Role of complement and antibodies in controlling infection with pathogenic simian immunodeficiency virus (SIV) in macaques vaccinated with replication-deficient viral vectors

Beside cellular immune responses, humoral immunity is considered a key component in AIDS vaccine development. Already during early stages of viral infection, anti-envelope (env) antibodies (Abs) are thought to reduce viremia [1–3]. Their effector functions are still not completely defined. Some of such neutralizing antibodies (nAbs) may inhibit viral entry either by interfering with structures of the gp120/gp41 complex [4] or with env-epitopes that bind to chemokine receptors. Alternatively, they may cross-link virus particles and induce clearance of immune-complexed viruses by phagocytosis. Additionally, antibody dependent cellular cytotoxicity (ADCC) is thought to appear early during acute infection [5] and can also be detected at later stages of disease progression. ADCC has been studied in the SIV monkey model, was associated with the control of HIV in infected humans [6–8] and may contribute to a slower disease progression in long-term non-progressors [9].

A further arm of the humoral immune response is the complement system as an important mechanism of innate immune defence. Complement (C) has been shown to enhance the activity of nAbs [10]. In synergy to the binding of Abs to viruses, C3 deposition, opsonization and immune complex formation are suggested to contribute to reduced viral infection rates. There is evidence that C-mediated lysis contributes mainly at early stages of HIV-1 infection to viremia control [11–13].

A major focus of current research is the design of safe and efficient vaccines providing a high level of protection against HIV. A promising approach is the application of replication-deficient single-cycle immunodeficiency viruses (SCIV) [14, 15]. Upon application, these viral constructs undergo only one single round of replication resulting in the production of non-infectious virus-like particles in vivo. The induced immune response is thought to protect from challenge by clearing infected cells.

A non-invasive application of live-attenuated SIV vaccines to the mucosa via the tonsils has been established. This approach induced protection against challenge with homologous SIV and SHIV, a SIV/HIV-1 hybridvirus containing HIV-1 envelope in the SIV backbone [16, 17]. Although effective, the delivery of attenuated retroviruses is not feasible in humans due to safety concerns [18, 19]. Thus, we adopted a heterologous prime-boost regimen through priming with SCIV and boosting with Adeno5 (Ad5)-SIV or SCIV. The vectors were either given systemically or exclusively mucosally.

The results of the systemic spread of SCIV after oral immunization, as well as analyses concerning the cellular immune responses, immunohistochemical and in situ hybridisation assays have been recently published by Stahl-Hennig et al. [21]. In the present study, we characterized the humoral immune response in immunized and challenged rhesus macaques and investigated the contribution of the induced neutralizing and non-neutralizing antibodies, C-deposition on the viral surface and C-mediated lysis with regard to the control of retroviral infection.

In this study we analyzed the efficacy of humoral immune responses induced by different vaccination strategies either combining a SCIV [VSV-G] prime with an adenoviral boost or administering SCIV only (Table 1). The used SCIV [VSV-G] vaccine provides a safer immunization strategy when compared to live-attenuated vaccines, as no replication-competent particles are generated [15]. Adenoviral vectors have been used in the past, but were usually applied intramuscularly [22] and not via the tonsils. Although our approach did not induce sterilizing immunity, the vaccinated animals had a significantly reduced peak viremia after challenge with the highly pathogenic SIVmac239 when compared to the non-immunized but infected control animals. Peak viral load levels were reduced between 1 log in group 3 and 2 log in groups 1 and 2 (Figure 1A) [21]. Similar reductions in the viral titre were achieved by an iv prime-boost strategy using SCIV as a vaccine [23]. As many studies have emphasised that the long-term prognosis is significantly improved the lower the peak viral load levels are [24, 25], the decrease of the viral load by oral administration of our vaccine may provide profound benefit.

While vaccination via the tonsils induced no nAb responses before challenge, the prime-boost application of the vaccine iv and intramuscularly, respectively, resulted in detectable nAb-titres in the animals of group 2. Similar to the animals of group 1, the monkeys in group 3, which were primed by SCIV [VSV-G] and boosted with the MLV-pseudotyped SCIV, developed hardly any nAbs upon immunization (Figure 4). The peak viremia of group 3 was tenfold higher when compared to animals in group 1 or 2. Surprisingly, animals in group 1 or 2 controlled the viral replication to a comparable extent upon challenge with pathogenic SIV, although the vaccination in the tonsillar group induced no detectable nAb titres in the serum. However, the Ab levels in this group increased rapidly after challenge and reached a constant high titre already 2 wpc. Additionally, the application of the vaccine via the tonsils may induce IgA or cytotoxic T-lymphocyte response at the mucosal site, which may contribute to the reduction of the viral titre upon tonsillar challenge with SIVmac239. Unfortunately, we were not able to measure IgA responses of these vaccinated animals. The presence of nAbs before challenge and/or their fast induction after challenge may contribute to the decrease of the virus in the plasma. This would be in line with reports indicating that only high concentrations of nAbs reduce the peak viremia [26, 27].

Along with the nAb titres, the levels of the total env-specific IgG were weak but mainly detectable in the systemically immunized animals of group 2 already 12 wpi. The detection of the Abs by FACS analysis using SIV-infected cells allows the detection of native, in vivo accessible epitopes only and may be less sensitive compared to ELISA detection systems. Stahl-Hennig et al. [21] used a gp130 ELISA with proteins expressed in E. coli for this animal study. However, these proteins do not reflect the in vivo conformation of the env-protein complex and may thus account for overestimated IgG titres and explain the controversial findings reported previously [21]. It is possible that neutralizing antibodies are not detected by FACS, but will be recognized in ELISA assays. One example is the monoclonal antibody 2F5 [28] which binds to the membrane proximal external region of gp41 during the fusion process but not in the native state. After infection with SIVmac239, the overall IgG response was dramatically boosted in all animals and ran parallel to the induction of nAbs. Interestingly, group 1 and 2 which both controlled the virus similarly well exhibited marked differences in the amount of total env-specific IgG. Due to the limited number of animals available for this study, these differences in the IgG titres reached significance only at week 28.

A neonatal macaque study showed that passively transferred non-nAbs did not protect the animals against oral challenge with SIVmac251 indicating that ADCC is not a main mechanism in reducing infection [29].

This is in contrast to recently reported findings which indicate that ADCC or the interaction of FcR with the Fc-region of the Abs may contribute to the elimination of retroviral infections [8, 30].

Furthermore, the data presented in the present study suggests that C activation is part of the humoral immune response. As shown by a virus capture assay, sera of the animals collected at 2 wpc induced C3-deposition on the viral surface. Although based on only four animals per group, a trend to an inverse correlation of C3-deposition on viral particles and viral load during peak viremia was observed at least within the individual groups of vaccinated monkeys (Figure 3). During the chronic phase of infection, sera of all vaccinated macaques induced C3 activation and opsonisation on SIV, independent of the viral load. C-mediated defence mechanisms have been discussed controversially in the literature. Opsonized virus particles may interact with C-receptor expressing cells, such as B-cells or dendritic cells [31–34], followed by an efficient transmission of opsonized HIV to autologous primary T-cells. At least in vitro, the infection is significantly enhanced by this mechanism. However, preliminary data indicate that in in vitro interaction assays the C-mediated increase of SIV infection is not observable in the monkey system using primary isolated macaque B- and T-cells and opsonised SIV (unpublished observation). A further mechanism of C to reduce infectivity of C-receptor-negative T-cells is the masking of viral epitopes due to the deposition of C3-fragments on the viral envelope [35, 36]. This neutralization mechanism has also been described for other viruses [37] and is an attractive hypothesis to explain, at least in part, the reduced viral loads observed during peak viremia.

A further result of C activation is the induction of the terminal C pathway, resulting in the destruction of pathogens. The in vitro lysis assays reduced the viral titres by a mean of 24.8% (range between 16 and 30%) when sera of immunized monkeys were tested before challenge (Table 2). Two weeks later, during peak viremia, mean lysis was 38.0% (ranging between 11 and 96%) tested in control and vaccinated monkeys. Lysis values increased further during chronic infection up to mean levels of 63.1% (range between 35 and 87%). Although C-induced lysis may contribute to the control of SIV replication, C-mediated destruction of the virus did not correlate with the control of the infection in vivo. Some animals had low peak viremia (#12127, #12142) but exhibited a poor induction of C-mediated lysis when compared to sera from other monkeys with extremely high lysis activities (#12133, #12140) but ten times higher viral loads. In line with earlier studies [11, 12, 38], no correlations between nAbs and C-mediated lysis was observed during the chronic phase of infection. Thus, Ab-mediated neutralization and C-induced lysis of retroviruses appear to represent two independent parameters which are not necessarily linked [38]. This does not exclude the possibility that lysis may play an important role during early phases of infection before or early after seroconversion [13].

Beside Abs, effective SIV-specific T-cell responses are important for controlling viremia [39]. Recently published INF-γ ELISPOT data from the present vaccination trial revealed increased cellular immune responses in cohort 2 compared to group 1 [21]. As both groups controlled the viral loads at comparable levels, it is presently unclear to which extent the cytotoxic T-lymphocyte response is the main contributor for the reduced peak viremia and viral load reduction in the chronic phase of infection.

With this rhesus macaque study it was demonstrated that priming with SCIV [VSV-G] and boosting with both Ad5-SIV vectors or SCIV [MLV] mount humoral immune responses comparable to that of live-attenuated immunodeficiency virus vaccines [40, 41], which may contribute to the significant reduction in viral load observed in animals of group 1 and 2 after challenge. This encourages tonsillar/mucosal immunization strategies which may simplify vaccine application in the future. Thus, more efforts in research further investigating this mucosal delivery route are warranted.