Next we evaluated neutralization of Jc
Next, we evaluated neutralization of Jc1, Jc1-N534A, Jc1-ΔHVR1 and Jc1-ΔHVR1-N534A by sera from individuals chronically infected with HCV GT1 and GT2 (Fig. 4A-C). Like the results of neutralization with monoclonal Kartogenin and CD81-LEL (Fig. 3), the sera poorly neutralized Jc1. Neutralization was enhanced by elimination of the N534 site in Jc1, further boosted by deletion of HVR1, and it was maximal when the deletion of HVR1 was combined with the inactivation of the N534 glycosylation site with minor interpatient variation, independent of the patients’ HCV GT (Fig. 4A-C). Thus, deletion of HVR1 together with inactivation of glycosylation site N534 enhanced antibody and CD81 binding to virus particles. As well, this had the most drastic effect on neutralization by bNabs, soluble CD81 and patient sera. As these viruses are infectious, they present well-folded and functional viral envelope proteins that display conserved epitopes in a more accessible manner. Thus, envelope proteins carrying these modifications, particularly the deletion of HVR1 combined with the N534A mutation, may constitute good candidates for focusing antibody responses on conserved viral epitopes.
Discussion All vaccination approaches induced robust cross-binding antibody responses as determined by ELISA assays involving recombinant proteins or cell extracts expressing E1-E2 heterodimers from all major HCV GTs (Fig. 5). Vaccination with E1-E2 heterodimers induced antibodies that competed with the binding of previously described bNAbs AR3B, 1:7, HC33.4 and AP33 and we detected virus neutralization in HCVpp assays (Fig. 7). However, inactivation of N417 or N532 glycosylation sites in the context of H77 E1E2 heterodimers did not grossly affect these responses although the N532Q mutation mediated increased binding of the recombinant E1E2 complex to CD81. Among vaccinations with 293T-derived E2 proteins immunization with J6-E2ΔHVR1 and the immunization series A and B mounted superior cross-binding antibodies compared with J6-E2 and the other examined approaches. However, this was only detected when recombinant E2 proteins from 293T cells were used as ELISA antigens and not when E1-E2 heterodimers partially purified from HCV replicating cells were employed. Thus, some of our E2-protein based vaccination approaches successfully improved production of cross-binding antibodies to the recombinant immunogens but not to the E1-E2 heterodimer expressed in infected cells. Thus, apparently there are structural differences between these proteins and E1-E2 heterodimers. These differences could impact on vaccine efficacy and they should be considered when standardizing tests between laboratories and when estimating cross-binding between different studies. It is also possible that such structural differences explain why vaccination with 293T-derived E2 proteins triggered only modest autologous neutralizing antibodies and no cross-neutralization. It has been shown that purified, UV-inactivated HCVcc particles induce cross-neutralizing antibodies in mice and non-human primates.,  Thus, it will be interesting to explore if vaccination with the HCVcc variants characterized here induces a focused and enhanced neutralizing antibody response. It is surprising that vaccination with E2 from 293T cells induced very robust, cross-binding antibodies that, however, did not cross-neutralize. One possible explanation was that these immunogens induced antibodies that interfere with cross-neutralization as has been described previously.,  However, the vaccine-induced antibodies did not interfere with cross-neutralization of HCV patient-derived polyclonal antibodies (Fig. 6). Thus, we believe that the 293T-derived E2 vaccines examined here efficiently induced antibodies that target conserved binding epitopes, but that targeting these epitopes is not cross-neutralizing. It is possible that this reflects in part a species-specific limitation of immunogenicity of recombinant E2 proteins in mice as recent results by other groups and us suggest that anti-HCV neutralizing antibody responses are more readily induced in guinea pigs.,  Although neutralizing antibodies were elicited, those were primarily strain-specific, thus explaining the neutralization of autologous Jc1 particles. Notably, the conserved CD81 binding site seems to be structurally highly flexible, and it was proposed that this flexibility may favor induction of non-neutralizing antibodies. In line with this, it is possible that the recombinant E2 proteins used here preferentially adopt a conformation that induces binding but non-neutralizing antibodies. Indeed, the conformation that induces neutralizing antibodies may be underrepresented in these recombinant proteins. Thus, epitope exposure alone may be insufficient and should be complemented by vaccine design aiming at rigidifying this region.