A novel non-integrative single-cycle chimeric HIV lentivector DNA vaccine

Novel HIV vaccine vectors and strategies are needed to control HIV/AIDS epidemic in humans and eradicate the infection. DNA vaccines alone failed to induce immune responses robust enough to control HIV-1. Development of lentivirus-based DNA vaccines deficient for integration and with a limited replication capacity is an innovative and promising approach. This type of vaccine mimics the early stages of virus infection/replication like the live-attenuated viruses but lacks the inconvenient integration and persistence associated with disease.

We developed a novel lentivector DNA vaccine “CAL-SHIV-IN−” that undergoes a single round of replication in the absence of integration resulting in augmented expression of vaccine antigens in vivo. Vaccine gene expression is under control of the LTRs of a naturally attenuated lentivirus, Caprine arthritis encephalitis virus (CAEV) the natural goat lentivirus. The safety of this vaccine prototype was increased by the removal of the integrase coding sequences from the pol gene.

We examined the functional properties of this lentivector DNA in cell culture and the immunogenicity in mouse models. Viral proteins were expressed in transfected cells, assembled into viral particles that were able to transduce once target permissive cells. Unlike the parental replication-competent SHIV-KU2 that was detected in DNA samples from any of the serial passage infected cells, CAL-SHIV-IN− DNA was detected only in target cells of the first round of infection, hence demonstrating the single cycle replication of the vaccine. A single dose DNA immunization of humanized NOD/SCID/2 mice showed a substantial increase of IFN–ELISPOT in splenocytes compared to the former replication and integration defective 4SHIV-KU2 DNA vaccine.

What should we know ?

Introduction HIV/AIDS pandemic was responsible of the death of more than 35 million people while as many individuals are still living infected with the virus. Although there have been enormous progress in HIV/AIDS prevention and therapy, highly active antiretroviral therapy (HAART) has many limitations that prevent virus eradication in treated individuals. It is believed that effective and safe therapeutic and/or prophylactic HIV vaccines are critical to stop this chronic infectious disease.

Enormous efforts have been made in developing an HIV-1 vaccine during these last 3 decades but to date there is still no safe and efficacious vaccine that can be used in human. Evidences of natural virus control in a tiny proportion of HIV-1 infected individuals, the long-term non-progressors along with possible successful cases of “human cure” following treatments provide the proof that HIV-1 can be defeated. To date, the only human trial that showed partial protection against acquisition (about 31%) is the Thai RV144 trial using a canarypox vector to prime and recombinant Env protein to boost.

In non-human primate models, the live-attenuated lentivirus vaccines were found to be highly effective in inducing protective immunity in macaques against pathogenic strains of SIV and SHIV.

However, their potential to randomly integrate in the host genome and revert to the pathogenic phenotype in neonates and some adults excluded their use in humans. Recent studies using a replication competent recombinant vaccine based on rhesus cytomegalovirus (RhCMV/SIV) expressing SIV proteins showed increased protection in about 50% of vaccinated macaques against a pathogenic SIVmac239 challenge, regardless of the route of challenge More remarkably, protected macaques progressively cleared the challenge virus and appeared cured. In this vaccine strategy, the viral clearance appears to correlate with CMV-maintained unconventional broad effector-memory CD8+ T cell immune responses.

These results demonstrate that a replication-competent and persisting live vaccine is capable at eliciting SIV-specific T cell responses that contribute to control viral replication and prevent the onset of AIDS in SIV-infected macaques. Similar results were previously reported using a herpes virus-based vaccine vector in macaques inducing similar proportion of protection against the challenge virus.

Other studies have clearly shown that a vaccine regimen could induce immune responses that are able to reduce viral replication. Among these vaccine strategies, replication-defective and non persisting plasmid DNA vaccines have shown promising results. These different approaches rely on the hypothesis that continuous antigen expression, provided by replicating vectors may not be the only way to promote the development of long-term T-cell based control of virus replication. Indeed, T cell-mediated immune responses that control infection and disease progression were shown to be generated after acute resolved infection, indicating that a moderate and limited viral expression can be critical in the setting of such responses. Interestingly, it has been found that a relatively low viral load environment during primary infection in LTNP/EC may be the crucial difference with progressors .

 Accordingly, early initiated HAART therapy in HIV patients allows long-term control of infection after the interruption of the treatment. Our focus was on the development of lentiviral-based DNA vaccines that mimic the early stages of lentivirus infection to stimulate natural and proportionate immune responses against virus antigens. We demonstrated with our first generation of non-replicating and non-integrating DNA vaccine, 4SHIV-KU2, that a single dose immunization in the mouse and NHP models allowed the development of long-lasting and polyfunctional SHIVspecific T cell responses against Gag and other HIV antigens in immunized animals. Importantly also, in earlier studies, macaques vaccinated with repeated doses ofthis DNA vaccine only, were protected from the challenge with SHIV89.6p.

To improve further the vaccine immunogenicity, we engineered a second generation of HIV lentivector DNA vaccines. This new type of vectors expresses all virus proteins that assemble into pseudo particles able to undergo a single cycle of replication. Consequently, antigen expression is increased in the absence of integration of the vaccine genome in that of the host. We hypothesized that SHIV antigens that can be expressed, maturated and assembled in pseudo-infectious viral particles in vivo will mimic an acute and limited viral replication.

This results in increase sources of antigens for antigen presentation and subsequently promotes optimal T cell-mediated immune responses. Thus, this synergizes the advantages of live-attenuated and DNA vaccines without inconvenient persistence and reversion to pathogenic state. In addition, we replaced the SIV LTRs with those of the Caprine arthritis encephalitis lentivirus (CAEV), a naturally attenuated lentivirus that does not cause AIDS like disease in infected goats and does not infect human cells. In contrast to primate lentiviruses, CAEV LTRs have Tat-independent constitutive promoters that express viral proteins at high level in all tested cells.

We have shown that CAEV 5 LTR promoter was fully capable at expressing high level of SHIV antigens in the genome of 4SHIVKU2, and inducing potent immune responses both in mice and macaques. Here we designed CAL-SHIV-IN−, driven by both 5 and 3 CAEV LTRs and deleted of SIV integrase (int) in the genome of SHIV-KU2 for increased safety.

This novel lentivector DNA vaccine was studied in cell culture system and in vivo in immunized mouse models. In this paper we report the functional and immunogenicity properties of this constitutively-expressed single cycle competent DNA vaccine.

How did they proceed ?

Cell culture reagents

SIV and HIV peptides

Overlapping15-merpeptides, with11-aa overlaps, spanning the entire molecules of SIV Gag, HIV Env, HIV Tat, HIV Rev, and SIV
Nef proteins were obtained from the National Institutes of Health (NIH) AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH (catalog nos. 6204, 6451, 5138, 6445 and 6443 respectively).

Cell lines

Human embryo kidney cell line (HEK-293) (Catalog no.103), as well as lymphocyte cell lines CEM-x-174 (Catalog no.103) and
M8166 (Catalog no.11395) were also kindly provided by the NIH AIDS Reagent Program, Division of AIDS, NIAID, NIH.

Vaccine plasmid DNA

To generate CAL-SHIV-IN−, we used the classical molecular cloning techniques and PCR to remove the 5 and 3 SIV LTRs
from the genome of the infectious pathogenic SHIV-KU2 (GenBank accession # AY751799) and replace them with those of the goat lentivirus CAEV co strain (GenBank accession # M33677).

These latter have Tat-independent constitutive promoters that highly express driven genes. In addition, the SIV int coding sequences in the 3 end of the pol gene were deleted following double digestion with Kpn1 and Acc1 that removes 314 bp. The viral genome was inserted into the pET9 plasmid (Novagen, EMD chemicals, San Diego,CA) and the recombinant plasmid introduced into Escherichia coli JM109.

Design of CAL-SHIV-IN− lentivector vaccine genome. CAL-SHIV-IN− was derived from the genome of the chimeric, highly pathogenic SHIV-KU2. Both the 5 and 3 SIV LTRs have been removed and replaced by the LTRs of the goat lentivirus CAEV, that are known to have Tat-independent constitutive promoters. In addition the SIV coding sequences for the int (int) were deleted from pol gene.
The resulting genome has the SIV gag, pol, vif, vpx, vpr and nef
(solid green arrows) and HIV tat, rev and env genes (solid orange arrows).

The resulting CAL-SHIV-IN− expresses Gag, Pol (without IN), Vif, Vpx, Vpr and Nef of SIVmac239 (GenBank accession# M33262) as well as Vpu, Tat, Rev and Env proteins of HIV-1SF2 (GenBank accession# K02007).

Single colony of recombinant bacteria was amplified using classical culture conditions and bacteria were used to isolate endotoxin free vaccine DNA using QIAamp DNA Maxipep Kit(Qiagen, Courtaboeuf, France), according to manufacturer’s instructions.

Expression of viral proteins and production of viral particles

Cell transfections

Transfection of HEK-293 cells was performed using a cationic polymer polyethylenamine, ExGenTM 500 (Euromedex, Souffelweyersheim, France) according to the protocol provided by the manufacturer for adherent cells. Monolayer cells at 80% confluence in 6 well plates were transfected with 5 microgrammes of each of CAL-SHIVIN−, SHIV-KU2 and 4SHIV-KU2 plasmids, respectively. A GFP expressing plasmid (pCG-GFP) was used as internal control and non-transfected cells were used as negative control. DNA was mixed with 16.5 microlitres of ExGen 500 and 350 microlitres of NaCl 150 mM, incubated 35 min at room temperature and then the mixtures were inoculated into the cell monolayers.

Transfection medium was removed after overnight incubation, cell monolayers washed with phosphate-buffered saline (PBS) and replenished with fresh medium. Supernatant fluids containing viral particles were harvested at 24 h, 48 h and 72 h post-transfection and stored at −80 ◦C for further experiments. Cells were used in different assays described below.


Viral proteins were radio-immunoprecipitated from the culture medium and cell lysates using the standard protocol. Briefly,
at 48 h post-transfection, transfected cells were starved by culture inMethionine/Cysteine-deprivedmediumduring 1 h, and then
labeled overnight with 100 Ci of 35S-methionine/cysteine added in the same medium.

Viral proteins were then immunoprecipitated from the cell lysate and supernatant compartments using a hyperimmune polyclonal macaque serum as previously reported.

Quantitative analysis of SIV Gag p27 by ELISA

Supernatant fluids were harvested from both transfected and infected cells: HEK-293-transfected cells, co-culture of HEK-293-
transfected monolayers with CEM-x-174, or from infected M8166 and CEM-x-174 cells as indicated in Section 3. All supernatant fluids were analyzed for their content in SIV Gag p27. ELISA tests were performed using a commercial kit and the manufacturer’s instructions (XpressBio life science, MD, USA). A standard curve was plotted for each assay and OD450 values oftested samples were inserted to determine the concentrations of SIV Gag p27 (ng/ml). Values were used to generate the kinetics of SIV Gag p27 production.

Electronic microscopy

At 48 h post-transfection with CAL-SHIV-IN−, HEK-293 cells were fixed 1 h in 2.5% glutaraldehyde diluted in 0.1 M sodium cacodylate buffer. Cells were washed 3 times 0.1 cacodylate buffer alone and then 1 h in 1% OsO4 in 0.1 M cacodylate buffer, pH 7.4. Fixed cells were stained overnightin dark with 2%uranyl acetate pH 4. Cells were then dehydrated following serial washes in 30%, 60%, 90%, and 100% ethanol respectively prior to embedding in epoxy resin. Ultra-thin sections were observed under a J.E.O.L. 1200 EX transmission electron microscope operating at 80 kev and selected views were photographed.

Determination of viral stock titers

Cleared culture medium samples harvested as described in Section 2.3.1 and stored at −80 ◦C were used for virus titer determination. For titer determination of SHIV-KU2, cells of the highly fusiogenic lymphoid human CD4+ T cell line (M8166) were seeded in quadruplets of 3 × 105 cells/well in 24 well plates and inoculated with 10 fold serial dilutions of viral stocks. Titers in TCID50/ml were determined based on CPE detection in limit dilution that induced 50% infected wells. Quantification of CAL-SHIV-IN− and SHIV-KU2 was also performed by ELISA evaluation of SIV Gag p27. (SHIV-KU2 stock was evaluated at 28 ng/ml using Gag p27 ELISAcorresponding to a titer of 106 TCID50/ml).

Assay to examine the single-cycle replication of CAL-SHIV-IN−

Co-culture of CAL-SHIV-IN− and SHIV-KU2 transfected HEK-293 with CEM-x-174 permissive cells

HEK-293 cells were seeded into 6 weel plates and 24 h later pre-transfection culture medium was harvested on day 0. At 24 h post-transfection, monolayers of HEK-293 cells, transfected with SHIV-KU2 and CAL-SHIV-IN− respectively, together with nontransfected cells were co-cultured with 106 CEM-x-174 cells in six well plates. Supernatant fluids of co-cultures were harvested daily during 9 days post-transfection, and used for quantitative SIV Gag p27 ELISA.

Serial infection of CD4+ T cell lines with CAL-SHIV-IN− and SHIV-KU2 viruses

Human lymphocytic CD4+ T cell lines (M8166 and CEM-x174) were cultured in 10% FBS and 1% penicillin/streptomycin supplemented RPMI at a density of 5 × 105 cells/ml. Triplicates of 2 × 106 cells were respectively infected with viral stock of CALSHIV-IN− and SHIV-KU2 (MOI = 0.2) and cells inoculated with non-transfected HEK-293 supernatant were used as negative controls. At day 4 cytopathic effects (CPE) were clearly diffused in SHIV-infected cells.

Cells were then washed twice with PBS and fresh medium was added. At day 7 post-infection, supernatant was harvested for monitoring viral protein release and subsequent round of infection of fresh target M8166 for infectivity assays. Cells were rinsed twice with 1× PBS, stored as dry pellet and then used for DNA isolation.

Detection of vaccine DNA in CEM-x-174 cells using nested-PCR

DNA extraction from cultured CEM-x-174 cells was performed with NucleoSpin Blood QuickPure (Macherey-Nagel, Hoerdt, France) according to the manufacturer’s instructions. The highly sensitive nested-PCR technique was used to detect the viral DNA in CEM-x-174 cells. Two successive rounds of PCR amplification were used to detect the SIV specific gag sequences by nested-PCR. The first round of PCR amplifies a 612 bp product using the external primers YCN-5 (5 -GAG TGG GAG ATG GGC GTG AG-3 ) and YCN-9 (5 -CTG CAT AGC CGC TTG ATG GTC TC-3 ).

The second round amplifies a 418 bp product using the internal primers YCN-6 (5 -AGT ATG GGC AGC AAA TGA AT-3 ) and YCN-8 (5 -CCT GGC ACT ACT TCT GCT CC-3 ). Amplification of the human housekeeping ˇ-actin gene was used as internal control. A set of external primers HHK–1 (5 – TCA TGT TTG AGA CCT TCA ACA CCC CAG-3 ) and HHK–2 (5 -CCA GGG GAA GGC TGG AAG AGT GCC-3 ) and two internal primers, HHK–3 (5 -TGG ACC TGG CTG GCC GGG ACC TG-3 ) and HHK–4 (5 -GCC TCA GGG CAG CGG AAC CGC TCA-3 ) were used for the first and second round of PCR, respectively. The expected sizes of PCR products are 436 and 246 bp for the first and second round of PCR respectively.

The first round of PCR was performed with 5 µl of 10× Taq buffer, 5 µl of MgCl2 (25 mM), 1 µl of dNTP mix (20 mM), 1 µl of both YCN5 and YCN9 (20 µM), 0.5 µl of both HHK–1 and HHK- -2 (20 µM), 500 ng of extracted DNA, 1.5 µl of TAQ Pol (5 U/µl, Euromedex, Souffelweyersheim, France), in a final volume of 50 µl adjusted with ultrapure water. The second round was performed using 2 µl ofthe first PCR with 1 µl of each of YCN6, YCN8, HHK–3 and HHK–4 (20 µM) in the same conditions as the first round. Amplification conditions were as follow: 5 min of initial denaturation at 94 ◦C followed by 35 cycles comprising successively 30 s at 94 ◦C, 30 s at 62 ◦C, 1 min at 68 ◦C; and a final incubation of 10 min at 68 ◦C. PCR products were separated by electrophoresis on 1.8% agarose gel and revealed under UV light after ethidium bromide staining.

Southern blot DNA analysis

DNAs were transferred from agarose gel to a nylon membrane Roti®-Nylon 0.2 (Roth-Sochiel, Lauterbourg, France). DNAs in the agarose gel were denatured by incubation 30 min in a solution containing 1.5 M NaCl and 0.5 M NaOH. Denaturation was neutralized following 30 min incubation in a solution containing 1.5 M NaCl, 0.5 M Trizma base, pH 7.5. The Gel was then rinsed 10 min in 3× SSC solutions (0.3 M Tri-sodium citrate, 3 M NaCl, pH 7-8) and placed overnight between two Roti®-Nylon membranes using a dry transfer system.

DNAs in the nylon membranes were then crosslinked 15 min under UV light. A specific DNA probe of SIV gag sequences was generated with the SIV external oligonucleotide primers using DIG High Prime DNA Labeling and Detection Starter Kit I (Roche, Meylan, France). 50 ng of a plasmid containing the SIVgag gene wereused asmatrix for PCR amplification with TAQ Pol DNA Polymerase Kit. PCR protocol was identicalto the first round ofthe nested PCR using YCN-5 and YCN-9 gag primers but in the absence of HHK–1 and HHK–2 actinspecific primers.

The PCR product was purified using NucleoSpin Gel and PCR Clean-Up (Macherey-Nagel, Hoerdt, France). The total purified PCR product was then used as template to produce random primed DNA probes following manufacturer’s instructions. Transferred membranes were hybridized 6 h at 42 ◦C with DIG-labeled SIV gag probe using the manufacture’s protocol. Membranes were then washed in 2× SSC 0.1% SDS and 0.1× SSC 0.1% SDS solutions, successively. Revelation was performed using an anti-digoxygeninalkaline phosphatase conjugate and NBT/BCIP substrate generating a blue color product as indicated by the manufacturer.

Assays for evaluation of immunogenicity

Animal models and immunizations

BALB/c mice were purchased from Harlan Laboratories and housed in the animal facility of the Jean Roget Institute, UJF, Grenoble. NOD/SCID/2 mice were sampled from the colony that is bred in this same facility. All protocols were approved by the regional Ethical committee prior to use and experimentation.

Animals were housed in cages and taken care in accordance with the guidelines of the Animal Care and Use of the EU and French regulations and the Local Ethic Committee recommendations (Agreement number: B3851610006). Human blood samples collected in sodium heparin-coated tubes were obtained from a healthy HIV-free non-infected blood donor, from the “Etablissement Franc¸ ais du Sang, EFS, Grenoble”. Peripheral blood mononuclear cells (PBMCs) were isolated using the classical Ficoll-Hypaque density gradient separation.

NOD/SCID/2 mice were exposed to a dose of 120 centigray gamma irradiations for a period of 50 s. At 24 h post-irradiation, mice were humanized by intra-peritoneal injection with 5 × 107 cells of freshly purified human PBMCs.

Six weeks old BALB/c and hu-NOD/SCID/2 mice were injected intra-muscularly with a single dose of naked plasmid DNA vaccine. 4SHIV-KU2 and CAL-SHIV-IN− DNAs were diluted in PBS to a final concentration of 1 µg/µl each. BALB/c mice were injected with a total of 100 µg of DNA vaccine, while hu-NOD/SCID/2 were given 50 µg only.

Control mice received PBS solution without DNA. Mice were sacrificed at week 1, 2, 3 and 4 post-immunization for immune response analyses. Isolated spleens of mice were aseptically smashed between two glass slides to harvest splenocytes. Cells were collected in PBS, 0.5 mM EDTA solution, erythrocytes were lysed using a red blood cell lysis solution (BD Pharmingen, Pont de Claix, France), and mononuclear cells were counted in a hemocytometer.

A portion of splenocytes was used for evaluation of human T cell engraftment.Another was used for stimulation with SIV/HIV antigens; a fraction of which was used in ELISPOT assay to evaluate cytokine-producing cells, and the other fraction was used for antigen-specific proliferation assay.


IFN-ɤ-producing cells were detected by quantitative ELISPOT assay using the mouse or human INF- antibody-precoated 96-well plates (Mabtech, Sophia Antipolis, France). Triplicates of 5 × 105 splenocytes were incubated 20 h at 37 ◦C with 2 µg/ml of pools of overlapping peptides corresponding to Gag, Env, and Tat + Rev + Nef (TRN).

Concanavalin A (1 µg/ml) or anti-CD3 stimulated cells were used as positive controls while only medium-cultured cells were used as negative control. The cutoff for positivity in this assay was 10 SFC/106 PBMCs. This corresponds to the average of spots obtained in cultured medium controls.

Multiparametric flow-cytometry assay monitoring SHIV-specific T cell functions

To assess proliferation and cytokine production of SHIV-specific T cells, splenocytes of hu-NOD/SCID-2 mice were first stained with carboxyl fluoroscein succinimidyl ester (CFSE) (107 cells/ml in 1 µM CFSE for 10 min at 37 ◦C; Molecular Probes, Cergy Pontoise, France). Fractions of 2 × 106 cells were separately seeded into 96 deep-well tissue culture plates with or without stimulation with overlapping peptides of Gag, Env, or TRN pools at final concentration of 2 µg/ml.

Four days later cells were stimulated again for 16 h with the same peptide pools in the presence of 0.5 µg/ml of costimulatory antibodies CD28 (clone 37.51) and CD49 (clone 9F10) and brefeldin A (Sigma–Aldrich, Saint-Quentin Fallavier, France). Cells were washed and stained for surface and intracellular compounds as previously described [30]. All Abs were purchased from BD Biosciences and flow-cytometry analysis was performed on a three-laser BD LSRII instrument with standard setup. Data files were collected and analyzed using the FACSDiva software program (version 4.1.2; BD Biosciences).

Statistical analyses

We used paired t-test and GraphPad software to examine the significance of the data. Error bars represent standard error of mean (s.e.m), *P < 0.05, **P < 0.01, ***P < 0.001, NS = not significant (P > 0.05).

What did they obtain ?

Design of CAL-SHIV-IN− lentivector DNA vaccine

CAL-SHIV-IN− lentivector DNA vaccine was derived from the genome of a replication-competent chimeric simian/human immunodeficiency virus, SHIV-KU2, as described in Section 2. Two major modifications were introduced in SHIV-KU2 genome. First, we removed the int coding sequences from the pol gene to abolish the integration potential resulting in an integration-defective lentivector genome.

Second, we replaced both 5 and 3 LTRs of SIVmac239 with those of the naturally attenuated caprine lentivirus (CAEV) to obtain CAL-SHIV-IN− genome. Based on the design of this novel lentivector DNA vaccine, we hypothesized that CAL-SHIV-IN− will provide multiple sources of antigen:

  1. viral proteins directly expressed by transfected myocytes and other somatic cells;
  2. VLPs produced by these transfected cells;
  3. target cells pseudo-infected with particles containing integration-deficient viral genomes will express viral proteins from the episomal unintegrated DNA as a third source.

These various cell-free and cell-associated viral antigens are potential sources for cross-presentation and optimal priming of T cell responses. This new lentivector DNA vaccine is differentfrom the former 4-SHIVKU2 and CAL-4-SHIV-KU2 since it carries both 5 and 3 LTRs of CAEV in both ends and the complete coding sequences for the rt gene. 4-SHIV-KU2 and CAL-4-SHIV-KU2 carry only the 5 LTR and lacks the complete coding sequences of rt.

Characterization of functional properties of CAL-SHIV-IN–in cell culture

Expression of viral proteins was examined by immunoprecipitation of virus-specific proteins from both cell lysates and supernatant fluids of HEK-293 cells transfected with CAL-SHIV-IN− and parental SHIV-KU2, respectively, using a polyclonal macaque serum anti-SHIV. All major proteins expressed by the vaccine genome (Gag p27, Gag polyprotein p55, Env surface glycoprotein gp160, gp120 and gp41) were detected in the cell lysates and mature proteins in the supernatant compartments. The protein profile of CAL-SHIV-IN− was similar to that of SHIV-KU2 indicating thatthe replacement of SIVLTRs and removal of SIVint didnot affect the expression of other major viral proteins.

Expression of CAL-SHIV-IN− lentivector proteins. (A) HEK 293 cells were transfected with either CAL-SHIV-IN− (2) or SHIV-KU2 (3) DNAs. At 48 h post-transfection, cells were starved using methionine/cysteine-free medium and then proteins were labeled with 100 Ci of 35S-methionine/cysteine. Viral proteins were immunoprecipitated from supernatant (S) and cell lysate (C) of transfected HEK-293 cells with a hyperimmune macaque polyclonal serum. Env precursor and mature glycoproteins (gp160, gp120) and Gag proteins (p55 and p27) were successfully immunoprecipitated from the cell lysates (C). Only mature proteins (Env gp120 and Gag p27) were immunoprecipitated from the supernatants (S) of CAL-SHIV-IN− (2) and SHIV-KU2 (3) transfected cells. Non transfected HEK-293 (1) used as control lacked any of these viral proteins

In addition, EM-observation of cells transfected with the CAL-SHIV-IN− showed assembly of virus proteins into virions budding at the cell surface as well as mature virions.

To further evaluate the production of the major SIV Gag p27, the culture medium of transfected HEK293 cells co-cultured with CEM-x-174 cells were collected for nine consecutive days and analyzed by ELISA (Fig. 3A). SIV Gag p27 protein was found to be efficiently expressed by the vaccine and rapidly released in the supernatant of HEK-293-transfected cells. On day 1 and 2 post-transfection, the same amount of SIV Gag p27 protein (≈30 ng/ml) was detected in the supernatants from both types of CAL-SHIV-IN− and SHIV-KU2 transfected co-cultured cells.

Electron micrographs of a HEK-293 cell transfected with CAL-SHIV-IN− DNA (a) showing budding and mature virions indicated by arrows.
A non-transfected HEK-293 cell (b) is shown as negative control. Magnification is 19,000×.

Supernatant from non-transfected HEK-239 did not contain any SIV Gag p27 protein. Subsequently, between day 3 and 6 post transfection, levels of released SIV Gag p27 started to drop in the supernatants of both SHIV-KU2- and CAL-SHIV-IN−-transfected cells. However, the concentration of SIV Gag p27 protein was significantly lower (11 fold) in day 6 samples from CAL-SHIV-IN− (≈1.4 ng/ml) than those from the parental SHIV-KU2 (≈15.9 ng/ml).

In addition, while at day 9 post transfection the concentration of SIV Gag p27 in the samples of CAL-SHIV-IN− remained very low similar to day 6 (1.4 ng/ml), that of SHIV-KU2 sample increased significantly to reach 28 ng/ml; This may result from active replication of SHIV-KU2 but not CAL-SHIV-IN− in co-cultured CEM-x-174.

Infectivity assay of CAL-SHIV-IN− virions. (A) Culture medium of non-transfected HEK-293 cells was harvested on day 0. Cells were transfected with either SHIV-KU2 or CAL-SHIV-IN− plasmid DNAs. At 24 h post-transfection (day 1), supernatant fluids were harvested before coculture of monolayers with CEM-x-174 cells. Supernatant fluids were then collected daily from the co-cultures till day 9 post-transfection. Kinetics of viral protein expression was monitored using quantitative Gag p27 ELISA. Results plotted in (A) were compared using paired t-test. Error bars represent standard error of mean (s.e.m), n = 3. *P < 0.05, **P < 0.01, ***P < 0.001, NS = not significant (P > 0.05). Production of Gag p27 by SHIV-KU2- or CAL-SHIV-IN−-transfected cells was significantly different from that of controls. Supernatant harvested at day 1 post-transfection was used to inoculate M8166 or CEM-x-174 cells at MOI of 0.2. (

CAL-SHIV-IN− undergoes one cycle of replication

To monitor the limited replication capacity of CAL-SHIV-IN− to a single cycle of replication, culture medium of transfected cells containing viral particles was harvested and tested for infectivity using the highly permissive/fusiogenic human CD4+ T cell lines M8166 and CEM-x-174.

Analysis showed that samples from both SHIV-KU2- and CAL-SHIV-IN−-transfected cells induced cytopathic infections in the indicator T cell lines that underwent syncytia formation with characteristic giant multinucleated cells .

Micrographs of M8166 cells after the first and second round of infection (R1, R2).
The black arrow (quadrant c) is pointing a sporadic cytopathic effect induced by CAL-SHIV-IN− in the first round of infection. Magnification is 400×.

Syncytia were larger and in greater number in cells infected with SHIV-KU2 than those with the vaccine. In these latter development of cytopathic effect was sporadic and did not progress with the time. When the supernatants of these infected cells were used to inoculate fresh CD4+ T cells.

Only the supernatant from SHIV-KU2-infected cells induced cytopathic effects. Cells inoculated with that of CAL-SHIV-IN−-infected cells failed to develop any syncytium remaining comparable to control cells. This indicated that CAL-SHIV-IN− was unable to undergo more than one cycle of replication and infection was restricted to the culture medium of transfected cells only. To confirm this observation, 4 successive rounds ofinoculation of M8166 and CEM-x-174 cells were performed using in parallel SHIV-KU2, CAL-SHIV-IN− viral stocks and the culture medium from non-transfected HEK-293 cells as negative control. In each inoculation experiment a fraction of supernatant was used to inoculate cells and the remaining was used to evaluate the amount of SIV Gag p27 proteins by quantitative ELISA.

Results presented in this Figure show that the release of SIV Gag p27 protein is restricted to the first round of inoculation with CAL-SHIV-IN−. In contrast, as expected SIV Gag p27 protein was found to be produced and released in the supernatant of cells inoculated with any passaged virus stock of SHIV-KU2.Accordingly, viral DNA was detected only in cellular DNA from the first round of infection with CAL-SHIV-IN− as reported by hybridization of the nested PCR products with an SIV gag specific probe.

Evaluation of viral protein expression by quantitative ELISA. Amount of released Gag p27 after first, second and third round (R1, R2 and R3) of infection of CEM-x-174 at weekly intervals are plotted.
Paired t-test, n = 3, was used as described above

DNA from cells inoculated with supernatant fluids from further rounds of CAL-SHIV-IN− failed to show the presence of viral DNA (lanes 7–9) similarly to negative controls (lanes 1–5). These data demonstrate that in contrast to the replication competent SHIV-KU2 (lane 10, Fig. S1), the integration defective CAL-SHIV-IN− has the potential to undergo only a single cycle of replication in target cells.

Southern blot analysis of nested PCR products obtained with cellular DNAs following the first, second, third and fourth rounds of serial infection using
supernatants fluids of M8166 cells.
Samples were separated in 1.8% agarose gel (supplemental Fig. S1). H2O without DNA (lane 1) and DNAs isolated from non-infected cells at each of the serial rounds (lanes 2–5, respectively) were used as negative controls. Sample of cells inoculated with each of the serial passage of CAL-SHIV-IN− rounds (lanes 6–9, respectively).
A sample of the fourth round of SHIV-KU2 (lane 10) was used as positive control.

Immune responses induced in the mouse models

To examine the profile of our DNA vaccine-induced immune responses, we first IM-injected BALB/c with a single dose of novel single-cycle-integration-defective CAL-SHIV-IN− in parallel with the former non-replicating non-integrating 4SHIV-KU2 DNA lentivector vaccines. Mice were euthanized at week 1 and 2 post-immunization and splenocytes were examined for detection of vaccine-specific immune T cells by ELISPOT. The results in this figure show that both vaccines have induced INF- producing T cells specific to all SHIV antigens (Gag, Env Tat + Rev + Nef) expressed by the vaccine constructs.

To explore further the immunogenicity induced by this new lentivector vaccine we used human PBMC-reconstituted NOD/SCID/2 mice to provide target cells that support SHIV replication. A single IM-immunization of humanized mice with CAL-SHIV-IN−, 4SHIV-KU2 and SHIV-KU2 plasmid DNAs was evaluated by ELISPOT.

While modest antigen-specific responses were induced by 4SHIV-KU2 DNA (35 spots/106 splenocytes as mean response against SHIVantigens),the responses induced by CAL-SHIV-IN− were substantially increased (190 spots/106 splenocytes). Importantly, the level of responses induced by this latter was almost equivalent to that induced by the replication-competent SHIV-KU2 (233 spots/106 splenocytes).

Altogether, these results provide the demonstration that the additional one cycle of replication in target human cells has significantly improved the immunogenicity of the novel vector. CAL-SHIV-IN−-specific T cell responses were further characterized using the multi-parametric FACS-based assay that we previously developed in the BALB/c model. This assay helps to examine simultaneously antigen-specific T cell proliferation and cytokine production in response to ex vivo stimulation with SHIV peptides.

Representative results indicate that Gag and TRN-specific T cell responses were mostly composed of proliferating CD3+ CD8+ T cells that increased by 3 fold from week 1 (5.9% of total CD8+ T cells) to week 2 (17.4% of total CD8+ T cells) post-immunization. The majority of these cells did not produce detectable level of IFN- in this five days in vitro culture assay as we previously described. In parallel to the T cell immune responses, we also examined vaccine-induced humoral responses in the immunized hu-NOD/SCID/2 mice.

We performed an ELISA against Gag p27 using serum samples collected at week 2 post-immunization. Similarly to the T cell responses, we found higher level of Gag-specific IgG with CAL-SHIV-IN− and SHIV-KU2 as compared with the response obtained with the replication-defective 4SHIVKU2 vaccine.


After thirty years of HIV discovery,there are still no efficient and safe enough prophylactic and/or therapeutic vaccines that can be used in humans. Live-attenuated vaccines showed promising levels of protection in NHP models; however,they were associated with irreversible integration in the host genome, persistence and reversion to pathogenic state in some vaccinated animals impelling safer strategies for clinical use.

Alternatively, DNA vaccines became more attractive tools for vaccination against HIV and other pathogens, particularly with effort being made to increase their immunogenicity via optimizing vaccine design, use of adjuvants, and delivery methods. In this paper we engineered and evaluated a new SHIV-based lentivector DNA prototype, CAL-SHIV-IN− as HIV vaccine.

This lentivector was conceived to be administered as plasmid DNA expressing all SHIV proteins except the integrase. Viral proteins assemble into mature viral particles lacking the integrase and able to perform a single cycle of replication in target cells, hence increasing antigen expression and presentation in vivo. As follows, our construct mimics the beneficial replication of attenuated viruses without their drawbacks.

The most important and unique features of CAL-SHIV-IN− lentivector vaccine are associated with:

  1. its design that includes both 5 and 3 LTRs from the genome of the naturally attenuated goat CAEV lentivirus,
  2. the constitutive gene expression driven by CAEV LTRs containing Tat-independent constitutive promoters,
  3. the fact that this design not only results in augmented and constitutive expression of all proteins in the vaccine genome, but also this expression is regulated sequentially and quantitatively mimicking that of the natural infection with the live-attenuated and the pathogenic viruses, and finally
  4. the protein assembly results into mature virions that undergo a single cycle replication in the absence of vaccine genome integration.

Our results provide the demonstration that CAL-SHIV-IN− vaccine is indeed a single-cycle genome since virus produced by HEK-293-transfected cells was able to transduce indicator T cells that underwent typical CPE and in which viral DNA was detected by conventional PCR.

In contrast, the culture medium of these cells failed to cause any further infection in inoculated new T cells that did not develop CPE and lacked detectable viral DNA. Altogether these data indicated that transfection of CAL-SHIV-IN− plasmid DNA into HEK-293 cells has produced viral particles containing CAL-SHIV-IN− genomic RNA that was reverse transcribed into viral DNA in transduced T cells, but this latter DNA failed to produce infectious viral particles able to undergo a new cycle of infection of target permissive cells.

Lentivirus-based vectors have been widely used for gene therapy for their non-oncogenic long-lasting expression of therapeutic genes . However, this type of gene therapy is associated with the integration of the vector genome in that of the host. In vaccine strategies integration of vaccine genome threaten its safety since it can be associated with insertion mutagenesis, increased probability of recombination with endogenous and/or exogenous viral genome, persistent antigen expression that could have negative outcome on the set-up of efficient memory immune responses .

For these and other reasons, some lentiviral vaccine vector prototypes were mutated to become integration defective . These types of vectors were shown to highly express HIV antigens thereby inducing potent immune responses both in mouse and macaque models of HIV vaccine. For the sake of increased safety, lentiviral vaccine vectors were mutated to abolish integration of their genome into that of the hosts.

This type of vectors was not only developed as prototypes for HIV vaccination but also malaria , Hepatatis and human papilliomavirus. To increase the immunogenicity of lentivirus-based vaccines they were engineered as one cycle replication to amplify in vivo the antigens . However, in the design of some of these vectors the integration capability of the genome was maintained.

These types of vaccine vectors were shown to be efficient at expressing vaccine antigens which induce potent immune responses. Immunization of macaques with SIVmac-based singlecycle virus demonstrated a reduction of 1–3 logs of viral load of the challenge virus in the acute phase in 3 out 4 immunized animals compared to the control non immunized infected animals.

Similarly, intramuscular injection of plasmid DNA expressing all SIV(Mne) antigens in a single-cycle design induced partial control of the viral load of the challenge virus at the initial peak of plasma viremia, and low to undetectable viremia in the long term. The originality of CAL-SHIV-IN− is the synergy of both the onecycle replication and non-integration of the vaccine genome that expresses all 13 proteins driven by CAEVLTRs. In addition,the CAEV integrase is highly specific for the “att” sequences atthe extremities of CAEV LTRs .

Similarly, SIV integrase is highly specific for “att” sequences of SIV and HIV-2 LTRs and does not interact with the closely related HIV-1 LTR “att” sequences. This adds another level of safety since if recombination occurs involving the genome of CAL-SHIV-IN− and that of HIV, the integrase of HIV will not be able to catalyze the integration of CAEV LTRs. More interestingly if HIV acquires any of the CAEV LTRs it will result in the suicide of the genome by inactivation of the specific integration.

Constructs expressing up to 15 viral antigens have been previously described; but none of these was driven by an LTR with a constitutive promoter and able to undergo a single cycle of replication like CAL-SHIV-IN−. Our 4-SHIV-KU2 and CAL-4-SHIV-KU2 that expressed 12 viral proteins lacked the rt coding sequences and the 3 LTR that abolish the reverse transcription and consequently there is no single-cycle of replication.

Interestingly, once injected to mice, CAL-SHIV-IN− induced specific humoral and T cell immune responses that were directed against all the proteins (Gag, Env and Tat + Rev + Nef) expressed by the vaccine as early as one week and these responses increased at week 2 post-immunization. The results showed that the extra cycle of replication in target human cells correlated with significant increase in antigen-specific immune responses in CAL-SHIV-IN−- immunized hu-NOD/SCID/2 mice compared to those immunized with the replication-defective 4-SHIV-KU2.

This increase was not observed in BALB/c mice in the cell of which the replication of SHIV is restricted. However, the increased ELISPOT responses observed with samples of mice immunized with 4-SHIV-KU2 compared to CAL-SHIV-IN− DNAs might be linked to the quality of isolated plasmid DNAs. Indeed, the proportion of supercoiled DNA is known to be determinant for efficient immunogenicity as this latter enters more efficiently the nucleus allowing increased gene expression.

Altogether, these data helped to conclude that the increase of induced immune responses observed in hu-NOD/SCID/2 immunizedmice (i) was associated withthe one cycle replicationintarget human cells, (ii) provided the demonstration of the effectiveness of this innovative strategy for augmenting the immunogenicity of DNA vaccines and (iii) induced both humoral and cellular arms of the immunity.

Further characterization of induced T cell responses reflected clearly a predominance of the CD8+ T cell responses. This type of immunity is known to play a pivotal role in the control of HIV/SIV replication by decreasing the viral load. In contrast to memory CD8+ T cells generated during chronic infection, responses linked to limited antigen expression observed during acute infection present enhanced capacity to be maintained in the absence of Ag and to undergo superior recall response. This was shown in mice using LCMV infection model.

Our nonintegrating and non-persisting vaccine will hence induce similar responses that persist over a long period. Interestingly, similarly to the data that we reported earlier in both mice and macaque models, most proliferating CD8+ T cells induced by CAL-SHIV-IN− do not produce INF- but do produce the lytic compound Granzyme B as characteristics of cytotoxicity. We observed that at the early time points post immunization (week 2 and 4) almost equivalent response were detected against Gag, Env and TRN.

As observed with 4SHIV-KU2 derivatives, we expect that this response will be preserved and become immunodominant for Gag at later time-point. Unfortunately hu-PLB-SCID model does not allow longitudinal evaluation of these responses due to rapid establishment of graft versus host disease . Recently, this limitation was pushed back by the introduction of new SCID models like NSG and NOG mice . Finally this construct showed interesting properties with ability to mimics the early stages of natural infection of primate lentivirus and to generate in vivo additional sources of antigens.

Those extra sources of antigens will be presented by APCs to lymphocytes T and B cells and should help to improve greatly the antigen-specific immune responses similarly to those seen in LAV-immunized animals and Nef-defective HIV-1-infected LTNP individuals that efficiently control their infections. This promising data promoted initiation of immunogenicity studies in cynomologus macaques a more appropriate model for HIV-1 vaccine studies. We recently reported the immunogenicity results of our first use of CAL-SHIV-IN− vaccine in cynomolgus macaques, and demonstrated that a single dose immunization with this vaccine induced long lasting humoral and T cell immune responses against all SHIV antigens.

Detailed characterization of induced T cell responses demonstrated that these persistent responses comprised circulating central and effector SHIV-specific memory T cell responses that should play a critical role in controlling viral replication after acquisition of infection.