Role of neutralizing antibodies in HIV infection.

Abstract

Introduction The year 2002 is considered by some as the year during which a revival in interest in anti-HIV neutralizing antibodies took place. It has been known for some years that neutralizing antibodies exert pressure on HIV during infection, resulting in escape variants Envelope mutations. New important information has emerged that increases our understanding of the protective role that neutralizing antibodies play during HIV infection, and has improved our knowledge of the ways in which HIV evades the immune pressure exerted by neutralizing antibodies. The epitope of a broadly reactive neutralizing antibody was defined, and new neutralizing antibodies that bind to occluded Envelope epitopes were isolated from HIV-infected patients. Novel concepts on the mechanisms of antibody-mediated HIV neutralization were put forth, and new assays were developed to evaluate the neutralizing activity of serum antibody. Vaccination studies in primates highlighted the importance of neutralizing antibodies in controlling infection. Finally, information concerning the antigenic and immunogenic structure of the HIV Envelope protein emerged that could lead to the design of more effective HIV Envelope immunogens. Protective role of neutralizing antibodies Whether or not neutralizing antibodies play a protective role during the early stages of HIV infection remains one of the most hotly debated issues in HIV research. Using the SIV/macaque experimental model, two groups examined whether neutralizing antibodies participate in the early control of infection, by artificially depleting the B cells from animals before their exposure to SIV [1,2]. Although both studies emphasized the role that neutralizing antibodies play during chronic infection, one group concluded that neutralizing antibodies play a limited role during acute infection [2], whereas the other group suggested that the role of neutralizing antibodies during this period is under-appreciated [1]. A direct role for Envelope antibodies in protection was demonstrated in a comparative SHIV challenge study in macaques [3]. The more rapid appearance of anti-Envelope antibodies recorded in the Gag–Pol–Env group, compared with the Gag–Pol group, resulted in more complete protection from CD4 T-cell loss. A more detailed analysis of the generation and evolution of neutralizing antibodies in recently infected HIV patients indicated that neutralizing antibodies are indeed involved in controlling viral replication during the first months after infection, and that the pressure they exert on the virus is significant [4]. Using a sensitive neutralization assay, it was demonstrated that potent autologous, but not heterologous, neutralizing antibodies become detectable within the first 2 months post-infection. Although the role of neutralizing antibodies during acute infection remains debatable, the early emergence of neutralizing antibodies forces HIV to mutate significantly. That study also suggested that the observed differences in the ability of an HIV-infected patient to elicit neutralizing antibodies might not be related to phenotypic differences of the infecting virus. As neutralizing antibodies are generated during infection, specific mutations are gradually but continuously incorporated in the HIV envelope gene, and some of these changes result in the modification of the glycosylation pattern of this viral protein [5]. The concept of an evolving ‘glycan shield’ on the viral Envelope as a potential mechanism of neutralization escape was proposed. Although the protective role that carbohydrate moieties have on HIV and SIV has previously been demonstrated, the study emphasized the fact that the glycosylation changes observed can be located outside the principal neutralization determinants themselves. Glycans can be strategically positioned and repositioned continuously throughout infection, in a way that affects minimally the interaction of the envelope with the receptor molecules on the cell surface, but hinders maximally the binding of neutralizing antibodies. Sugar moieties Although sugar moieties on the HIV surface can mask neutralization epitopes, in rare instances they become targets of neutralizing antibodies. Two research groups reported that the epitope recognized by the broadly neutralizing antibody 2G12 is composed of terminal mannose residues present at asparagines located on either side of the base of the V3 loop and at the carboxyl terminal side of the C3 region, with limited involvement of the underlying amino acids [6,7]. This recent finding may lead to the identification of additional neutralizing antibodies that recognize oligosaccharide molecules on the surface of HIV virions. It may even assist in the development of more effective HIV Envelope-based immunogens. Binding of non-neutralizing antibodies to the HIV surface Several groups previously demonstrated that certain antibodies interact with virion-associated Envelope molecules without preventing infection, even though a recent study [8] indicated that their epitopes may share extensive overlap with that of neutralizing antibodies. Two recent studies [9,10] reported that the binding of non-neutralizing antibodies to the virion surface may hinder the binding of neutralizing antibodies, but does not decrease their neutralizing activity. The hypothesis put forward to explain this observation is that although neutralizing antibodies and non-neutralizing antibodies bind to non-functional Envelope molecules on the virion surface, neutralizing antibodies alone bind to functional Envelopes. The nature of such non-functional Envelope molecules is unclear at the moment, but probably includes various Envelope forms that do not support virus–cell fusion. Isolation of neutralizing antibodies from HIV-infected patients Several novel neutralizing monoclonal antibodies that preferentially bind to their epitopes during HIV Envelope-CD4 cell binding (CD4 cell-induced epitopes) were isolated and characterized. One of them, termed X5, was isolated from a phage display library from an HIV-infected patient with relatively high cross-clade neutralizing antibody titers, using a trimolecular complex formed from the association of gp120, CD4 and CCR5 cells as a screening molecule [11]. CD4-binding to gp120 and CCR5-binding to the gp120–CD4 cell complex facilitates the binding of X5 to its epitope. Three additional neutralizing antibodies, 23e, 49e, and 21c, which also recognize CD4 cell-induced epitopes, were isolated from Epstein–Barr virus-transformed B cells from a long-term non-progressor HIV-1-infected patient [12]. Their epitopes overlap and are centered around a region of the fourth conserved domain on gp120, which participates in gp120–CCR5 binding. However, the fact that these four antibodies neutralize both X4 and R5-tropic isolates suggests that their epitopes are not specific for CCR5 binding. The epitope accessibility of such antibodies is hindered by a ‘thermodynamic barrier', which is alleviated once the Envelope binds to CD4 cells or the chemokine receptors [13]. Despite this, the recent isolation of several antibodies that recognize the relatively inaccessible CD4 cell-induced epitopes, suggests that these epitopes are immunogenic in at least certain HIV-infected patients. The finding that antibody binding to these epitopes results in HIV neutralization makes these epitopes targets for vaccine development. The V3 loop Another prominent region of the HIV Envelope that contains neutralization epitopes is the V3 loop. The majority of anti-V3 loop antibodies elicited by vaccination recognize linear epitopes and do not have cross-neutralizing activity. In a recent study [14], anti-V3 loop neutralizing antibodies that recognize conformational epitopes on virion-associated subtype A, B, C, D and F Envelope molecules were isolated from HIV-infected patients. These results imply that the conformation of the V3 loop may share common features among isolates of subtypes A, B, C, D and F. The existence of common V3 loop structures, irrespective of amino acid composition, is supported by a recent report proposing that isolates displaying the same chemokine receptor specificity share common V3 loop structural features [15]. Structural similarities between B and C-derived Envelopes were also identified in regions other than the V3 loop, such as the region encompassing the epitope recognized by monoclonal antibody IgG1b12 [16]. Interestingly, other Envelope regions may differ in structure or glycosylation between clade B and C isolates, because the latter but not the former isolates are resistant to neutralization by the anti-gp41 monoclonal antibody 2F5 and the anti-gp120 monoclonal antibody 2G12. Such studies suggest therefore that HIV Envelope immunogens properly expressing conserved structural elements could in theory elicit cross-clade neutralizing antibodies targeting non-V3 loop epitopes. The appearance of a mutation in the proximal limb of the V3 loop, 313-4 PM, in the R2 isolate renders it susceptible to neutralization by the anti-V3 loop monoclonal antibody 19b, an anti-CD4 cell-induced monoclonal antibody, and the patient's serum [17]. Key changes within and around the V3 loop arising in vivo in the course of infection in some patients may thus be important in the generation of broad neutralizing antibodies. Whether these sequences will be better immunogens in vaccines remains an open question. Cross-reactive neutralizing antibodies and vaccines Although natural HIV infection results in the generation of cross-reactive neutralizing antibodies, immunization with HIV Envelope proteins elicits primarily autologous neutralizing antibodies. A recent study [18] described the development of cross-reactive neutralizing antibodies that were elicited in mice and rabbits immunized with alphavirus-derived particles expressing the Envelope of a primary HIV isolate (R2), isolated from a patient whose serum displayed broad neutralizing activity as described above. Such studies are interesting because they not only demonstrate the generation of cross-neutralizing antibody responses by vaccination, but these responses were elicited without HIV Envelope protein boosting. Whether the generation of cross-reactive neutralizing antibodies is caused by the method of immunization, the immunogen, or both needs to be further examined. In general, the titers of antibodies elicited by vaccination with HIV Envelope immunogens are low, and several methods to enhance the immunogenicity HIV Envelope immunogens are currently being evaluated. It was recently demonstrated that chimeric molecules between gp120 and the C3d component of the complement cascade, when used as immunogens in mice and rabbits, elicit higher titers of antibodies than the gp120 protein alone [19]. These antibodies target autologous but not heterologous viruses. In an effort to increase the ability of HIV Envelope immunogens to elicit neutralizing antibodies against conserved neutralization epitopes, modifications within and around hypervariable regions have been introduced. To date, immunization with such modified envelope constructs has resulted in inconclusive and contradictory results. The immunization of macaques with the Envelope gp140 form of the SF162 isolate lacking the central region of the V2 loop (ΔV2gp140), or immunization of mice with the HXB2 Envelope gp120 lacking the V1, V2 and V3 loops, elicited modest neutralizing antibody responses against heterologous primary HIV isolates [20,21]. In contrast, the immunization of mice with the DH12-derived Envelope lacking hypervariable regions, or with the 89.6-derived Envelope lacking multiple potential N-linked glycosylation sites from the V1V2 region, or with the BRU Envelope lacking a conserved N-linked glycosylation site in the V3, not only failed to generate cross-reactive neutralizing antibodies, but in certain cases not even autologous neutralizing antibodies [22–24]. The different results cannot simply be explained by differences in the animal species used during vaccination and differences in the method of immunization. The effect that a modification has on the Envelope structure and immunogenicity is probably Envelope-background- dependent. A different approach to enhance the immunogenicity of ‘cryptic’ HIV neutralization epitopes consists of engineering stable complexes between the HIV Envelope and CD4 cell receptor derivatives. The binding of CD4 cells to gp120 or gp140 proteins induces Envelope conformational changes during which the exposure of certain neutralization epitopes increases. It was demonstrated that such chimeric HIV Envelope–CD4 immunogens elicit low titers of neutralizing antibodies against a variety of primary HIV-1 isolates [25]. Additional approaches are currently being evaluated by several groups to improve the immunogenicity of such promising immunogens further. Neutralizing antibodies elicited in human vaccine trials With much of the vaccine focus in recent years on cell-mediated responses, there has been little progress in the development and inclusion of new Envelope subunits designed explicitly to increase neutralizing antibodies. Nonetheless, Envelope gp120 monomer has been advanced both as a single agent and as a boosting agent. In phase I trials published in 2003, vaccination with canarypox boosted with gp120 led to the development of neutralizing antibodies in 95% of seronegative recipients [26]. However, phase III trials conducted by VaxGen using Envelope gp120 monomer failed to achieve the 30% level of efficacy needed for significance in this arm of the study [27,28]. It will be valuable to determine how well the vaccine primed the vaccinees for neutralizing antibodies, particularly those who became infected, relative to sham-vaccinated individuals in the control arm. Neutralizing antibodies block infection Studies in primate and murine models have consistently shown that high levels of neutralizing antibodies are required to block infection, regardless of the route of exposure. In a study with polyclonal IgG prepared from HIV-infected chimpanzees [29], plasma titers of passively transferred neutralizing antibodies at the time of virus inoculation were correlated with protection from infection in SHIV-challenged animals. A neutralizing antibody titer of 1:38 was 99% effective in neutralizing virus in vitro correlated with protection from infection by SHIV-DH12, a dose similar to other hyperimmune Ig clinical applications. Such studies have further underscored the critical role of neutralizing antibodies in providing sterilizing immunity, but the high levels needed have further undermined the confidence that vaccines can generate, let alone sustain, such high levels of neutralizing antibodies. Efforts to determine whether neutralizing antibodies might be useful as blocking agents at the vaginal surface were tested in macaques when monoclonal antibody IgG1b12 was applied mucosally in gel [30]. Protection was observed when high concentrations of the monoclonal antibody were applied and when SHIV exposure occurred shortly after treatment. Human monoclonal antibodies show promise A number of studies published in recent years have shown that neutralizing monoclonal antibodies of the IgG class alone can be effective in blocking the infection of non-human primates by mucosal challenge with SHIV. Such studies provided a rationale for testing groups of monoclonal antibodies with synergistic neutralizing antibodies in vitro as immediate post-exposure prophylaxis, modeling for perinatal exposure in infants. Cocktails of human IgG1b12, 2G12, 2F5, and 4E10 neutralizing monoclonal antibodies prevented disease in newborn macaques and prevented the establishment of SHIV89.6P infection in half of the animals when given within an hour of exposure [31]. An important step was taken in 2003 towards the use of monoclonal antibodies in humans with the first phase I study of monoclonal antibodies 2F5 and 2G12. After repeated intravenous infusions to asymptomatic HIV-1-infected individuals, there were no adverse effects, and a transient reduction in viral loads was observed in five out of seven patients [32]. Vaccines and monoclonal antibodies to limit perinatal transmission will need to function in the face of maternal anti-HIV neutralizing antibodies. In studies in macaques [33], it was shown that maternal antibody did not limit the ability of neonates to respond to vaccination. Finally, phase I/II trials were initiated in Uganda to test increasing doses of HIV immune globulin in the maternal–infant setting [34]. New assays for measuring HIV neutralizing antibodies Neutralization assay standardization, reproducibility, and high throughput remain key objectives for HIV research, despite nearly two decades of development. The year 2003 was one of the more productive years in expanding the repertoire and scope of the available assays. The precise and direct enumeration of HIV-infected peripheral blood mononuclear cells by flow cytometry was shown to be feasible as soon as one day post-infection by staining for intracellular p24 antigen. The resulting 2-day assay was highly sensitive and specific [35]. Another assay included ‘plaque’ formation in U87.CD4-CCR5 and U87.CD4-CXCR4 cells, in which infected cells form syncytia that are stained with hematoxylin and enumerated by light microscopy [36]. Importantly, the development and utilization of an assay for autologous plasma HIV as a population showed the importance of examining the susceptibility of these Envelope-pseudotyped virions in proportion to their presence in the plasma [4]. The combination of highly sensitive assays and relevant target virus isolates should further enhance our understanding of the development of autologous and heterologous neutralizing antibodies. In conclusion, the research highlighted in this review is meant to be representative rather than comprehensive. We believe that the significant advance seen in the area of neutralizing antibodies resulted from the synthesis of studies encompassing diverse disciplines. The information gained will not only enhance our understanding of HIV pathogenesis and transmission, but will also assist in the development of more effective vaccination and therapeutic approaches.