|Home | About | Journals | Submit | Contact Us | Français|
The development of a preventative HIV/AIDS vaccine is challenging due to the diversity of viral genome sequences, especially in the viral envelope (Env160). Since it is not possible to directly match the vaccine strain to the vast number of circulating HIV-1 strains, it is necessary to develop an HIV-1 vaccine that can protect against a heterologous viral challenge. Previous studies from our group demonstrated that a mixture of wild type clade B Envgp160s were able to protect against a heterologous clade B challenge more effectively than a consensus clade B Envgp160 vaccine. In order to broaden the immune response to other clades of HIV, in this study rhesus macaques were vaccinated with a polyvalent mixture of purified HIV-1 trimerized consensus Envgp140 proteins representing clades A, B, C, and E. The elicited immune responses were compared to a single consensus Envgp140 representing all isolates in group M (Con M). Both vaccines elicited anti- Envgp140 IgG antibodies that bound an equal number of HIV-1 Envgp160 proteins representing clades A, B and C. In addition, both vaccines elicited antibodies that neutralized the HIV-1SF162 isolate. However, the vaccinated monkeys were not protected against SHIVSF162p4 challenge. These results indicate that consensus Envgp160 vaccines, administered as purified Envgp140 trimers, elicit antibodies that bind to Envgp160s from strains representing multiple clades of HIV-1, but these vaccines did not protect against heterologous SHIV challenge.
One of the greatest struggles for developing a preventative human immunodeficiency virus (HIV)/acquired immunodeficiency syndrome (AIDS) vaccine is overcoming the diversity of viral isolates . The Envgp160 sequences can differ up to 35% between clades and ~15% within a specific clade . Viruses classified as clade B are responsible for ≥40% of infections in the Americas and Europe, but in Asia and sub-Saharan Africa, where most new infections are recorded each year, other clades are dominant. Most new infections in these regions are classified as clades A, C, or A/E viruses [1,3]. Any HIV vaccine that will prevent infection must be able to overcome the diversity of HIV sequences.
To overcome the HIV sequence diversity, polyvalent mixture of antigens and consensus proteins were designed [4-7]. Polyvalent vaccines increase breadth by including multiple copies of a target (s) or epitopes into a single formulation. Polyvalent vaccine strategies have been employed to increase the breadth of the humoral and cellular immune responses [8,9]. Polyvalent mixtures of Envgp140 or HIV proteins (Gag-Pol, Tat and trimeric Envgp140) elicit a degree of protection against heterologous challenge [8,10]. Consensus-based vaccines rely on a centralized antigen designed to reduce sequence diversity by using the most common amino acid at each position of the protein. Consensus vaccines are designed to reduce the genetic differences between the vaccine and the primary isolate and increase the breadth of immune responses [11-14].
To overcome the diversity in Envgp160 sequences and to design a more effective AIDS vaccine, consensus Envgp140 sequences were designed for 4 clades of HIV-1 (A, B, C, and E), as well as a single consensus Envgp160 representing isolates from all of Group M. For the first time, in the same study, consensus A, B, C, and E Envgp140 sequences were used in a polyvalent vaccine mixture, and compared to a Con M Envgp160, to assess the ability to elicit a broadly reactive anti-Envgp160 immune response. The immunological responses of the polyvalent mixture in vaccinated rhesus macaques were compared to that of the single Con M Envgp140 vaccine. Both vaccines elicited anti-Env immune responses against multiple clades of HIV; however neither vaccine strategy efficiently protected monkeys against a SHIVSF162p4 challenge.
The goal of this study was to design a HIV Envgp160 vaccine that elicits broadly reactive immune responses in an effort to overcome the inherent diversity in the Envgp160. Therefore, an HIV-1 group M consensus Envgp140 vaccine was compared to a polyvalent mixture of clade consensus Envgp140s representing 4 individual clades of HIV-1 (A, B, C, and E). The env gene sequences were then truncated at the transmembrane domain, and the cleavage site mutated, to generate a Envgp140. To stabilize the truncated Envgp140 trimers, the bacteriophage fibronectin domain (FT) was added to the 3’ end of the Envgp140 sequence, as previously described .
Purified trimerized Envs were detected at ~480kDa size indicating oligomerization as trimer proteins (Figure1A). Some Env dimers were observed in consensus C, E and M Envgp140 protein fractions. To probe the antigenic structure, the broadly reactive monoclonal antibody b12  was used to determine binding kinetics to each consensus envelope by surface plasmon resonance (SPR) on a Biacore 3000 (Figure1B and Additional file 1: Figure S1). The rate of association between the consensus Envgp140 trimers and b12 was similar to the rate of association between b12 and primary Envgp140 trimers. The rate of dissociation of b12 from all the Envgp140 trimers was similar, except for consensus B, which had a slower rate of disassociation. Each Envgp140 bound to the primary HIV receptor, human CD4 (hCD4) (Figure1C). The MAb b12 is a monoclonal isolated from an infected patient with a consensus B wild-type Envgp160 on its viral surfaces and therefore may recognize clade B Envgp160 with more efficiency than non-clade B Envgp160s. In addition, consensus B envelopes are bound with higher affinity antibodies to clade B Envgp160s, but less so with consensus C Envgp140. The reverse is true as well. A polyclonal serum, HIV-Ig, was obtained from the AIDS Reference and Reagent Program. This polyclonal serum, collected from a clade B infected person, may not recognize the consensus E Envgp160 with the same affinity and efficiency as the consensus B Envgp160.
To determine the ability of the vaccines to induce a protective response in non-human primates, rhesus macaques were vaccinated intramuscularly with 300 μg of total Envgp140 protein either Con M or the polyvalent Con Mixture (equal amounts of each Envgp140 protein). Animals were treated according to the guidelines of the IACUC of the University of Pittsburgh. All the protocols used were approved by the IACUC of the University of Pittsburgh (#1002617). Monkeys were administered 3 vaccinations of Envgp140 formulated with Imject® alum adjuvant. Mock vaccinated monkeys were administered PBS formulated with adjuvant (Table1). Monkeys vaccinated with Con M or polyvalent consensus developed anti-Envgp140 antibody titers greater than 1:400 dilution against all consensus Envgp160s following the three vaccinations (Figure2A). Collected sera were tested for the ability to bind to a set of primary Envgp160s representing clades A, B, C, and E (Table2). The diversity of the Envgp160s chosen, as well as their similarity to the consensus Envgp160s, is displayed in the phylogenetic tree where Envgp160s cluster into their identified clades (Figure2B). The Con M did not cluster with any one clade, but as expected, was located in a more central position on the tree.
Both Con M and polyvalent consensus Envgp140 vaccines elicited anti- Envgp140 antibodies that recognized Envgp160s from clade A, B and C. However, the Envgp160 SC42, THRO4, PVO4, (clade B), DU172 (clade C) and 93TH975 (clade E) were not significantly recognized by sera collected from vaccinated animals (Figure2C). Overall, there was no binding preference of the elicited anti-Envgp160 antibodies to primary Envgp160s based on clade, location, or year of Envgp160 isolation.
To evaluate the protective efficacy of each vaccine, monkeys were challenged rectally four weeks after final vaccination with SHIVSF162p4 (dose: 640TCID50) to ensure all mock vaccinated animals become infected. All mock vaccinated and polyvalent consensus vaccinated monkeys were infected following challenge (Figure3A and B). Viral loads peaked at day 14 post-challenge at ~1 × 10e+6 RNA copies/ml and then declined to undetectable levels by day 80 post-infection as seen in previous studies using the same stock of virus (unpublished and . Two out of four monkeys (M1 and M4) vaccinated with Con M vaccine had no detectable virus at any time point post-challenge (Figure3C). Monkeys M2 and M3 had a similar viral pattern as mock vaccinated monkeys with a peak at day 14, followed by a rapid decline.
To determine whether Con M Envgp140 vaccinated monkeys were protected from infection, at day 70 post-challenge all Con M Envgp140 vaccinated monkeys were depleted of CD8+ cells by administering monoclonal antibody M-T807R1 intravenously . Seven days following antibody administration, no CD8+ cells were detected in the peripheral blood that was sustained for an additional 18 days (Figure4A). Previously infected monkeys M2 and M3 had a re-emergence of virus during this CD8+-depletion period (Figure4B). Monkey M4 who was initially aviremic after challenge had an emergence of virus after CD8+-depletion. In contrast, monkey M1 maintained undetectable viral loads following CD8+ depletion.
At 2 weeks after final vaccination (day -14 prior to challenge), anti-EnvSF162 IgG antibodies were detected in monkeys vaccinated with either vaccine (Figure2C). In addition, these antibodies were able to neutralize the ability of HIVSF162 to infect cells in vitro (Figure5A). The presence of neutralization titers to all Envgp160s used in the ELISA binding assay were investigated, however, only titers to HIVSF162 Envgp160 was observed. All vaccinated monkeys had a neutralizing titer of 1:40, except for monkey M4 that had a titer of 1:320. Two weeks following challenge, only monkey P4 had an increase in neutralizing titers. There were detectable neutralization titers in 3 of the 4 mock vaccinated monkeys two weeks post-challenge (Figure5A). No monkey had antibodies that recognized a set of overlapping SF162 Envgp160 peptides. These pools of peptides represented the SF162 Envgp160 regions V1/V2, V1 only, or V2 only (data not shown). Interestingly, there was no detectable IFN-γ Envgp160 or gag-specific T-cell responses prior to challenge or 7 days post-challenge. At day 14 and 21 post-challenge, IFN-γ T cell responses were detected in all monkeys (Figure5B), with no significant differences in the number of cells between groups.
HIV-1 Envgp160 based vaccines protect monkeys against a homologous SHIV challenges [18,19]. The ability to match the gene sequences used in the vaccine to the possible exposure virus in humans is not possible. Therefore, studies that use a matched Envgp160 in the vaccine to the challenge strain is appropriate for proof-of concept studies, but our group set a more challenging goal to protect against a challenge virus with a mismatched vaccine with a limited number of vaccinations. We report here that each consensus sequence representing clade A, B, C, E, in a polyvalent mixture or as a single consensus Envgp160 representing the entire Group M elicited anti-Envgp140 antibodies that bound to a broad panel of HIV-1 Envgp160s. The immunization schedule used was not optimal for antibody affinity maturation; nonetheless, the regimen did induce modest neutralizing antibody titers to the challenge Envgp160. However, the elicited immunity did not prevent infection by SHIVSF162p4.
In previous studies, consensus sequences designed for clades B and C Gag and Envgp140 elicited increased breadth of humoral and cellular immune responses [12,20-22]. Consensus Envgp160 sequences representing Group M, termed CON-S, elicited antibodies that neutralized multiple Envgp160s, as well as eliciting cross-clade cellular immune responses [13,14]. However, viral challenges of CON-S vaccinated monkeys were not reported and therefore the efficacy of the induced immune response elicited by these vaccines is unknown.
Compared to Envgp120 monomers, Envgp140 trimers may expose binding and neutralizing epitopes that are present only in Env’s quaternary state [23,24]. Our consensus Envgp140 trimers have similar antigenic properties as wild-type Envgp160s, as demonstrated by attaching to human CD4 and binding to the monoclonal antibody b12. The b12 antibody recognizes a conserved region on gp120 mapped to a discontinuous epitope overlapping the CD4 binding site .
Following three intramuscular vaccinations, all monkeys seroconverted by day 14 following the final vaccination. Nonetheless, there were differences in the vaccine efficacy following challenge between the two vaccine groups. Both non-neutralizing and neutralizing antibodies have been implicated in reducing rates of infection by HIV-1 [26,27]. A report based on the analysis of the sera samples of vaccinated volunteers in the RV144 clinical trial stated that the vaccine elicited antibodies against the V2 region of the HIV-1 Envgp160 were correlated with lower rates of HIV infection . Antisera collected from these vaccinated individuals did not neutralize the infection in vitro. Additionally, vaccine induced protection against a neutralization resistant virus in macaques was correlated with antibodies to the V2 region of Envgp160. Whether antibodies that bind to the V2 region are correlated with protection against SHIVSF162p4 infection in this study is unclear. There were no antibodies elicited in monkeys vaccinated with Con M or polyvalent consensus Envgp140 vaccines that recognized SF162 Envgp160 linear peptides, including those specific to V2 (data not shown). Further studies are necessary to determine if the two protected animals in the Con M Envgp140 group elicited antibodies recognizing conformational epitopes, such as the V1/V2 scaffold proteins. The V1/V2 scaffold was used to analyze human sera collected from vaccinated volunteers in the RV144 clinical trial [30,31]. Determining if antibodies specific to various conformational epitopes on Envgp160 may explain the differences observed in vaccinated animals following SHIV challenge.
Two monkeys that had no detectable viral levels following SHIV infection were M1 and M4. M2 and M3 had detectable viral levels and therefore were not protected against infection. Monkeys vaccinated with either polyvalent consensus or Con M Envgp140 trimers had neutralizing titers to HIV-1SF162. Neutralizing antibodies against Envgp160 can protect monkeys against viral challenge [32,33]. However, only one monkey (M4) in the present study had high neutralizing antibodies (1:320) against SHIVSF162p4 and had undetectable viral titer 14 days after challenge. However, following CD8+ T cell depletion, virus was detected (<1×105 RNA copies /ml) in the blood indicating that infection was not blocked, but may have been controlled by the vaccine elicited antibodies. T cell responses did not appear to play a role in protecting the monkeys from infection. There was no difference in the number or kinetics in the elicitation of Envgp160 or Gag specific IFNγ producing cells following challenge in any of the vaccinated monkeys compared to mock vaccinated animals.
Upon CD8+ depletion, it was not unexpected that M2 and M3 had a rebound in blood titer virus, but the detection M4 was unexpected. In contrast to monkey M4, no virus was detected in monkey M1 even after depletion of CD8+ T cells. Both IFNγ specific T cells and neutralizing antibodies were detected, but it is unclear which of these immune responses may have contributed to the protection. In addition, the MHC class I haplotype did not appear to correlate with protection. Even though no viremia was ever detected in monkey M1, it is possible that virus could be located in reservoirs, such as the bone marrow or gut mucosa . The M-T807R1 monoclonal antibody used for CD8+ T cell depletion is specific for cells in the serum and lymph nodes , therefore, it may have not depleted cells in reservoirs of hidden virus. In an effort to identify possible reasons for M1 protection the animals’ halotypes were determined. Monkey M1 had a Mamu-B*008 MHC class I haplotype, which has been associated with control of SIVmac239 virus; the parent virus of the challenge SHIVSF162p4. Therefore, a combination of the neutralizing antibodies, non-neutralizing antibodies and the Mamu-B*008 MHC class I haplotype may have resulted in “sterilizing” protection after viral challenge. However, the Mamu-B*008 MHC class I haplotype was also present in monkey M3, which had similar binding and neutralizing antibody titers as monkey M1, but was not protected from SHIV infection. Interestingly, two monkeys vaccinated with the polyvalent consensus vaccine, P2 and P4, had a rebound in viral titers at day 40 post-infection before returning to undetectable levels (Figure3B). The rebound virus could have been a variant that escaped the vaccine elicited immune response, however, sequencing of the virus in the blood collected at day 40, did not show any significant variation of the viral sequence compared to the input virus on day 0.
While Envgp140 only vaccines have been successful against homologous challenge, both the human RV144 trial and previous monkey studies showed significant protection from heterologous challenge, included other HIV protein components [8,37,38]. Including Tat in the vaccine formulation induces strong and persistent CD4+ T cells  and broadens T cell responses directed against Gag and Envgp160[40,41]. Gag is known for inducing strong cellular responses that may lead to reduced viral loads [42,43]. Addition of Gag and/or Tat to our Con M vaccine may have prevented infection or controlled undetectable virus in vaccinated animals more effectively than Con M Envgp140 alone. Even though some of these studies use Envgp140 proteins, they are combined with other HIV proteins to elicit a broadly reactive response. In our vaccine presented here, the purified VLPs only have Gag and Envgp140 expressed in different modalities than VLPs and we achieve a broadly reactive anti-Envgp140 response using our consensus Envgp160s. For example, viral vectors are used to express Gag and Envgp160 independently in the RV144 human trial, which really does not allow for comparison with our VLP strategy.
Rhesus macaques were vaccinated with trimerized Envgp140 proteins representing consensus sequences for clade A, B, C, E, in a polyvalent mixture or as a single consensus Envgp160 representing the entire Group M. These consensus Envgp140 elicited antibodies with cross-clade anti-Envgp140 binding against a panel of HIV-1 Envgp160s. However, this breadth of antibody binding to HIV-1 Envgp160s only partially correlated with the prevention of infection by SHIVSF162p4.
The consensus sequences represent the most common amino acids found at each position of the aligned envelope sequences used. One hundred Envgp160s sequences per clade were used to design consensus sequences of clades A, B, C, and E. The design of the group M consensus Envgp160 was based on two hundred Envgp160 sequences representing clades A, B, C, D, E, F and H. Envgp160s chosen were isolated following mucosal transmission, within weeks after infection and included a diversity of viruses that were isolated in different locations around the world between 1995 and 2005. Each Envgp160 used the CCR5 co-receptor. Vaccine immunogens were designed as Envgp140 trimers as previously described .
Human embryonic kidney (HEK) 293T cells were transiently transfected with DNA (8 μg) expressing one of the consensus HIV-1 Envgp140 proteins or wild type Envgp120 6X-HIS tagged proteins. Following DNA transfection as previously described , secreted Envgp140 proteins were purified using lectin columns made from agarose galanthus nivalis (snowdrop) lectin (Vector laboratories, Burlingame, CA, USA)  and the Envgp120 6X-HIS tagged proteins were purified using nickel columns . Other purified Envgp160s used for ELISAs were purchased from eEnzyme (Gaithersburg, MD, USA). Each purified consensus Envgp140 trimer protein (1 μg) was loaded on to NativePAGE native gel (Invitrogen, Carlsbad, CA, USA) and separated by electrophoresis in the manufacture’s recommended buffers. After separation, the proteins were detected using the ProteoSilver Sliver Stain kit (Sigma, St. Louis, MO, USA) following manufacturer’s protocol .
The CD4 binding assay was performed to demonstrate Envgp160 binding to its primary receptor using a similar protocol as previously described . Protein G Dynabeads (Invitrogen) were mixed with anti-his antibody. The tubes with the mixture were placed on the magnets to remove all unbound antibody. Then soluble human CD4-6XHIS tagged protein (eEnzyme, Gaithersburg, MD, USA) and consensus Envgp140 mixtures were then mixed with the beads. Following pellet fractionation, samples were separated on a 10% SDS PAGE gel, transferred unto a nitrocellulose membrane and probed for sCD4 or Envgp160 using specific antibodies. Following secondary IgG-HRP antibodies which were used to detect proteins by Western blot.
MAb b12 binding kinetic analyses of the HIV rgp140(s) was performed by surface plasmon resonance (SPR) on a Biacore 3000 (GE/Biacore AB, Inc., Uppsala, Sweden) as previously described .
Animals were treated according to the guidelines of the IACUC of the University of Pittsburgh. All the protocols used were approved by the IACUC of the University of Pittsburgh (#1002617). Rhesus macaques (Macaca mulatta) were used for all non-human primate experiments. All animals were cared for adhering to USDA guidelines for laboratory animals. Rhesus macaques were anesthetized using 10-20 mg/kg ketamine and vaccinated intramuscularly in the quadriceps and formulated with Imject® alum adjuvant (Imject® Alum, Pierce Biotechnology; Rockford, IL, USA). Vaccinations were completed at weeks 0, 4 and 8. Twelve animals were divided into three (3) groups, four animals per group (Table1). For blood sample collection animals were anesthetized with a mixture of ketamine/xylazine. Sera was then harvested and stored at -80°C until needed.
Total anti-Envgp160 IgG was detected by enzyme-linked immunosorbent assay (ELISA) using Concanavalin A (50 μg/μl) per well as previously described . End point titer for assay was determined as the reciprocal of the dilution at which the optical density reading was above the mean plus two standard deviations of naïve sera. For In vitro neutralization, antisera were tested for the ability to neutralize virus infection in vitro using TZM-Bl indicator cells . The sera dilution necessary to neutralize virus was calculated by the following formula (relative light units (RLU) of virus only-RLU of cell only)/2 + RLU cell only. For assessment of T cell responses, NHP IFN-γ ELISPOT was used to enumerate anti-Envgp160 specific cellular responses. The number of anti-Envgp160 (SF162p3 and Con M) and Gag (SIVmac239) specific IFN-γ secreting cells were determined using the non-human primate enzyme-linked immunospot (ELISPOT) assay (R&D Systems, Minneapolis, MN, USA).
Real time PCR-based SIV viral detection assay was used to determine the viral titers post-challenge as described in . cDNA (10 μl) generated by the RT-RCR reaction was then used for PCR using the ABI 7000 Gene detection system (Applied Bioscience, Carlsbad, California, USA).
All animals in the Con M group were depleted of CD8+ T cells. The antibody M-T807R1 (NIH NHP Reagent Source, Beth Israel Deacones Medical Center, Boston, MA, USA) was administered subcutaneously (50 mg/Kg) on day 0 (Day 70 post infection). CD8+ T cell depletion was verified using TruCOUNT tubes (BD Bioscience, San Jose, CA, USA).
Statistical tests were performed using Graph Pad Prism software. Statistical significance of antibody test was determined by two-way analysis of variance (ANOVA) followed by the Bonferroni’s post-hoc test. Post-test was used to analyze differences between the vaccine groups. Significance was determined to be a p<0.05.
The authors declare that they have no competing interests.
HSE designed and performed most of the experiments analyze data and wrote the paper. BRP was instrumental in animal care and processing of animal samples. JKC performed and analyzed all Surface plasmon resonance data. She was also involve in discussion on results and implications and commented on the manuscript. Finally TMR was the supervisor on the project and was involved in discussions on the results and its implications. He was also instrumental in editing of the manuscript at all stages. All authors read and approved the final manuscript.
Representative sensograms of IgG b12 SPR data. SPR binding isotherms detailing the interaction of each recombinant trimeric Envgp140 with the monoclonal antibody b12 are displayed. Indicated concentrations of each Envgp140 were flowed over captured b12 at 37°C on CM5 chips as detailed in Materials and Methods. Association rates, dissociation rates, and affinity constants were calculated with BIA evaluation 4.1.1 software (GE/Biacore AB, Uppsala, Sweden). A, ConA; B, ConB; C, ConC; D, Con M; E, ConE; F, ADA; G, PV0.04; H, R2; I, SC42; and J, YU2. Black lines are determined 1:1 Langmuir kinetic fits. RU, resonance units.
This research was primarily supported by an award from the National Institute of Health/National Institute of Allergy and Infectious Diseases R01AI068507 to T.M.R. The HIV and SIV specific peptides and reagents were obtained through the AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH including TZM-bl cells (also called JC57BL-13) (#8129). The authors would also thank investigators that provided specific HIV-1 isolates (Rob Center, Dennis Ellenberger, Phalguni Gupta, Beatrice Hahn, David Montefiori, Gary Nabel, Yiming Shao, and Carolyn Williamson). We are also grateful to Amara Luckay and Jim Smith at the CDC for determination of viral loads. This work was supported, in part, under a grant with the Pennsylvania Department of Health. The department specifically disclaims responsibility for any analyses, interpretations or conclusions.