Mucosal immunization against ovine lentivirus using PEI–DNA complexes and modified vaccinia Ankara encoding the gag and/or env genes

Vaccine 26 (2008) 4494–4505

Contents lists available at ScienceDirect

Vaccine journal homepage: www.elsevier.com/locate/vaccine

Mucosal immunization against ovine lentivirus using PEI–DNA complexes and modified vaccinia Ankara encoding the gag and/or env genes R. Reina a,1 , C. Barbezange b,1 , H. Niesalla c,1 , X. de Andrés a , H. Arnarson d , E. Biescas e , M. Mazzei f , C. Fraisier g , T.N. McNeilly c , C. Liu c , M. Perez e , M.L. Carrozza f , P. Bandecchi h , C. Solano a , H. Crespo a , I. Glaria a , C. Huard i , D.J. Shaw c , I. de Blas e , D. de Andrés a , F. Tolari f , S. Rosati h , M. Suzan-Monti g , V. Andrésdottir d , S. Torsteinsdottir d , G. Petursson d , L. Lujan e , M. Pepin i , B. Amorena a , B. Blacklaws b , G.D. Harkiss c,∗ a

CSIC-Public University of Navarra, Pamplona, Spain University of Cambridge, Cambridge, UK c University of Edinburgh, Edinburgh, UK d University of Iceland, Reykjavik, Iceland e University of Zaragoza, Zaragoza, Spain f University of Pisa, Pisa, Italy g Faculté de Médecine, Marseille, France h University of Turin, Italy i AFSSA, Sophia Antipolis, France b

a r t i c l e

i n f o

Article history: Received 14 June 2007 Received in revised form 4 June 2008 Accepted 13 June 2008 Available online 9 July 2008 Keywords: Sheep Lentivirus DNA vaccination Mucosal

a b s t r a c t Sheep were immunized against Visna/Maedi virus (VMV) gag and/or env genes via the nasopharynxassociated lymphoid tissue (NALT) and lung using polyethylenimine (PEI)–DNA complexes and modified vaccinia Ankara, and challenged with live virus via the lung. env immunization enhanced humoral responses prior to but not after VMV challenge. Systemic T cell proliferative and cytotoxic responses were generally low, with the responses following single gag gene immunization being significantly depressed after challenge. A transient reduction in provirus load in the blood early after challenge was observed following env immunization, whilst the gag gene either alone or in combination with env resulted in significantly elevated provirus loads in lung. However, despite this, a significant reduction in lesion score was observed in animals immunized with the single gag gene at post-mortem. Inclusion of IFN-␥ in the immunization mixture in general had no significant effects. The results thus showed that protective effects against VMV-induced lesions can be induced following respiratory immunization with the single gag gene, though this was accompanied by an increased pulmonary provirus load. © 2008 Elsevier Ltd. All rights reserved.

1. Introduction Visna/Maedi virus (VMV) and caprine arthritis encephalitis virus (CAEV), also known as small ruminant lentiviruses (SRLV), are monocyte/macrophage tropic lentiviruses that infect sheep and goats, causing a lifelong infection despite the immune response generated against them. SRLV are endemic in sheep and goat industries in European countries and cause significant economic losses

∗ Corresponding author at: Division of Veterinary Biomedical Sciences, Royal Dick School of Veterinary Studies, University of Edinburgh, Summerhall, Edinburgh EH9 1QH, UK. Tel.: +44 131 650 6177; fax: +44 131 650 6511. E-mail address: [email protected] (G.D. Harkiss). 1 These authors made an equal contribution to this work. 0264-410X/$ – see front matter © 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.vaccine.2008.06.065

due to premature culling, reduced milk yields, reduced numbers of offspring, reduced lamb/kid weights, and poor quality of meat [1]. The economic impact of SRLV infections tends to be hidden due to the slow insidious nature of the disease, and the true cost is usually only appreciated when a careful analysis of flock performance over a period of years is undertaken. Approaches based on diagnosis and culling strategies are currently being applied to control SRLV infection, but prophylactic methods such as immunization are still being developed. The immune response in VMV infections involves T cell responses and antibody production, including that of neutralising antibodies [2–4]. Cellular responses might help in controlling viremia and disease development, whilst humoral responses might prevent viral entry into permissive cells and control early viral burden. However, it is still not clear what arm of the immune

R. Reina et al. / Vaccine 26 (2008) 4494–4505

response might associate with protection from SRLV infection or disease. Immunization against lentiviruses has been used to prevent infection or disease and to study the immune response to other lentiviruses such as the immune deficiency viruses; FIV in cats, SIV in simians and HIV in humans [5–7]. Immunizations against CAEV reveal that the use of recombinant env plasmids for immunization leads to lower viral replication and load and diminished disease [8], whilst using gag gene, peptide or inactivated virus enhancement of disease or proviral load were observed [9–11]. Genes used in sheep for VMV immunization have involved those of a complete viral clone [12] and the env gene, the latter administered mucosally [13]. In both cases a significant protective effect (diminished viral load in blood) has been found in immunized animals after challenge. The cellular (proliferative and cytotoxic) responses were not investigated in the VMV immunization studies and the association between the immune response and protection against infection needs to be further assessed. DNA vaccination is one of the immunization strategies that could be applied against SRLV, with potential ability to induce both humoral and cell-mediated immune response [14]. Enhancement of immune responses by DNA vaccination relies on the plasmid itself, genes that encode specific viral proteins, delivery systems that may attract professional antigen presenting cells (APCs), and genes that encode cytokines, chemokines or T cell costimulatory molecules. For example, IFN-␥ has been shown to be effective in modulating the immune response towards a Th1 profile against lentiviruses such as HIV-1, enhancing the production of specific antibody (IgG2a), cytotoxic activity and T helper proliferation [15,16]. Some of the DNA vaccination strategies, such as the inclusion of IFN-␥ [17,13] or IL-12 [18] have been explored against SRLV. However, in goats, experiments based on env-immunization against CAEV have shown that IFN-␥ does not enhance Th1 cytokine responses and appears to inhibit antibody production against the encoded antigen [17]. Vehicles may also play an important role in DNA vaccination, since they can promote increased plasmid transfection efficiency and protect DNA from degradation. Particularly, the use of DNA–polycation complexes to improve DNA immunization appears to be of potential use. The cationic polymer, polyethylenimine (PEI 25 kDa) is an efficient vehicle of plasmid DNA that enhances transfection of cells in vitro and in vivo [19,20]. It provides a global positive charge that allows DNA cellular uptake, protects DNA from enzymatic degradation, and allows DNA to traverse the endosomal membrane, reaching the cytoplasm and the nucleus [21]. Organs such as lungs can be transfected efficiently and safely [22]. In this study, we aimed to investigate whether new approaches to vaccination may provide another avenue of immune modulation and control of SRLV infections. This study involved the use of plasmid PEI–DNA complexes, containing gag and/or env genes in the presence or absence of the gene encoding ovine IFN-␥. This approach allowed us to determine primarily whether the core or envelope genes provided the better targets for vaccination. Immunization was performed via the respiratory tract with recombinant plasmids for priming and first boosting, and a live attenuated viral vector for second boosting. The animals were then challenged with live VMV via the lung, a major target site for natural transmission. Cellular and humoral responses in blood were assessed and the degree of protection conferred by immunization was investigated. The results showed that env mucosal immunization responses decreased the viral load in blood transiently, whilst gag immunization protected against lesion development measured 12 weeks post-challenge despite being associated with elevated lung proviral loads.

4495

2. Materials and methods 2.1. Sheep for vaccination experiments One to 2-year-old Lleyn sheep (females and castrated males) were obtained from a VMV-free accredited flock. Prior to immunization, absence of VMV was confirmed by ELISA using a commercially available diagnostic kit (Elitest MVV/CAEV, Hyphen Biomed) following the manufacturer’s instructions and by realtime PCR as described below. The animal experiments were performed at three European centres according to national regulations and institutional guidelines. 2.2. Viruses and plasmids VMV strain EV1 [23] was grown and titred as described previously [24]. The same stock of virus was used to challenge sheep in all three centres. The same strain of VMV (kept at low passage, 10% above that from heterologous infected minus mock infected cells [30], in which the killing of mock infected and heterologous cells was less than 10%. 2.7.3. IFN-gamma expression measured by quantitative real-time RT-PCR Sheep PBMC’s in 10% RPMI were stimulated for 4 h at 37 ◦ C in 5% CO2 with 5 ␮g/ml Concanavalin A, with whole virus at a concentration of 1:400, or with recombinant GAG antigens at a concentration of 25 ␮g/ml. Unstimulated (medium), and mock EV1 stimulated cells were used as controls. Stimulated PBMCs were resuspended in RNAlater (ABgene) and stored at −80 ◦ C. RNA was isolated using a commercial kit (RNeasy Mini kit, Qiagen Ltd.) following the manufacturer’s instructions, treated with RNase-free DNase under the on-column conditions (RNasefree DNase set, Qiagen Ltd.), and eluted in 50 ␮l RNase-free H2 O. Single-strand cDNA was synthesized as follows: briefly, 9 ␮l of DNase-treated RNA was mixed with 1 ␮l oligo(dT) (500 ␮g/ml), incubated at 90–100 ◦ C for 2 min, transferred on ice and subjected to the reverse transcription reaction using 100 U of M-MLV reverse transcriptase (Invitrogen) and 20 U RNase OUT as RNase inhibitor (Invitrogen) in a 20 ␮l final volume, for 20 min at 50 ◦ C. Real-time PCR experiments were performed using the Light Cycler Fast Start DNA Master SYBR Green I (Roche). Two microliters of cDNA was amplified by real-time PCR using primers specific to ovine IFN-␥, and ovine GAPDH as the “housekeeping” gene. Primer sequences were the following: IFN-␥ forward 5 -TACACAAGCTCCTTCTTAGC3 ; reverse 5 -AGAAGGAGACAATTTGGTC-3 (Accession number: X52640); GAPDH forward 5 -GTTTGTGATGGGCGTGAAC-3 ; reverse 5 -GTCTTCTGGGTGGCAGTGAT-3 . The PCR cycles were as follows: denaturation 95 ◦ C 10 min; amplification 95 ◦ C 8 s, 60 ◦ C 7 s, 72 ◦ C 20 s for 40 cycles. Melting curves were performed by cooling to 62 ◦ C 15 s, then heating to 95 ◦ C by 0.1 ◦ C/s continuously and finally cooling 40 ◦ C 20 s. Results of the real-time PCR data were taken as Ct values where the Ct is defined as the threshold cycle of the PCR at which amplified product was first detected. Analysis was performed using the mathematical model described by Pfaffl [31]. Relative quantification of IFN-␥ target gene transcript was made in comparison to GAPDH reference gene transcript using the equation: ratio =

(Etarget )Ct target(control−sample) (Eref )Ct ref(control−sample)

where Etarget and Eref are the real-time PCR efficiencies of IFN-␥ and GAPDH gene transcripts, respectively, calculated as E = 10[−1/slope]; the slopes for IFN-␥ and GAPDH assays were calculated from standard curves where Ct values were plotted against serial dilutions of RNA from ConA stimulated lymphocytes. Mean E values from three different sheep lymphocytes were Etarget = 1.924

4498

R. Reina et al. / Vaccine 26 (2008) 4494–4505

and Eref = 2.628. Ct is the difference in Ct between unstimulated (mock antigen) − stimulated (antigen) samples of the IFN-␥ and GAPDH gene transcripts. Due to technical difficulties samples from sheep immunized with gag-env or gag-env-ifn that were stimulated with whole virus were not analyzed. 2.8. Measurement of proviral load 2.8.1. Real-time PCR assays Provirus load was measured using real-time PCR. Assays were developed to detect, respectively, gag and pN3-gag sequences. Primers and probes were selected using Beacon Designer (Premier Biosoft International). 2.8.1.1. Gag assay. Amplicon length was 106 bp. Primers and probe sequences were as follows: sense primer: 5 -TCAACAGGCATCACAGGCTAATA-3 (nt. 1249–1271); antisense primer: 5 GTTACCTGGCCTATGCGACAT-3 (nt. 1334–1353); dual-labelled antisense probe (Operon): 5 -(6-FAM) ACCGCTCTCAAGGCTGTTATGACCCA(BHQ-6-FAM)-3 (nt. 1301–1325). Nucleotide positions refer to the published EV1 strain sequence (S51392) [22]. Choice of amplicon and set-up of the procedure will be described elsewhere. Serial dilutions of plasmid (pDRIVE, Qiagen Ltd.) carrying a 567 bp EV1 gag fragment (nt. 963–1529) [22] spanning the real-time amplicon, were used in each assay to generate the standard curve. Plasmid copy numbers ranged from 1 × 106 to 3 per reaction. Reaction volume was 25 ␮l (Quantitect Probe PCR kit, Qiagen Ltd.), programme was: 15 min at 95 ◦ C to allow for Taq polymerase activation, then 15 s at 94 ◦ C and 1 min at 58 ◦ C for 50 cycles. 2.8.1.2. pN3-gag assay. pN3-gag amplicon length was 130 bp. Primers and probe sequences were as follows: sense primer: 5 TGAACCGTCAGATCCGCTAG-3 (nt. 576–595 in U57609), located within the CMV promoter/MCS of pN3 plasmid sequence; antisense primer: 5 -AGCTCGGGGTATCCCTTTT-3 (nt. 531–549 of S51392) located within gag; dual-labelled antisense probe (Operon): 5 (6-FAM) CCTTGAGCCTTGCTTCGCCATACGAC (BHQ-6-FAM)-3 (nt. 506–526 of S51392) spanning the ATG of p17 gag with 5 bases derived from cloning the gag fragment at the 3 end. Assays were performed as in Section 2.8.1.1. 2.8.2. Proviral load in blood Eight milliliter EDTA blood were collected at each time point. Following erythrocyte lysis with ammonium chloride, the leukocytes were washed 2 times in cold PBS and genomic DNA extracted with DNeasy tissue kit (Qiagen Ltd.) according to the manufacturer’s instructions. DNA was quantified by fluorimetry with PicoGreen dsDNA Quantitation Kit (Invitrogen). DNA concentrations varied between 40 and 500 ng/␮l. Five microliters of freshly prepared DNA were analyzed with real-time PCR assays specific for EV1 gag p25 coding sequences and for pol sequences. DNA samples were tested in triplicate; two assays were performed for each DNA. Negative and positive controls were included in each assay. Reactions were carried out in an I Cycler (Bio-Rad) thermal cycler with the programmes described above. The coefficient of variation (COV) of the proviral copy number per reaction of the six replicates for each DNA was calculated. A threshold value of 0.26 was established for DNAs having more than 10 copies per reaction and outliers were excluded from further analysis. No COV analysis was applied to DNA samples carrying less than 10 copies per reaction. Mean copy number per reaction was converted to mean copy number per microgram of template DNA for each animal.

2.8.3. Proviral load in tissues At autopsy 100 mg pieces of tissue from the same mediastinal lymph node and lung samples as for the pathological evaluation (see below) were taken and stored at −80 ◦ C until they were processed. Tissue samples were minced with a scalpel and homogenized with a pestle in ATL buffer (DNeasy tissue kit, Qiagen Ltd.). Following overnight incubation with proteinase K at 56 ◦ C, DNA was extracted according to the manufacturer’s instructions. DNA was quantified as above and the gag copy number measured for each lung piece and lymph node with the gag and the pN3-gag assays, as described above. The lung tissue sample with the maximum proviral load for each animal was used for statistical analysis. Representative lung samples from the gag, gag-ifn, gag-env, and gag-env-ifn groups were assayed simultaneously using the pN3GAG assay to detect input gag plasmid. The number of samples assayed for each gag immunized group was as follows: gag, n = 9; gagifn, n = 10; gagenv, n = 7; gagenvifn, n = 9. The number of copies/␮g obtained from each animal using the pN3gag assay was expressed as a percentage of the number of copies/␮g obtained using the GAG assay. Also, the copies/␮g obtained in the gag assay was compared statistically with those obtained in the pN3–gag assay for each gag vaccinated group. 2.9. Determination of pathological changes The lungs were removed aseptically and samples taken from 4 different lung lobes (the right accessory (cranial) lobe, the right apical (medial) lobe, the right caudal lobe and the left cardiac (cranial) lung lobe). In addition, a sample from the mediastinal lymph node was also taken. Tissues were fixed in 10% phosphatebuffered formalin (Sigma), embedded in paraffin and sectioned according to standard procedures. Sections were stained with Hematoxylin–Eosin and scored blind for pathology by two independent pathologists. Histological sections from one lung lobe were stained for VMV p25 antigen using immunocytochemistry. The sections were fixed in Bouin solution and paraffin embedded. Following dewaxing and rehydration, sections were unmasked by immersing in citrate buffer, pH 6 (DAKO) and autoclaved for 3 min (∼100 ◦ C at 1 bar) in a (20/10) pressure cooker. Sections were cooled in distilled water at room temperature for 5 min and after they were treated with 3% hydrogen peroxide (DAKO® ) for 5 min at room temperature followed by several washes in Tris-buffered saline (TBS) (DAKO® ). Sections were then incubated with the VPM70 monoclonal antibody (anti-GAG p25) as supernatant [32] for 1 h at 37 ◦ C. After incubation the sections were washed with TBS and incubated with EnvisionTM peroxidase mouse (DAKO® ) for 30 min at room temperature. After incubation, sections were washed with TBS and incubated with 3-3-diaminobenzidine (DAB) (DAKO® ) for 10 min at room temperature. The sections were then washed with distilled water and the slides rinsed and counterstained with DAKO Hematoxylin and mounted under DPX for microscopy. Sections in which the specific primary antibodies were replaced by TBS were used as negative controls. Sections from a known positive animal were used as positive controls. 2.10. Statistical analysis Data were analysed using non-parametric Mann–Whitney tests. Data were considered significantly different if p ≤ 0.05. The experiment was performed in 3 centres and the statistical analysis first determined if the control groups in each centre were different. When the control groups were shown not to be different, they were pooled and compared to the relevant immunized groups.

R. Reina et al. / Vaccine 26 (2008) 4494–4505

Fig. 2. Expression of recombinant p55 GAG protein. Expression of p55 GAG precursor from plasmid (A) and rMVA (B) was detected by anti-p25 Western blotting. HEK 293 cells were transfected with plasmid. (A) Track 1, pN3-gag; 2, pN3-gag.inv (gag gene in the inverse orientation to promoter); 3, pN3-gag + CTE (gag gene with the constitutive transport element of Mason-Pfizer Monkey virus); 4 and 5, lysates from two separate transfections with pN3-p25 (pN3 with the p25 gag region only); and 6, pN3 alone as control. Recombinant MVA infected cell lysates (B) track 1, MVApSC11; 2, MVAp25 (recombinant MVA expressing p25 gag region only); 3, MVAgag. Cell lysates were analysed by Western blotting and proteins were probed with a rabbit anti-p25 serum diluted 1:5000. The expression of p55 GAG is indicated from the gag and gag + CTE constructs and MVAgag, whilst p25 is expressed from the p25 constructs.

3. Results 3.1. Expression of IFN- and p55 GAG and gp150 ENV precursor proteins The expression of the gag-CTE, env, and ovine IFN- genes was driven from the eukaryotic immediate early CMV promoter. The production of GAG and ENV proteins was confirmed prior to use by transfection of human embryonic kidney 293 cells and Western blotting. Fig. 2A shows expression of p55 GAG precursor from the recombinant plasmid using the MPMV CTE-fragment in HEK293 cells 48 h after transfection in vitro. The recombinant GAG protein was detected using anti-p25 (Fig. 2A) and anti-p17 (data not shown) specific antibodies. Expression of gp150 ENV precursor protein either from plasmid or from MVAenv was reported elsewhere [33]. The equivalent p55 GAG protein in MVA was shown to be expressed in vitro using anti-p25 antiserum (Fig. 2B) and antip14 antiserum (data not shown). Expression of ovine IFN-␥ was demonstrated using RT-PCR and ELISA in supernatants of transfected HEK293 cells and by RT-PCR in ovine fibroblasts (data not shown).

4499

transient with the env and env-ifn groups not varying from the controls at week 20. However at week 24 there was a reduced antibody response in these animals compared to controls (p = 0.023 for both individual groups and p = 0.003 for the pooled groups) (Fig. 3B). Plasmid immunization with gag or gag and env together with or without IFN- did not induce an increased antibody response compared to controls using the whole virus ELISA. Similarly, recombinant MVA administration did not boost the antibody response in these groups above background (week 12 post-immunization). After challenge the antibody response in these animals was similar to controls (Fig. 3B). In all cases, the antibodies were of the IgG1 subclass, with no IgG2 antibody responses being observed (data not shown). 3.3. Induction of T cell immunity 3.3.1. Cytotoxic T cell responses Very few animals in any group (usually less than 1–2 per group) induced detectable CTL responses systemically after the prime boost vaccination and challenge (data not shown). Therefore statistical analysis was not performed. 3.3.2. T cell proliferative responses T cell proliferation assays on PBMC were performed using whole virus as a source of ENV and GAG and with recombinant GAG proteins. The median stimulation indices obtained from using two or three antigen concentrations were used for the analyses. The results are shown in Fig. 4. Proliferative responses were not elevated in any

3.2. Induction of antibody Sheep were immunized using a prime boost vaccination method with plasmid and recombinant MVA before challenge with VMV (Fig. 1) and anti-virus antibody measured by ELISA. After two immunizations with plasmid (week 7 post-immunization), only the env and env-ifn immunized groups induced a significant antibody IgG response systemically (measured by a whole virus ELISA) with titres greater than those of the control group (p = 0.047 and 0.003, respectively) (Fig. 3A). These two groups were not different to each other and when pooled because there was no overall effect of IFN- immunization at any time point, the difference from controls was significant at p = 0.009. The antibody responses seen in the env and env-ifn immunized groups were transient and not boosted systemically by recombinant MVA administration intra-tracheally. Indeed at week 12 post-immunization, the antibody response was not different to controls (Fig. 3A). However, there was an increased response early after challenge (week 16 post-immunization, i.e. week 4 post-challenge) with both groups showing increased antibody titres compared to controls (p = 0.036 and 0.046, respectively, and p = 0.012 when the groups were pooled) (Fig. 3B). Again this was

Fig. 3. Antibody titre determined using whole virus ELISA. Serial two-fold dilutions of sera from sheep were tested on a whole VMV ELISA before and after challenge with VMV. The results are given as the mean titre (log2 [titre/50]) ± standard error of the group. Time post-immunization: () week 0; () week 7; () week 12; () week 16; () week 20; (䊉) week 24. Panel (A) shows week 0–12, pre-challenge results on an expanded scale, panel (B) shows week 16–24 results after challenge. Asterisk (*) indicates titre greater than the control group (p < 0.05, groups env and env-ifn week 7 and week 16) and symbol (§) indicates less than the control group (p < 0.05, env and env-ifn week 24) using Mann–Whitney tests.

4500

R. Reina et al. / Vaccine 26 (2008) 4494–4505

Fig. 4. T cell proliferation responses. T cell proliferative results have been expressed as the median of the group’s median stimulation index for each antigen against time. Antigen used: (A) EV1; (B) p25; (C) p17; (D) p14. Groups: () control; (䊉) gag; () gag-ifn; () env; () env-ifn; () gag-env; (♦) gag-env-ifn. For clarity, significant differences have only been indicated after challenge: asterisk (*) indicates significantly greater than controls (env-ifn: EV1 week 16; gag-env-ifn: p17 week 20; gag-env: p14 week 16) and symbol (§) indicates significantly less than controls (§ gag: EV1 week 24; p25 and p17 weeks 16, 20 and 24; p14 weeks 16 and 24; § gag-ifn: EV1 week 24; p25 week 20; p17 and p14 week 16; § env: p25 week 20). Analysis by Mann–Whitney test.

of the immunized groups compared to controls prior to challenge except in the gag-env immunized group at week 7 which showed a weak proliferative response with p25 (p = 0.044). Following challenge, the gag-env and gag-env-ifn groups tended to show higher proliferative responses to p25, p17 and p14 compared to the controls, with significant elevations being observed at week 20 in p17 (p = 0.04) and p14 (p = 0.034) responses, respectively (Fig. 4C and D). The T cell proliferative responses to EV1 antigen in the gag-env and gag-env-ifn groups were not available for technical reasons. In animals immunized with env or env-ifn-, T cell proliferation to EV1 antigen was observed at week 24, though the levels were not different to controls. In contrast, the gag and gag-ifn groups tended to show depressed responses at one or more time points after challenge compared to controls. This was evident with all four antigens (Fig. 4A–D). Combining the gag and gag-ifn groups gave significantly reduced responses compared to controls using EV1 at week 24 (p = 0.006), p25 at week 20 (p = 0.003) and 24 (p = 0.026), p17 at week 16 (p = 0.003), and p14 at week 16 (p = 0.001) and 24 (p = 0.023). The combined gag and gag-ifn groups were significantly lower than the combined gag-env and gag-env-ifn groups at week 16 in response to p17 (p < 0.001) and p14 (p = 0.001), at week 20 in response to p25 (p = 0.001), p17 (p = 0.002), and p14 (p = 0.019), and at week 24 in response to p25 (p = 0.002), p17 (p = 0.003), and p14 (p = 0.002). The responses in the combined gag and gag-ifn groups were also significantly lower than the combined env and env-ifn groups at one or more time points when p17, p14 or EV1 was used as antigen in the assays (data not shown). 3.3.3. IFN- expression by PBMC Expression of IFN- mRNA was measured by real-time PCR using cultures of PBMC stimulated with whole virus or GAG recombinant proteins. By Mann–Whitney test, IFN- mRNA expression in response to EV1 antigen was elevated at week 7 for gag immunized

animals compared to controls (p = 0.032). The combined gag and gag-ifn groups showed elevated IFN␥ expression levels at week 16 in response to p17 antigen (p = 0.047). For env immunized animals IFN- expression induced by EV1 stimulation at week 16 was higher than in control animals (p = 0.049). However, these responses were transient as there was no difference to controls at week 20. Apart from the above no significant differences in IFN␥ responses were noted when recombinant p14, p17, or p25 were used as stimulating antigens (data not shown). 3.4. Blood proviral load To determine the level of systemic infection, the proviral load in blood was measured using quantitative real time PCR. Two assays were developed, based, respectively, on EV1 pol and gag sequences. Preliminary control experiments performed on plasmids diluted in genomic ovine DNA showed that the gag assay was more sensitive than pol, and therefore the gag assay was chosen for subsequent routine analysis of blood. All samples before challenge were negative (week 0, 7 and 12, data not shown). Proviral load was not affected by having ifn- present in the plasmid immunization at any time point. Results are shown in Fig. 5. When the env and env-ifn groups were analysed together they were found to have a lower virus load than control animals 4 weeks after challenge (p = 0.002) and the individual groups were also significantly different to controls (p = 0.004 and p = 0.024, respectively). By 8–12 weeks after challenge, there was still a trend for the pooled env and env-ifn immunized groups to show a reduced blood proviral load (8 weeks p.i., p = 0.075; individually env, p = 0.053 and env-ifn, p = 0.333 compared to controls) with decreasing significance with time (12 weeks p.i., p = 0.206 for the pooled groups compared to controls). The gag-env and gag-env-ifn groups were not significantly different to one another and, although individually they were not different to controls, when pooled they showed an increased provi-

R. Reina et al. / Vaccine 26 (2008) 4494–4505

4501

tion of the assay. As such they did not affect the results shown in Fig. 6. The lung lobe with the maximum value for each animal was taken for analysis of the immunized groups (Fig. 6A). Again there was no effect on proviral copy number if ifn- was present in the plasmid immunization. Animals immunized with gag, i.e. gag, gagifn, gag-env and gag-env-ifn showed increased levels of proviral DNA in the lung (Fig. 6A, p ≤ 0.024). Immunization with env or env-ifn either individually or pooled did not affect lung proviral load compared to controls (Fig. 6A). Thus, immunization with gag mucosally caused increased proviral load in the lung after challenge, whilst env was not protective by this measure and did not result in an increased proviral load. 3.5.2. Mediastinal lymph node proviral load The major draining lymph node for the lung is the mediastinal lymph node. Samples were taken from each animal at week 12–13 post-challenge and the proviral load was determined. The proviral load was not significantly different between any group and the controls whether groups were analysed individually or grouped: immunogen and immunogen + ifn-; as all animals with either gag or env in the immunogen; or as all animals with ifn- in the plasmid immunization (Fig. 6B). Thus, the significant increases in virus load seen in the lung in gag and gag-env immunized animals were not seen in the draining lymph node at week 12–13 post-challenge. Fig. 5. Blood proviral load after challenge. The proviral load in white blood cells was measured using a VMV p25 specific real-time PCR assay. Results have been expressed as mean copy number per ␮g extracted DNA ± standard error of the group. Time post-immunization: (A) week 16; (B) week 20; (C) week 24. Asterisk (*) indicates proviral load greater than the control group (p < 0.05, pooled gag-env and gag-env-ifn group week 16) and symbol (§) indicates proviral load less than the control group (p < 0.05, env week 16 and 20; env-ifn week 16). Analysis by Mann–Whitney test.

3.6. Pathology Sections from 4 different lung lobes and the mediastinal lymph node of each animal were taken at week 12–13 post-challenge. Pathological changes typical of VMV infection (Fig. 7A and B) were

ral load compared to controls 4 weeks after challenge (p = 0.049). Indeed, when groups immunized with gag in the immunogen were compared to those without gag, there was a significant increase in proviral load at weeks 4 and 8 post-challenge (weeks 16 and 20; Fig. 5, p = 0.001 and p = 0.005, respectively) in the gag immunized animals. Thus, early after infection, immunization with env caused a reduction whilst the presence of both gag and env in the immunization mixture resulted in an increase in systemic proviral load. These effects were short lived as there was no significant difference between groups by week 12 post-challenge. 3.5. Tissue proviral load 3.5.1. Lung proviral load Each animal was challenged intra-tracheally with 103 TCID50 EV1 VMV. At week 12–13 post-challenge the number of copies of proviral DNA per ␮g DNA extracted from tissue was measured in 4 different lung lobes for each animal with the gag assay, although the gag gene was present in the plasmids used for immunization, on the grounds that input plasmid disappears rapidly from blood and lung tissues [34,35]. To control for the possible presence of residual input gag plasmid, representative lung samples from the gag, gag-ifn, gag-env, and gag-env-ifn groups were assayed simultaneously using the pN3-gag assay. In general, the input gag plasmid copy numbers (data not shown) were very low. The copies/␮g obtained in the pN3-gag assay were expressed as a percentage of those obtained in gag assay. The mean% ± S.E. of pN3-gag/gag for the gag, gag-ifn, gag-env, and gag-env-ifn groups was 0.06 ± 0.06, 3.03 ± 1.35, 2.73 ± 1.49, 3.06 ± 1.47, respectively, i.e. ranging from 0 to 3%. Therefore, plasmid levels remaining were approximately 100-fold less than the gag levels and were on the limits of detec-

Fig. 6. Tissue proviral load. The proviral load in lung and mediastinal lymph node was measured using a VMV p17 specific real-time PCR assay at week 12–13 postchallenge (week 24–25 post-immunization). Results have been expressed as mean copy number per ␮g extracted DNA ± standard error of the group. (A) Mean of the maximum lung section. (B) Mean of the mediastinal lymph node proviral load. Asterisk (*) indicates proviral load greater than the control group (p < 0.05, in lung all groups immunized with gag). Analysis by Mann–Whitney test.

4502

R. Reina et al. / Vaccine 26 (2008) 4494–4505

Fig. 7. Tissue pathology. Characteristic pathological changes in the lung and lymph node are shown. Haematoxilin and Eosin stained lung sections showing (A) bronchiole associated lymphoid tissue hyperplasia and (B) interstitial pneumonitis. (C) Pathological changes in the lung and mediastinal lymph node were scored and the values for each animal calculated. Results have been expressed as the mean of the pathology score ± standard error of the group. Symbol (§) indicates pathological score less than the control group (p < 0.05, pooled gag and gag-ifn group). Analysis by Mann–Whitney test.

screened and scored: in the lung, lymphoid follicle hyperplasia and interstitial pneumonia as well as BALT hyperplasia, perivascular infiltrates, congestion and oedema were scored on a scale of 0–4; and in the mediastinal lymph node, lymphoid follicle reactivity and cortical hyperplasia were scored on a scale of 0–2. The mean of all scores was used for each animal. There was no effect of the presence of ifn- in the plasmid immunogen. Individually groups were not different to controls but when single immunogen groups were pooled (once they had been shown not to be different) the gag with gag-ifn group showed decreased pathological changes (Fig. 7C, p = 0.047). The env with env-ifn group and gag-env with gag-env-ifn group were not different to controls. So immunization with gag alone reduced the levels of pathology induced by infection with VMV. Staining for viral p25 GAG showed only a few positive cells present in sections. When sections were scored as either virus antigen positive or negative there was no correlation with lesion score (data not shown). 4. Discussion In this study, we investigated whether a prime boost regimen of immunization against VMV given via the respiratory mucosa would induce immune responses, and modulate virus loads or lesion formation in a challenge model of ovine lentivirus infection. We utilized the ability of PEI–DNA complexes to transfect cells in vivo and induce immune responses in rodents [19,36], and coupled this with administration of a live virus vector, MVA, to deliver a “protein” boost [37]. The genes encoding the GAG and ENV precursor proteins of VMV were used individually or together, with and without the gene encoding ovine IFN-␥ with a view to determining whether the gag or env genes provided the best immu-

nization targets or whether both structural genes were required for protection. Humoral responses were found to be elevated in sheep immunized with env or env-ifn plasmids prior to challenge, but low in all other immunized groups. The elevated anti-ENV antibody responses were transient but seen again early after challenge (week 16). The antibody responses in these animals were reduced at week 24 following challenge with VMV compared to controls. This reduction may have been due to sequestration of anti-ENV producing plasma cells to sites of virus production such as the lung and draining MLN. It is well established that virus levels increase transiently following experimental infection then decrease to low or undetectable levels for prolonged periods thereafter [38,24]. It is not known whether this reduction is due to humoral or cell-mediated responses or to a combination of both. Anti-ENV antibodies could conceivably be involved in the reduction of virus leading to the restricted replication phase, and might be expected to return to control levels once virus levels have dropped. No antibody response was seen in blood to the gag or gag-env immunizations. This may reflect local antibody production at the mucosal surface after immunization or skewing of the response to IgM or IgA but this was not tested. There was no systemic anamnestic antibody response in the immunized groups after challenge; these animals showed similar antibody responses to control infected animals in blood. We did not test for locally produced antibody but intra-nasal immunization in a variety of systems is known to induce mucosal antibody [39]. It is of interest to note that the antibodies generated by this immunization method were all of the IgG1 subclass (data not shown), as has been found in naturally infected animals [40,41]. IgG2 antibody responses were not induced. This is in contrast to results obtained in goats immunized intra-dermally with the env

R. Reina et al. / Vaccine 26 (2008) 4494–4505

gene of CAEV, where a bias towards IgG2 antibodies was observed after plasmid immunization. However, after immunization with a vaccinia recombinant, a bias towards IgG1 antibody production was seen [42]. The difference in the antibody subclass produced in the present study may relate to the mucosal site of priming and boosting versus the intra-dermal route or may reflect a species specific response to plasmid immunization. In goats IgG2 is linked to a Th1 response [43] whilst here in sheep, even in the presence of IFN- at priming, there was no IgG2 response. Co-immunization of goats with CAEV env and IL-12 or IFN- has also failed to potentiate the antibody response to protein [18,17]. In sheep lungs and nasal secretions there are high concentrations of IgG and IgA with IgM present at low concentrations. There is evidence that all three subclasses may be produced within the lung as well as there being transudation of serum IgG into the lung fluid [44,45]. After intranasal immunization, raised IgA and IgG1 levels may be detected in serum but there is less evidence for induction of good IgG2 or IgM responses in serum in sheep [27]. The respiratory prime boost regime used also did not induce neutralizing antibodies (not shown). Other SRLV vaccination protocols have resulted in such antibodies being induced [46], but in general neutralizing antibodies are usually not detected after vaccination [47,8,13]. Indeed, neutralizing antibody production after challenge was inhibited by mucosal env immunization in the study by Gonzalez et al. [13]. Neutralizing antibodies are also slow to be induced after natural infection: they are usually present by 3–6 months post-infection but can take up to 2 years to develop [4]. Thus, the lack of induction of neutralizing antibody is not unexpected given the length of time used here. To assess cell-mediated immune responses following vaccination, attempts were made to determine T cell cytotoxicity, proliferation, and IFN␥ production in peripheral blood. One of the perceived benefits of plasmid DNA vaccination in animal models is the induction of CD8+ cytotoxic T cells [48]. Cytotoxic T cell responses have been documented following DNA vaccination protocols in the SIV [49,50] and FIV [51] models. However, it has proven difficult to induce significant levels of CTL by plasmid DNA vaccination against other lentiviruses such as EIAV [52]. In the present study, systemic cytotoxic T cell responses were very low both before and after VMV challenge in all groups of animals. The paucity of CTL precursors in the blood could either be due to poor induction by the immunization regimen or to sequestration of such cells in the lungs and MLN following respiratory challenge with VMV. T cell proliferative responses were also not elevated compared to controls prior to challenge except for one time point with the gag-env immunized groups. Transiently elevated responses to GAG antigens in animals immunized with gag-env or gag-env-ifn were observed after challenge. This suggests that combining the two antigens enhances proliferative responses. This may simply be an additive effect of individual responses to the GAG and ENV antigens or possibly due to linked recognition between GAG- and ENV-specific T helper cells following administration of virus [53]. Recent studies on immune responses to HIV gag and env genes have revealed a suppressive effect of env on GAG-specific T cell responses in mice [54]. No such effect was observed here, where responses to GAG antigens in animals receiving the combined gag and env immunogens were not lower than controls. In contrast to the T cell proliferative responses found in animals immunized with both gag and env, proliferative responses were significantly lower in gag or gag-ifn immunized animals than in controls after challenge. This could be explained by peripheral GAG-specific T cells being sequestrated in the lungs and MLN following challenge. Alternatively, the immunization with gag may have resulted in specific inhibition of GAG-specific T cell responses. Lack of IFN-␥ producing cells in GAG antigen stimulated cells is con-

4503

sistent with the low proliferative responses obtained to the recombinant GAG antigens. The frequency of antigen-specific T cells in the lungs and MLN was not determined in this study, and further work is required to determine if the above hypothesis is correct or not. Cytokines such as IFN-␥ and IL-12 have been used in many studies to skew immune responses towards a Th1 phenotype, whereby cell-mediated responses such as T cell cytotoxicity tend to dominate over antibody responses. In one such study, both ifn- and IL-12 augmented immune responses to HIV in mice [55]. However, the inclusion of ifn- in the respiratory immunization protocol used here clearly failed to have a pronounced effect on either antibody levels or on cell-mediated responses. This may have been due to expressed IFN-␥ inhibiting the CMV promoter in the ifn- plasmid [56]. However, the ovine ifn- plasmid was shown to express in both HEK293 cells and primary ovine fibroblasts in vitro. Also, it has been shown that feline ifn- under the control of an RSV promoter had an enhancing effect in FIV vaccination [57], suggesting that IFN-␥ silencing of virus promoters does not always occur. Other results obtained using an attenuated provirus DNA vaccine against FIV showed that ifn- did not enhance the efficacy of the vaccine [58]. Similarly, ifn- did not potentiate the activation of Th1 lymphocyte responses to DNA vaccination against CAEV, though antibody responses were significantly inhibited [17]. The results in sheep and goats suggest that ifn- is not a useful molecular adjuvant in plasmid DNA vaccine approaches to controlling SRLV. Analysis of proviral DNA levels showed different results depending on the compartment examined. In blood, animals immunized with env or env-ifn showed significantly reduced proviral loads in the early phase after VMV challenge. No alterations in proviral loads were observed in animals receiving gag or gag-env immunogens. The protective effects of env immunization were, however, transient and the provirus levels returned to control levels 12 weeks after challenge. This early protective effect may have been related to the immune response produced in env immunized groups. Mucosal immunization with env has previously shown transient reduced blood proviral loads in sheep [13]. However, in the latter study, blood provirus levels were depressed for over 18 months. This may have been due at least in part to the efficacy of mucosal DNA vaccination using a gene gun in the vulva compared to the PEI–DNA complexes used here in the NALT-lung. At post-mortem (12–13 weeks after VMV challenge), proviral DNA levels in lung and MLN in env or env-ifn immunized animals were similar to controls. In contrast, the proviral DNA levels in animals receiving the gag immunogen were significantly elevated in the lung. This was not due to interference by residual input gag plasmids, since the signal obtained using a real-time assay that spanned the vector and gag sequences was either zero or very low. This result is consistent with other studies showing that the level of plasmid introduced into the lungs drops by approximately 3 logs over a 28-day period [35]. Similarly, plasmid administered parenterally disappears very rapidly from blood and from tissues such as the lung within 10 days [34]. In our sheep, the period between the last plasmid administration and post-mortem when the lung tissues were analyzed was 4 months, and very little residual plasmid would be expected. It is unlikely that anti-GAG antibodies enhanced infection. It is probable, however, that gag immunization primed T cells for recall activation following challenge. Such activated T cells could upregulate virus production during antigen presentation by infected dendritic cells and macrophages, as it is known that CD4+ T lymphocytes are necessary for efficient viral infection [59] and may up-regulate viral infection of dendritic cells [60]. Certainly activated T cells produce GM-CSF which is known to up-regulate virus expression in macrophages [61]. It is unclear, however, why this effect was observed only with the gag immunogens and not with

4504

R. Reina et al. / Vaccine 26 (2008) 4494–4505

the individual env gene. It is possible that ENV-specific cytotoxic responses counteracted the virus enhancing effects of the gag genes in the lung ablating the increase. Again, there was no consistent effect on provirus loads of including the ifn- gene in the immunizing mixture. The SRLV cause chronic inflammatory lesions in target organs such as the lungs and lymphoid tissue. Here, we scored the presence and degree of severity of lesions in four lung lobes and in the MLN 12–13 weeks after respiratory challenge with VMV in control and immunized animals. Lesions were found in most animals after challenge. However, the mean lesion scores were reduced in animals receiving the single gag immunogen, with the combined gag and gag-ifn groups showing a significant reduction at postmortem. No reduction in lesion frequency or severity was observed with the env or gag-env immunized groups. In goats, by contrast, priming with plasmid DNA encoding CAEV env and boosting with the ENV surface protein in Freund’s incomplete adjuvant via the intradermal route suppresses the development of arthritis following intravenous challenge [8]. Longer term studies are required to determine if the reduced lesions observed here persist. The immunizations in the present study thus had differing effects on virus levels and pathology scores, with env immunization reducing blood virus levels but not lesions, and gag increasing tissue virus levels but reducing lesion formation. A strong correlation between virus levels and lesion formation in sheep and goats has been established in several studies. In goats infected with CAEV, the frequency of virus isolation was shown to be associated with lesion development in joints [62]. Similarly, in sheep infected with VMV, a correlation has been found between the degree of lung inflammation and levels of proviral DNA and virus RNA, the frequency of virus isolation and antigen expression, and the number of cells expressing the gag p25 core protein [63–66]. However, it is thought that lesions are immune mediated [67]. One possible explanation for the results could, therefore, be that gag immunization in the absence of env inhibits T cell responses to VMV antigens, leading to a reduction in T cell-mediated tissue damage. In contrast, the presence of env in the immunization leads to functional T cell responses and expected levels of lesion formation. The observation that provirus levels are higher in tissues following gag immunization is, however, more consistent with viral replication being driven by active T cell activation. The latter would require virus protein expression. In fact, p25 protein expression levels in lung tissue of gag or env immunized animals were similar to controls. Further work is required to unravel the immunological and virological mechanisms operating in the target tissues following mucosal gag DNA immunization. In conclusion, mucosal immunization via the respiratory tract against VMV env results in transiently reduced levels of proviral DNA in the blood, whilst immunization with the single gag gene induces a degree of protection against lesion development despite increasing proviral load in the lung. Having identified the gag gene as a potentially useful immunogen for respiratory tract immunization, future work will focus on determining local immune responses to gag in the lung, and the long-term effects of vaccination on lesion formation and virus replication. Acknowledgements This study was supported by grants from European Union (QLK2-CT-2002-00617) and Spanish CICYT (AGL2003-08977-C0301 and AGL2006-13410C-06/01/GAN). Ramsés Reina and Ximena de Andrés were supported by a fellowship FPI from the Spanish Ministry of Science and Education. Tom McNeilly was supported by a Ph.D. studentship from the Royal (Dick) College of Veterinary Studies, University of Edinburgh. We thank Margaret Ross, Pauline

Love, Paolo Caldato, Paul Tonks, Santiago Becerra, Rosario Puyó, Margherita Profiti, Katia Ricci, Giorgia Tozzini, Angela Wheatley, Sigridur Matthiasdottir, and Paul Wright for technical help. References [1] Peterhans E, Geenland T, Badiola J, Harkiss G, Bertoni G, Amorena B, et al. Routes of transmission and consequences of small ruminant lentiviruses (SRLVs) infection and eradication schemes. Vet Res 2004;35:257–74. [2] Reyburn HT, Roy DJ, Blacklaws BA, Sargan DR, Watt NJ, McConnell I. Characteristics of the T cell-mediated immune response to Maedi-Visna virus. Virology 1992;191:1009–12. [3] Singh I, McConnell I, Blacklaws B. Immune response to individual Maedi-Visna virus gag antigens. J Virol 2006;80:912–9. [4] Petursson G, Nathansson N, Georgsson G, Panitch H, Palsson PA. Pathogenesis of Visna. I. Sequential virologic, serologic, and pathologic studies. Lab Invest 1976;35:402–12. [5] Yamamoto JK, Pu R. Feline immunodeficiency virus pathogenesis and development of a dual-subtype feline-immunodeficiency-virus vaccine. AIDS 2007;21:547–63. [6] Amara RR, Patel K, Niedziela G, Nigam P, Sharma S, Staprans SI, et al. A combination DNA and attenuated simian immunodeficiency virus vaccine strategy provides enhanced protection from simian/human immunodeficiency virusinduced disease. J Virol 2005;79:15356–67. [7] Goonetilleke N, Moore S, Dally L, Winstone N, Cebere I, Mahmoud A, et al. Induction of multifunctional human immunodeficiency virus type 1 (HIV-1)specific T cells capable of proliferation in healthy subjects by using a primeboost regimen of DNA- and modified vaccinia virus Ankara-vectored vaccines expressing HIV-1 gag coupled to CD8+ T-cell epitopes. J Virol 2006;80:4717–28. [8] Cheevers WP, Snekvik KR, Trujillo JD, Kumpula-McWhirter NM, Pretty On Top KJ, Knowles DP. Prime-boost vaccination with plasmid DNA encoding caprinearthritis encephalitis lentivirus env and viral SU suppresses challenge virus and development of arthritis. Virology 2003;306:116–25. [9] Nenci C, Zahno M-L, Vogt H-R, Obexer-Ruff G, Doherr MG, Zanoni R, et al. Vaccination with a T-cell-priming Gag peptide of caprine arthritis encephalitis virus enhances virus replication transiently in vivo. J Gen Virol 2007;88:1589–93. [10] McGuire TC, Adams DS, Johnson GC, Klevjer-Anderson P, Barbee DD, Gorham JR. Acute arthritis in caprine arthritis-encephalitis virus challenge exposure of vaccinated or persistently infected goats. Am J Vet Res 1986;47:537–40. [11] Torsteinsdottir S, Carlsdottir HM, Svansson V, Matthiasdottir S, Martin AH, Petursson G. Vaccination of sheep with Maedi-Visna virus gag gene and protein, beneficial or harmful? Vaccine 2007;25:6713–20. [12] Petursson G, Matthiasdottir S, Svansson V, Andresdottir V, Georgsson G, Martin AH, et al. Mucosal vaccination with an attenuated Maedi-Visna virus clone. Vaccine 2005;23:3223–8. [13] Gonzalez B, Reina R, Garcia I, Andres S, Glaria I, Alzueta M, et al. Mucosal immunisation of sheep with a Maedi-Visna virus (MVV) env DNA vaccine protects against early MVV productive infection. Vaccine 2005;23:4342–52. [14] Donnelly JJ, Wahren B, Liu MA. DNA vaccines: progress and challenges. J Immunol 2005;175:633–9. [15] Kim JJ, Yang JS, vanCott TC, Lee DJ, Manson KH, Wyand MS, et al. Modulation of antigen-specific humoral responses in rhesus macaques by using cytokine cDNAs as DNA vaccine adjuvants. J Virol 2000;74:3427–9. [16] Kim JJ, Yang JS, Manson KH, Weiner DB. Modulation of antigen-specific cellular immune responses to DNA vaccination in rhesus macaques through the use of IL-2, IFN-gamma, or IL-4 gene adjuvants. Vaccine 2001;19:2496–505. [17] Cheevers WP, Beyer JC, Hotzel I. Plasmid DNA encoding caprine interferon gamma inhibits antibody response to caprine arthritis-encephalitis virus (CAEV) surface protein encoded by a co-administered plasmid expressing CAEV env and tat genes. Vaccine 2001;19:3209–15. [18] Cheevers WP, Hotzel I, Beyer JC, Kumpula-McWhirter N. Immune response to caprine arthritis-encephalitis virus surface protein induced by coimmunization with recombinant vaccinia viruses expressing the caprine arthritis-encephalitis virus envelope gene and caprine interleukin-12. Vaccine 2000;18:2494–503. [19] Boussif O, Lezoualc’h F, Zanta MA, Mergny MD, Scherman D, Demeneix B, et al. A versatile vector for gene and oligonucleotide transfer into cells in culture and in vivo: polyethylenimine. Proc Natl Acad Sci USA 1995;92:7297–301. [20] Kasturi SP, Sashaphibulkij K, Roy K. Covalent conjugation of polyethyleneimine on biodegradable microparticles for delivery of plasmid DNA vaccines. Biomaterials 2005;26:6375–85. [21] Bieber T, Meissner W, Kostin S, Niemann A, Elsasser HP. Intracellular route and trasncriptional competence of polyethylenimine–DNA complexes. J Control Release 2002;82:441–54. [22] Goula D, Becker N, Lemkine GF, Normandie P, Rodriguez J, Mantero S, et al. Rapid crossing of the pulmonary endothelial barrier by polyethylenimine/DNA complexes. Gene Ther 2000;7:499–504. [23] Sargan DR, Bennet ID, Cousens C, Roy DJ, Blacklaws BA, Dalziel RG, et al. Nucleotide sequence of EV1, a British isolate of Maedi Visna virus. J Gen Virol 1991;72:1893–903. [24] Bird P, Blacklaws B, Reyburn HT, Allen D, Hopkins J, Sargan D, et al. Early events in immune evasion by the lentivirus Maedi-Visna occurring within infected lymphoid tissue. J Virol 1993;67:5187–97.

R. Reina et al. / Vaccine 26 (2008) 4494–4505 [25] Sutter G, Moss B. Nonreplicating vaccinia vector efficiently expresses recombinant genes. Proc Natl Acad Sci USA 1992;89:10847–51. [26] Sanchez AB, Rodriguez D, Garzon A, Amorena B, Esteban M, Rodríguez JR. Visna/Maedi virus env protein expressed by a vaccinia virus recombinant induces cell-to-cell fusion in cells of different origins in the apparent absence of env cleavage: role of glycosylation and of proteoglycans. Arch Virol 2002;147:2377–92. [27] Stanley AC, Buxton D, Innes EA, Huntlry JF. Intranasal immunisation with Toxoplasma gondii tachyzoite antigen encapsulated into PLG microspheres induces humoral and cell-mediated immunity in sheep. Vaccine 2004;22:3929–41. [28] McNeilly TN, Tennant P, Lujan L, Perez M, Harkiss GD. Differential infection efficiencies of peripheral lung and tracheal tissues in sheep infected with Visna/Maedi virus via the respiratory tract. J Gen Virol 2007;88:670–9. [29] Houwers DJ, Gielkens AL, Schaake Jr J. An indirect enzyme-linked immunoassay (ELISA) for the detection of antibodies to Maedi-Visna virus. Vet Microbiol 1982;7:209–19. [30] Blacklaws BA, Bird P, Allen D, McConnell I. Circulating cytotoxic T lymphocyte precursors in Maedi-Visna virus-infected sheep. J Gen Virol 1994;75: 1589–96. [31] Pfaffl MW. A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res 2001;29:2002–7. [32] Reyburn H, Roy DJ, Blacklaws BA, Sargan DR, McConnell I. Expression of MaediVisna virus major core protein, p25: development of a sensitive p25 antigen detection assay. J Virol Methods 1992;37:305–20. [33] Fraisier C, Arnason H, Barbezange C, Andresdottir V, Carrozza ML, de Andres D, et al. Expression of the gp150 Maedi Visna Virus envelope precursor protein by mammalian expression vectors. J Virol Methods 2007;146:363–7. [34] Oh Y-K, Kim J-P, Yoon H, Kim JM, Yang J-S, Kim C-K. Prolonged organ retention and safety of plasmid DNA administration in polyethylenimine complexes. Gene Therapy 2001;8:1587–92. [35] Pringle IA, Raman S, Sharp WW, Cheng SH, Hyde SC, Gill DR. Detection of plasmid DNA vectors following gene transfer to the murine airways. Gene Therapy 2005:1206–14. [36] Huang X, Xu J, Qiu C, Ren L, Liu L, Wan Y, et al. Mucosal priming with PEI/DNA complex and systemic boosting with recombinant Tian Tan vaccinia stimulate vigorous mucosal and systemic immune responses. Vaccine 2007;25:2620–9. [37] Woodland DL. Jump-starting the immune system: prime-boosting comes of age. Trends Immunol 2004;25:98–104. [38] Nathanson N, Petursson G, Georgsson G, Palsson PA, Martin JR, Miller A. Pathogenesis of Visna. IV. Spinal fluid studies. J Neuropathol Exp Neurol 1979;38:197–208. [39] Gherardi MM, Perez-Jimenez E, Najera JL, Esteban M. Induction of HIV immunity in the genital tract after intranasal delivery of a MVA vector: enhanced immunogenicity after DNA prime-modified vaccinia virus Ankara boost immunization schedule. J Immunol 2004;172:6209–20. [40] Mehta PD, Thormar H. Neutralizing activity in isolated serum antibody fractions from Visna-infected sheep. Infect Immun 1974;10:678–80. [41] Bird P, Reyburn HT, Blacklaws B, Allen D, Nettleton P, Yirrell DL, et al. The restricted IgG1 antibody response to Maedi Visna virus is seen following infection but not following immunization with recombinant gag protein. Clin Exp Immunol 1995;102:274–80. [42] Beyer JC, Chebloune Y, Mselli-Lakhal L, Hotzel I, Kumpula-McWhirter N, Cheever WP. Immunization with plasmid DNA expressing the caprine arthritisencephalitis virus envelope gene: quantitative and qualitative aspects of antibody response to viral surface glycoprotein. Vaccine 2000;19:1643–51. [43] Cheevers WP, Beber JC, Knowles DP. Type 1 and type 2 cytokine gene expression by viral gp135 surface protein-activated T lymphocytes in caprine-arthritis encephalitis lentivirus infection. J Virol 1997;71:6259–63. [44] Alley MR, Wells PW, Smith WD, Gardiner AC. The distribution of immunoglobulin in the respiratory tract of sheep. Vet Pathol 1980;17:372–80. [45] Scicchitano R, Sheldrake RF, Husband AG. Origins of immunoglobulins in respiratory tract secretion and saliva of sheep. Immunology 1986;58:315–21. [46] Kemp RK, Knowles DP, Perry LL, McGuire TC, Besser TE, Cheevers WP. Crossreactive neutralizing antibodies induced by immunization with caprine arthritis-encephalitis virus surface glycoprotein. Vaccine 2000;18:1282–7.

View publication stats

4505

[47] Cheevers WP, Knowles DP, McGuire TC, Baszler TV, Hullinger GA. Caprine arthritis-encephalitis lentivirus (CAEV) challenge of goats immunized with recombinant vaccinia virus expressing CAEV surface and transmembrane envelope glycopriteins. Vet Immunol Immunpathol 1994;42:237–51. [48] Ulmer JB, Otten GR. Priming of CTL responses: direct transfection of antigen presenting cells versus crosspriming. Dev Biol (Basel) 2000;104:9–14. [49] Yasutomi Y, Robinson HL, Lu S, Mustafa F, Lekutis C, Arthos J, et al. Simian immunodeficiency virus-specific cytotoxic T-lymphocyte induction through DNA vaccination of rhesus monkeys. J Virol 1996;70:678–81. [50] Hanke T, Samuel RV, Blanchard TJ, Neumann VC, Allen TM, Boyson JE, et al. Effective induction of simian immunodeficiency virus-specific cytotoxic T lymphocytes in macaques by using a multiepitope gene and DNA prime-modified vaccinia Ankara boost vaccination regimen. J Virol 1999;73:7524–32. [51] Flynn JN, Hosie MJ, Rigby MA, Mackay N, Cannon CA, Dunsford T, et al. Factors influencing cellular immune responses to feline immunodeficiency virus induced by DNA vaccination. Vaccine 2000;18:1118–32. [52] Cook RF, Cook SJ, Bolin PS, Howe LJ, Zhou W, Montelaro RC, et al. Genetic immunization with codon-optimized equine infectious anaemia virus (EIAV) surface unit (SU) envelope protein gene sequences stimulates immune responses in ponies. Vet Microbiol 2005;108:23–37. [53] Janeway CA, Travers P, Walport M, Shlomchik MJ. Immunobiology: the immune system in health and disease. Edinburgh: Churchill Livingston; 2004. [54] Toapanta FR, Craigo JK, Montelaro RC, Ross TM. Reduction of anti-HIV-1 Gag immune responses during co-immunization: immune interference by the HIV1 envelope. Curr HIV Res 2007;5:199–209. [55] Abaitua F, Rodríguez JR, Garzon A, Rodríguez D, Esteban M. Improving recombinant MVA immune responses: potentiation of the immune response to HIV-1 with MVA and DNA vectors expressing env and the cytokines IL-12 and IFNgamma. Virus Res 2006;116:11–20. [56] Harms JS, Splitter GA. Interferon-gamma inhibits transgene expression driven by SV40 or CMV promoters but augments expression driven by the mammalian MHC I promoter. Hum Gene Ther 1995;6:1291–7. [57] Hosie MJ, Flynn JN, Rigby MA, Cannon C, Dunsford T, Mackay NA, et al. DNA vaccination affords significant protection against feline immunodeficiency virus infection without inducing detectable antiviral antibodies. J Virol 1998;72:7310–9. [58] Gupta S, Leutenegger CM, Dean GA, Steckbeck JD, Cole KS, Sparger EE. Vaccination of cats with attenuated feline immunodeficiency virus proviral DNA vaccine expressing gamma interferon. J Virol 2007;81:465–73. [59] Eriksson K, McInnes E, Ryan S, Tonks P, McConnell I, Blacklaws B. CD4(+) T cells are required for the establishment of Maedi-Visna virus infection in macrophages but not dendritic cells in vivo. Virology 1999;258:355–64. [60] Ryan S, Tiley L, McConnell I, Blacklaws B. Infection of dendritic cells by the Maedi-Visna lentivirus. J Virol 2000;74:10096–103. [61] Zhang Z, Harkiss GD, Hopkins J, Woodall CJ. Granulocyte macrophage colony stimulating factor is elevated in alveolar macrophages from sheep naturally infected with Maedi-Visna virus and stimulates Maedi-Visna virus replication in macrophages in vitro. Clin Exp Immunol 2002;129:240–6. [62] Knowles Jr D, Cheevers W, McGuire T, Stem T, Gorham J. Severity of arthritis is predicted by antibody responses to gp135 in chronic infection with caprinearthritis encephalitis virus. J Virol 1990;64:2396–8. [63] Brodie SJ, Marcom KA, Pearson LD, Anderson BC, de la Concha-Bermejillo A, Ellis JA, et al. Effects of virus load in the pathogenesis of lentivirus-induced lymphoid interstitial pneumonia. J Infect Dis 1992;166:531–41. [64] Staskus KA, Couch L, Bitterman P, Retzel EF, Zupancic M, List J, et al. In situ amplification of Visna virus DNA in tissue sections reveals a reservoir of latently infected cells. Microb Pathog 1991;11:67–76. [65] Lujan L, Begara I, Collie D, Watt NJ. Ovine lentivirus (Maedi Visna virus) protein expression in sheep alveolar macrophages. Vet Pathol 1994;31:695–703. [66] Zhang Z, Watt NJ, Hopkins J, Harkiss G, Woodall CJ. Quantitative analysis of Maedi-Visna virus DNA load in peripheral blood monocytes and alveolar macrophages. J Virol Methods 2000;86:13–20. [67] Nathanson N, Panitch H, Palsson PA, Petursson G, Georgsson G. Pathogenesis of Visna. II. Effect of immunosuppression upon early central nervous system lesions. Lab Invest 1976;35:444–51.