Introduction

Human rotavirus (HRV) is one of the most important infectious agents worldwide, causing severe dehydrating diarrhea in infants and young children [1] with roughly 500,000 deaths annually, mostly in developing countries [2, 3]. In 1998, RotaShield was the first tetravalent live-attenuated vaccine which, however, was withdrawn from use because of a number of intestinal intussusception cases in vaccinated infants [4]. Since 2006, two live-attenuated HRV vaccines (RotaTeq and Rotarix) have been approved. Despite rigorous safety tests and good protection correlates, live vaccines still have potential risks including reversion of virulence [5].

Protection against rotavirus infection still lacks a clear understanding [4, 6–8]. Components of the immune system including neutralizing antibodies (NAbs), non-neutralizing antibodies (Abs), secretory Abs in the intestine and cell-mediated immune (CMI) response have all been proposed as correlates of protection [4, 7]. However, total serum rotavirus-specific IgAs and NAbs are currently considered as the best markers of protection against rotavirus [4, 8].

The two major viral surface structural proteins termed VP4 and VP7 are involved in HRV cell binding and entry [9, 10]. These proteins are the main targets of the host immune system for the induction of NAbs [4], and therefore have been selected by many investigators for the development of subunit vaccines [11–13]. However, unlike live vaccines, subunit vaccines are often inefficient at inducing optimal immune response [5]. As a result, the use of adjuvants [14] and/or of vectors like viral vectors [15] to induce optimal immune response is required.

Adenovirus vectors investigated for years as part of gene therapy applications are among the most exploited viral vectors for vaccine purposes [15, 16]. In 2008, the failure of Merck’s STEP clinical trial using replicating-defective adenovirus vector for human immunodeficiency virus (HIV) vaccine has revealed some weaknesses of such a system [17]. Therefore, many efforts were made to improve the efficacy of these vectors [18–20]. One of them relies on a strategy to enhance the immunogenicity of replication-defective adenovector -based HIV vaccine by complementing for expression of the adenovirus E1 gene through DNA inoculation allowing virus replication in vivo [20]. However, the complementation strategy with the simultaneous use of two different vectors is more likely limiting for vaccine administration and efficiency.

A replicating but non disseminating adenovector (AdV) has been described [21]. This vector is based on a replication-competent platform consisting of a human adenovirus type 5 with a deletion in the protease (PS) gene [22], the product of which is necessary for viral assembly [23]. In the absence of the PS gene, the AdV replicates its DNA normally in infected cells but fails to form infectious particles, thus preventing its dissemination in the host environment [22, 23]. Bourbeau et al. (2007) demonstrated that such AdVs not only increased transgene expression in infected cells when compared to non-replicating adenovectors, but were also as efficient as a replicating disseminating adenovector system in cancer therapy [21].

In this study, the AdVs were evaluated as a proof-of-concept vaccine vector using HRV VP4 and VP7 proteins as immunogens. Since the assembly of rotaviral epitopes has been shown to improve immune responses [24, 25], we also sought to fuse a truncated form of VP4 (VP4∆) with VP7 and evaluate the immunogenicity of the resulting fusion protein delivered by AdVs.

Materials and Methods

Cell Cultures

The human alveolar adenocarcinoma A549, human embryonic kidney 293A (HEK 293A), and simian epithelial kidney MA104 cell lines were maintained in Dulbecco minimal essential medium with 8 % fetal bovine serum (FBS) at 37 °C in a humidified atmosphere of 5 % CO2. The 293-PS-CymR cell line, a clone derived from the HEK 293A cells that was developed by Bernard Massie, was propagated as previously described [22, 26].

Generation of Gene Constructs for Expression in Mammalian Cells

VP7- and VP4-encoding genes from the HRV Wa strain (GenBank accession no. AAA4734 and AAA66953, respectively) were optimized for codon usage in mammalian cells and synthesized through GeneArt services (Regensburg, Germany). VP7- [326 amino acids (aa)], VP4- (776 aa), VP4Δ- (N-terminal 336 aa of VP4), and VP4Δ::VP7-encoding sequences were amplified by PCR and cloned into pcDNA3.0 (Invitrogen, Carlbad, CA) as described elsewhere [27]. HEK 293A cells (plated in wells of six-well plates at a density of 5 × 105 cells per well) were transfected with PolyFect transfection reagent (Qiagen, Mississauga, ON, Canada) with each plasmid construct and incubated for 48 h. Total cell protein concentrations were quantified with a detergent-compatible protein assay (Bio-Rad, Mississauga, ON, Canada).

Construction of AdVs

AdVs were constructed using the AdenoVator system (MP Biomedicals, Irvine, CA). GFP-, VP7-, VP4-, VP4Δ-, and VP4Δ::VP7-encoding sequences were first subcloned into the BglII restriction site of the modified shuttle vector pAdenoVator-CMV5(CuO)-IRES-E1A [21] in which the early domain 1A (E1A) sequence is inserted downstream of an internal ribosome entry site (IRES) to be co-expressed with the transgene. Expression of adenovirus E1A enables viral DNA replication, thereby increasing viral genome copy number in infected cells [23]. Each construct was confirmed by DNA sequencing. The various recombinant AdVs (AdVΔPS-CuO-transgene-IRES-E1A) were generated using the AdEasy system (MP biomedicals) with the AdVΔPS backbone according to the manufacturer’s protocol. Recombinant AdVs were propagated and produced at high titers in 293-PS-CymR cells. They were then purified by double cesium chloride (CsCl) gradient as described [21]. The recombinant AdV titers were determined and calculated as the median tissue culture infective dose (TCID50) per milliliter.

AdV-delivered protein expression was monitored in A549 cells that were plated in wells of six-well plates at a density of 5 × 105 cells per well and transduced with each recombinant AdV at an MOI of 10. After 48 h of incubation, cell extracts were prepared and total cell protein concentrations were determined as described above.

Recombinant Protein Production

Recombinant VP4Δ and VP7Δ (amino acids 84–332 of VP7) proteins were produced in E. coli and purified as previously described [28]. Briefly, genes encoding VP4Δ and VP7Δ proteins were cloned into pTrcHisB expression vector (Invitrogen), and proteins were purified on Ni–NTA-His-Bind® resin (Novagen, Madison, WI, USA) under denaturing conditions according to the manufacturer’s manual. Proteins were dialyzed against a phosphate-buffered saline solution (PBS, pH 7.3) and qualitatively evaluated by SDS-PAGE followed by immunoblotting using Horseradish peroxidase (HRP)-coupled goat anti-His Abs (Qiagen). The identity of the proteins was confirmed by MALDI time-of-flight tandem mass spectrometry (McGill University and Genome Quebec Innovation Center), using the MASCOT software (Matrix Science Inc., Boston, USA). Purified His-tagged VP4Δ and VP7Δ proteins were quantified as above and stored at −80 °C for further use as antigens in the immunological tests.

Western Blot Assay

For each sample, 50 μg of total cell extract prepared as described [29] was fractionated on 12 % SDS–polyacrylamide gels, electroblotted onto a nitrocellulose membrane, and probed with a 1:10,000 dilution of either rabbit VP7- [28], mouse monoclonal VP8*- (for VP4 and VP4Δ detection) kindly provided by Dr. Harry B. Greenberg (Stanford University) or AdV-E1A gene product-specific Abs (Millipore, Billerica, MA). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) immunostaining (Sigma–Aldrich, St. Louis, MO, USA) was used as a loading control. Detection of the proteins was performed as previously described [28].

Rotavirus Wa Strain Propagation and Purification

Propagation and titration of the human rotavirus (HRV) Wa strain (kindly provided by Dr. Dongwan Yoo from the University of Illinois) were performed in MA104 cells as described [30]. The virus was purified using CsCl gradients [30] and titered using a Focus Fluorescent Unit assay (FFU) [31]. Virus concentration (μg/ml) was determined by multiplying OD260 × dilution factor × 185 μg/ml [30].

Immunization of Mice

Animal protocols were approved by the University’s Animal Protection Institutional Committee according to the regulations of the Canadian Council for Animal Care. Six-week-old female BALB/c mice were purchased from Charles River (St-Constant, QC, Canada). Mice were inoculated intramuscularly in both legs of each mouse with a total volume of 50 μl per leg with either AdVs or plasmid DNA. Blood samples and spleens were collected as previously described [32].

Preliminary Experiment

In a preliminary experiment, mice of five groups (five mice per group) were immunized on day 0 and on days 14 and 35 post primary immunization (PPI). At each time point, mice of a control group received PBS and mice of the second group received AdV expressing GFP (1 × 108 TCID50 per dose). Mice of the third and fourth groups were inoculated with AdV expressing VP4Δ (AdV-VP4Δ) or plasmid DNA (100 μg per dose) also expressing VP4Δ (pcDNA3.0-VP4Δ), respectively. Mice of the fifth group received AdV-VP4Δ on day 0 and pcDNA3.0-VP4Δ on days 14 and 35 PPI. Animals were euthanized on day 66 PPI.

Experiment 1

Mice of five groups (eight mice per group) were inoculated on day 0 with 1 × 108 TCID50 of AdV and boosted on day 14 with 100 μg of plasmid DNA expressing the same protein as AdV. Mice received the following immunogens: GFP, VP7, VP4, VP4Δ, or VP4Δ::VP7. An additional plasmid DNA-boost was realized 5 days before euthanasia of animals on day 40 PPI. Preliminary data showed that this final boost increases the in vitro spleen lymphocyte proliferation to HRV (not shown).

Experiment 2

Four groups of six mice were used. Mice of the first two groups were immunized on day 0 with 2 × 108 TCID50 AdV expressing either GFP (group 1) or VP4Δ::VP7 (group 2) and boosted on days 14 and 35 PPI with 200 μg of plasmid DNA expressing the same proteins. Mice were euthanized on day 40 PPI. Mice of the last two groups received a mix of vectors expressing VP4Δ and VP7 (group 3), or VP4 and VP7, respectively (group 4). Mice of these two groups, namely VP4Δ/VP7 and VP4/VP7, were injected on day 0 with 1 × 108 TCID50 AdV-VP4Δ and 1 × 108 TCID50 AdV-VP7, or with 1 × 108 TCID50 AdV-VP4 and 1 × 108 TCID50 AdV-VP7, respectively. On days 14 and 35 PPI, mice received the appropriate mix of 100 μg of each recombinant plasmid DNA expressing the same proteins as for AdVs for a total of 200 μg of plasmids.

Indirect Immunofluorescence Assay (IFA)

MA104 cells were plated into wells (1.0 × 104 cells per well) of a 96-well plate and infected the next day with HRV (200 FFU per well). Cells were fixed in PBS-4 % formaldehyde for 15 min and blocked for 1 h at 37 °C in PBS-5 % bovine serum albumin (BSA). The assay was performed using mouse sera as primary Abs and Cy3-coupled anti-mouse IgG secondary Abs (1:1,000), each incubated for 2 h at 37 °C. IFA titers were expressed as the log2 of the reciprocal of the highest serum dilution producing a positive fluorescent signal.

Indirect ELISA

96-well Costar ELISA plates (Fisher Scientific) were coated overnight at 4 °C with 1 μg recombinant VP4Δ or VP7Δ produced in bacteria [28] or 100 ng CsCl-purified, UV-inactivated [33] HRV per well diluted in 100 mM carbonate–bicarbonate buffer (pH 9.6) to a final volume of 100 μl. Following a blocking treatment with BSA, Ab titration was performed using 100 μl of twofold serial serum dilutions and an incubation time of 2 h at 37 °C. HRP-conjugated goat anti-mouse IgG1 (1:10,000), IgG2a (1:5,000), or IgA (1:5,000) (Santa Cruz Biotechnologies) were added for 1 h at 37 °C. HRP signal was detected as described [32]. ELISA titers of serum samples from each individual mouse were expressed as the log2 of the last dilution of sample giving a mean OD higher than twofold the mean OD obtained from negative control mouse serum.

Rotavirus Neutralization Assay

The virus neutralization test was performed according to the FFU reduction assay [25, 31] with modifications. Briefly, serially diluted heat-inactivated sera were mixed with the virus (250 FFU) in DMEM supplemented with porcine pancreatic trypsin (5 μg/ml) and incubated at 37 °C for 1 h. MA104 cells in 96-well tissue culture plates were infected during 2 h with each of the serum-virus mixtures (two wells per mixture). Cells were then incubated in DMEM supplemented with porcine pancreatic trypsin (5 μg/ml) and were fixed in PBS-formaldehyde 4 % after 18 h. Residual infectious virus was detected by IFA using VP4Δ- and VP7-specific antisera [28] (1:500) and Alexa green 488 coupled anti-rabbit secondary IgGs (1:1,000) (Invitrogen). The presence of virus-infected cells was detected using a TE-300 inverse microscope (Nikon, Mississauga, Canada) coupled to the confocal MRC-1024ES system (Bio-Rad). NAb titers were expressed as the log2 of the reciprocal of the highest serum dilution producing a 50 % reduction in FFU. The virus neutralization assays were repeated twice.

Spleen Lymphocyte Proliferation Assay

Spleen white cells from each mouse were prepared as previously described [32]. Quadruplicate cell cultures were stimulated for 3 days with CsCl-purified, UV-inactivated HRV (200 ng/ml). Cells were labeled with radioactive H3-thymidine incorporated in cell DNA for 24 h prior to cell harvesting [32]. The lymphocyte proliferation results were expressed as stimulation index (SI) which is the ratio of mean counts per minute (CPM) of cells stimulated with the antigen divided by the mean CPM of cells without antigen.

Statistical Analysis

One-way analysis of variance (ANOVA) followed by Tukey’s post-test was carried out for statistical analyses between multiple groups using GraphPad Prism software (Windows Version 5.0, Graphpad Software, Lajolla, CA, USA). Significance was set at P < 0.05.

Results

Expression of VP7, VP4, VP4Δ, and VP4Δ::VP7 Constructs in Mammalian Cells

Sequences encoding GFP, VP7, VP4, VP4Δ, and VP4Δ::VP7 proteins were inserted each into AdVs or pcDNA3.0 plasmids. VP4Δ is a truncated version of VP4 that contains the major neutralizing epitopes of VP4 and the sequence motif for virus attachment to target cells [10, 34, 35].

Expression of VP7 (38 kDa), VP4 (86 kDa), VP4Δ (38 kDa), and VP4Δ::VP7 (76 kDa) was evaluated by Western blot either in AdV-infected A549 cells (Fig. 1a) or in pcDNA3.0-transfected HEK 293A cells (Fig. 1b). Molecular weights of the expressed proteins were those predicted for the corresponding genes. However, additional VP4Δ and VP7 bands were observed in cells treated with the VP4Δ::VP7-expressing AdV and DNA plasmid vectors (Fig. 1a and b). These bands were attributed to spontaneous cleavage of a fraction of VP4Δ::VP7. As expected, 38 and 40 kDa E1A proteins that allow adenoviral DNA replication were detected in the cellular extracts of AdV-infected cells (Fig. 1a).

Fig. 1
figure 1

Expression in mammalian cells of VP7, VP4, VP4Δ, and VP4Δ::VP7 constructs through replicating but non disseminating adenovector (AdV) or plasmid DNA delivery systems. Total cell proteins (50 μg) were fractionated on 12 % SDS–polyacrylamide gels, electroblotted onto a nitrocellulose membrane, and probed with VP7-, VP8*-, or AdV-E1A gene products–specific antibodies. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) immunostaining was used as a loading control. The molecular masses of the expressed proteins are indicated in the central margin. a Expression of VP7, VP4Δ, or AdV E1A proteins was detected by Western blot using total cell protein extracts of A549 cells infected with the various AdVs. Mock-infected or A549 cells infected with an AdV-GFP construct were used as negative controls. b Expression of VP7 and VP4Δ was detected by Western blot using total cell protein extracts of HEK 293A cells transfected with the various pcDNA3.0 constructs. Non-transfected (NT) HEK 293A cells and cells transfected with pcDNA3.0-GFP were used as negative controls

Full size image

Immunogenicity of HRV VP4Δ Delivered by AdVs

In a preliminary experiment, the single immunogen VP4Δ was selected to be delivered with AdVs inoculated in mice on days 0, days 14 and 35 post primary immunization (PPI) to ensure the validity of the AdV system (Fig. 2). This procedure was also compared with pcDNA3.0-VP4Δ inoculations. The kinetics of serum IgG responses to HRV, as determined by IFA, is shown in Fig. 2. Although HRV-specific Abs were detected in both groups of mice, the highest Ab levels were observed in the AdV-VP4Δ-inoculated group. Abs were detected as early as 14 days PPI in this group and their level was stable for over 1 month after the last immunization at day 35 PPI. In contrast, Abs were first detected on day 35 PPI in the plasmid DNA-inoculated group and their level slightly decreased 1 month after the last immunization. No HRV-specific Abs were detected in mice of the PBS- or AdV-GFP-inoculated negative control groups.

Fig. 2
figure 2

Preliminary experiment design and kinetics of anti-HRV IgG antibody production as determined by indirect immunofluorescence assay (IFA). Mice of four groups (five mice per group) received three intramuscular injections on day 0 and on days 14 and 35 post primary immunization (PPI) of (i) PBS, (ii) AdV-GFP, (iii) AdV-VP4Δ, and (iv) pcDNA3.0-VP4Δ. Mice of a fifth group, namely AdV/pcDNA3.0-VP4Δ, received AdV-VP4Δ on day 0 and pcDNA3.0-VP4Δ on days 14 and 35 PPI. Sera were collected on day 0 and on days 14, 35, 56, and 66 PPI. IFA titers of pooled sera from each group of mice were expressed as log2 of the reciprocal of the highest serum dilution giving positive fluorescent signal. Titers are representative of three independent ELISAs.

Full size image

One inconvenience of viral vector-based vaccines is the immunogenicity of the vector itself, which may limit their use in repeated immunizations. In order to circumvent possible AdV neutralization, prime/boost immunization regimens with AdV and plasmid DNA, respectively, were developed [18, 20]. Mice of a fifth group, namely the AdV/pcDNA3.0-VP4Δ group, were then immunized with AdV-VP4Δ on day 0 and boosted with pcDNA3.0-VP4Δ on days 14 and 35 PPI (Fig. 2). This prime/boost regimen resulted in VP4Δ-specific Ab levels that were about 4 times higher than those obtained from mice immunized with AdV-VP4Δ alone from day 35 PPI.

VP7, VP4, VP4Δ or VP4Δ::VP7 Immunogens Induce HRV-Specific Ab Responses

The results of the preliminary experiment prompted us to compare the immunogenicity of the VP7, VP4, VP4Δ, and VP4Δ::VP7 proteins administered with the AdV-prime/plasmid DNA-boost regimen. Mice were immunized with AdV on day 0 and boosted with plasmid DNA on days 14 and 35 PPI. On day 40 PPI, the IgG response specific to HRV was evaluated by IFA in individual mouse sera (Fig. 3A). While each group of mice had significant levels of HRV-specific IgGs, the highest Ab levels were observed in the VP4Δ::VP7 group of mice (Fig. 3a). Ab titers in VP4 and VP4Δ groups of mice were similar and were higher than those of the VP7 group. An indirect ELISA was also performed using either purified HRV (Fig. 3b), VP4Δ (Fig. 3c), or VP7Δ (Fig. 3d) recombinant proteins as antigen. While the highest Ab levels were again observed in the VP4Δ::VP7 group, no significant difference was noted between the VP4, VP4Δ, and VP7 groups. This was attributed to the technique itself where mature viral particles were used as antigen such that it may not have detected the IgG response directed to some conformational epitopes exposed on unfolded VP4 proteins present in the infected cells used in IFA.

Fig. 3
figure 3

Antibody (Ab) response in mice immunized intramuscularly with vectors expressing various HRV immunogens. In experiment 1, mice were inoculated with AdVs on day 0 and boosted on days 14 and 35 post primary immunization (PPI) with 100 μg of plasmids expressing the same proteins as for AdVs. The groups of mice (eight mice per group) were named according to the proteins the animals received: GFP, VP7, VP4, VP4Δ, and VP4Δ::VP7. In experiment 2, mice of the first two groups (six mice per group), namely GFP and VP4Δ::VP7, were inoculated with 2 × 108 TCID50 AdVs on day 0 and boosted on days 14 and 35 PPI with 200 μg of recombinant plasmid expressing the same proteins as for AdVs. Mice from two other groups, namely VP4Δ/VP7 and VP4/VP7, were inoculated on day 0 with a mix of AdV-VP4Δ and AdV-VP7 or with a mix of AdV-VP4 and AdV-VP7, respectively. Mice were then boosted on days 14 and 35 PPI with a mix of plasmids expressing the same proteins as for AdVs. In both experiments, serum was collected from each mouse on day 40 PPI (prior to the sacrifice of animals) for HRV-specific IgG response determination either by IFA (a) or ELISA (b). Levels of serum IgGs specific to VP4Δ (c) or VP7Δ (d) were also evaluated by ELISA. For each group of mice, columns represent the mean value of IgG titers (±standard deviation). Labeling of data with superscripts of different letters indicates significant differences at P < 0.05

Full size image

A second set of experiments was conducted to determine whether VP4Δ and VP7 need to be delivered as a VP4Δ::VP7 fusion protein to obtain optimal Ab response. Mice were immunized on day 0 with AdVs expressing either GPF or VP4Δ::VP7 fusion protein and boosted with the plasmid DNA vectors on days 14 and 35 PPI. Two other groups of mice, namely VP4Δ/VP7 and VP4/VP7, were immunized with a mixture of AdV and subsequently plasmid DNA vectors encoding the proteins separately. Results showed that the IgG responses obtained on day 40 PPI were higher in the VP4Δ::VP7 group than both the VP4Δ/VP7 and VP4/VP7 groups regardless of the antigen used in the ELISA (Fig. 3b–d). The immune response was further analyzed in terms of Ab isotypes and IgG subclasses on sera collected from individual mice at day 40 PPI (Fig. 4a). Significant levels of serum HRV-specific IgG1 (typical of a Th2 immune response), IgG2a (typical of a Th1 immune response), and IgA were detected by ELISA in each group of animals immunized with the HRV immunogens in both experiments 1 and 2, the highest levels being obtained in mice of the VP4Δ::VP7 group. This result also correlated with the highest levels of HRV-specific NAbs observed in mice of the VP4Δ::VP7 group (Fig. 4b). Taken together, the results show that VP4Δ::VP7 fusion protein represents an immunogen of choice for immunization against HRV.

Fig. 4
figure 4

Characterization of antibody (Ab) isotypes and neutralizing Ab response specific to HRV on day 40 post primary immunization from the same groups of mice described in the legend of Fig. 3. a IgA, IgG1, or IgG2a specific to HRV was detected by ELISA using purified HRV particles as antigen. Columns represent the mean value of Ab titers (±standard deviation). b Mean (±standard deviation) neutralization Ab titers specific to HRV. Labeling of data with superscripts of different letters indicates significant differences at P < 0.05

Full size image

CMI Response to HRV

The CMI response to HRV was evaluated in individual mice at day 40 PPI using a rotavirus-specific lymphocyte proliferation test. As shown in Fig. 5, the VP4 and VP4Δ::VP7 groups of mice from experiment 1 showed positive lymphoproliferative response with similar mean SI values of 1.93 (±0.30) and 2.10 (±0.42), respectively. The control GFP and VP7 and VP4Δ groups of mice did not show significant HRV-specific lymphocyte activity.

Fig. 5
figure 5

HRV-specific lymphoproliferation response from spleen cells of mice on day 40 post primary immunization. Columns represent the mean stimulation index (SI) values (±standard deviation) obtained in mice of the same groups described in the legend of Fig. 3. Labeling of data with superscripts of different letters indicates significant differences at P < 0.05

Full size image

The VP4Δ::VP7 and VP4/VP7 groups of animals from experiment 2 also showed similar positive lymphocyte response to HRV stimulation. This response was higher than that observed in the VP4Δ/VP7 group.

Discussion

In this study, replicating but non disseminating adenovectors were generated and found to be suitable expression vectors for HRV immunogen delivery in vivo. The effectiveness of replicating but disseminating adenovectors has been shown in vaccine studies [15, 36]. These vectors provide elements of innate immunity and induce immunogen-specific Ab and CMI responses [15, 19]. Another advantage is that AdV genome replication, enabled by the adenovirus E1A genes, increases transgene expression as compared to non-replicating AdVs [21], a characteristic that is important in vaccinology. Expression of E1A-encoding sequence also offers the advantage of activating the innate immune response through Toll-like receptor (TLR)-dependent and independent pathways, resulting in enhanced type I interferon (IFN) production and, presumably, enhanced vaccine potency [23, 37, 38]. However, safety is a major concern of using these AdVs for vaccination. Here, dissemination was not possible as the replication-competent platform was used with a deletion in the PS gene, preventing the formation of adenovirus infectious particles [22].

The mature triple-layered icosahedral rotavirus particle has two surface structural proteins termed VP4 and VP7 that determine the P and G genotypes, respectively [39]. Rotavirus structural VP4 and VP7 proteins are suitable immunogens for vaccine development because they carry the major epitopes eliciting the production of NAbs in the immunized host [11, 40–42]. It is noteworthy that approximately 90 and 65 % of HRV strains circulating worldwide share cross-reactive neutralizing epitopes on VP4 and VP7, respectively. In addition, the G1P[8] Wa strain was selected in our study because it was shown that the monovalent Rotarix vaccine which is of the G1P[8] genotype is protective against homotypic, partially heterotypic, and heterotypic HRV strains [43, 44]. Nevertheless, as new rotavirus strains may emerge, incorporation of VP4- and VP7-encoding sequences from other viral strains in replicating but non disseminating AdVs may need to be considered for HRV vaccine development. Finally, another way to improve HRV vaccines might be the incorporation of the VP6 protein because of its high degree of conservation among various HRV strains and its ability to induce protection in animal models [4, 7].

Systemic DNA vaccination targeting VP4 and VP7 proteins was shown to correlate with significant production of protective Abs in mice following rotavirus challenge [40, 45] although failure of this approach to allow protection was also reported [11]. Non-replicating AdVs expressing VP7 or VP4 were also used to immunize monkeys [41] and mice [42, 46], resulting in rotavirus-specific Ab response. Here, vaccination with either a plasmid DNA or an AdV expressing VP4Δ resulted in HRV-specific Ab responses. Moreover, the AdV-prime/plasmid DNA-boost regimen resulted in VP4Δ-specific Ab titers that were greater than those obtained from mice immunized with the AdV or plasmid DNA alone (Fig. 2).

It is known that non-replicating AdVs are capable of either priming or boosting the immune response to an immunogen when used in heterologous immunization regimens [47]. Here we demonstrated that replicating but non disseminating AdVs are suitable vectors for priming the Ab response which is currently considered as the best marker of protection against rotavirus [4, 8]. The plasmid DNA-prime/vector boost also is a common regimen in vaccinology to achieve protection predominantly mediated by a CD8 + T cell response. Such an immunization regimen was indeed effective against hepatitis C virus, HIV, and malaria [18, 47, 48]. Since the CMI response has been proposed as a correlate of protection against HRV infection [4, 7], the plasmid DNA-prime/AdV boost immunization regimen might need to be considered for future HRV vaccine development.

By using an AdV-prime/plasmid DNA-boost regimen, we were able to investigate further various HRV immunogens, namely VP7, VP4, VP4Δ, and VP4Δ::VP7 in order to determine the best approach to induce optimal HRV-specific immune response. Each of the HRV proteins induced serum HRV-specific IgG and IgA responses capable of neutralizing HRV in vitro. Mixed Th1/Th2 responses were also elicited with all immunogens used as shown by the presence of anti-HRV IgG1 and IgG2a isotypes. This aspect of the immune response was not studied when immunizing mice with VP7- or VP4-expressing AdVs [41, 42]. Our results are thus in agreement with the Th1/Th2 response observed during HRV infection in a pig model [49] and with the fact that AdVs, like other viral vectors, are able to induce both Ab and CMI responses [15].

Fusion of VP4Δ and VP7 proteins resulted in the highest immune responses in terms of eliciting HRV-specific serum IgGs, IgAs, Nabs, and systemic CMI. Although measurement of serum HRV-specific NAbs and IgAs is the current standard for assessing protective immune response following rotavirus vaccination in humans, CMI is also believed to be at play in protection [4, 7]. Indeed, VP7 of various rotavirus strains was shown to induce CMI response through CD4+/CD8+ T cell activation [50–52].

This study is the first report using VP4Δ::VP7 fusion protein as immunogen to induce a HRV-specific immune response. Other fusion strategies in attempts to develop a rotavirus vaccine have been investigated. The assembly of several copies of a major rotavirus VP8* epitope was previously shown to improve the immune response [25]. Studies in mice also have shown that a fusion between VP8* (the N-terminal subunit of VP4) and a truncated version (first N-terminal 92 aa) of the VP2 protein induces high VP4-specific Ab titers [24]. Remarkably, the VP4Δ::VP7 fusion protein generated an Ab immune response that was higher than that obtained with VP4 or VP7 alone, or with a mix of both latter proteins, suggesting a synergetic effect for VP4Δ::VP7. Such an immune synergetic effect in using fusion proteins was also observed in the influenza virus [53] or Plasmodium falciparum [54] systems.

Induction of mucosal immunity in the intestine in terms of secretory IgA response is crucial in preventing or clearing rotavirus infection [4]. Nevertheless, high serum IgG and/or IgA levels have been associated with protection against HRV infections in some studies [4, 7]. Here we determined for the first time that replicating but non disseminating AdVs used either alone or in combination with plasmid DNA in a prime-boost strategy are suitable vectors for immunogen delivery. We also showed that both the Ab and CMI responses to rotavirus were induced when using these vectors. This study lays the foundation for developing subunit vaccines against rotavirus infection, targeting more particularly the VP4Δ::VP7 fusion protein, and to use the replicating but non disseminating adenovector system for mucosal administration to induce protective local immunity. Additional work is needed to determine the efficacy of this vaccine strategy in a rotavirus infection model.