was greater than 9 and 11, which might explain why the binding property of antisera induced by 13 was better than the other two. A similar tendency was observed in the candidates with Tn glycosylation (8, 10, 12).

Download:
Fig. 4. Chemical structure of vaccines cluster presenting full length glycosylated MUC1 tandem repeat.

Richard J. Payne et al. reported an efficient fragment condensation strategy for the convergent assembly of novel di- or tri-component vaccine candidates based on a full MUC1 VNTR sequence peptide (GVTSAPDTRPAPGSTAPPAH, 20 amino acids) with five-site glycosylation [25, 26]. Two kinds of TLR2 ligands, Pam3CS and macrophage-activating lipopeptide-2 (MALP2) were used in his group to construct a series of fully synthetic vaccines [25-28] (Fig. 5). MALP-2, whose structure contains a Pam2C moiety and a peptide moiety (GNNDESNISFKEK), firstly found in Mycoplasma fementans, can stimulate the immune response of macrophages efficiently through TLR2/6 heterodimers pathway. Therefore, the immunostimulatory ability of MALP-2 was recruited in this study [28, 29]. Similar as the above discussion, the Th epitopes conjugated in 16a-c and 17a-c was important for inducing IgG class switch. The antibodies elicited by 16a did not cross react with the MUC1-Tn and MUC1-T epitopes. The antibodies elicited by 16b did not cross react with MUC1-T epitopes. Only the antibodies elicited by 16c had appreciable cross reaction with MUC1-Tn epitopes. This observation indicated that IgG antibodies can recognize the GalNAc motif shared by Tn and T antigens. Fortunately, the antisera induced by 16a-c all showed strong binding to MCF-7 cancer cells expressing tumor-associated MUC1, but rarely exhibited binding to native MUC1 protein isolated from human breast milk. This result indicated that the MUC1 protein on normal cells with heavily glycosylation could not be recognized by these antibodies.

Download:
Fig. 5. Chemical structure of vaccines using Pam3CS or MALP2 as build-in adjuvant.

In addition to the above endeavor, they also investigated the formulation details about these conjugates. These amphiphilic molecules could self-assembly into micelles or particles with a diameter around 12-20 nm in aqueous and 17-25 nm in DMSO/water (1/9). This micelle formulation could promote the B cell responses because of the clustering presentation of the antigens on the particle surfaces. The advantage of this formulation is that an external adjuvant is no longer required. Although the conjugates induced high levels of class-switched IgG1, IgG2b and IgG3 antibodies, the results of cytotoxic activity assay suggested that there was no increasing in IL-4, IFN-γ or CD25+ CD4+ or CD8+, which suggested that the immune response was T cell-independent way.

Recently, Wei Zhao's group firstly employed fibroblast stimulating lipopeptide 1 (FSL-1, Pam2CGDPKHPKSF) to construct selfadjuvanting anticancer vaccines. FSL-1 as a novel TLR2/6 agonist is derived from Mycoplasma salivarium [30] (Fig. 6). The full length MUC1 VNTR domain with or without Tn-glycosylated was synthesized by SPPS method, which then coupled with the peptide motif of FSL-1, and finally conjugated with Fmoc-Pam2Cys-OH. The antibody investigation revealed that the unglycosylated vaccine candidate 19 elicited higher levels of all isotype antibodies than glycosylated conjugate 20. The high levels of IgG1 isotype suggested that significant Th2 response was elicited by 19 and 20, respectively. The cytokine levels, especially IL-4, IL-5, and IL-10, were increased significantly after the immunization. This result confirmed that both the vaccine candidates 19 and 20 provoked Th2 immune responses. The increasing level of IFN-γ showed that CD8+ cytotoxic T cells or Th1 immune response were activated, as well. The antisera induced by the two vaccine candidates could bind with the MCF-7 cells but not SKMEL-28 cells, which indicated the elicited antibodies, were MUC1-specific. This research proved that lipopeptide FSL-1 was a promising build-in adjuvant which can be used for further investigation to construct fully synthetic vaccines.

Download:
Fig. 6. Chemical structure of vaccine using FSL-1 as build-in adjuvant.

To investigate the structure-activity relationship and further promote the immunogenicity of the anticancer vaccines, many research focused on clustering antigen-presentation strategy which was mentioned in above Li's work. In Lbachir BenMohamed and co-worker's study, a B cell epitope made of regioselectively addressable functionalized templates (known as RAFT molecules), which presented a cluster of four Tn tumor-associated antigens, was recruited to construct the antitumor vaccines [31] (Fig. 7). The vaccine construct included a fully synthetic linear or a branched HER-2-glycolipid (GLP), which were regioselectively assembled with a CD8+ Th epitope peptide from HER-2, a well-defined CD4+ Th epitope (AKXVAAWWTLKAAA), known as PADRE, and one palmitic acid moiety (Pam) as an build-in adjuvant. The cellular and molecular mechanisms of GLPs immunogenicity, the position of the lipid moiety on immunogenicity and the protective efficacy of GLPs were the main issues they studied. They found that branched 23 induced stronger RAFT-specific IgGs that can bind to human tumor cell lines expressing the Tn antigen. However, linear conjugate 22 provoked stronger and more long-lasting HERspecific CD8+ T cell that can produce IFN-γ than 23. Both 22 and 23 could protect mice from death after implanting subcutaneously with NT2 tumor cells in the mammary fat pad. To further investigate the mechanisms of the immunogenicity, the up-taking kinetics of conjugates with Alexa Fluor 488 labelling by immature dendritic cells (DCs) were studied. Both 22 and 23 could be efficiently taken up by the DCs, and be accumulated in the cytoplasmic which could be visualized within 10 min of incubation. Conjugate 22 exhibited relatively faster delivery properties comparing with 23. They also observed that 21 could not be taken up by DCs indicated that the palmitic acid was a critical moiety for helping the conjugates internalized into the cytoplasm of DCs. The phenotypic maturation of DCs was only occurred in the conjugates with lipid moiety, which further suggested that the Pam moiety played an important role in the immune response through the TLR-2 signaling pathway. The position of the lipid moiety was another profound factor influencing the cross-presentation pathway. Conjugates 22 and 23 appeared to follow two different antigenpresentation pathways. The processing pathway of 22 was dependent on the proteasomal pathway followed by loading HER epitope on MHC class Ⅰ molecules, while the processing of 23 took the lysosomal acidification pathway as well as loading HER epitope on MHC class Ⅰ and transporting from endosome to cell surface. In summary, conjugates 22 and 23 both can elicit strong TLR-2 dependent T cell immune responses and reduce the established tumors.

Download:
Fig. 7. Chemical structure of vaccines using RAFT to cluster present Tn antigens and bearing with HER, PADRE and Pam.

Another example of clustering antigen-presentation strategy was achieved by using calix(4, 8)arene macrocycle skeletons to construct multicomponent vaccines developed in Geraci's group. Calix(4, 8)arene macrocycle skeletons were suitable platform for molecular recognition application, with four or eight PDRRP presenting at the wide rim (conjugates 24 and 25) [32] (Fig. 8). The PDTRP motif has been identified as the immunodominant B cell epitope by monoclonal antibody studies in mice, humoral and cellular immune studies using breast cancer patients with MUC1-expressing tumors. To further promote the immunogenicity of the conjugates, one unit of Pam3CS was also employed to present at the narrow rim of the calixarene macrocycle as shown in Fig. 8. Both 24 and 25 elicited effective immune response, while the conjugate 25 provoked 4 fold higher antibody responses than that of 24. The result indicated that the non-restricted conformational flexibility was more suitable for the ligand-receptor recognition, and clustering presentation of multivalent antigens was critical for eliciting strong antigen-specific immune responses.

Download:
Fig. 8. Chemical structure of vaccines using calix(4) or (8)arene skeleton to cluster present antigens and bearing Pam3CS as build-in adjuvant.

In Istvan Toth's group, the cluster strategy was carried out through the construct of three Tn antigens presented in tri-serine residues as B cell epitopes. In their vaccine design, a universal CD4+ Th epitope, the Tn cluster, and a TLR2 agonist, lipoamino acids (LAAs), were conjugated together to generate a self-adjuvanting vaccine 26 [33] (Fig. 9). The conjugate 26 elicited strong immune response which indicated that the LAAs were an efficient immunostimulant for constructing vaccines. The antibodies elicited by conjugate 26 neither bind to the peptide antigen nor the mono-Tn peptide antigen. This result suggested that the antibodies were specific to the peptide with cluster Tn antigen. Cluster Tn antigen residues were proved to be crucial for the binding properties with anti-Tn monoclonal antibodies. It was also demonstrated that the spatial arrangement of Th epitopes had important impact on the specificity of antibodies provoked by Tnbased multicomponent vaccines.

Download:
Fig. 9. Chemical structure of vaccine using LAAs as build-in adjuvant and bearing CD4+ T-helper epitope.

Taking together, to design effecient fully synthetic multicomponent vaccines, several factors have to be considered. First, the chemical structure of the build-in adjuvant is the most essential part for the vaccine design. For example, Pam3CSK4 showed better immunogenicity than Pam2CSK4, while tetralysine (K4) is important for the construct. Second, the Th epitope, either CD4+ or CD8+ Th epitope, is important for inducing IgG class switch and the covalent attachment of these three components is a necessary for vaccine construct. Third, the formulation of liposome or micell can present the B epitopes multivalently and in a high local concentration, which facilitates B-cell receptor clustering and thus could result in stronger immune responses. These formulations also replaced the use of external adjuvant formulation to elicit strong immune response. Finally, using antigen clustering strategy to present the antigens to APC cells is another useful approach to enhance immunogenicity.

2.3. TLR4 ligands

Lipopolysaccharide (LPS), which is the major component of the outer monolayer of Gram-negative bacteria cell outer membrane, has immunostimulatory ability through TLR4 pathway. Lipid A, the conserved hydrophobic core of LPS, is the functional moiety to bind with TLR4 and further induces immune activity [34]. They have a β-1, 6-linked disaccharide of D-glucosamine (GlcNH2) with two phosphate groups linked to the O-1 and O-4'-positions, and several lipids, varying in chain length, saturation, number and distribution, linked to the N-2-, N-2', O-3- and O-3'-positions. The structureactivity relationship studies revealed that the number, structure, position of lipids and the degree of phosphorylation are important factors influencing the bioactivity of lipid A [35] (Fig. 10). The nature form of lipid A is very toxic and could cause serious proinflammatory reactions. Very interesting, removing one of the phosphate groups could significantly reduce the toxicity and result in monophosphoryl lipid A (MPLA) derivatives with strong immunostimulatory ability. Thus, the MPLA has been widely investigated as an ideal adjuvant in vaccine formulation. Application of MPLAs both in adjuvant system and in vaccine formulation has been widely explored, including anti-infections, autoimmune regulation and cancer immunotherapy. To date, two vaccines containing MPLA adjuvant component have been successfully licensed for human use: CervarixTM vaccine against human papillomavirus and FENDrixTM vaccine against hepatitis B virus [36].

Download:
Fig. 10. Schematic representation of lipid A structure-activity relationships. The number near the chemical group changes indicated the factor of bioactivity reduction compared to natural lipid A.

Guo's group developed a number of vaccines based on MPLA as build-in adjuvant, including anti-cancer, anti-bacteria and anti-fungi vaccines. At first, this kind of fully synthetic vaccine was developed with the combination of a novel cancer immunotherapy strategy based on glycoengineering [37]. This strategy was developed to overcome the challenging immunotolerance problem due to immune system unresponsiveness to TACAs [38]. Briefly, a vaccine employing unnatural TACA derivatives was used to induce strong immune response, and then a precursor with the same artificial modification was applied to treat the target cells and enforce the cells to express the corresponding unnatural TACAs on the surface. Finally, the pre-elicited antibodies could recognize the unnatural TACAs on the tumor cells and induce immune cytotoxicity against tumor cells [37] (Fig. 11).

Download:
Fig. 11. The immunotherapy strategy based on glycoengineering.

As a proof of concept, this immunotherapy strategy was carried out based on GM3 antigen. GM3, which is overexpressed on melanoma cancer cell surface, is one kind of TACAs with a sialic acid residue in the terminal. Based on GM3 structure, N-phenylacetyl GM3(GM3NPhAc) was synthesized to construct vaccine and N-phenylacetyl-D-mannosamine (ManNPhAc) was the precursor to engineering the melanoma cancer cell to express artificial GM3NPhAc antigens. The whole strategy was confirmed in the development of the vaccine candidates with keyhole limpet hemocyanin (KLH) as the carrier protein, and the glycoengineering process was successfully proved as well [38-40]. To further investigate this strategy, improve the immunogenicity of the vaccine candidate, and find an easy-prepared formulation of vaccine, the fully synthetic vaccine construct with MPLA as buildin adjuvant was developed. The synthetic GM3NPhAc was coupled to a MPLA analogue derived from lipid A structure of Neisseria meningitidis as vaccine candidates (conjugate 27a) [37] (Fig. 12). The results of immunological studies demonstrated that the conjugate 27a could successfully elicit a strong immune response against GM3NPhAc without external adjuvant regardless of different linkers. The antibodies elicited by 27a could only recognize the corresponding GM3NPhAc antigen, but have no cross-reactivity to natural form of GM3.

Download:
Fig. 12. Chemical structure of vaccines using MPLA as build-in adjuvant.

To further explore the structure-activity relationships of MPLAs, Guo's group chemically synthesized four monophosphoryl derivatives of Neisseria meningitidis lipid A (a-d) as shown in Fig. 12, and investigated their immunological activities [41]. The four MPLAs were conjugated with sTnNPhAc as a model antigen to generate conjugates 28a-d which could be further used in the metabolic-engineering immunotherapy. Both the ELISA results and cell binding assay revealed that the immunogenicity of the four conjugates followed the order: 28b > 28c28d > 28a. It was concluded that the free hydroxyl groups on the lipid chains were critical for activity (28b vs. 28a), while the length of lipids and additional lipid chains at 3-O- and 3'-O-positions (28b vs. 28c, 28c vs. 28d) had less impact on the activity. The optimized structure was further applied in late vaccine design.

Another successful application of MPLA-based vaccine was the development of anticancer vaccine based on Globo H antigens [42] (Fig. 12). Globo H, which is a kind of TACAs in glycolipid form, was first discovered on the surface of human breast cancer cell line MCF-7, and later was widely found on other epithelial tumor cells, such as lung, colon, ovarian, and prostate cancer [43]. The Globo HKLH conjugate showed promising immunotherapeutic ability and have been well developed for the treatment of breast and prostate cancers in the formulation with an external adjuvant QS-21, which is now under Phase Ⅲ clinical trial [43]. The improved anticancer vaccine based on Globo H glycoprotein with other adjuvant was investigated in clinical trial, as well [43-45]. Based on these examples, the Globo H was considered as a one of the most potential target for the development of anticancer vaccine [16]. Therefore, the Globo H was synthesized in Guo's group and coupled with the optimized MPLA to form conjugate 29b [42, 46]. The immunological results revealed that 29b in liposome formulation elicited higher titers of both total and IgG antibodies than that of Globo H-KLH conjugate. Moreover, the Globo H-specific antibodies elicited by conjugate 29b increased more quickly than that of Globo H-KLH conjugate. Even after four immunizations, the antibodies induced by Globo H-KLH conjugate were approximately half as that induced by Globo H-MPLA conjugate. IgG1 was the main IgG subclasses provoked by both MPLA and KLH conjugate, which indicated that both conjugates induced the similar patterns of immune responses. The analysis of cytokine releasing suggested that T cell immunity was involved to enhance the immune response and help antibody switching to IgG1. The antibodies elicited by Globo H-MPLA conjugate was further proved that they could recognize MCF-7 cells overexpressing Globo H antigens and induce CDC against MCF-7 cells. It was for the first time discovered that a MPLA conjugate could elicit a quicker and stronger immune response than the corresponding KLH conjugate.

The concept of constructing self-adjuvant vaccine based on MPLA was further expanded to anti-bacteria vaccines in Guo's group. One of the example was the fully synthetic vaccines against serotype C Neisseria meningitidis according to the capsular polysaccharides (CPSs) structure, which has a typically α-2, 9-polysailic acid structure [47]. Consequently, a series of α-2, 9-oligosialic acids, from di- to penta-sialic acids were synthesized and coupled to MPLA to give conjugate 30-34b [48] (Fig. 12). Compared with the corresponding oligo-sialic acid-KLH conjugate, the MPLA conjugate elicited similar pattern of antibody response, mainly IgG2b, IgG2c, and less amount of IgG1 and IgG3. MPLA conjugates elicited lower IgG1, but higher IgG3 antibody titers than KLH conjugate. The IgG2b and IgG2c elicited by both conjugates were in similar high levels.

Recently, a mimic of lipid A, with hexa-acyl lipid chain was proved to stimulate strong immune response including the production of proinflammatory cytokines by monocytic THP-1 cells with the mediation of TLR4 [36]. Based on this result, Jiang's group designed and synthesized lipid A mimic-ThomsenFreidenreich (TF) antigen conjugates as a self-adjuvanting cancer vaccines [49] However, the immunological studies were not reported yet.

2.4. Other TLR ligands

The CpG oligodeoxynucleotides (CpG-ODNs), an agonist of TLR9 receptor, was widely found in the microbial genomes but rarely in vertebrate ones. The CpG-ODNs are short single-stranded unmethylated DNA fragments and can elicit efficient immune response through TLR9 receptor pathway. Using CpG-ODNs as external adjuvant has shown that CpG-ODNs increased the anti-cancer immunity elicited by MUC1-based vaccines [50] (Fig. 13). The covalent conjugation of CpG to an antigen has the potential to further promote the efficient cellular uptake and increase immune response [51]. For example, Boons' group designed and synthesized a build-in adjuvant cancer vaccine MUC1-CpG (ODN 1826) conjugate 35 with Th epitope in the middle [52]. However, the immunological results indicated that 35 did not have superior immunological properties than Pam3CSK4 conjugate, which may be due to the immunotolerance of MUC1. In spite of the underperformance of CpG-ODN 1826 in MUC1-based vaccine construct, it is still an attractive immunostimulatory reagent [51].

Download:
Fig. 13. Chemical structure of vaccine using TLR9 ligand, CpG (ODN 1826), as build-in adjuvant.

In another example, a TLR7 agonist, an analogue of imidazoquinoline, was employed to conjugate with protein or maltoheptaose as vaccine candidates. The humoral immune response against protein was increased significantly after conjugating with the TLR7 agonist. The maltoheptaose modification of imidazoquinoline resulted in less TLR7-stimulatory activity. The humoral immune response elicited by maltoheptaose modified imidazoquinoline was not evaluated [53]. The TLR 7/8 agonists was also used in polymer-based vaccine and showed the abilities to enhance the immunogenicity [54]. Therefore, design, synthesis and immunological studies of carbohydrate-based vaccine incorporating TLR7/8 agonists as build-in adjuvant need further investigation in the future.

3. Conclusion and Outlook

The TLR ligands play important roles in the immune system and have attracted strong interests in vaccine development and immunotherapy against pathogen diseases and cancer. The application of TLR ligands in fully synthetic vaccine constructs is just explored in this decades and it has shown great promise in the ability to increase immunogenicity of carbohydrate-based antigens, especially in the area of developing anti-tumor immunotherapy strategy. These novel full synthetic vaccines with build-in adjuvant catalyzed the development of organic synthesis methods, improved the antigen immunity. Further structure-activity relationship studies by elucidating the vaccination mechanism will guide new constructs with enhanced biological activity to overcome the problems in current vaccine and immunotherapy development. Inspired by these researches, some other small molecules with the similar abilities to stimulate immune responses, such as α-GalCer and its analogues were discovered to apply in vaccine construct as well. With the help of new materials and nanotechnologies, the semi- and full-synthetic vaccine molecules are further assembled to polymers or nano-particles and the immunological properties of these vaccine candidates are further promoted. A more accurate and precious vaccination strategy developed by multi-technologies may be the future direction and be the way to design and synthesize more powerful vaccines to combat with diseases.

Acknowledgments

This work was supported by the National Natural Science Foundation of China (Nos. 21472070, 21602084), the project for Jiangsu Scientific and Technological Innovation Team, the fund for Jiangsu Distinguished Professorship Program, the project funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions, the 111 Project (No. 111-2-06), the Fundamental Research Funds for the Central Universities (No. JUSRP11729) and the Open Foundation of Key Laboratory of Carbohydrate Chemistry & Biotechnology Ministry of Education (No. KLCCB-KF201608).

References
[1]
R.D. Astronomo, D.R. Burton, Nat. Rev. Drug. Discov. 9(2010) 308-324. DOI:10.1038/nrd3012
[2]
B.L. Bartlett, A.J. Pellicane, S.K. Tyring, Dermatol. Ther. 22(2009) 104-109. DOI:10.1111/dth.2009.22.issue-2
[3]
Z. Guo, Q. Wang, Curr. Opin. Chem. Biol. 13(2009) 608-617. DOI:10.1016/j.cbpa.2009.08.010
[4]
D. Feng, A.S. Shaikh, F. Wang, ACS Chem. Biol. 11(2016) 850-863. DOI:10.1021/acschembio.6b00084
[5]
T. Buskas, Y. Li, G.J. Boons, J. Chem. Eur. 10(2004) 3517-3524. DOI:10.1002/(ISSN)1521-3765
[6]
T. Buskas, S. Ingale, G.J. Boons, Angew. Chem. Int. Ed. 44(2005) 5985-5988. DOI:10.1002/(ISSN)1521-3773
[7]
C. Anish, B. Schumann, C.L. Pereira, P.H. Seeberger, Chem. Biol. 21(2014) 38-50. DOI:10.1016/j.chembiol.2014.01.002
[8]
R.L. Coffman, A. Sher, R.A. Seder, Immunity 33(2010) 492-503. DOI:10.1016/j.immuni.2010.10.002
[9]
M. Hedayat, K. Takeda, N. Rezaei, Med. Res. Rev. 32(2012) 294-325. DOI:10.1002/med.2012.32.issue-2
[10]
T.P. Hopp, Mol. Immunol. 21(1984) 13-16. DOI:10.1016/0161-5890(84)90084-1
[11]
B.D. Livingston, C. Crimi, H. Grey, et al., J. Immunol. 159(1997) 1383-1392.
[12]
A. Vitiello, G. Ishioka, H.M. Grey, et al., J. Clin. Invest. 95(1995) 341-349. DOI:10.1172/JCI117662
[13]
J. Metzger, K.H. Wiesmuller, R. Schaude, W.G. Bessler, G. Jung, Int. J. Pept. Protein Res. 37(1991) 46-57.
[14]
T. Toyokuni, S. Hakomori, A.K. Singhal, Bioorg. Med. Chem. 2(1994) 1119-1132. DOI:10.1016/S0968-0896(00)82064-7
[15]
V. Kudryashov, P.W. Glunz, L.J. Williams, et al., Proc. Natl. Acad. Sci. U. S. A. 98(2001) 3264-3269. DOI:10.1073/pnas.051623598
[16]
M.A. Cheever, J.P. Allison, A.S. Ferris, et al., Clin. Cancer Res. 15(2009) 5323-5337. DOI:10.1158/1078-0432.CCR-09-0737
[17]
N. Gaidzik, U. Westerlind, H. Kunz, Chem. Soc. Rev. 42(2013) 4421-4442. DOI:10.1039/c3cs35470a
[18]
S. Ingale, M.A. Wolfert, J. Gaekwad, T. Buskas, G.J. Boons, Nat. Chem. Biol. 3(2007) 663-667. DOI:10.1038/nchembio.2007.25
[19]
V. Lakshminarayanan, P. Thompson, M.A. Wolfert, et al., Proc. Natl. Acad. Sci. U. S. A. 109(2012) 261-266. DOI:10.1073/pnas.1115166109
[20]
P. Thompson, V. Lakshminarayanan, N.T. Supekar, et al., Chem. Commun. 51(2015) 10214-10217. DOI:10.1039/C5CC02199E
[21]
H. Cai, Z.Y. Sun, Z.H. Huang, et al., Chem. Eur. J. 19(2013) 1962-1970. DOI:10.1002/chem.v19.6
[22]
H. Cai, M.S. Chen, Z.Y. Sun, et al., Angew. Chem. Int. Ed. 52(2013) 6106-6110. DOI:10.1002/anie.201300390
[23]
H. Cai, Z.Y. Sun, M.S. Chen, et al., Angew. Chem. Int. Ed. 53(2014) 1699-1703. DOI:10.1002/anie.201308875
[24]
H. Cai, Z.H. Huang, L. Shi, et al., Chem. Eur. J. 17(2011) 6396-6406. DOI:10.1002/chem.201100217
[25]
B.L. Wilkinson, S. Day, L.R. Malins, V. Apostolopoulos, R.J. Payne, Angew. Chem. Int. Ed. 50(2011) 1635-1639. DOI:10.1002/anie.201006115
[26]
B.L. Wilkinson, L.R. Malins, C.K. Chun, R.J. Payne, Chem. Commun. 46(2010) 6249-6251. DOI:10.1039/c0cc01360a
[27]
B.L. Wilkinson, S. Day, R. Chapman, et al., Chem. Eur. J. 18(2012) 16540-16548. DOI:10.1002/chem.201202629
[28]
D.M. McDonald, B.L. Wilkinson, L. Corcilius, et al., Chem. Commun. 50(2014) 10273-10276. DOI:10.1039/C4CC03510K
[29]
P.F. Muhlradt, M. Kiess, H. Meyer, R. Sussmuth, G. Jung, J. Exp. Med. 185(1997) 1951-1958. DOI:10.1084/jem.185.11.1951
[30]
Y. Liu, W. Zhang, Q. He, et al., Chem. Commun. 52(2016) 10886-10889. DOI:10.1039/C6CC04623A
[31]
O. Renaudet, G. Dasgupta, I. Bettahi, et al., PloS One 5(2010) e11216. DOI:10.1371/journal.pone.0011216
[32]
C. Geraci, G.M. Consoli, G. Granata, et al., Bioconjug. Chem. 24(2013) 1710-1720. DOI:10.1021/bc400242y
[33]
A.B.M. Abdel-Aal, D. El-Naggar, M. Zaman, M. Batzloff, I. Toth, J. Med. Chem. 55(2012) 6968-6974. DOI:10.1021/jm300822g
[34]
B.S. Park, D.H. Song, H.M. Kim, et al., Nature 458(2009) 1191-1195. DOI:10.1038/nature07830
[35]
E.T. Rietschel, T. Kirikae, F.U. Schade, et al., FASEB J. 8(1994) 217-225.
[36]
C.H. Marzabadi, R.W. Franck, J. Chem Eur 23(2017) 1728-1742. DOI:10.1002/chem.v23.8
[37]
Q. Wang, Z. Zhou, S. Tang, Z. Guo, ACS Chem. Biol. 7(2012) 235-240. DOI:10.1021/cb200358r
[38]
Q. Wang, J. Zhang, Z. Guo, Bioorg. Med. Chem. 15(2007) 7561-7567. DOI:10.1016/j.bmc.2007.09.005
[39]
Z. Zhou, G. Liao, S. Stepanovs, Z. Guo, J. Carbohydr. Chem. 33(2014) 395-407. DOI:10.1080/07328303.2014.933483
[40]
L. Qiu, J. Li, S. Yu, et al., Oncotarget 6(2015) 5195-5203.
[41]
Z. Zhou, M. Mondal, G. Liao, Z. Guo, Org. Biomol. Chem. 12(2014) 3238-3245. DOI:10.1039/C4OB00390J
[42]
Z. Zhou, G. Liao, S.S. Mandal, S. Suryawanshi, Z. Guo, Chem. Sci. 6(2015) 7112-7121. DOI:10.1039/C5SC01402F
[43]
S.J. Danishefsky, Y.K. Shue, M.N. Chang, C.H. Wong, Acc. Chem. Res. 48(2015) 643-652. DOI:10.1021/ar5004187
[44]
F. Burkhart, Z. Zhang, S. Wacowich-Sgarbi, C.H. Wong, Angew. Chem. 40(2001) 1274-1277. DOI:10.1002/(ISSN)1521-3773
[45]
Y.L. Huang, J.T. Hung, S.K. Cheung, et al., Proc. Natl. Acad. Sci. U. S. A. 110(2013) 2517-2522. DOI:10.1073/pnas.1222649110
[46]
S.S. Mandal, G. Liao, Z. Guo, RSC Adv. 5(2015) 23311-23319. DOI:10.1039/C5RA00759C
[47]
G. Liao, Z. Zhou, Z. Guo, Chem. Commun. 51(2015) 9647-9650. DOI:10.1039/C5CC01794G
[48]
G. Liao, Z. Zhou, S. Suryawanshi, M.A. Mondal, Z. Guo, ACS Cent. Sci. 2(2016) 210-218. DOI:10.1021/acscentsci.5b00364
[49]
J.D. Lewicky, M. Ulanova, Z.H. Jiang, Chem. Select 1(2016) 906-910.
[50]
J. Schettini, A. Kidiyoor, D.M. Besmer, et al., Cancer Immunol. Immunother. 61(2012) 2055-2065. DOI:10.1007/s00262-012-1264-y
[51]
P. Daftarian, R. Sharan, W. Haq, et al., Vaccine 23(2005) 3453-3468. DOI:10.1016/j.vaccine.2005.01.093
[52]
A.M. Abdel-Aal, V. Lakshminarayanan, P. Thompson, et al., Chembiochem 15(2014) 1508-1513. DOI:10.1002/cbic.201402077
[53]
N.M. Shukla, T.C. Lewis, T.P. Day, et al., Bioorg. Med. Chem. Lett. 21(2011) 3232-3236. DOI:10.1016/j.bmcl.2011.04.050
[54]
G.M. Lynn, R. Laga, P.A. Darrah, et al., Nat. Biotechnol. 33(2015) 1201-1210. DOI:10.1038/nbt.3371