The HIV-1 Tat protein plays important roles in the virus life cycle and maintenance of HIV-1 infected CD4+ T cells 12. It is a trans-activating regulatory protein that stimulates efficient transcription of the viral genome, which requires structural changes of Tat to bind to a RNA stem-loop structure called TAR 34. However, Tat differs from other HIV-1 regulatory proteins because it is rapidly secreted by CD4⁺ T cells following HIV-1 infection, and extra-cellular Tat is suspected to be directly involved in the collapse of the cellular immune response against HIV-infected cells 2 and directly contributes to the pathology of AIDS 5. Extra-cellular Tat inhibits macrophage responses by binding to the Fas ligand membrane receptor 6 and inhibits cytotoxic T cell (CTL) responses due to its ability to cross cell membranes and induce apoptosis of uninfected T cells 78 via interaction with tubulin 8910. In addition, a number of studies have shown that the presence of antibodies against Tat blocks the replication of HIV-1 in vitro and is related to non-progression to AIDS 111213. Moreover, it has been shown that a HIV-1 Tat-specific cytotoxic T lymphocyte response is inversely correlated with rapid progression to AIDS 14. Further studies have emphasized the hypothesis that anti-Tat CTLs are important in controlling virus replication early after primary infection 1415.

The discovery of the extra-cellular functions of Tat in the inhibition of the cellular immune response against HIV-infected cells constitute the rationale to develop a vaccine against HIV targeting Tat 16. However, the development of a Tat vaccine may face the same problems encountered with HIV-1 envelope proteins as Tat exists in different sizes (86 to 101 residues) and mutations exist that induce structural heterogeneity 17. The 2D NMR studies of two active Tat variants from Europe and Africa confirmed this structural heterogeneity, although a similar folding appears to exist among Tat variants 181920. Currently, there are five main HIV-1 subtypes in the world: subtypes A (25 %) and C (50 %) are predominant and are found mainly in Africa, India and South America; subtype B (12 %) is found mainly in Europe and North-America; subtype D (6%) is found in Africa and subtype E (4 %)(a recombinant form known as CRF_01AE), is found mainly in South East Asia 21. Tat variability follows this geographical diversity with mutations of up to 38 % observed among Tat variants from A, B, C, D and E HIV-1 subtypes that do not alter Tat functions but do not allow cross recognition with Tat antibodies 22.

Up to now, the two main vaccine strategies against Tat use a recombinant protein corresponding to a short 86 residue version of a subtype-B European Tat variant that is either inactivated 11 or has full activity 23. These two approaches were tested on macaques followed by a homologous SHIV challenge 2425. A significant decrease of viremia was observed in these two studies carried out respectively on Cynomolgus 24 and Rhesus macaques 25, without showing complete protection during primary infection. A recent study showed long term control of infection following homologous SHIV challenge on Tat-vaccinated Cynomolgus macaques 26. However, immunization with a subtype B Tat variant of 86 residues does not stimulate an efficient response against subtype A and C Tat variants 27. Moreover, most Tat variants found in the field are of 101 residues 4.

Over the last 20 years, several HIV vaccine studies have been tested using a homologous SHIV/macaque model and some have met with success 28. However, these were not followed by success in clinical trials 29, possibly due to the high genetic diversity of HIV-1. This is why heterologous SHIV challenge in macaques, using a genetically distinct virus, is now recommended to determine if a vaccine can be effective against HIV-1 infection in humans and corresponds to the most significant in vivo experiment after clinical trials 28.

The interest to develop a Tat vaccine rose with the discovery that seropositive long-term non-Progressor (LTNP) patients had a higher level of Tat antibodies than seropositive Rapid Progressor (RP) patients 13. However, LTNP patients are unable to eradicate HIV since they still have HIV released from reservoir cells. Another category of patients, the highly exposed persistently seronegative (HEPS), appears to be more interesting since they were in contact with the virus, they have developed a strong cytotoxic T lymphocyte (CTL) response against viral proteins and have retro converted to become seronegative 30. There is a very low prevalence of HEPS among adults and it could be possible that the HEPS phenotype is due to innate immunity 31.

Although HEPS patients have normally no detectable virus, it was possible to isolate and clone a HIV-1 strain from patients in a cohort in Gabon 32 that could be now classified as HEPS. This strain called HIV-1 Oyi has genes similar to regular HIV-1 strains except the tat gene, which had mutations never found in other Tat variants 16. The epidemiological survey was carried out on a sample of 750 pregnant women and 25 were identified as seropositive 32. From these 25 seropositive women, 23 rapidly retro converted and became HEPS. All the HEPS women were infected with HIV-1 Oyi. The high proportion of HEPS phenotype in this cohort (92%) indicated that the retro conversion was probably due to an acquired immunity and not an innate immunity. Ten years after the publication of this epidemiological survey, the 23 women were in good health and the HIV was no longer detectable in their blood 22. Immunization with Tat Oyi raises antibodies in rabbits that are able to recognize different Tat variants even with mutations of up to 38 %, which is not possible with other Tat variants 22. Tat Oyi appears to induce a humoral immune response against three-dimensional epitopes that are conserved in Tat variants in spite of 38% mutations 22. Moreover, Tat Oyi has a similar structure to active Tat but is unable to trans-activate 20.

This study is the first step of pre-clinical studies of a vaccine using a synthetic protein of 101 residues. Synthetic vaccines are developed for many years because they could be safer regarding biological vaccines, i e vaccines made from inactivated pathogens or recombinant proteins. However, most of the vaccines commercially available up to now have a biological origin. Very few synthetic vaccines were able to demonstrate their efficacy in vivo against a pathogen such as bacteria or virus due to the short size of the peptides that can constitute only linear epitopes, while 3D epitopes are the most susceptible to trigger an immune response that neutralize a pathogen. This is why, one of the objective of this study was to determine a vaccine formulation suitable for human use to prepare clinical trials, as a previous study with Tat Oyi was carried out using complete Freund adjuvant 22. We evaluated the antibody responses raised in rabbits by Tat Oyi complemented with adjuvants authorized for human use and we determined formulations providing similar results previously obtained with the Freund adjuvant 22. Vaccination with Tat Oyi on seven Rhesus macaques provided an excellent model to test in vivo the efficacy of this synthetic vaccine before clinical trials. Furthermore, the vaccinated macaques were challenged with a European SHIV. This was a heterologous SHIV challenge and no success in heterologous SHIV were published until now.

We selected four adjuvants (Calcium phosphate, Montanide, Adju-Phos and Alhydrogel) to develop different vaccine formulations with our synthetic protein Tat Oyi. The usual dose of aluminium for human vaccines is around 0.5 mg 33 and at this concentration, approximately 90 % of 100 μg of Tat Oyi adsorbed to both aluminium containing adjuvants (Adju-Phos and Alhydrogel). For these two reasons, we decided to carry out our inoculations at 0.5 mg Al per dose of vaccine for both Adju-Phos and Alhydrogel. For the calcium phosphate gel, we achieved 92 % adsorption using 1 mg Ca per 500 μl dose while only 62% adsorption using 0.5 mg Ca in the same volume. Montanide adjuvant (70 %) was used because it is a metabolizable oil that can be used for human vaccination and has chemical properties similar to those found in the Freund adjuvant as used in our first vaccination studies 22.

Twelve rabbits were immunized with the four formulations (three rabbits for each formulation) and we analyzed the antibody responses against five Tat variants representative of subtypes A, B, C, D, and E (Table I). No antibody response was observed using the calcium phosphate gel and the aluminium phosphate adjuvants at 60 days post-inoculation. However, at 90 days post-inoculation, a strong antibody response was observed using these two adjuvants against five Tat variants (Table I). The best humoral response against Tat oyi was obtained using Montanide ISA720 (titer: 128,000 against Tat Oyi) at both 60 and 90 days post-inoculation. However, Montanide ISA720 and Calcium phosphate appear to be the most suitable adjuvants to complement the synthetic protein Tat Oyi, due to the absence of toxicity and the heterologuous immunity compared with different Tat variants observed after vaccination (Table I).

A heterologous SHIV-BX08 challenge carried out on seven macaques vaccinated with Tat Oyi/Montanide ISA720 and four control macaques vaccinated with β-galactosidase that were used also as control for another vaccine trial 34. Figure 1 shows the viremia as revealed by SHIV RNA copy number in the sera of macaques after SHIV challenge. Similarly to what is observed in human a couple of months after HIV infection, both Tat Oyi vaccinated macaques and controls had an undetectable viremia 63 days after the SHIV challenge (Fig 1). In addition, virus isolation and cytoviremia was measured by co-cultivation of PBMC's with non-infected human cells at the day of challenge and each week afterwards and allow to estimate the level of reservoir cells (Fig 2). Five on seven Tat Oyi vaccinated macaques showed a better control of viremia compared to control macaques (Fig 1). Reservoir cells were not detectable at 56 days post-challenge in all Tat Oyi vaccinated macaques but not in the controls (Fig. 2).

It has been shown in SHIV challenge that plasma viremia in the first peak does not correlate with survival whereas plasma viremia levels of the second peak at or about six weeks post-infection were highly predictive of relative survival 35. In our vaccine trial, panel C in figure 1 shows that plasma viral RNA levels were significantly lower in the vaccinated Macaques compared to the controls at nine weeks post-infection (p = 0.009 using Mann-Whitney test). While we did not observe major differences in the level of CD4 cells between vaccinated and non vaccinated macaques (data not shown), we did observe an augmentation of the number of CD8 lymphocytes in Tat Oyi vaccinated macaques (Fig. 3). However, we did not determine if these CD8 are HIV specific CTL. It is interesting to observe that before the SHIV challenge, control macaques had a higher level of CD8 compared to Tat Oyi vaccinated macaques. Control macaques were immunized with the Semliki Forest Virus (SFV) lac Z expressing β-galactosidase that boost the CD8 response 34. This high level of CD8 were not HIV specific in control macaques and they had no antibodies against Tat. Therefore, we think that the decreased level of CD8+ cells in control macaques after the SHIV challenge could be due to extracellular Tat, since the SHIV infection should have increased the CD8 response as observed for SFV.

All Tat-vaccinated macaques, with the exception of Macaque 969, developed a strong anti-Tat antibody response (Fig 4), which correlated with an efficient reduction in viremia at nine weeks post-infection (Fig 1C). This was best demonstrated by monkey 965, which had a strong anti-Tat antibody titer and a significantly reduced viremia nine weeks post-infection despite a high viremia in the primary phase (Fig 1C). To a lesser extent, macaque 9711 shows the same relationship between the level of anti-Tat antibody and the viremia at nine weeks (Fig 1C). Moreover, the control of viremia in Tat Oyi vaccinated macaques was not due to antibodies raised against the HIV envelope proteins since the four SHIV challenged control macaques had high anti-gp120 antibody titers. Overall, gp120 antibody titres were similar in control and Tat Oyi vaccinated macaques (Fig 5).

Macaque 966 did react differently from the other Tat Oyi vaccinated macaques and is the most interesting. It was the one to have an almost complete immunity against SHIV BX08 with a viremia peak around 300 RNA copies per ml whilst most of the others macaques had viremia peaks between 100 000 and 3 000 000 RNA copies per ml (Fig 1). Interestingly, almost no antibodies against gp120 were detectable and no virus could be isolated from cultured PBMC's (Fig 2). To verify this strong immunity, macaque 966 was challenged a second time with another heterologous SHIV 162P 3.2 seven weeks after the SHIV BX08 challenge (Roger Legrand, Personal communication). This second challenge explains the higher viremia peak at nine weeks post-infection compared to the other Tat Oyi vaccinated macaques (Fig 1C), which rapidly decreased to an undetectable level. It is interesting also to note that antibodies against gp120 were observed with macaque 966 following the second SHIV challenge that also rapidly declined (Fig 5). Results observed with macaque 966 are very important and constitute the best proof of concept for the Tat Oyi vaccine and its rational as previously described 22. Macaque 966 had the highest titer of anti-Tat antibody (Fig 4), the lowest viremia (Fig 2) and no detectable virus from cultured PBMCs (Fig 1). Macaque 965 had nearly identical level of anti Tat antibodies but was not able to control its viremia as macaque 966. It is possible that innate immunity helped macaque 966, but it is interesting to note that antibodies against gp120 disappeared rapidly for macaque 966 (Fig 5), similarly to what was observed with the patients infected by HIV-1 Oyi in Gabon 32 and HEPS patients 30.

Conflicting results appears in Tat vaccine studies in nonhuman primate viral challenges models ranging from no protection 34363738 to significant 392425, long term protection 26. Although these conflicting results could be explained by differences in immunization regimen, viral stock, route of viral challenge and animal species, the result of two studies using similar viral vector expressing Tat, Env and Gag and giving opposite conclusion is puzzling 3639. One study shows the efficacy of vectored Tat but not Gag and Env 39, while another study showed efficacy of vectored Gag and Env but not Tat 36. These conflicting results could be due to a homologous challenge in the first study 39 and a heterologuous challenge in the second study, since the second study use the Tat Jr sequence instead of the homologuous Tat Bru sequence for the vaccine 36. HIV-1 Jr and HIV-1 Bru are B subtypes but their Tat sequences have non conservative mutations inducing conformational changes 16. The mutations between the vaccine and the challenge virus might explain the lack of efficacy of the Tat vectored vaccine in the second study 36. Of course, the second study more closely resembled reality since a vaccinated person will not likely be exposed a homologous virus infection. It is possible that the study by Silvera et al. would have had an different outcome had heterologous gag and env genes been used in the SHIV challenge 36. These studies outline how mutations can affect Tat cross recognition as shown in former studies 2227.

Three adjuvants authorized for human use trigger an immune response with Tat Oyi similar to what was observed with the complete Freund adjuvant in a former study 22. No local or systemic toxicity or adverse effects were observed in rabbits and macaques with vaccine doses superior to those planed for clinical trials. Furthermore, the synthetic protein Tat Oyi is pharmacologically stable in solution for at least a period of one month, which is a requirement for mass vaccination (data not shown). Although a low viremia was not achieved for all macaques, reservoir cells were no longer detectable 56 days after a heterologuous challenge. Taken together, these results suggest that a Tat Oyi synthetic protein could be an excellent component of a vaccine targeting HIV-1 and could provide an appropriate treatment against HIV-1 in both developing and industrial countries. On a fundamental point of view, the decreased level of CD8 cells in the control macaques suggests an important role of extra cellular Tat in the immunodeficiency induced by the HIV-1. We hope to be able to confirm in phase I/II clinical trial with seropositive patients that a therapeutical effect can be obtained from the Tat Oyi vaccination. This therapeutic effect might result, firstly, in a reduced viremia and stable CD4 cells level following an interruption of the antiretroviral treatment. We believe this vaccine will not prevent sero negative people from HIV infection, however it could avoid the collapse of the cellular immunity, and therefore a therapeutic effect could be expected with the eradication of the virus titres and viral reservoir as is observed with HEPS patients. This vaccine could be also the only affordable therapy for millions of seropositive patients that have no access to antiretroviral treatment.

HIV, human immunodeficiency virus

PBMC, Peripheral Blood Mononuclear Cell

Tat, Trans activator protein

The author(s) declare that they have no competing interests.

JDW carried out ELISA test on rabbits, interpreted the SHIV challenge's results and participated to the redaction of the manuscript. DE participated to ELISA test on rabbits. GC participated to ELISA test on rabbits and the redaction of the manuscript. SL, SO and JM participated in the immunization protocol of the preformulation's studies. DE, GC, SO, SA synthesized the proteins for rabbit ELISA. JPS interpreted SHIV challenge results and participated to the redaction of the manuscript. EPL immunized rabbits, synthesized and provided Tat Oyi for macaque immunization, and wrote the manuscript.