- An adenovirus prime/plasmid boost strategy for induction of equipotent immune responses to two dengue virus serotypes

Creation of an EDIII-based bivalent antigen-encoding adenovirus vector, rAd-Bg

Two rAd vectors, shown in Figure 1A, were constructed for this study. The test vector, rAd-Bg, carries the human cytomegalovirus (CMV) promoter-driven chimeric EDIII-4/2 antigen gene (derived by linking the EDIII-encoding cDNAs generated from DEN-4 and DEN-2 viral genomic RNAs) fused to a 3' green fluorescent protein gene (GFP) tag, inserted into the Ad5 genome, in place of its E1 region. The control vector, rAd-C, is similar in design, with the exception that it carries an empty CMV expression cassette. Aliquots of the rAd viral DNA were subjected to PCR analysis using Ad E1- and insert-flanking primers as described earlier [42]. The results are shown in Figure 1B. The E1-specific primers produced the predicted ~0.47 Kb amplicon when the template was DNA from wild type (wt) Ad (lane 2), but not when it was from either the rAd-C (lane 3), or rAd-Bg (lane 4) viruses. PCR with insert-flanking primers resulted in the amplification of a ~1.5 Kb cDNA only from rAd-Bg (lane 8) but not from the wt Ad DNA template (lane 6). From rAd-C viral DNA as the template, the insert-flanking primers amplified a ~0.1 Kb vector derived cDNA product (lane 7). Extensive restriction mapping of the rAd-C and rAd-Bg viral DNAs were found to be consistent with the construction strategy (data not shown).

Figure 1. The rAds constructed for this study. (A) Schematic representation of the linear genomes of the rAd-C and rAd-Bg viruses. The dashed line, close to the left end represents the position of region E1. The open box located towards the right end represents a ~2.7 Kb deletion in the non-essential E3 region (ΔE3). The hatched boxes at either end represent the left (L) and right (R) inverted terminal repeats (ITRs), and the open circle close to the L-ITR represents the packaging signal (ψ). In the rAds, region E1 is replaced either by an empty expression cassette (rAd-C) or the EDIII-4/2 expression cassette (rAd-Bg). Components of the expression cassettes are indicated by the following abbreviations, P: CMV promoter (the arrow indicates its orientation); 4 and 2: DEN-4 and DEN-2 EDIII-encoding regions, respectively, of the bivalent EDIII-4/2 gene; g: enhanced green fluorescent protein gene (fused in-frame to the EDIII-4/2 gene) and pA: SV40 polyadenylation signal. (B) Characterization of the rAd genomes by PCR. The picture depicts agarose gel analysis of PCR products generated using wt Ad5 (lanes 2, 6), rAd-C (lanes 3, 7) or rAd-Bg (lanes 4, 8) viral DNA templates. Primers used were either E1-specific (lanes 2–4) or specific to insert-flanking sequences (lanes 6–8). DNA size markers were run in lanes 1 and 5. Sizes of the lane 1 markers (in Kb) are indicated to the left of the panel. The position of the 1.6 Kb band in lane 5 is indicated by a white dot for orientation. The arrows to the right indicate the positions of the amplicons corresponding to the bivalent insert (top), E1 region (middle) and the insert-less vector sequences (bottom).

We next examined the ability of the rAd-Bg virus to express the bivalent antigen by direct fluorescence microscopy to detect its GFP tag. In the experiment shown in Figure 2A, HEK 293 cells were observed ~24 hours after infection with rAd-Bg (panel ii). Detection of the C-terminal GFP tag suggests successful rAd-mediated expression of the bivalent antigen. On the other hand, rAd-C virus-infected cells (panel i), analyzed in parallel, did not display any green fluorescence. That the rAd-C virus had efficiently infected the cells was evident from the cytopathicity observed by visible light microscopy (data not shown). In addition, we carried out indirect immunofluorescence assays (IFA) to detect the EDIIIs corresponding to DEN-2 and DEN-4 in the bivalent antigen. To this end, we probed rAd-Bg infected cells, separately, with monoclonal antibodies (mAbs) 3H5 and MAB8704 to identify EDIIIs corresponding to DEN-2 and DEN-4, respectively. As the bivalent antigen was GFP-tagged, we used anti-murine IgG-rhodamine conjugate as the secondary antibody in the IFA experiment. 3H5 is a DEN-2 serotype-specific mAb, which specifically recognizes an epitope that maps to amino acid (aa) residues 386–397 of the E protein [44]. As this epitope falls within the DEN-2 EDIII region, this mAb is expected to recognize the rEDIII-4/2 bivalent antigen as well. MAB 8704 is a DEN-4 virus-specific mAb. We found that it was able to specifically recognize E. coli-expressed DEN-4 EDIII (our unpublished data). Thus, this mAb also is expected to recognize the rEDIII-4/2 bivalent antigen. Consistent with this, probing rAd-Bg infected cells with mAbs 3H5 (panel iii) and MAB8704 (panel iv) resulted in red immunofluorescence. We also identified the bivalent antigen in rAd-Bg virus-infected cells using an immunoprecipitation approach (Figure 2B). In this experiment, we used mAb 3H5 to immunoprecipitate the bivalent antigen expressed in infected HEK 293 cells. Consistent with the direct fluorescence and IFA data, the 3H5 mAb specifically identified a protein of the predicted size (~55 kDa) in rAd-Bg infected (lane 4), but not in rAd-C infected (lane 2) and mock infected (data not shown) HEK 293 cells. As a positive control, a parallel immunoprecipitation was performed using cells infected with a previously reported rAd [45], expressing aa 1–395 of the DEN-2 virus E protein (lane 3). Taken together, these data showed that the rAd-Bg virus is capable of expressing the intended chimeric bivalent antigen in infected mammalian cells.

Figure 2. rAd-mediated expression of rEDIII-4/2 protein as aGFP fusion. (A) Detection of rAd-mediated bivalent antigen expression by fluorescence microscopy. HEK 293 cells were either infected with rAd-C (panel i), or rAd-Bg (panel ii-iv) for 24 hours and examined either directly for GFP fluorescence (panels i and ii) or analyzed by IFA, using either mAb 3H5 (panel iii) or MAB8704 (panel iv), as the primary antibody in conjunction with rhodamine-conjugated anti-murine IgG as the secondary antibody. (B) Immunoprecipitation of the bivalent protein from infected HEK 293 cells. Cells were infected with rAd-C (lane 2), a rAd expressing DEN-2 E (lane 3 [ref. 45]) or rAd-Bg (lane 4) viruses. Infected cells were metabolically labeled with [35S]-methionine, lysed, immunoprecipitated with 3H5 mAb, and analyzed on a denaturing 15% polyacrylamide gel. Pre-stained protein markers were run on the same gel (lane 1); their sizes (in kDa) are indicated on the left. The arrows on the right show the positions of the bivalent protein (upper) and the E protein (lower).

The bivalent antigen-induced antibodies recognize and neutralize the infectivity of DEN-2 and DEN-4, but not DEN-1 and DEN-3 viruses

A major objective of this study was to investigate if a single rAd vector encoding an EDIII-based bivalent DEN antigen would elicit antibodies specific to both the constituent serotypes. To this end, we undertook mice immunization experiments. As anti-Ad antibodies resulting from the initial immunization of the mice can potentially interfere with the efficacy of the rAd vector to function as a vaccine carrier during booster immunizations, we adopted a heterologous prime/boost strategy that entailed priming the animals with the rAd vector, followed by boosting with a plasmid vector encoding the same bivalent antigen gene. Sera were collected periodically after priming and after each boost. We performed two series of ELISAs using these sera. In the first experiment, we determined antibody titers after each immunization, using the bivalent r-EDIII-4/2 protein as the capture antigen, as shown in Figure 3A. It is evident from the data presented that initial antibody titers which were relatively low, progressively increased after each boost, with maximal titers being observed after the final plasmid boost. In the next experiment, we analyzed the capacity of sera obtained after the final boost, to specifically react with monovalent EDIII proteins corresponding to each of the four DEN virus serotypes. Accordingly, in this experiment antibody titers were determined using using rEDIII-1, rEDIII-2, rEDIII-3 and rEDIII-4 proteins, as capture antigens. The results of this experiment, depicted in Figure 3B, showed that serum from the immunized animals manifested high levels of reactivity towards the rEDIII proteins corresponding to DEN virus serotypes 2 and 4, but not to serotypes 1 and 3.

Figure 3. Reactivity of immune sera towards monovalent and bivalent rEDIII proteins. (A) Antibody titers in serial dilutions of murine sera drawn 1 week after priming with rAd-Bg (blue) and 1 week following the first (red), second (green) and final (violet) boosts with pVAX-EDIII-4/2 plasmid were determined in an ELISA using the rEDIII-4/2 bivalent protein as the capture antigen. The black dashed curve indicates ELISA reactivity of control sera, drawn from rAd-C primed/pVAX1-boosted mice after the final boost. (B) Antibody titers in serial dilutions of murine sera drawn from rAd-Bg primed/pVAX-EDIII-4/2 boosted mice, one week after the final boost. ELISA was performed using rEDIII-1 (blue), rEDIII-2 (green), rEDIII-3 (red) and rEDIII-4 (violet) proteins as capture antigens. The black dashed curve indicates ELISA reactivity of control sera using a mixture of the four rEDIII proteins as capture antigen. All ELISA values in this experiment represent the average of two separate determinations.

To address the question if these antibodies would display this serotype preference toward the DEN viruses themselves, we carried out an immunofluorescence-based assay. BHK cells were separately infected with each of the four DEN virus serotypes and probed with the test antiserum (obtained from rAd-Bg primed/bivalent plasmid boosted animals) in conjunction with a secondary antibody-FITC conjugate. The prediction was that the test antiserum should light up cells infected with DEN-2 and DEN-4, but not DEN-1 and DEN-3 viruses. The results of this experiment which substantiate this are presented in Figure 4A. Panels in the first (a, e, i and m) and second (b, f, j and n) columns depict the negative and positive controls, respectively. DEN-1 (panel b), DEN-2 (panel f), DEN-3 (panel j) and DEN-4 (panel n) viruses were detected using murine polyclonal antibodies raised against E. coli-expressed rEDIII-1, rEDIII-2, rEDIII-3 and rEDIII-4 proteins, respectively. The detection of immunofluorescence in panels b, f, j and n demonstrates that DEN-1, 2, 3 and 4 viruses, respectively, had successfully infected the cells in this experiment. Mock-infected cells did not fluoresce with any of these antibodies (panels a, e, i and m), indicating that the antibodies used are virus-specific. The control serum (obtained from rAd-C primed/empty plasmid boosted animals) also resulted in no fluorescence in DEN virus-infected cells (panels c, g, k and o). This suggests that the vaccine vectors per se do not induce a DEN virus-specific immune response. Consistent with the ELISA data, the use of test serum as the source of primary antibodies in the assay resulted in the identification of DEN-2 (panel h) and DEN-4 (panel p), but not DEN-1 (panel d) and DEN-3 (panel l) viruses, as predicted. Taken together, the data strongly suggest that the bivalent EDIII-4/2 antigen elicits antibodies that specifically recognize DEN-2 and DEN-4 viruses with little evidence of cross-reactivity towards DEN-1 and DEN-3 viruses.

Figure 4. Simultaneous induction of antibodies that recognize and neutralize DEN virus serotypes 2 and 4. (A) IFA of DEN virus-infected cells using murine antisera. BHK cells were either mock infected (panels a, e, i and m) or infected separately with DEN-1 (panels b, c, d), DEN-2 (panels f, g, h), DEN-3 (panels j, k, l) and DEN-4 (panels n, o, p) viruses. One day after infection, cells were fixed and probed separately with different antisera as the source of primary antibodies. The antisera were obtained from mice, immunized with purified recombinant proteins [rEDIII-1 (panels a, b), rEDIII-2 (panels e, f), rEDIII-3 (panels i, j), and rEDIII-4 (panels m, n)] and rAd/plasmid [rAd-C/pVAX1 (panels c, g, k, o) or rAd-Bg/pVAX-EDIII-4/2 (panels d, h, l, p)]. The resultant virus/antibody complexes were visualized using anti-mouse IgG-FITC conjugate. (B) LLCMK2 monolayers were separately infected with each of the four DEN viruses (DEN-1: filled circles; DEN-2: filled squares; DEN-3: open circles; & DEN-4: open squares) that had been pre-incubated with serial two-fold dilutions of pooled immune serum drawn from the rAd-Bg primed/plasmid pVAX-EDIII-4/2 boosted animals. The resultant plaque counts, expressed as percent inhibition of virus infectivity (with reference to the number of plaques generated in the absence of antiserum, which was taken to represent 100% infectivity) are plotted as a function of antiserum dilution. Each data point shown represents the mean of four replicate assays (the error bars represent SD).

While the data obtained so far indicated that the EDIII-4/2 antigen induces a bivalent response targeting DEN-2 and DEN-4 viruses, it does not provide any information regarding the capacity of these antibodies to neutralize their infectivity. To this end, we performed plaque reduction neutralization tests (PRNTs), by plaqueing the antibody-incubated DEN viruses on LLCMK2 cells. Figure 4B presents the percent reduction in virus infectivity as a function of test serum dilution for each of the four DEN virus serotypes. The data show that the test serum displayed PRNT50 titers of ~1:80 towards both DEN-2 and DEN-4 viruses, with no significant neutralizing antibody titers towards DEN-1 and DEN-3 viruses.

Induction of DEN-2 and DEN-4 virus-specific T cell responses

Spleen cells in culture, derived from control and test mice, were stimulated in vitro with DEN-2 and DEN-4 viruses, separately, and monitored for proliferation responses and cytokine (IFN-γ and IL-4) production, as depicted in Figure 5. Splenocytes from test mice manifested significant and similar levels of proliferation in response to in vitro stimulation with both DEN-2 and DEN-4 viruses, compared to mock-stimulated splenocytes (Figure 5A). Control splenocytes did not manifest any proliferative response regardless of whether they were stimulated or not. The proliferative response of the test splenocytes was accompanied by IFN-γ secretion (Figure 5B). The elevated levels of IFN-γ, first evident at 48 hours post stimulation, persisted for the duration of the experiment. Again, as seen in the thymidine uptake experiment, both DEN-2 and DEN-4 virus stimulation resulted in comparable levels of IFN-γ production. At all time points tested, mock-stimulated test cells did not secrete discernible levels of IFN-γ. When control mouse spleen cell cultures were tested, IFN-γ was barely detectable in mock- and virus-stimulated culture supernatants. With regard to IL-4, the splenocytes manifested an increase in the levels of IL-4 secretion in response to virus stimulation, but it was not dramatic (as seen for IFN-γ), as basal IL-4 levels in control splenocytes were somewhat elevated. The magnitude of stimulation of IL-4 secretion in test splenocytes with respect to mock-stimulated cells, which was about 2 fold higher at 24 hours, increased to about 5 fold at 96 hours (Figure 5C). Taken together, these experiments demonstrate that the bivalent antigen, elicits DEN-2 and DEN-4 virus-specific T-cell responses. Interestingly, the magnitude of these responses was more or less equivalent for the two serotypes.

Figure 5. Analysis of T cell responses elicited by the bivalent antigen. Splenocytes were obtained from mock-immunized (open bars) and test-immunized (gray, black and hatched bars) animals, 7 weeks after the final immunization and placed in culture. They were either mock-stimulated (open and gray bars) or stimulated in vitro, either with DEN-2 (black bars) or DEN-4 (hatched bars) for 96 hours for performing the T cell assays. Tritiated [3H] thymidine uptake was determined in a scintillation counter (A). Aliquots of the culture supernatant withdrawn at the indicated time points were assayed for the presence of IFN-γ (B) and IL-4 (C) by solid phase ELISA. Data depicted represent the mean value of three separate determinations (the error bars represent SD). In panel A, two dosage levels of virus were used for in vitro stimulation [L, low (0.015 PFU/cell); and H, high (0.03 PFU/cell); N, no antigen]. Data in panels B and C were generated using splenocytes stimulated with the high dose.

rating: 0.00 from 0 votes | updated on: 20 Jul 2009 | views: 9784 |

Rate article:

excellent! bad…