|Home | About | Journals | Submit | Contact Us | Français|
Previously, we demonstrated that type I interferon (IFN-α/β) or a combination of IFN-α/β and type II IFN (IFN-γ) delivered by a replication-defective human adenovirus 5 (Ad5) vector protected swine when challenged 1 day later with foot-and-mouth disease virus (FMDV). To gain a more comprehensive understanding of the mechanism of protection induced by IFNs, we inoculated groups of six swine with Ad5-vectors containing these genes, challenged 1 day later and euthanized 2 animals from each group prior to (1 day postinoculation [dpi]) and at 1 (2 dpi) and 6 days postchallenge (7 dpi). Blood, skin, and lymphoid tissues were examined for IFN-stimulated gene (ISG) induction and infiltration by innate immune cells. All IFN-inoculated animals had delayed and decreased clinical signs and viremia compared to the controls, and one animal in the IFN-α treated group did not develop disease. At 1 and 2 dpi the groups inoculated with the IFNs had increased numbers of dendritic cells and natural killer cells in the skin and lymph nodes, respectively, as well as increased levels of several ISGs compared to the controls. In particular, all tissues examined from IFN-treated groups had significant upregulation of the chemokine 10-kDa IFN-γ-inducible protein 10, and preferential upregulation of 2′,5′-oligoadenylate synthetase, Mx1, and indoleamine 2,3-dioxygenase. There was also upregulation of monocyte chemotactic protein 1 and macrophage inflammatory protein 3α in the skin. These data suggest that there is a complex interplay between IFN-induced immunomodulatory and antiviral activities in protection of swine against FMDV.
Foot-and-mouth disease virus (FMDV) is the causative agent of one of the most contagious and economically devastating diseases affecting cloven-hoofed livestock worldwide. The virus, a member of the Picornaviridae family, rapidly replicates in the host and spreads via aerosol and by direct contact (3). Foot-and-mouth disease (FMD) is characterized by fever, lameness, lymphopenia, and the appearance of vesicular lesions on the mouth, tongue, nose, feet, and teats (19, 28) and is controlled by inhibition of susceptible animal movement, slaughter of infected animals, and possibly vaccination with an inactivated whole-virus vaccine. However, both the conventional inactivated vaccine and a replication-defective human adenovirus 5 (Ad5) FMDV subunit vaccine that we have recently developed require approximately 7 days to induce full protection in swine and cattle (25, 38, 44). In FMD outbreaks in disease-free countries the induction of rapid protection, prior to the development of vaccine-induced adaptive immunity, is necessary to inhibit or limit disease spread and thus potentially reduce the number of animals that have to be slaughtered.
Interferons (IFNs) are the first line of the host innate immune defense against viral infection (49), and it has been demonstrated that type I and type II IFN (IFN-α/β and IFN-γ, respectively) have antiviral activity against many viruses (5, 29). After the IFN pathway is stimulated transcriptional upregulation of hundreds of effector genes occurs (17, 52). Our group has demonstrated that pretreatment of cell cultures with type I and type II IFNs can dramatically inhibit FMDV replication (12, 14, 37), and at least two type I IFN-stimulated genes (ISGs), double-stranded-RNA-dependent protein kinase (PKR) and 2′,5′-oligoadenylate synthetase (OAS)/RNase L, are involved in this process (12, 16). Furthermore, swine pretreated with an Ad5 vector expressing porcine IFN-α (Ad5-pIFN-α) or porcine IFN-γ (Ad5-CI-pIFN-γ) are sterilely protected when challenged with FMDV 1 day later (13, 37) and protection lasted 3 to 5 days (36). Interestingly, the action of type I IFN in combination with type II IFN can synergistically block virus replication in vivo since swine inoculated with a combination of Ad5-CI-pIFN-α and Ad5-CI-pIFN-γ, at doses that alone do not protect against FMDV, are completely protected against clinical disease and do not develop viremia (37). The protection conferred in IFN-treated pigs correlated with an increase of the 10-kDa IFN-γ-inducible protein 10 (IP-10), indoleamine 2,3-dioxygenase (INDO), and OAS mRNA expression in peripheral blood mononuclear cells (PBMCs), and this upregulation was synergistic in the group of animals treated with the combination of Ad5-CI-pIFN-α and Ad5-CI-pIFN-γ. Nevertheless, this approach has only been partially successful in cattle, since Ad5-pIFN-α-treated animals developed clinical disease after FMDV challenge, although the disease was delayed and less severe compared to control animals (60).
Type I and type II IFNs have some overlapping biologic activities but unique functional roles in the innate and adaptive immune response. Type I IFNs, are primarily responsible for inducing direct antiviral responses in infected cells and do so with more potency than IFN-γ (53). IFN-γ is mainly produced by activated T lymphocytes and natural killer (NK) cells predominantly activating components of the cell-mediated immune system such as cytotoxic T lymphocytes, macrophages, and NK cells, and inducing Th1 differentiation, but it can also display antiviral activity (42). However, it has also been shown that type I IFNs can stimulate dendritic cell (DC) maturation and NK cell activation (56). Furthermore, other cytokines and chemokines, including IP-10 (35, 37), monocyte chemotactic protein 1 (MCP-1) (8), macrophage inflammatory protein 1α (MIP-1α) (31), and MIP-3α (50), that do not have direct antiviral activity but are involved in chemoattraction of various immune cells to the site of infection are also induced by IFNs.
DCs, professional antigen-presenting cells, and NK cells play a key role in the initiation and regulation of the immune response. After priming by pathogen-derived products, their reciprocal interactions may result in a potent activating cross talk that regulates both the quality and intensity of the innate immune response (15). Swine DCs include distinct subsets with specific biological functions (41, 51, 61). Langerhans cells (LCs) are DCs resident in the epithelium, particularly the epidermis. The skin and mucosa are the predilection sites of FMDV replication (3, 28). LCs migrate selectively to MIP-3α (via CCR6) (10), a chemokine that is regulated by the expression of IFN-γ (50) and is secreted by keratinocytes (10). Other IFN-γ-regulated chemokines, including MCP-1, MIP-1α, and IP-10, are involved in epidermal DC maturation (24) and NK cell recruitment and activation (54). Furthermore, upon maturation DCs are able to produce cytokines, including interleukin-12 (IL-12) (47), IL-15 (34), and IL-18 (4), that are involved in NK cell proliferation and activation (46, 57). It has also been shown that murine DCs produce IL-2, and this cytokine can activate NK cells (27). Recently, Pintaric et al. (46) and Toka et al. (57) demonstrated that human IL-2 can activate swine NK cells. Clearly there is a complex interplay between IFN-induced chemokine and cytokine activation of various immune cells and their possible role in protection against FMD.
In the present study we examined in more detail the effects of IFN treatment on induction of protective mechanisms against FMDV. Groups of animals were inoculated with either Ad5-CI-pIFN-α or Ad5-CI-pIFN-γ alone or in combination, challenged with FMDV 1 day later, and two animals from each group were euthanized 1 day postinoculation (dpi) and 1 and 6 days postchallenge (dpc). We examined ISGs in skin, the main site of virus replication, PBMCs, and lymphoid tissues and also evaluated possible immune cell recruitment to the skin and lymph nodes. We found that protection correlated with recruitment of DCs and NK cells to the skin and lymph nodes, respectively, and upregulation of a number of cytokines, some of which have been shown to block FMDV replication in cell culture (12), and chemokines that are involved in chemoattraction of DCs and NK cells. The chemokine IP-10 was significantly upregulated in all tissues examined, whereas MIP-3α, and MCP-1 were mainly induced in the skin. These data suggest that IFN-induced immunomodulatory, as well as antiviral, activities may be involved in protection of swine against FMD.
Human 293 cells were used to generate, grow, and titer recombinant Ad5s (26, 38). Baby hamster kidney cells (BHK-21, clone 13) were used to measure FMDV titers in plaque assays. IBRS-2 cells (swine kidney) were used to measure antiviral activity in plasma from inoculated animals by a plaque reduction assay (12). The recombinant viruses Ad5-CI-pIFN-α and Ad5-CI-pIFN-γ were constructed as described by Moraes et al. (37). FMDV serotype A24 (strain Cruzeiro, Brazil, 1955) was isolated from vesicular lesions of an infected bovine.
Twenty-eight circovirus-free Yorkshire gilts (ca. 40 to 60 lb) were housed in the secure disease agent isolation facilities at the Plum Island Animal Disease Center according to a protocol approved by the Institutional Animal Care and Use Committee. In this experiment, the animals were divided into five groups containing six animals per group (unless otherwise stated), and each group was housed in a separate room: one group inoculated with 2 ml of 109 PFU Ad5-CI-pIFN-α, one group inoculated with 2 ml of 1010 PFU Ad5-CI-pIFN-γ, one group inoculated with 2 ml of 109 PFU Ad5-CI-pIFN-α and 1010 PFU Ad5-CI-pIFN-γ, and two control groups, one inoculated with 2 ml of 1010 PFU Ad5-Blue (four animals; two animals in this group were euthanized prior to the start of the experiment because of health reasons) and one inoculated with 2 ml of phosphate-buffered saline (PBS). The animals were monitored clinically for adverse effects from Ad5-CI-pIFN-α and Ad5-CI-pIFN-γ administration, including fever and lethargy. One day postinoculation two animals in each group were slaughtered, and the remaining animals were challenged with 3.5 × 105 PFU of FMDV serotype A24, at two sites in the heel bulb of the right rear foot. One day postchallenge (or 2 dpi) another two animals from each group were slaughtered and the remainder of the pigs were monitored for clinical signs until 6 dpc (7 dpi) when they were also slaughtered. Lesion scores of the animals were determined as described in Moraes et al. (37).
Blood was obtained daily to assay for antiviral activity and the presence of pIFN-α and pIFN-γ by enzyme-linked immunosorbent assay (ELISA) and to extract PBMC (see below). Blood and nasal swab specimens were collected daily for the first 7 days after challenge for those animals that were kept alive until the end of the experiment. After euthanasia and exsanguination of the animals at set time points, tissue samples (three different anatomical locations of skin, as well as inguinal and popliteal lymph nodes, and tonsils) were taken and divided in two sets. One set was immersed in RNA later (Ambion, Austin, TX) and stored at −70°C for RNA extraction, and the other was snap-frozen in liquid nitrogen-chilled isopentane and placed in a cryomold embedded in optimal cutting temperature compound (OCT; Sakura Finetek, Torrance, CA) and stored at −70°C for analysis by histopathology and immunohistochemistry (IHC).
Viremia was determined by a standard plaque assay of total blood on BHK-21 cells. Plasma was obtained by centrifugation of heparinized blood at 2,500 rpm for 10 min and examined for antiviral activity by a plaque reduction assay (13). The level of pIFN-α and pIFN-γ in plasma was determined by ELISA (37). Nasal swab specimens were obtained on the day of challenge and daily for 7 days after challenge, and virus was isolated from the swab samples as previously described (37).
Frozen 4-μm sections were mounted onto electrostatically charged glass slides (SuperFrost Plus; Fisher Scientific, Worcester, MA) and fixed for 10 min in acetone at −20°C. Thereafter, the slides were kept at −70°C for up to 8 weeks, until they were stained. For immunostaining, the slides were incubated with the following primary antibodies: mouse monoclonal antibody (MAb) anti-porcine SLA-II (AbD Serotec, Raleigh, NC), mouse MAb anti-porcine CD172 (Southern Biotech, Birmingham, AL), mouse MAb anti-porcine CD1 (Southern Biotech), mouse anti-human iNOS (BD Bioscience, Franklin Lakes, NJ), and mouse MAb anti-human Mx1 (kindly provided by Otto Haller, University of Freiburg), which labels porcine Mx1 protein (30). The bound primary antibodies were detected by the avidin-biotin-peroxidase complex technique (Vectastain ABC Kit Elite; Vector, Burlingame, CA) according to the manufacturer's instructions and developed either with 3,3′-diaminobenzidine (Dako, Glostrup, Denmark) or Fast Red TR/naphthol (Sigma, St. Louis, MO). Slides were counterstained with Harry's hematoxylin and coverslipped by using routine methods. To control the specificity of antibody binding, a duplicate negative control serial section treated with nonspecific primary antibody was used. The positive cells were counted in 10 consecutive fields of 1 mm2 each per slide. Means and standard deviations of the numbers of labeled cells were calculated, and differences were tested for significance.
Four percent paraformaldehyde fixed inguinal and popliteal lymph node and skin frozen sections were pretreated with 0.1% saponin (wt/vol) in PBS, blocked with blocking buffer (PBS, 2% fetal bovine serum, 10% porcine serum), and then incubated overnight at 4°C with the primary antibodies mouse MAb anti-porcine CD3 (Southern Biotech), anti-porcine-CD8 (Southern Biotech), rat MAb anti-porcine CD2 (Antigenix America, Huntington Station, NY), mouse anti-porcine CD1 (Southern Biotech), and mouse anti-human CD207/Langerin (Denritics, Dardilly, France). Alexa Fluor 488, Alexa Fluor 594, and Alexa Fluor 647 (Molecular Probes/Invitrogen, Carlsbad, CA)-conjugated secondary antibodies were used for detection. Nuclei were visualized by DAPI (4′,6′-diamidino-2-phenylindole) staining included in ProLong Gold Antifade mounting medium (Invitrogen). Sections were examined in a Nikon 90i fluorescence microscope, and the images were taken with a DSQi1Mc digital camera and NIS-Elements Software, version 3.0 (Nikon, Melville, NY).
The number of positive cells for CD2 and CD8 and negative for CD3 primary antibodies were counted in 10 consecutive fields of 34 mm2 each per slide, and the percentage of these cells was calculated in relation to the total number of cells per field. Mean and standard deviation of the number of labeled cells were calculated, and differences were tested for significance.
Expression of several ISGs (see Table S1 in the supplemental material) was analyzed in PBMCs and different tissues isolated from IFN-treated animals. PBMCs were purified from heparinized blood with Lymphoprep (Axis-Shield, Oslo, Norway). RNA was extracted from ~107 cells by utilizing an RNeasy miniprep kit (Qiagen, Valencia, CA). RNA from skin was extracted by using an RNeasy miniprep kit with a modified protocol. Then, 4- to 8-mm3 sections of skin were pulverized by cooling in liquid nitrogen and crushing. Pulverized tissue was transferred to a tube with RLT buffer (Qiagen), with β-mercaptoethanol (10 μl per ml of RLT buffer) and proteinase K (20 mg/ml), incubated for 10 min at 55°C, and then centrifuged at 10,000 × g. Supernatants were mixed with 100% ethanol and passed through an RNeasy minicolumn (Qiagen), and the extraction was completed according to the manufacturer's protocol. Lymphoid, spleen, and tonsils tissue RNA was extracted by mechanical homogenization with a Tissuemiser (Fisher Scientific) in RLT with β-mercaptoethanol. After a homogenate was obtained, the solution was passed through Qiashredder (Qiagen), and RNA was obtained by using an RNeasy extraction kit according to the manufacturer's instructions.
A quantitative real-time reverse transcription-PCR (RT-PCR) was used to evaluate the mRNA levels of several ISGs (16). 18S rRNA or porcine GAPDH (glyceraldehyde-3-phosphate dehydrogenase) were used as the internal control to normalize the values for each sample. The sequences of primers and probes that were used are listed in Table S1 in the supplemental material. Reactions were performed in an ABI Prism 7000 sequence detection system (Applied Biosystems). Relative mRNA levels were determined by comparative cycle threshold analysis (user bulletin 2; Applied Biosystems) utilizing as a reference the samples at 0 dpi from the control groups. We only considered genes upregulated if there was a twofold or greater induction in both animals.
Data handling, analysis, and graphic representation was performed by using Prism 2.01 (GraphPad Software, San Diego, CA) or Microsoft Excel. Statistical differences were determined by using a Student t test (*, P < 0.05; **, P < 0.01; ***, P < 0.005).
Based on previous experiments (13, 37), we inoculated animals with 109 PFU of Ad5-CI-pIFN-α, 1010 PFU of Ad5-CI-pIFN-γ, or a combination of 109 PFU of Ad5-CI-pIFN-α and 1010 PFU of Ad5-CI-pIFN-γ, respectively. We included two control groups, one inoculated with 1010 PFU of Ad5-Blue and one inoculated with PBS. In previous studies these doses of IFN were sufficient to sterilely protect all animals when challenged with 105 PFU of FMDV A24 (13, 38). In the present study, animals were challenged with a higher dose (3.5 × 105 PFU of FMDV A24). All of the control animals inoculated with PBS developed viremia at 1 dpc and clinical signs starting at 2 dpc with a significant lesion score by 3 to 4 dpc (Fig. (Fig.1a).1a). Similarly, the other control animals inoculated with 1010 PFU of Ad5-Blue also developed viremia by 1 dpc (data not shown).
One of the IFN-α-treated pigs never showed clinical signs and only developed a low level viremia at 6 dpc, while the other animal showed a significant delay in both the onset of viremia and clinical disease. i.e., at 4 and 5 dpc, respectively, with a lower lesion score and 1,000- to 10,000-fold lower levels of viremia compared to the controls (Fig. (Fig.1b).1b). Animals treated with Ad5-CI-pIFN-γ also showed a delay in the onset of clinical signs, i.e., 3 or 5 dpc, and in the development of viremia, i.e., 2 or 3 dpc, and the level of viremia was 10- to 20-fold lower than the controls (Fig. (Fig.1c).1c). Finally, animals that were treated with the combination of Ad5 expressing IFN-α and IFN-γ showed a delay in the onset of the disease comparable to what was observed in the animals inoculated with Ad5-CI-pIFN-α and a lower lesion score compared to the controls and the IFN-γ-inoculated group. However, the levels of viremia were ~10-fold higher than in the IFN-α treated group (Fig. (Fig.1d).1d). Although viremia correlated with FMDV in nasal swabs, virus shedding was significantly reduced and delayed in all of the animals treated with IFNs, especially in the Ad5-CI-pIFN-α group, which did not have detectable virus in nasal swabs until 5 dpc (data not shown).
When we analyzed the skin of these animals we observed that, 1 day after the challenge, all animals from the control groups had ballooning degeneration and increased cytoplasmic eosinophilic staining of the cells in the stratum spinosum with acantholysis and areas of necrosis at the site of inoculation, where we could identify the start of vesicle formation (Fig. (Fig.2i).2i). Interestingly, in skin from the heel bulb of a foot other than the one inoculated (skin NIS) we also observed vacuolization of keratinocytes of the stratum spinosum of the epidermis and lymphocytic infiltration (Fig. (Fig.2e),2e), although macroscopically we did not detect any lesions at this time point. On the other hand, none of the IFN-treated animals showed any microscopic lesions in skin at a site other than the inoculation site (Fig. 2f to h) 1 day after challenge. However, at this time point, IFN-treated animals showed microscopic lesions in the skin at the inoculation site, including cellular vacuolization and karyopyknosis of keratinocytes and the presence of acantholysis in the epidermis (Fig. 2j to l). One day after challenge we could detect the presence of virus by real-time RT-PCR at the inoculation site in all of the control animals, low levels of virus in both animals in the IFN combination group, and in one of two animals in the IFN-α and IFN-γ groups. However, only the control animals and the IFN combination group showed a virus-positive signal in skin at a different location than the inoculation site at 1 dpc (see Table S2 in the supplemental material).
All of the animals were assayed for the presence of IFN-α, IFN-γ, and antiviral activity in plasma. Only plasma from the groups inoculated with Ad5-CI-pIFN-α or the combination had antiviral activity and the level was the highest in the Ad5-CI-pIFN-α group (data not shown). The levels of IFN-α detected in animals 1 day after treatment with Ad5-CI-pIFN-α were low (ca. 3.5 × 103 to 5.3 × 103 pg/ml) (Fig. (Fig.3)3) compared to previous experiments in which 1.1 × 105 to 3.0 × 105 pg of protein/ml was detected in the plasma of swine treated with the same dose of Ad5-pIFN-α (28). Surprisingly, animals from the combined treatment group expressed lower levels of IFN-α in plasma than the animals inoculated with Ad5-CI-pIFN-α alone (Fig. (Fig.3).3). Animals inoculated with Ad5-CI-pIFN-γ did not express significant levels of IFN-α protein in plasma (Fig. (Fig.3)3) or detectable levels of IFN-γ (data not shown). However, the levels of IFN detected seemed to be sufficient to produce a delay in the onset of the disease in a dose-dependent manner.
As mentioned above, the level of IFN-γ was undetectable in plasma and IFN-α levels were very low compared to previous experiments. Therefore, we were concerned that we would not be able to detect the upregulation of ISGs induced by the IFN treatment. In order to confirm that the animals inoculated with Ad5-IFNs were systemically expressing and responding to the cytokines, we performed IHC in different tissues, including skin and various lymphoid organs, using iNOS and Mx1 as markers of IFN-γ and -α stimulation, respectively.
One day after inoculation (0 dpc), we observed the presence of positive cells for iNOS in the tonsils of all treated animals compared to controls (Fig. (Fig.44 compare panels e, i, and m to panel a). At the same time points, skin of treated animals also showed a diffuse iNOS positive signal in the stratum basale of the epidermis (Fig. 4f, j, and n). The tonsils of Ad5-CI-pIFN-α-treated animals, alone or in combination with Ad5-CI-pIFN-γ, were also reactive against anti-Mx1 MAb compared to control animals or animals inoculated with Ad5-CI-pIFN-γ (Fig. (Fig.4,4, compare panels g and o to panels c and k). In animals treated with Ad5-CI-pIFN-α the positive reaction was not only localized in the epithelium of the tonsils but also in the lymphoid follicles (Fig. (Fig.4g).4g). Skin of animals inoculated with Ad5-CI-pIFN-α showed an increase in the Mx1 signal localized to the stratum basale and basal cells of the stratum spinosum (Fig. (Fig.4h).4h). However, animals inoculated with Ad5-CI-pIFN-γ did not show any positive staining for Mx1 (Fig. (Fig.4l)4l) and resembled the skin of control animals (Fig. (Fig.4d).4d). Although Mx1 is not directly involved in the inhibition of FMDV replication (14), and we currently do not know whether iNOS has anti-FMDV activity, distribution of these proteins in the skin is indicative of systemic antiviral responses induced by type I and II IFNs.
The major FMDV replication organ and the main site of macroscopic lesions in FMD is the skin (3). Although we have reported that either IFN-α or IFN-γ treatment can protect against FMDV replication, and their action is synergistic when inoculated in combination (13, 37), we have never examined the possible role of various immune cells in the IFN-induced FMDV-protective response. Swine DCs, including skin DCs, are CD1+, SLAII+ and CD172+ (6, 51) and skin DCs also express Langerin (41). To study the possible infiltration of DCs in the epidermis after IFN treatment, skin cross-sections were stained using different antibody combinations: CD1/SLAII, CD1/CD172, CD172/SLAII, or each antibody separately. Serial skin sections from a naive pig demonstrated that epidermal DCs, mainly localized in the stratum basale of the epidermis, were CD1+, SLAII+, and CD172+. To facilitate the analysis of this cell population, we examined IHC staining against the CD1 cellular marker (Fig. (Fig.5),5), since the three antibodies hybridized to the same cell in the epidermis (data not shown). Skin from Ad5-CI-pIFN-α-treated animals had a higher number of CD1+ cells at the time of challenge (1 dpi/0 dpc) (Fig. (Fig.5a,5a, panel 3) compared to control animals (either the PBS [Fig. [Fig.5a,5a, panel 1] or the Ad5-Blue-inoculated groups [data not shown]). LCs of animals inoculated with Ad5-CI-pIFN-α not only increased in number but also showed more dendrites, and they were spread throughout the stratum basale and stratum spinosum (Fig. (Fig.5a,5a, panels 3 and 4, note the arrowheads). Similar findings were observed in animals inoculated with Ad5-CI-pIFN-γ at either 1 or 2 days after the treatment (0 dpc or 1 dpc) (Fig. (Fig.5a,5a, panels 5 and 6), although the number of CD1+ cells was slightly lower. Skin from animals treated with the combination of Ad5-CI-pIFN-α and Ad5-CI-pIFN-γ resembled the pattern of skin DC distribution in the two other treated groups (Fig. (Fig.5a,5a, panels 7 and 8). When we double stained for CD1 and Langerin, we observed that all CD1+ cells from the epidermis were Langerin-positive cells (Fig. (Fig.5b).5b). In order to analyze the significance of these findings, CD1+ cells from 10 consecutive 1-mm2 fields were counted in the skin samples of each animal examined (see Materials and Methods). The difference between the control group and the IFN-treated animals was significant (P < 0.05) (Fig. (Fig.5c).5c). Interestingly, ex vivo skin DCs from animals inoculated with Ad5-CI-pIFN-α and Ad5-CI-pIFN-γ showed upregulated expression of CD80/CD86, molecules that are directly correlated with the maturation of DCs (data not shown).
To determine whether NK cells might be infiltrating the skin, we also examined skin utilizing triple IFA. Porcine NK cells are included in the CD2+/CD8+/CD3− cell compartment (62). We could not detect any CD2+/CD8+/CD3− cells infiltrating the epidermis, nor in proximal areas of the dermis prior to (1 dpi/0 dpc) and after (2 dpi/1 dpc) challenge in IFN-treated animals (data not shown), suggesting that direct interaction between DCs and NK cells induced by IFN treatment, if any, does not occur in the skin.
Although we could not detect the presence of NK cells in the skin, we hypothesized that IFN treatment could induce upregulation of cytokine production by skin DCs that might stimulate the recruitment, proliferation or activation of NK cells into the draining lymph nodes. To test this hypothesis, we performed CD2/CD3/CD8 IFA in inguinal and popliteal lymph nodes (Fig. 6a and b). The majority of cells distributed in the paracortex of the inguinal and popliteal lymph nodes, either in the central or peripheral deep cortical units, were CD3+ (Fig. (Fig.6a)6a) but some CD2+/CD8+/CD3− cells (NK cells) could be observed (Fig. (Fig.6a,6a, asterisks). Moreover, in agreement with previous data published by Ferlazzo et al. (22), naive animals showed a low percentage of NK cells present in peripheral lymph nodes, mainly colocalized with T cells in the paracortex (data not shown) (21). Quantification of the percentage of NK cells per 34-mm2 field in inguinal and popliteal draining lymph nodes of IFN-treated animals and control pigs showed a statistically significant increase in the number of NK cells in all treated animals compared to the controls at 1 (0 dpc) and 2 dpi (1 dpc) (Fig. (Fig.6b6b).
Upon activation, DCs secrete a number of cytokines, including IL-15, IL-18, and IL-12, which upregulate NK cell cytotoxicity (34). As explained above, we hypothesized that the increase of NK cells in lymph nodes would be a consequence of secretion of proinflammatory cytokines by DCs that could result in recruitment and proliferation/activation of NK cells in lymph nodes. To test this hypothesis, we examined skin and lymph nodes for the upregulation of IL-15, IL-18, and IL-12 mRNA and their receptors (IL-15R, IL-18R, and IL-12R) after IFN treatment. In the skin we found that IL-18 was induced in all IFN-treated groups by 1 dpi, but there was no or reduced induction by 2 dpi (Fig. (Fig.7a).7a). In contrast, IL-15, which only slightly increased in the IFN-α/γ group by 1 dpi, showed higher expression levels in the IFN-α group by 2 dpi concomitant with enhanced IL-15R levels. The slightly increased levels of IL-15 and IL-15R were maintained in the IFN-α/γ group at 0 and 1 dpi (Fig. (Fig.7a).7a). At the same time, there was an upregulation of IL-12R in all IFN-treated groups at 1 dpi, possibly induced by IL-15 or IL-18 (Fig, 7a), as previously described (20, 40). In correlation with the upregulation of mRNA levels of IL-15 and IL-18 in skin we observed upregulation of IL-15, IL-18, and IL-12 and IL-15R and IL-18R in the popliteal lymph node in the group inoculated with IFN-α and IL-15 and IL-15R in the combination group at 1 dpi, but the induction decreased by 2 dpi (Fig. (Fig.7b).7b). In the inguinal lymph node we observed more consistent changes by 2 dpi, with upregulation of IL-18R in the IFN-α group; IL-18R and IL-12R in the IFN-γ group; and IL-15R, IL-18R, and IL-12R in the IFN-α/γ group (Fig. (Fig.7c).7c). Although there is a variable effect among IFN treatments, the correlation between the upregulation of these cytokines and the increased presence of NK cells in the lymph nodes suggest that NK cell proliferation might be induced by these cytokines.
We also examined the effect of IFN treatment on the induction of a number of ISGs, including genes with direct antiviral activity, as well as chemokines that are involved in attraction and activation of various immune cells (Fig. (Fig.88 and and9).9). Previously, we demonstrated that inoculation of pigs with Ad5-CI-pIFN-α in combination with Ad5-CI-pIFN-γ induced the synergistic upregulation of INDO and IP-10 in PBMCs (37). In the present study we observed similar results in PBMCs at both 1 dpi (Fig. (Fig.88 and and9a)9a) and 2 dpi (Fig. (Fig.88 and and9b).9b). Mx1 and OAS were upregulated in PBMCs upon IFN-α treatment, but there was also upregulation of INDO in the IFN-γ group (Fig. (Fig.8).8). In the inguinal lymph node we observed consistent upregulation of Mx1 and OAS induced by IFN-α, and synergistic upregulation of INDO by the combination of type I and II IFNs (Fig. (Fig.8).8). The popliteal lymph node showed a similar upregulation pattern as in the inguinal lymph node, but in addition there was an upregulation of iNOS (Fig. (Fig.8).8). Interestingly, we observed a slight upregulation of the chemokines, MIP-3α and MIP-1α in the inguinal and popliteal lymph nodes, respectively, which were not upregulated in the PBMCs (Fig. (Fig.8).8). In the skin, the effect of mRNA upregulation of chemokines involved in maturation/chemoattraction of DCs was evident, particularly in the heel bulb (skin IS and skin NIS), with MCP-1 upregulated by IFN treatment and MIP-3α synergistically upregulated at both 1 dpi and 2 dpi (Fig. (Fig.8).8). At the same time points skin showed upregulation of Mx1 and OAS by IFN-α and synergistic upregulation of INDO by the combination treatment at 1 and 2 dpi (Fig. (Fig.88).
It has been shown that after IFN treatment, induced IP-10 has several roles in the innate immune response, including chemoattraction of NK cells and DCs (32, 58). One day after treatment, IP-10 was upregulated by type I and type II IFN or the combination in lymph nodes and all anatomical locations of skin examined and was synergistically upregulated in PBMCs (Fig. (Fig.9a),9a), as previously described (37). Two days after treatment there was a similar induction of IP-10 in the IFN-α-treated group and synergistic upregulation in PBMCs (Fig. (Fig.9b).9b). However, induction of IP-10 was reduced in lymph nodes from IFN-γ- and combination-treated animals compared to 1 dpi (Fig. (Fig.9b).9b). At 6 days after challenge all control animals showed several genes induced in all organs examined but predominantly in the skin, probably due to the effect of viral replication (Fig. (Fig.88 and and9c).9c). Interestingly, IP-10 was never upregulated in the control animals in any organ examined, but it was upregulated in all IFN-treated animals, with synergistic effects in PBMC and the popliteal lymph node at 7 dpi (Fig. (Fig.9c).9c). These results suggest that the expression of IP-10 in response to IFN treatment, and its correlation with the increase in the number of cells from the innate immune system (DCs and NK cells) in the skin and lymph nodes, respectively, might be involved in the tissue-specific mechanism of protection against FMDV conferred by IFNs.
In this study we have attempted to gain a comprehensive understanding of the IFN-induced mechanisms that result in the rapid protection of swine against FMDV challenge. Previously, we described that protection of animals inoculated with Ad5-pIFN-α correlated with the level of IFN-α expression in plasma (13), while in animals inoculated with Ad5-pIFN-γ alone or with the combination of both IFNs protection correlated with upregulation in the PBMC of ISGs that have either direct antiviral activity, INDO and OAS, or immunomodulatory roles, IP-10 (37). In the present study, we show for the first time proliferation/migration of immune cells (DCs and NK) to skin and lymphoid organs in response to IFN treatment. In addition, we confirm the IFN-induced upregulation in the PBMC of several ISGs involved in antiviral responses, as previously described, but we also show tissue-specific effects of IFN treatment in the skin and lymph nodes with increased expression of cytokines and chemokines involved in activation of the innate immune system.
Ad5 vectors containing porcine type I and II IFN alone or in combination were used to treat swine. Two animals in each group were euthanized prior to and at 1 and 6 dpc. Clinically, all animals in the IFN-inoculated groups had delayed and less severe disease compared to the control animals, and one animal in the IFN-α-treated group had no clinical disease at 6 dpc. These results appear to differ from our previous data since a dose of 109 PFU of Ad5-pIFN-α completely protected swine from FMDV challenge 1 to 3 days after administration (13, 36). However, in these earlier experiments the level of IFN-α expression in protected swine was 11 to 30,000 pg/ml, which is considerably higher than the ~5,000 pg/ml expressed in the Ad5-CI-pIFN-α-inoculated group in this experiment. At this lower level of IFN-α expression we previously found that swine have delayed and less severe clinical disease after challenge, similar to our present experiment. Furthermore, in this experiment the FMDV challenge dose is ~20-fold higher than recommended by the OIE and 3.5-fold higher than used previously (37). The relatively low levels of IFN-α detected in this experiment suggested that the Ad5 vector was not expressing the transgene as efficiently possibly because of the use of a cytomegalovirus (CMV) promoter with an enhancer (CI promoter ) compared to only the CMV promoter previously used (13, 36) or the use of a higher passage level of the Ad5 vectors, compared to the earlier study (13), may contain higher amounts of WT Ad5 lacking the pIFN genes.
Since the level of IFN-α was low and we could not detect IFN-γ expression in plasma after Ad5-CI-pIFN-γ administration, we examined whether ISG products could be detected. We demonstrated systemic expression of proteins induced by type I and type II IFNs. The detection of Mx1 and iNOS in tonsils and skin of IFN-treated animals indicates that, although expression of the delivered IFNs by Ad5 is low, especially IFN-γ, it is still sufficient to induce ISG expression.
We found that all animals inoculated with IFN have a statistically significant increase in the number of Langerhans DCs in the skin and NK cells in the lymph nodes. These changes correlated with (i) the upregulation of mRNA levels of IL-15 and IL-18 in the skin, two cytokines that have been shown to be produced by activated DCs (4, 34) and that are involved in porcine NK cell proliferation and activation (46, 57); (ii) a slight increase of IL-15, IL-18, and IL-12 in the popliteal lymph node at 1 dpi and a upregulation of IL-15R and IL-18R at 1 dpi; (iii) the increase in the mRNA levels of IL-18R in the inguinal lymph node, the upstream draining lymph node, only at 2 dpi in all IFN-treated groups. It is important to mention the absence or low levels of IL-12 upregulation in these tissues, a cytokine involved in activation of NK cells (7). However, there is evidence that for other viral infections, such as lymphocytic choriomeningitis virus, there is no induction of IL-12-dependent IFN-γ production by NK cells (43). In the case of FMDV infection, this might be compensated for by the observed upregulation of IL-18, which has two IL-12-independent effects on NK cells, the stimulation of IFN-γ and production and enhancement of perforin-dependent cytotoxicity (2). On the other hand, IL-15, which was also upregulated in our study, stimulates NK cell activity against virus-infected cells (1), which may be physiologically relevant in the innate immune response during the course of FMDV infection. Whether the increase in the number of LCs in skin and NK cells in the lymph nodes is directly correlated with the upregulation of proinflammatory cytokines or whether these factors play a role in protection against FMDV requires further investigation. However, it has been demonstrated that NK cells are rapidly recruited to lymph nodes upon DC activation (33). Furthermore, preliminary data indicate that there is maturation of skin DCs in IFN-α- and IFN-γ-treated animals compared to controls, as measured by upregulation of CD80/CD86 molecules. In addition, preliminary studies showed that NK cell killing activity is activated in IFN-α- and IFN-α/γ-treated animals at 1 to 2 dpi, while there is a delay and a lower level of killing activity in IFN-γ-treated animals. Hence, the pathways involving NK cell cytokine responses due to possible cross talk with activated/mature DCs and the resultant induction of an FMDV-antiviral state are intricate and perhaps even more complex than currently appreciated. Nevertheless, the data presented here have yielded important findings that will presumably help focus our examination of the role of NK cell responses in protection against FMDV infection.
Concomitant with these effects, IFN treatment generated a significant upregulation of the chemokine IP-10 in PBMCs, lymph nodes, and skin. This chemokine is involved in the recruitment, proliferation, and activation of NK cells (54) and has been demonstrated to have a protective effect in mice infected with a number of viruses, including mouse hepatitis virus (58), coxsackievirus B3 (62), dengue virus (11), and respiratory syncytial virus (32), but not in Theiler's murine encephalomyelitis virus (TMEV) (59). It has also been shown that IP-10 is rapidly but transiently induced by Ad5 vectors delivered systemically or within the liver or lung (39, 54). Although we have noted that our control vector, Ad5-Blue, induces a rapid upregulation of IFN-α protein by 4 h after intramuscular administration (36) and induction of a number of cytokines/chemokines, including IP-10, by 6 h postinfection, these effects were not detectable by 24 h postinfection (36) (Fig. (Fig.88 and and9).9). Furthermore, we have shown that Ad5-Blue-inoculated swine are not protected from challenge with FMDV (13, 45). Thus, the rapid but transient induction of numerous genes by the Ad5 vector does not appear to be involved in the protection of swine against FMDV induced by Ad5-pIFN.
The chemokines MIP-3α and MCP-1 were also upregulated after IFN treatment, mainly in the skin. These chemokines are involved in the recruitment and proliferation of DCs and NK cells and have been shown to be coordinately induced after murine cytomegalovirus infection (48). In particular, MIP-3α has been described to control the migration of skin DCs (9). Therefore, the upregulation of these chemokines may be involved in the observed increase of DCs in IFN-treated animals which may have a role in the protection of the animals against FMDV.
Interestingly, we did not detect upregulation of IP-10 in control animals challenged with FMDV at any time or in any of the samples examined even at 7 dpi/6 dpc when numerous other genes, including MCP-1, MIP-3α, and MIP-1α, were induced. These results are in contrast to the significant induction of IP-10 observed in mice infected with a variety of viruses, including dengue virus (11), respiratory syncytial virus (32), and mouse hepatitis virus (58), as well as two members of the picornavirus family, coxsackievirus B3 and TMEV (59, 62). In a recent study, the cytokine IL-10 has been proposed as an immunosuppressive factor involved in the regulation of the adaptive response to FMDV infection (18). Increased amounts of IL-10 inhibit the action of monocytes, macrophages, and NK cells during the immune response to viral infection and inhibit the synthesis of proinflammatory cytokines (23, 55). Furthermore, it has been shown that IL-10 can suppress IP-10 gene transcription by inhibiting lipopolysaccharide-induced synthesis of type I IFNs (55). It will be interesting to examine whether the induction of IL-10 during FMDV infection plays a role in the suppression of IP-10 transcription.
Based on our results, we hypothesize that one or more of the above chemokines (or others not yet examined) may be involved in the recruitment/proliferation and activation of DCs and NK cells in IFN-treated animals and that the maturation of these cells has a role in the control of FMDV replication. However, at this time, we cannot confirm this hypothesis nor can we speculate about the sequence of events leading to the increased numbers of DCs and NK cells. To follow these events in more detail we plan to pursue several approaches, including examining earlier time points after Ad5-IFN administration to swine and utilizing model systems that will allow us to demonstrate direct involvement of various cytokines/chemokines in inhibition of FMDV replication.
It is possible that both the IFN-induced antiviral and immunomodulatory activities have roles in protection of swine against FMD. Therefore, it will be of interest to identify the specific players in these processes with the goal of developing a more robust strategy to induce rapid protection in both swine and cattle.
This research was supported in part by the Plum Island Animal Disease Research Participation Program administered by the Oak Ridge Institute for Science and Education through an interagency agreement between the U.S. Department of Energy and the U.S. Department of Agriculture (appointment of Fayna Diaz-San Segundo and Camila C. A. Dias), by CRIS project number 1940-32000-053-00D, ARS, USDA (M. J. Grubman and T. de los Santos) and by reimbursable agreement 60-1940-7-047 with the Department of Homeland Security (M. J. Grubman).
We thank Noemi Sevilla, CISA-INIA, Valdeolmos, Madrid, Spain, for helpful discussions and suggestions. We also thank Harry Dawson, USDA, ARS, Nutrient Requirements and Function Laboratory, Beltsville, MD, for creating the PIN library with the recommendations of RT-PCR conditions for measuring swine gene expression. Finally, we thank the animal care staff at the Plum Island Animal Disease Center for their professional support and assistance.
Published ahead of print on 2 December 2009.
†Supplemental material for this article may be found at http://jvi.asm.org/.