Both CD4 and coreceptor binding sites on gp120 form potential targets for antibody-mediated intervention. The primary isolate-neutralizing antibodies identified thus far that are specific for gp120 are directed to epitopes that appear to be present on the oligomeric Env complex before contact with CD4. They are believed to act, at least in part, by binding to the virion surface and sterically obstructing the interaction between virus and target cell (37
). For these antibodies, the larger whole antibody molecules are more effective than the corresponding Fab fragments at neutralization due to greater steric obstruction but probably also because of increased avidity (bivalency) (57
). Antibodies to the CD4i epitopes generally do not display primary isolate neutralizing activity at relevant concentrations.
The major finding of our study is that, for some HIV-1 isolates, the size of CD4i-specific (antibody-derived) neutralizing agents is inversely correlated with neutralization efficiency. Thus, antibody fragments are more effective than whole antibody molecules in neutralization. Further, scFv (25 kDa) are generally more effective than Fab (50 kDa) fragments. The temperature-regulated (Fig. ) and SOS-arrested (Fig. ) neutralization data presented here are consistent with a model in which CD4i scFvs or Fabs, but not IgGs, can neutralize after attachment of the virus to CD4 (43
). Because the difference in size is expected to have limited effects on the diffusion rates of the CD4i antibody fragments, these results strongly suggest that the restriction against CD4i antibody neutralization is steric, not temporal. Therefore, the likeliest explanation for the observed sensitivity of neutralization to antibody size is that, after CD4 binding to the virus, the available space between the virus and the target cell surface is not enough to accommodate a whole antibody molecule but is sufficient for antibody fragments. This is consistent with the estimated size of the antibody fragments and predictions of the distance between the CD4i epitope on the CD4-bound envelope glycoprotein trimer and the target cell membrane (Fig. ). Recent findings demonstrating the inaccessibility of the 17b epitope at the fusing cell interface are in agreement with this hypothesis (10
FIG. 6. Steric restrictions on CD4i antibody neutralization after CD4 attachment. (A) HIV-1 viral spike attached to cell surface CD4. The gp120 trimer (brown) (27) is shown with the threefold axis oriented perpendicular to the target cell surface. Cα (more ...)
The higher sensitivity to CD4i MAbs observed for some isolates is most likely a reflection of the exposure of the CD4i epitope on the oligomer prior to CD4 binding, as suggested previously (37
). In addition, our data suggests that when the CD4i epitope is exposed, any advantage of Fab or scFv fragments disappears (Tables and ). It is, however, noteworthy that exposure of the CD4i epitope, through deletion of the V1/V2 stem-loop structure, not only facilitates virus neutralization by the whole IgG molecule but also increases the potency of the Fab and scFv fragments (Fig. ). Comparison of the relative neutralization potencies of 17b antibody and antibody fragments under different assay conditions (Fig. ) implies that the V1/V2 variable loops continue to play a role in obstructing antibody binding even after CD4 attachment. Thus, our results suggest that CD4i MAb-mediated neutralization is biphasic, with (i) a “preattachment” phase that may not be antibody fragment size dependent but is dependent on the accessibility of the CD4i epitope on the resting oligomer, primarily governed by the V1/V2 variable loops, and (ii) a “postattachment” phase that is antibody fragment size dependent due to steric restrictions imposed by the cellular membrane and the V1/V2 variable loops. The fact that we observe more effective neutralization by X5 of HIV-1JR-CSF/IR
cells than on CXCR4+
cells (Fig. ) and the fact that scFv X5, but not scFv 17b, was more effective than the corresponding Fab fragments in neutralization of HIV-1HxB2
(Table ), however, indicates that there are additional variables influencing these mechanisms. One could speculate that the precise orientation of the epitope may influence the susceptibility of the antibody fragments to steric constraints by the surrounding protein and carbohydrate structures at the virus-cell interface.
Attempts to improve Env immunogens include strategies to better expose the coreceptor binding site and the overlapping CD4i epitopes (11
). It is argued that increasing CD4i epitope exposure may elicit more-potent CD4i MAbs or higher serum antibody titers. Our data, however, suggest that the inability of CD4i MAbs to neutralize primary isolates is not due to lack of potency per se but to spatial constraints. Since non-syncytium-inducing, R5 HIV-1 variants establish primary infection in humans (39
), the inability of CD4i-specific whole antibodies to neutralize the majority of R5 viruses used in the present study raises concerns as to whether the CD4i epitopes would be useful targets for antibody neutralization in vaccine design.
In conclusion, we show that CD4i-specific MAbs do not neutralize some primary isolates due to steric constraints. We propose that HIV-1 can exclude whole antibody molecules from the CD4i epitopes due to the close physical proximity of the cellular membrane and obstruction by the V1/V2 variable loops. The constraints were especially apparent for the primary R5 isolates tested. This raises questions about the utility of CD4i epitopes as targets for antibody-mediated neutralization in vaccine design. Understanding these viral defenses may suggest new strategies to circumvent them. The fact that the smaller antibody fragments were able to neutralize does, however, suggest that the CD4i epitopes could be used as targets for small molecule entry inhibitors.