Here we describe: i) a quantitative change in the Env-specific plasma Ab titers in four out of five persistently aviremic vaccinees, ii) subtractive biopanning with recombinant phages encoding random peptides as a new tool to dissect qualitative (virus-induced) changes in the humoral immune responses of vaccine-protected animals, iii) an increase of vaccine-induced anti-V3 binding Abs and induction of novel, cross-neutralizing anti-V3 Abs due to live SHIV-C encounters in animals that had remained virus-free.
Different immunization studies performed by our group yielded a cohort of five vaccinees that resisted multiple live SHIV-C exposures completely. Here, we dissected the Env-specific Ab responses in these monkeys. More specifically, we investigated the consequences of live-virus challenges on the strength and specificity of vaccine-induced Ab responses in the absence of any detectable viral load. Unexpectedly, we detected higher Env-specific binding Ab titers in four out of five aviremic RMs after they had been exposed to live SHIV-C. The boosting of Ab responses was independent of the challenge route. Similar effects were observed after oral (for animal RAt-9 [8–10]) and intrarectal challenges (for all the other monkeys tested).
To verify that the increased anti-Env Ab titers were a direct consequence of the virus challenges, we compared the dynamics of anti-gp140 Ab responses in vaccinated RMs with or without subsequent exposures to live virus. As expected, we detected a vaccine-induced boost of Ab responses about two weeks after the last protein immunization in both groups of animals. Remarkably, these titers kept increasing during the repeated virus challenges although no viremia ensued. In contrast, immunization without subsequent virus challenges resulted in the expected increase in anti-gp140 Ab levels that peaked at week 2 post-immunization and continuously declined thereafter. These data confirmed a live-virus induced increase of pre-existing, vaccine-induced Ab responses.
We sought to examine the Ab repertoire in two of the protected animals further and determined whether the live-virus challenges only boosted the pre-existing, vaccine-induced responses or also induced new Abs against neo-antigens represented by the Env proteins on the challenge virus particles. Phage display  has been used to analyze humoral immune responses in the context of chronic infections with immunodeficiency viruses [23–27]. Here, we designed a novel subtractive biopanning strategy to probe the paratopes of virus-induced Abs in vaccine-protected animals. We hypothesized that depleting phages recognized by Abs present after immunization but before any virus exposures would select for phages recognized by live virus-induced Abs only. We performed subtractive biopannings using polyclonal plasma samples from completely protected vaccinees and analyzed the selected phagotopes for similarities to Env sequences. This approach allowed us to identify the V3 crown as an epitope specifically recognized by week 7 but not week 0 Abs and shared by the two aviremic vaccinees, although subtractive biopanning revealed other Env regions for these vaccinees individually (data not shown).
Recently, the V3 crown and its conserved structural elements that are involved in co-receptor binding were shown to be important targets for bnAbs, including mAbs 447/52-D, 2219, 3074, 33B2, 33C6 [19, 28–30] and HGN194 ; the latter also provided complete cross-clade protection against SHIV-C acquisition in vivo. Together with the novel bnAbs, PG9 and PG16, targeting quaternary epitopes formed by the V2 and V3 loops in trimeric Env , the V3 loop is now considered again as target for vaccine development (reviewed in ).
By phage ELISA, we confirmed that the V3 mimotopes isolated by subtractive biopanning were only recognized after but not before the first virus challenge. Importantly, these responses were also boosted by the re-challenge with a high-dose of the same challenge virus. These data indicate that both low- and high-dose challenges can induce a similar boosting effect.
Our detailed analysis of anti-V3 responses using the actual V3 peptides revealed a significant boosting of immunogen-induced anti-HIV 1084i V3 binding Abs after the multiple low-dose challenges in five out of five vaccinees tested. Importantly, the same boosting was noticed in the two completely protected animals. This is remarkable considering that no viremia was ever detected for >3 years. In addition to these altered binding Ab titers, we also observed an increase in neutralizing activity against the challenge virus in the same aviremic RMs. Peptide absorption linked these increased nAb titers to anti-V3 responses. Interestingly, most of the induced cross-neutralizing anti-V3 Abs targeted the antigenically different version of the V3 crown presented by the challenge virus. Based on these results, we propose that the multiple live-virus exposures boosted anti-HIV-1 responses and altered their specificity in the absence of any detectable viremia.
Pre-existing antiviral immunity has been considered problematic for the recognition of antigenically diverse strains by the “primed” immune system. This idea of a compromised immune system was first discussed in the context of influenza virus antigens  and termed Original Antigenic Sin (OAS) . Ab formation during initial influenza infections in childhood was believed to greatly influence future Ab formation against newer strains encountered later . More specifically, vaccination with one strain of influenza virus and subsequent exposure to a heterologous strain would induce anamnestic Ab responses against the initial strain . Later, Nara and coworkers expanded the concept of OAS to the model of deceptive imprinting and defined it as a mechanism leading to a fixed state of immunity which in turn fails to adapt to a changing, but similar pathogen [36, 37]. Moreover, they described the induction of clonally restricted B-cell responses due to immunodominant epitopes found on the original antigen (reviewed in ). With regard to HIV-1 infections, it was postulated that initial vaccine-induced nAb responses would target immunodominant epitopes of variable gp120 regions (especially the V3 loop), which mutate due to immune selection pressure . Thus, nAb responses would be limited to the primary virus variant, which is considered potentially problematic during chronic HIV-1 infection, as well as for the formation of effective nAb responses against non-homologous HIV-1 strains in vaccine recipients [36–39]. Yet, we  and others [41–43] gave evidence that the occurrence of OAS is not absolute.
Investigating the anti-V3 binding Ab repertoire in vaccine-protected RMs, we indeed detected a statistically significant boosting of the initial vaccine-induced binding Abs after live-virus encounters with the heterologous strain, supporting the concept of OAS. However, our data do not imply a limitation in subsequent Ab responses against non-homologous virus strains as suggested by the model of deceptive imprinting [36, 37]. Based upon the initial selection of V3 mimotopes and their specific recognition of plasma samples after but not before live SHIV-C exposures, we demonstrate a change in the Ab repertoire as a consequence of live-virus encounters. This neo-antigen reactivity was then confirmed indirectly by peptide absorption analysis showing the induction of nAbs against an antigenically different version of the same V3 epitope. Thus, despite the pre-existing anti-V3 Abs present after immunization, virus-challenged RMs that never developed viremia produced cross-neutralizing anti-V3 Abs targeting the epitope version presented by the challenge virus.
Of note, the boosting of virus-specific Ab responses as a consequence of live-virus exposures that failed to cause detectable viremia was not restricted to Env only. According to our recent work , multiple live-virus exposures also affected anti-Tat Ab responses; we showed that mimotopes displaying the N-terminus of HIV-1 Tat were only recognized by Abs of protected vaccinees, RRi-11 and RTr-11, after the live-virus exposures (week 7), but not on the day of challenge (week 0). Subsequent quantitative ELISAs using the full-length HIV-1 Tat protein and Tat peptides confirmed an increase of anti-Tat Abs after multiple exposures to SHIV-C in three out of four aviremic vaccinees (RRi-11, RTr-11 and RAt-9). These data indicate that the virus challenges altered Ab responses against at least two different HIV-1 proteins, Env and Tat, in the absence of systemic infection.
Animals ROb-12, RAt-9, RRi-11 and RTr-11 were never viremic. How did the multiple low-dose virus challenges alter the vaccine-induced anti-Env and anti-Tat Ab repertoire in these animals? Theoretically, three mechanisms could be responsible: i) cryptic infection of target cells and their subsequent lysis. This may have produced sufficiently high concentrations of challenge virus proteins to induce Abs with altered specificities; ii) formation of immune complexes of virions and/or soluble protein with pre-existing, vaccine-induced Abs followed by efficient binding to and presentation by Fc-gamma receptor (FcγR)-expressing cells; iii) a combination of the two mechanisms.
In order to estimate potential mechanisms that may have been involved, we first employed an ultrasensitive HIV-1 gp120 antigen capture ELISA and used parental SHIV-1157ip gp120  as reference protein. Total gp120 concentration of the virus stock solubilized in disruption buffer was only 35 pg per challenge virus dose. This total amount of gp120 is about 106-107 less than what has been used for standard immunizations in humans or macaques [44–48], making boosting via soluble Env a highly unlikely mechanism – especially since no adjuvant was involved in contrast to the standard vaccination/boosting protocols [44–48]. Moreover, the virus-induced boosting effect was not only detected in animals from one study  but in four animals derived from three different studies ([8, 12] and unpublished). Thus, we propose that this observation is not virus-stock dependent since the challenge virus strains varied among the different studies.
Cryptic infection is expected to boost antiviral cellular immunity. This was observed in some of the vaccine-protected RMs (RTr-11, RAt-9, ROb-12). In contrast, vaccinee RRi-11 had no anamnestic cellular immune responses and thus fulfilled the criteria for sterilizing immunity. For this animal, we propose that vaccine-induced nAbs not only blocked initial infection of target cells but also led to the formation of virion-antibody complexes (reviewed in ). Such opsonized virions may then have been taken up by follicular dendritic cells in lymph nodes, which in turn resulted in more effective presentation of Env epitopes to B cells (reviewed in [50, 51]). Thus, we suggest that virions – although unable to cause systemic infection – might have acted as effective immunogens in the form of antigen-antibody complexes.
Ab-covered virions also altered the anti-Tat Ab responses. According to Monini et al. , HIV-1 Tat can form a specific molecular complex with trimeric Env and hence is found on virion Env spikes. This observation would explain the boosting of pre-existing or induction of novel anti-Tat Ab responses.
In the current study, we examined the Env-specific Ab responses in five completely protected vaccinees and chose two of them to investigate the virus-induced consequences on pre-existing Ab responses further. At this point, it should be emphasized that well studied individual cases, as the protected monkeys described above, can shed light onto important basic mechanisms and potential solutions to the problems of prevention and/or cure of retroviral infections. To illustrate the power of case reports, we point to the description of a monkey with breakthrough SIV infection , HIV-1 superinfection in a relatively recently infected individual , and the “Berlin patient” .
In summary, we described a boosting of pre-existing, vaccine-induced Ab responses in immunized macaques that remained aviremic throughout all heterologous SHIV-C challenges. Furthermore, detailed epitope mapping revealed newly induced Abs specific for epitopes formed only in the challenge virus.