Date Published: September 26, 2013
Publisher: Public Library of Science
Author(s): Kathryn A. K. Finton, Kevin Larimore, H. Benjamin Larman, Della Friend, Colin Correnti, Peter B. Rupert, Stephen J. Elledge, Philip D. Greenberg, Roland K. Strong, Alexandra Trkola.
The broadly-neutralizing anti-HIV antibody 4E10 recognizes an epitope in the membrane-proximal external region of the HIV envelope protein gp41. Previous attempts to elicit 4E10 by vaccination with envelope-derived or reverse-engineered immunogens have failed. It was presumed that the ontogeny of 4E10-equivalent responses was blocked by inherent autoreactivity and exceptional polyreactivity. We generated 4E10 heavy-chain knock-in mice, which displayed significant B cell dysregulation, consistent with recognition of autoantigen/s by 4E10 and the presumption that tolerance mechanisms may hinder the elicitation of 4E10 or 4E10-equivalent responses. Previously proposed candidate 4E10 autoantigens include the mitochondrial lipid cardiolipin and a nuclear splicing factor, 3B3. However, using carefully-controlled assays, 4E10 bound only weakly to cardiolipin-containing liposomes, but also bound negatively-charged, non-cardiolipin-containing liposomes comparably poorly. 4E10/liposome binding was predominantly mediated by electrostatic interactions rather than presumed hydrophobic interactions. The crystal structure of 4E10 free of bound ligands showed a dramatic restructuring of the combining site, occluding the HIV epitope binding site and revealing profound flexibility, but creating an electropositive pocket consistent with non-specific binding of phospholipid headgroups. These results strongly suggested that antigens other than cardiolipin mediate 4E10 autoreactivity. Using a synthetic peptide library spanning the human proteome, we determined that 4E10 displays limited and focused, but unexceptional, polyspecificity. We also identified a novel autoepitope shared by three ER-resident inositol trisphosphate receptors, validated through binding studies and immunohistochemistry. Tissue staining with 4E10 demonstrated reactivity consistent with the type 1 inositol trisphosphate receptor as the most likely candidate autoantigen, but is inconsistent with splicing factor 3B3. These results demonstrate that 4E10 recognition of liposomes competes with MPER recognition and that HIV antigen and autoepitope recognition may be distinct enough to permit eliciting 4E10-like antibodies, evading autoimmunity through directed engineering. However, 4E10 combining site flexibility, exceptional for a highly-matured antibody, may preclude eliciting 4E10 by conventional immunization strategies.
An effective prophylactic AIDS vaccine will need to generate anti-HIV neutralizing antibodies (Abs) that target the HIV envelope glycoprotein (Env) – and broadly neutralize as many HIV isolates as possible (bNAbs). The bNAb 4E10 – recognizes an epitope that is highly conserved across HIV-1, HIV-2, and SIV and displays one of the widest breadths of any anti-HIV bNAb, neutralizing 98% of HIV-1 strains , . These properties have made 4E10 an attractive vaccine target, but previous attempts to elicit 4E10 or equivalent Abs through vaccination have failed.
When applying “reverse vaccinology”  to the design of an HIV vaccine, it is important to determine whether or not the targeted bNAbs, e.g. 4E10, are autoreactive, and, if so, whether autoantigen recognition is separable from antigen recognition and neutralization. Despite usage of a breadth of murine light chains, 4E10H knock-in mice exhibited a profound blockade during B cell development consistent with the induction of central tolerance mechanisms, in which immature B cells expressing an autoreactive BCR are deleted (Fig. 1). In addition, the small residual population of B cells exported to the periphery, while able to respond to a variety of mitogenic signals, was unable to proliferate in response to IgM cross-linking. This phenotype substantiates the hypothesis that 4E10 is autoreactive. The questions then arise as to what autoantigen/s are mediating deletion and anergy and whether the mechanisms of 4E10 recognition of HIV and autoantigen epitopes are interdependent or separable, permitting the design of immunogens that generate Abs functionally equivalent to 4E10 with focused specificities that evade central and peripheral tolerance mechanisms.