Live attenuated rubella vaccine has a proven record of safety and immunogenicity in humans and rhesus macaques. This small RNA virus would be an attractive viral vector, if it could present the major vaccine antigens of other viruses as well as its own. This is the first report showing that live attenuated rubella vectors replicate robustly in vivo while expressing SIV and HIV vaccine antigens. The new vectors infected six out of six rhesus macaques, while eliciting high-titered antibodies to the SIV Gag insert in all of them and to HIV MPER in five out of six. The anti-Gag antibody titers elicited by immunization were greater than or equal to those induced by natural infection with SIV. The antibodies to both inserts have persisted for over 9 months, and they have declined at the same rate as antibodies to rubella, which protect for life. The anti-Gag antibody response was boosted by re-exposure to the vector after six months, indicating the induction of memory B cells. Since rhesus macaques are also susceptible to SIV infection, they will provide an ideal model for testing immunogenicity of novel rubella vectors and protection against SIV or SHIV challenge.
At the start of these studies, rubella vectors faced a series of questions before they could be considered a vaccine candidate. These included: location of insertion sites, size and stability of inserts, adaptation to the live attenuated vaccine strain, replication in vivo, immunogenicity, and concurrent immunization with more than one vector. We previously reported that rubella vectors could grow to high titers in cell culture while expressing SIV and HIV antigens at either of two insertion sites [18, 19]. Inserts at the Not I site in the nonstructural region included zGFP, SIV Gag epitopes, and HIV MPER [17, 18]. These inserts were limited in size and diversity, probably because they were expressed as fusion proteins with rubella nonstructural protein P150, which performs essential viral functions.
However, when the genes were inserted in the structural region, between envelope glycoproteins E2 and E1, insert expression was uncoupled from essential viral functions. This allowed the expression of larger inserts . These inserts were expressed as part of the structural polyprotein under control of the strong subgenomic promoter, leading to high-level antigen expression for a prolonged period. When we switched to the rubella vaccine strain RA27/3, the new vectors were able to express the same inserts as wild type rubella, with little or no loss in vector titer, antigen expression, or insert stability . The vectors with Not I deletion did not replicate well in vivo. This was solved by restoring the deleted sequence, which resulted in replication of type 3 vectors in six out of six macaques. The deleted sequence is part of the “Q” domain in nonstructural protein P150 with unknown function . This region may be important for interferon signaling or suppression. Vectors with the Not I deletion replicate well in Vero cells, which are incapable of interferon production, but they do not replicate in normal WI38 or BSC-1 cells (Virnik et al., manuscript in preparation). Future studies will address this phenomenon.
Live replicating rubella vectors expressing SIV Gag or HIV MPER at the structural insertion site were highly immunogenic in macaques. These vectors could contribute to vaccine potency at each stage of the immune response: by simulating acute infection and triggering innate immunity, they could initiate a stronger immune response to the inserts. Subsequently, exponential growth of the vector would expose the host to increasing doses of antigen each day. Finally, prolonged antigen expression can mimic an ongoing infection [14, 15], leading to germinal center formation, which is needed for immunoglobulin class switching, somatic hypermutation to produce high-affinity antibodies, and maturation of memory B cells . Unlike their non-replicating homologues, live rubella vectors elicited anti-Gag antibodies after a single dose, and the antibody titers continued to rise for four to seven weeks after immunization. These strong stimuli, lasting two weeks or more, explain how live vectors could achieve the same high titers of anti-Gag antibodies as natural SIV infection.
With some other vaccines and vectors, the immune response to SIV and HIV antigens has been short-lived and lacked memory B cells [35, 36]. Transient antibody responses are considered one of the major obstacles to HIV vaccine development . Using live rubella vectors, the anti-Gag and anti-MPER antibodies have persisted for over nine months. They are declining with nearly the same half-life as antibodies to rubella proteins, which protect for life. In general, persistent antibody titers are thought to depend on long-lived plasma cells, while boosting depends on memory B cells, and both are signs of germinal center function during immunization . The primary immune response to these vectors included memory B cells, as shown by boosting 6 months later. The secondary response depended on successful priming, and it could overcome the inhibitory effects of rubella antibodies. Potentially, two or more doses of live rubella vectors, given several years apart, could boost and update immunity to circulating strains of HIV. In addition, the ability to prime and boost memory B cells would allow us to combine rubella vectors with other viral vectors bearing similar HIV vaccine inserts [37, 38], and this will be tested.
Rubella is a small RNA virus that replicates exclusively in the cytoplasm. This location is ideal for eliciting T cell immunity, since it delivers antigens directly into the proteasomal pathway leading to antigen presentation with MHC class I . Priming with DNA vaccine and boosting with the vector gave high levels of Gag specific CD8+ T cells that were comparable to natural infection (Virnik, et al., manuscript in preparation). When two rubella vectors were given simultaneously, they both replicated side by side, as shown by RT-PCR, and they elicited antibodies to both Gag and MPER inserts. Prolonged expression of MPER antigen by rubella vectors may contribute to its immunogenicity and may improve on natural infection, which elicits these antibodies in less than half of the cases [40–43].
The safety and immunogenicity of live attenuated rubella vaccine were demonstrated in rhesus macaques  and in children [5, 6, 16]. Macaques have shown no signs of disease during successful immunization with rubella vectors. The safety of rubella vectors should be comparable to rubella vaccine: if a rubella vector lost its insert, it would revert to the vaccine strain. Given the overlapping host range, rhesus macaques will provide an ideal animal model for testing rubella vectors for immunogenicity and protection against SIV or SHIV challenge. Novel vectors that demonstrate vaccine potency in macaques could be quickly translated into human vaccine design.