III. ABSTRACTS AND POSTERS / ORAL PRESENTATIONS
6. Mice deficient in interferon regulatory factor 4 (IRF4) are more susceptible to infection with mouse-adapted influenza A/Aichi/2/68
2.6.1. Enhancement of SARS Spike DNA vaccine efficacy fused to the integrin binding motif
DNA vaccination shows great promise for eliciting protection against infectious diseases and other immune disorders although the efficacy of DNA vaccines can vary widely due to different transfection efficiency in target tissues (Tuting, Austyn et al. 2000). DNA vaccines are mostly delivered through intramuscular route which are poorly populated by immune cells such as dendritic cells, macrophages or lymphocytes. Myocytes are the dominant cell types that express DNA vaccines. These are not antigen-presenting cells (APC) and the expressed antigens must be presented to APCs, such as dendritic cells, which migrate to draining lymph nodes to induce immune responses (Condon, Watkins et al. 1996, Tuting, Austyn et al. 2000). Mechanisms that facilitate APC uptake of antigens are expected to enhance the efficacy of vaccines. In the present study, the integrin binding motif Arg-Gly-Asp (RGD) was fused to the C-terminal end of the SARS-CoV spike protein antigen. This is to evaluate whether the short RGD tag enhances immune responses against the spike protein antigen. Dendritic cells phagocytose apoptotic cells through the v5 integrin which promotes cross-presentation of antigens. v5 is a RGD- binding integrin. In a previous study, RGD fusion to synthesized peptide antigens has been shown to enhance the immunogenicity of the peptide antigens which may involve αvβ5–mediated antigen uptake (Yano, Onozuka et al. 2003, Yano, Onozuka et
66 al. 2005) Antigen-specific antibodies were induced by intranasal immunization with RGD-tagged antigens without adjuvants.
The sequence and composition of a DNA vaccine may contribute some role in its immunogenicity (Arnon 2011). However its efficacy is directly related to the expression levels of the encoded protein in eukaryotic cells. Alteration of codon usage of viral antigens, to remove regulatory sequences and deletion of functional sequence elements are known to be useful for augmenting the immunogenicity of DNA vaccines Codon-optimized cDNA constructs increase protein antigen expression (Babcock, Esshaki et al. 2004). The gene encoding the S protein of SARS-CoV contains many codons used infrequently in mammalian genes for efficiently expressed proteins (Haas, Park et al. 1996, Marra, Jones et al. 2003, Rota, Oberste et al. 2003).
A study was done to compare different codon-optimized forms of the S-protein gene and its expression with its native viral sequence (Moore, Dorfman et al. 2004). Indeed the codon-optimized gene expressed more than twice as much protein as the native viral sequence.
As such we used the same codon-optimized SARS-CoV spike protein cDNA to modify and express the RGD-tagged antigen (Yang, Kong et al. 2004, Zhong, Zhong et al. 2006). Codon-optimized S gene SARS-CoV has been incorporated in the construction of DNA vaccine plasmids expressing either the full-length or segments of S protein (Wang, Chou et al. 2005). High titer S-specific IgG antibody responses were elicited in rabbits immunized with DNA against various segments of the S protein.
The spike protein antigen is expressed without the transmembrane and C- terminal domains with RGD being fused to the C-terminal end followed by a His tag for antigen purification and analysis. A spike protein DNA vaccine without tags was
67 shown to generate neutralizing antibody and protective immunity in mice (Yang, Kong et al. 2004). They also managed to demonstrate that humoral response in mice vaccinated with an expression vector encoding the spike glycoprotein which includes its transmembrane domain elicited neutralizing antibodies. However, it was concluded that protection was mediated by a humoral but not a cell-mediated immune mechanism. A secreted recombinant expression plasmid of the pVAX family that encoded a partial S glycoprotein fused to a signal peptide was constructed and immunized into BALB/C mice. A decrease in CD4+ and CD8+ T-lymphocytes accompanied by an increase in IFN-γ level in serum where levels of IL-4 were undetectable, indicated that humoral and cellular immune responses were elicited (He, Tang et al. 2005).
In our experiments, the spike protein vaccine was also potent in antibody induction. When the antigen is tagged with RGD, it showed decreased rather than increased antibody induction suggesting that either the RGD tag rendered spike protein less immunogenic or it polarizes immune responses away from humoral immunity. This became clearer when cellular responses were assessed. With some epitopes, S-RGD induced higher CTL activation than S-His. When IFN-γ induction was examined, it was found more effectively induced by S-RGD as compared with S- His. Therefore, the RGD tag appears to polarize the host from humoral to cellular responses which is in line with that initially hypothesized.
An unexpected observation after S-RGD DNA vaccine immunization is that the mice exhibit increasing fur loss with each booster dose leading to complete alopecia following the last boost. Alopecia was reported before in human scalp xenografts upon laminin-10 inhibition using antibodies specific for either (Li, Rao et al. 2003, Li, Tzu et al. 2003). In the latter settings, laminins are major components of
68 basement membranes to which epithelial cells attach and interruption of these attachments are expected to cause skin abnormalities (Hynes 1992, Engvall and Wewer 1996, McGowan and Marinkovich 2000). It is unclear how immunity against the RGD-tagged spike protein antigen or the antigen itself may cause skin abnormalities.
In addition, Zhi et al. (2005) demonstrated that codon optimization of the SARS-CoV S gene can greatly enhance S-specific CD8 T-cell responses in mice immunized with the recombinant S protein expressed by a simian adenoviral vaccine vector. In the present study, pcDNA-SS was more effective in eliciting humoral and cellular immune responses in vaccinated mice than pLL70. Although pcDNA-SS expressed codon-optimized S gene from the Urbani SARS-CoV strain and pLL70 noncodon optimized S gene from the Tor-2 strain, there is only one amino acid residue difference between the S protein sequences of these strains (serine to alanine in the position 577). It is unlikely that substitution of one amino acid will dramatically affect immunogenicity of the SARS-CoV S protein. Therefore, we suggest that genetic codon optimization and=or foreign signal sequence may play a major role in the enhancing of the immunogenicity of a DNA vaccine based on the SARS-CoV S protein.
A number of studies have found the receptor-binding domain to be the major immunodominant portion and to harbor potent neutralizing epitopes on the S protein (Zhang, Wang et al. 2004, Zhou, Wang et al. 2004, Chen, Zhang et al. 2005, He, Lu et al. 2005, Keng, Zhang et al. 2005, Wang, Chou et al. 2005). The spike protein is necessary for the infectivity and pathogenicity of coronaviruses. Neutralization domains have been identified on the S protein, one at the N-terminus (Ser12-Thr535) and the other at the C-terminus (Arg797-Ile1192) (Wang, Chou et al. 2005). B-cell
69 epitopes were also identified from the S glycoprotein. A multi-epitope DNA vaccine containing S437-459 and M1-20 were intramuscularly immunized into mice where a long-term memory humoral response was generated. The authors concluded that a DNA-prime-protein boost strategy greatly enhanced the antibody generation (Wang, Xu et al. 2008).
Identification of functional dominant epitopes for T cells is crucial for the understanding of cellular immune responses elicited by SARS-CoV DNA vaccine.
Zhi et al. (2005), identified the H-2b-restricted S436- and S525-epitopes, and concluded that the latter epitope to be more dominant in C57BL/6 mice. The strength of these epitopes appeared to correlate with the binding affinity to H2-Kb predicted by the SYFPEITHI algorithm. In contrast to these data by Zhi et al. (2005), it was found that the S436 epitope is a relatively weak MHC binder with an MFI of only 1.23.
Although S436 was ranked the highest MHC binder by the Rankpep and SYFPEITHI predictions, the S497, S525 and S884 epitopes exhibited higher MFI values (6.58, 6.6 and 4.81, respectively) than S436. Two spike-specific CD8 T-cell epitopes in mouse hepatitis virus (MHV)-infected C57BL/6 mice were previously identified. These were the S510-518 (dominant) and S598-605 (subdominant) (Castro and Perlman 1995, Bergmann, Yao et al. 1996). The dominant S510-518 epitope is located in a hypervariable region of the spike protein that appears to be readily deleted without loss of viability of MHV (Parker, Gallagher et al. 1989). In comparison, the S436-443 dominant epitope resides in the minimal region of S1 required for interaction with its cellular receptor, ACE2 (Babcock, Esshaki et al. 2004, Tan, Lim et al. 2005). In a study with SARS patient sera, Lu et al. 2004 discovered a fragment of spike protein (amino acid 441 to 700) to be the major immunodominant epitope.
70 This study demonstrated that prime-immunization of mice with SARS-CoV spike DNA vaccine constructs S-RGD/His and S-His induced antigen-specific cellular immune responses. In fact the cellular and humoral immune responses elicited by different combinations of gene-based and inactivated viral particles with various adjuvants were assessed (Kong, Xu et al. 2005). T cell responses were found to be altered by different prime-boost immunization, where optimal CD8+ T cell response induced by DNA priming and replication-defective adenoviral vector boosting.
Humoral immune response is known to be enhanced most effectively through the use of inactivated virus with adjuvants, associated with CD4+ T cells stimulation. After a prime-immunization of a SARS-CoV DNA vaccine, both CD4+ and CD8+ T cell responses are induced (Huang, Ma et al. 2006). Boosting mice intramuscularly with a SARS-CoV DNA vaccine enhances CD4+ and CD8+ T cell responses in both lymphoid and non-lymphoid organs persistent over months. In fact a single-plasmid DNA vaccine encoding the S glycoprotein was given in three priming-boost doses to humans to evaluate safety and immune responses (Martin, Louder et al. 2008).
Antibody against the SARS-CoV was detected in 80% of subjects where all mounted neutralizing antibody response. CD4+ T-cell specific response was detected in all subjects but only 20% mounted a CD8+ T-cell specific response.
While the S-RGD/His vaccine was not more efficacious than the S-His vaccine, Figure 3 appears to suggest that S-RGD is more potent in IFN-γ induction and Figure 2C/D suggests S-RGD also induced more CTL. Babcock et al. (2004) speculated that amino acids 1 to 510 of the spike glycoprotein represented a unique domain containing the receptor-binding site. Additionally, He et al. (2005) characterized the minimal receptor-binding domain to be between residues 318-510.
71 Therefore this gives more evidence that a vaccine targeting this region to be promising.
Based on the present study, the S436- and S497- epitopes which lie within these regions are highly immunogenic. CTLs induced by both S-His and S-RGD/His DNA vaccines kill RMA/S cells which were pulsed with the S436, S497, S525, S884 and S1116 epitopes. These epitopes also stimulate IFN-γ secretion. Besides the S436 and S525 epitopes (Zhi, Kobinger et al. 2005), the S884 epitope was discovered to be another strong MHC binder which elicits cytotoxic T-cell responses. The S1116 peptide was predicted to be a weak MHC binder in this study, but it induced the highest cytotoxic response of all peptide tested. It is also induced the strongest IFN-γ production. The S884 and S1116 epitopes are close to the carboxyl end of spike protein where the S2 subunit resides. A S1167 epitope (RLNEVAKNL) within the S2 subunit was found to be able to induce peptide-specific CTLs in both HLA-A2.1/Kb transgenic mice and human PBMCs (Wang, Sin et al. 2004). Two more S2 subunit epitopes S978 and S1203 were recognized by peripheral blood cells of recovered SARS patients with HLA-A*0201 (Wang, Chen et al. 2004, Chen, Hou et al. 2005).
Additionally, Tsao et al. (2006) reported that the S787 and S1042 epitopes were HLA-A*0201 restricted, with the latter being able to induce recall responses in PBMC harvested from SARS-recovered-patients when in vitro incubated against their cognate antigen (Tsao, Lin et al. 2006). HLA-A*0201 transgenic mice were primed intramuscularly with a DNA vaccine encoding the S glycoprotein and boosted subcutaneously with its haplotype-restricted peptides (Zhao, Yang et al. 2010). CD8+ T cell responses in the transgenic mice were found to be elicited by epitopes from the S glycoprotein.