Introduction
Bluetongue (BT) is a haemorrhagic disease of ruminants, which is caused by Bluetongue virus (BTV), a member of the orbivirus genus within the family
Reoviridae. BTV consists of seven structural proteins (VP1 - VP7) organised in a double-capsid structure. Two of the seven proteins (VP2, VP5) make up the outer capsid of the virus and the remaining five proteins (VP1, VP3, VP4, VP6 and VP7) are located in the inner capsid or core together with the double-stranded RNA genome consisting of ten segments. Three non-structural proteins that are not associated with the virion are also expressed (NS1-3) in the infected cells. To date, 24 different serotypes have been officially recognised and an additional serotype has recently been identified by sequence analysis [
1,
2]. BTV is an insect-borne virus, which is transmitted from animal to animal by blood feeding midges (
Culicoides spps) and has been endemic mainly in tropical and sub-tropical countries. Although BTV infects a wide variety of domestic and wild ruminants, classically, BT is considered predominantly as a sheep disease and indeed BTV infection in certain breeds of sheep may cause severe morbidity and high mortality. In recent years, BTV has emerged in northern Europe and re-emerged in the Mediterranean basin causing severe disease and high mortality in naïve ruminant populations. Outbreaks have affected not only sheep, but also other livestock such as cattle and goats [
3,
4]. The clinical symptoms of BTV infection are thought to be associated with virus-induced vascular injury and endothelial cell-derived inflammatory responses [
5‐
8] and apoptosis [
9], although host responses at a cellular level that result in the pathogenesis caused by BTV infection have not been investigated thoroughly.
BTV induces apoptosis both in cultured cells and in target tissues
in vivo and one current hypothesis is that apoptosis plays a major role in the pathogenesis of BTV infection [
10‐
12]. Virus infected cells that undergo apoptosis show highly characteristic morphological changes, including shrinkage, blebbing of the plasma membrane, chromatin condensation and DNA fragmentation. In a previous report, we showed that extracellular treatment with a combination of both the cellular receptor binding protein VP2 and the cell penetration protein VP5, is sufficient to trigger apoptosis through the activation of executioner caspase-3 [
11]. Subsequent to this report, others have reported that both the extrinsic and intrinsic pathways are involved in the induction of apoptosis by BTV [
9,
10,
12]. However, the results in these reports have contradictory conclusions, particularly in relation to caspase-8 activation. While Li
et al.[
10] reported that BTV infection does not cause caspase-8 cleavage [
10], a subsequent publication by others presented the cleavage data of caspase-8 [
12]. Further, the interrelationship between the intrinsic and extrinsic pathways in triggering apoptosis has not been investigated.
Previously, we have also identified the translocation of NF-κB into the nucleus from cytoplasm during BTV infection of mammalian cells and we had postulated, based on the finding with reovirus [
13] that NF-κB activation by BTV infection was involved in induction of cellular apoptosis [
11]. Activation of NF-κB by a viral infection promotes the expression of a variety of genes that are involved either in regulating the host survival immune responses or in apoptosis. However, certain virus infection such as African swine fever virus inhibits NF-κB activation, which results in enhancement of virus replication and thereby contributing to virus-induced pathogenesis [
14].
Activation of NF-κB is generally a rapid response to an inducer, including virus infection [
15]. The NF-κB exists as a heterodimer which is sequestered in the cytosol of unstimulated cells via non-covalent interactions with a class of inhibitor proteins, called IκBs (
viz., IκBα, β, γ etc; [
16]). These inhibitor proteins mask the nuclear localisation signal of NF-κB. Signals that induce NF-κB activity cause the phosphorylation of IκBs, their dissociation and subsequent degradation, allowing NF-κB proteins to enter the nucleus and induce gene expression. The specific phophorylation and degradation of the IκBs determines whether the activation of NF-κB is a rapid and transient response. In this report, we investigated the activation and role of NF-κB during BTV infection, demonstrating that it has a role in initiating an antiviral environment as a part of the innate immune response.
Another mechanism of the innate immune system in response to virus infection is the activation of interferon regulatory factors (IRF) that results in the secretion of antiviral cytokines, particularly, interferon (IFN). Since BTV induces strong cytokine responses, BTV may also trigger innate immune pathways via IRF that are responsible for regulating cytokine production [
17‐
19]. In particular, IRF-3 is ubiquitously expressed and accumulates in the cytosol to enable a rapid response to viral infection and up-regulate the type 1 IRF [
18]. Although the induction of IRF in the other members of the family
Reoviridae has been documented, to date there is no published report on IRF activities in BTV infection.
In this study, we have examined the interrelationship between the intrinsic and extrinsic pathways of apoptosis in BTV infected cells and demonstrated that the intrinsic and extrinsic apoptotic pathways are independently triggered, and a second executioner caspase, caspase-7, is activated. The role of apoptosis or necrosis in the development of BTV pathogenesis is currently not determined, however, it has been suggested by some that it is due to cell necrosis [
6] while others believe that cell apoptosis is the major cause for the disease [
9]. The data that we obtained here demonstrates that BTV infection triggers strong apoptotic response indicating that it has some role in the disease. Since translocation of high mobility box group-1 (HMBG-1) from nucleus to cytoplasm is a distinct feature of necrosis, we examined if HMBG-1 was translocated during BTV infection [
20‐
22]. The data demonstrated that HMBG-1 was not translocated from the nucleus thereby indicating that in cell culture apoptosis is the predominant cause of cell death. We also present evidence that NF-κB, although is activated, is unlikely to be responsible for BTV apoptosis. Indeed both BTV replication and cytopathic effect are enhanced in the presence of a chemical inhibitor of NF-κB activation. These results suggest that NF-κB activation may be involved in the development of the cytokine response. In addition, we observed that IRFs, particularly IRF-3, were up-regulated during BTV infection. Thus, it is likely that BTV infection triggers the NF-κB and IRF activations to establish an antiviral state in order to control BTV replication.
Discussion
The infection of mammalian cells by many viruses induces apoptosis and a variety of signal transduction pathways either to promote cell survival or to enhance the cell death. BTV infection of mammalian cells also triggers apoptosis. There are two common pathways for the induction of apoptosis and it appears that BTV may trigger both intrinsic and extrinsic pathways [
9‐
12]. The extrinsic pathway is primarily initiated by virus attachment to receptors, while the intrinsic pathway is mediated by damage to the mitochondria. In this report, we have undertaken a series of stepwise experiments to examine the cellular activations of various caspases and thereby induction of apoptosis during BTV infection. Further, we investigated if the caspase activities via intrinsic and extrinsic pathways during BTV infection are interdependent. Data obtained in this report conclusively confirmed the activation of caspase-8 in BTV infected cells by two different methods. The cleavage of caspase-8 was observed from 12 h p.i. by western analysis and the active, cleaved form was visualised by confocal microscopy. As the primary activation of caspase-8 is due to receptor binding, it was expected that caspase-8 activation would occur relatively early after infection with BTV. However, we noted that the caspase-8 activation in BTV infection was somewhat delayed compared to other viruses (e.g., reoviruses). Although we have not investigated further, it can be hypothesised that it is regulated by c-FLIP, an apoptosis inhibitor, as in reovirus infection [
27]. The control of caspase-8 by c-FLIP could be regulated by the activity of MAPK p38 [
31,
32], which has been shown to play a role in BTV-induced apoptosis and endothelial damage [
33,
34].
Our data also demonstrated the involvement of the intrinsic pathway in BTV infection as documented by the disruption of the mitochondria and release of cytochrome C which is known to be responsible for activating a number of events that leads to the cleavage of caspase-9. Thus, our data supports the recent report that cytochrome C is released from the mitochondria during an early stage of the BTV replication cycle [
10]. The direct evidence of caspase-9 cleavage was confirmed by western analysis. Moreover, the timing of mitochondrial damage precedes the cleavage of caspase-9 as presented in this report. Although the activation of an apoptotic response by the initiator caspase-8, -9 and executioner caspase-3 in response to BTV replication has been reported previously, to date the relationship between the two pathways remains unknown.
In this study, the relationship between the intrinsic and extrinsic pathways was determined using a series of assays including cell markers, pharmacological inhibitors and knock-out cell lines. Unlike reoviruses [
35], there was no cleavage of BID by caspase-8 in BTV infected cells. Therefore, the extrinsic pathway is unlikely to induce the mitochondrial damage as observed in our studies. Furthermore, the use of a chemical inhibitor and the caspase-9 deficient cell line (Jurkat ΔC9) clearly confirmed that caspase-8 activation was also independent of caspase-9 during BTV infection of mammalian cells. Thus, the activation of each caspase pathway during BTV-1 infection appears to be independent.
In addition to the activation of executioner caspase-3, we found that BTV infection also induces the activation of caspase-7, another executioner caspase which is closely related to caspase-3, sharing similar structure and substrates, but less promiscuous [
36]. Caspase-3 and -7 share a number of target proteins and both recognise caspase death substrate PARP. We investigated PARP cleavage in BTV infection and the cleavage of PARP was detected in response to BTV infection of mammalian cells, adding further support that BT clinical signs are mainly mediated by apoptosis.
DeMaula
et al.[
5,
6] have hypothesised that the main component of BT clinical signs and lesions in endothelial cells is due to cellular necrosis. Their hypothesis was based on the observation at the late stage of BTV infection. In this study, we examined cellular necrosis by using a cellular marker HMGB-1. When necrosis is induced it causes the loss of cellular membrane integrity and HMGB-1 is translocated from nucleus to cytoplasm and rapidly released into the extracellular space, which results in inflammation and tissue damage [
22,
37,
38]. When we examined the cytosolic and nuclear fractions of BTV infected HeLa at 48 h p.i., there was no HMGB-1 translocation to the cytosol. Therefore, the cell death and damage observed in BTV infection of mammalian cells is not due to necrosis but probably due to the very late timing of events and the cell disruption caused by the inflammatory response.
The data presented in our paper demonstrates that BTV infection of mammalian cells induces caspase cascade, resulting in apoptosis. These results indicate that apoptosis is a major cause of cellular damage in the host animal, supporting previous reports [
9,
12]. However, it will be imperative to investigate the caspase activation and the role of apoptosis in BT disease in susceptible sheep.
Previously, we reported the activation of NF-κB in BTV infected cells [
11] and hypothesised that it has a role in BTV-induced apoptosis similar to that of reovirus T3 [
13]. However, here we found that the inhibition of NF-κB had no effect in virus-induced apoptosis. Indeed, our data clearly demonstrated that chemical inhibition of the p50 subunit of NF-κB enhanced the early onset of visible cytopathic effect in BTV infected cells, and thus contradicted the report of Mortola and Larsson [
25].
Therefore, it was necessary to further investigate the response of NF-κB to BTV infection. The type of response by NF-κB is related to the phosphorylation and degradation of the IκB complex, an inhibitory protein complex, which masks the nuclear translocation signal of NF-κB. Our data showed the degradation of IκBα but not IκBβ in response to BTV infection, which indicates that the classical NF-κB pathway is activated resulting in a controlled transient response. Furthermore, we examine the activation of NF-κB by using a NF-κB dependent firefly luciferase reporter assay. Our data showed an initial early period of NF-κB activation that was not sustained, as the NF-κB activity appeared to be inhibited at later stages of virus replication. The NF-κB response observed by BTV infection was similar to that of the T3 strain of reovirus [
39].
From our preliminary data, we postulate that the transient nature of the NF-κB response was due to BTV proteins suppression of NF-κB activation, which in turn allowed efficient virus replication. BTV NS1 and NS2 limited NF-κB activation when the response was stimulated by poly I:C. Both of these proteins have the ability to bind RNA and therefore could act by sequestering dsRNA and inhibiting the pathway, or could interact with cellular proteins to inhibit the cascade (i.e. rotavirus [
40]), or inhibit mRNA export from the nucleus (i.e. rotavirus NSP3 [
41]).
Thus, our data showed that activation of NF-κB by BTV infection had a minimal role, if any, in the induction of apoptosis. However, it may play a role in initiating an antiviral state through the induction of the innate immune response. To this end, we investigated the effect of NF-κB activation on BTV replication and generation of infectious virions. In the presence of an NF-κB inhibitor, BTV titres were higher at the early stages (up to 24 h p.i.) of virus replication than in the absence of inhibitor. This result would indicate that NF-κB activation acts to control virus replication. At the late times, virus titres were similar in both cases, which could be indicative of the virus proteins controlling the NF-κB response (i.e. NS1 and NS2). The induction of NF-κB response in controlling virus replication through the induction of an antiviral response needs to be investigated further, especially as a recent report demonstrated that BTV replication was significantly high in IFNα deficient mice resulting in rapid death of these animals [
42].
BTV infection has been reported to produce strong cytokine responses [
7,
43‐
45], however the mechanisms that trigger their production has not been investigated. NF-κB activation could be a candidate to initiate a cytokine response during BTV infection, although there are other pathways, including IRF that could induce cytokine response. In this report we were able to identify the translocation of IRF-3 to the nucleus in response to BTV infection, indicating the induction of IRF-3 by BTV infection. As well as IRF-3, which is sequestered in the cytosol, the production of IRF-7 and its subsequent translocation to the nucleus were also detectable in BTV infected cells. Moreover, using an IRF dependent firefly luciferase reporter assay we confirmed that IRF activation was stimulated during BTV infection. Interestingly, the kinetics of IRF activation mirrored NF-κB activation. This early response of IRF was suppressed over the course of BTV infection similar to that of NF-κB. Despite the activation of IRF and NF-κB, it is clear that BTV infection triggers robust apoptosis in mammalian cells, which most likely play a role in BTV pathogenesis.
Materials and methods
Cells, viruses and antibodies
Human cervical epithelial carcinoma (HeLa) and BSR cells were maintained in Dulbecco's modified Eagle's medium (DMEM; Lonza) supplemented with 10% fetal calf serum (FCS), 100 U/ml penicillin and 100 mg/ml of streptomycin (Lonza) and incubated at 35°C, with CO2. Jurkat T-cells were purchased from ATCC and caspase-9 deficient Jurkat T-cells (Jurkat ΔC9) and complemented caspase-9 Jurkat T-cells (Jurkat compΔC9) were kindly provided by Klaus Schulze-Osthoff (University of Dusseldorf, Germany). The cells were grown in RPMI media (Sigma Aldrich) supplemented with 6% FCS, 100 U/ml penicillin and 100 mg/ml of streptomycin (Lonza) and incubated at 35°C, with CO2.
BTV serotype 1 (BTV-1) was titred by plaque assay as described [
46] and multiplicity of infections (MOI) between 0.1 and 1. For all experimental infections, cell monolayers were washed with FCS free growth medium, incubated with BTV-1 at the appropriate MOI in serum-free medium for 30 mins, the virus inoculums were aspirated, the cells washed twice in excess DMEM and further incubated in DMEM supplemented with 2% FCS and 100 U/ml penicillin and 100 mg/ml of streptomycin for the time course of the experiment. The procedure was similar for the infection of Jurkat T-cells with the exception that RPMI media was used.
Antibodies for detection of the caspase-8 (ab25901), caspase-9 (ab25758), caspase-3 (ab17819; ab90437) and anti-HMGB were purchased from Abcam. Active caspase-8 (18C8 #9496), caspase-9 (#9501) and caspase-7 (#9492) antibodies were purchase from Cell Signaling. Antibodies to detect the cellular proteins (tubulin and actin; AC-15), caspase-3 (clone 4H334) and PARP (clone C-2-10) were purchased from Sigma Aldrich. Specific antibodies for BTV were generated in house. Inhibitors of caspase cleavage were purchased from Calbiochem and Sigma Aldrich. Anti-IkBα (E:130), anti-IkBβ (ab7547), anti-NF-κB (p50/p105) and anti-NF-κB (p65) polyclonal antibodies were purchased from Abcam. Inhibitor of NF-κB (SN50; Calbiochem) was resuspended in distilled water was used at final concentration of 32 uM. Anti-IRF-3 (FL-425, sc9082) and anti-IRF-7 (H-246, sc9083) polyclonal antibodies were purchased from Santa Cruz.
Alkaline phophatase conjugated secondary antibodies were purchased from Millipore (mouse) and Sigma Aldrich (guinea pig and rabbit). Tetramethyl rhodamine isothiocyanate (TRITC) and fluorescein isothiocyanate (FITC)-conjugated secondary antibodies were purchased from Sigma Aldrich.
Preparation of cell lysates for the immunoblot
Whole cells (detached and adhered) were harvested and rinsed with cold 150 mM phosphate-buffered saline (PBS) and resuspended in SDS PAGE lysis buffer and processed for western immunoblot. For preparation of different cellular fractions, cells were harvested and incubated in Hepes buffer, pH 6.8 (10 mM Hepes 10 mM KCl, 1.5 mM Mg2Cl, 340 mM sucrose, 10%(v/v) glycerol, 0.1%(w/v) Triton X-100, 1× protease-inhibitor cocktail and 10 μM phenylmethylsulfonyl fluoride (PMSF) on ice for 30 min and then centrifuged at 4°C, 5 min at 3 000 xg (nuclear fraction). The supernatant was further centrifuged at 4°C for 45 min at 20 000 xg to separate the cytosolic fraction from the pellet (mitochondria) and the cell extracts were analysed by western immunoblot.
Western immunoblot analysis
The proteins were resolved on 10-15% SDS-PAGE gels according to target protein molecular weight, and transferred to a nitrocellulose membrane (0.45 micron, Amersham) by the standard semi-dry transfer protocol. Each blot was developed with specific primary and secondary antibodies following the manufacturer protocols.
Immunofluorscence by confocal microscopy
BTV-infected cells were fixed with 4% (w/v) paraformaldehyde on coverslips, washed in PBS, permeabilised and blocked as described [
47]. Subsequently, cells were immuno-labelled with primary antibodies, diluted (1:100-1:200), washed and incubated with appropriate secondary antibodies conjugated to TRITC (1:64) or FITC (1:128). Cover slip were mounted in Vectashield mounting media (Vector Laboratories, Burlingham) and were examined with a Zeiss Axiovert 200 M laser-scanning microscope fitted with a helium-argon laser. Images were acquired and analysed using LSM 510 confocal software (Zeiss).
NFκB and IRF dependent luciferase reporter gene assays
The NF-κB dependent luciferase reporter plasmid, pGL3NF-κB, was constructed by ligating the 4 repeats of the of the decameric sequence (GGGAATTTCC) recognised by NF-κB into the
Nde 1-
Bgl II sites of pGL3 basic luciferase. The IRF dependent luciferase reporter plasmids were kindly provided by Takashi Futija (Kyoto University, Japan) and were previously described [
48,
49].
HeLa cells grown in 96 well plates were transfected with 0.1 μg/well of the reporter using Lipofectamine2000™ according to manufacturer's protocols. At 24 h post transfection, cells were either infected with BTV-1 or transfected with 1 μg/well of poly I:C (Amersham Pharmacia) using Oligofectamine™ as per manufacturer's protocol or treated with 0.6 μg/ml DOX (doxorubicin.HCl; Sigma Aldrich) or mock infected. Each treatment was undertaken in triplicates and the assay performed no less than 5 times. The firefly luciferase activity in each well was quantified at different times using the Dual-Luciferase Assay Kit (Promega) to detect both firefly and Renilla luciferase.
For the inhibition assays using NF-κB (SN50) inhibitor, cells were transfected with the reporter plasmid 24 h prior to treatment with 32 μM SN50 as above and then either infected with BTV, or transfected with poly I:C or treated with DOX.
To study the effect of the BTV proteins on the NF-κB or IRF activation, HeLa cells grown in 96-well plates were co-transfected with 25 ng of T7 generated BTV transcripts or Renilla transcript with either the NF-κB or IRF dependent reporter construct using Lipofectamine2000™. T7 mRNA transcripts for VP2, VP4, VP6 and NS1-3 were generated from the BTV-1 exact copy constructs as previously described [
50]. At 24 h post transfection, the NF-κB or IRF response was induced by transfecting the cells with 1 μg poly I:C. The firefly luciferase activity was measured 6 and 24 h post induction with the Dual-Luciferase assay kit (Promega).
Virus titration by TCID50/ml
The growth of BTV in the presence of NF-κB (SN50) inhibitor was determined using TCID50 (tissue culture infectious dose of 50%) endpoint titration. BSR were grown in 96-well plates and 10-fold serial dilution of the BTV at different times p.i. were made in DMEM supplemented with 2% FCS in replicates of 5. The cells were fixed with 1% (w/v) paraformaldehyde, CPE was visualised by crystal violet stain and the titre determined as a TCID50/ml.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
PR conceived the project and designed experiments. MS designed experiments and carried out research. PR and MS wrote the manuscript and prepared the figures. Both authors approved the final manuscript.