Comparative in silico analysis of two vaccine candidates for group A streptococcus predicts that they both may have similar safety profiles

Background: Concerns of immune cross-reactivity, between epitopes of the group A streptococcal (GAS) M-proteins and host proteins have hindered the progress of an effective GAS vaccine. An ideal M-protein based subunit vaccine should not elicit heart tissue crossreactive antibody responses and should not activate M-protein specific $CD4^+$ T-cells. In the current study we used a bioinformatic and immunoinformatic approach to assess the safety of J8 and J14, chimeric vaccine constructs containing a GAS derived M-protein epitope embedded in flanking GCN4 region. We demonstrate that at the primary amino acid level J8 and J14 show very little homology to human proteins. ProPred, RANKPEP and HLABIND algorithms failed to predict significant binding between the M-protein specific regions of J8 and J14 and class II binding alleles. A single peptide was predicted to bind to HLA class I allele B_2705. This data was supported by cellular proliferation assays demonstrating few periferal blood mononuclear cells (PBMCs) from donors respond to J8 and J14. Reassuringly, there was no correlation between proliferation to these peptides, and proliferation to host proteins. This data suggests that J8 and J14 are unlikely to induce cross-reactive immune responses, and will be safe for use in humans