**3. RNA motifs in the viral genome can trigger innate responses**

Accurate discrimination of self from non-self is critical to avoid immune triggering against self that leads to autoimmunity [32]. In that sense, it has been proposed for hepatitis C virus (HCV) that a combinatorial non-self signature in the viral genome for PRR binding may lead to accurate PAMP discrimination [33].

TLRs involved in recognition of the viral genomes are TLR3, TLR7/8 and TLR9, all of them localized to the endosomal compartment [34]. TLR3 is widely expressed in innate immune cells with the exception of neutrophils and pDCs and responds to dsRNA, a common viral PAMP, and its synthetic analog polyriboinosinic-polyribocytidylic acid (poly I:C) [35]. TLR7 and TLR8 are closely related receptors that recognize nearly any long ssRNA with some differences between them. Short ssRNA containing certain motifs preferentially activate TLR7, and activation with synthetic agonists specific to TLR7 or TLR8 trigger different cytokine profiles [36]. TLR9 is highly expressed in pDCs and responds to the unmethylated deoxycy‐ tidylate-phosphate-deoxyguanylate (CpG) motifs in viral and bacterial DNA [37].

Different features have been defined for RIG-I recognition as RNA PAMPs, including the presence of a free 5´-triphosphate, absent from eukaryotic cytoplasm due to RNA metabolism in the nucleus, length (longer than 19 nt), secondary structure characteristics (a base-pairing region of 10-20 nt near the 5´-ppp) [38] and nucleotide sequence motifs (such as a 3´-poly U/UC tract in the HCV genome) [33]. Panhandle structures adopted by Sendai virus DI- genomes or self-complementary influenza virus genome have been described as potent PAMPs sensed by RIG-I [39]. Data on MDA5 ligands are scarce. MDA5 seems to sense dsRNA analog poly I:C in mice [40] and higher-order RNA structures present in infected cells have been found to activate MDA5 [41]. A recent report shows the direct interaction of MDA5 with dsRNA replicative intermediate forms of positive strand RNA viruses [14]. RLRs have evolved to sense the presence of largely different sets of viruses but not always acting in a mutually exclusive way [13, 42].

proposed that, as the virus must maintain this motif in the 3´NCR for its viability, the host takes advantage of this and targets this region as a discriminator of PAMP RNA through RIG-I interaction [45]. Thus, HCV infection seems to be regulated by hepatic immune defenses triggered by the cellular RIG-I helicase. For FMDV, we also found that RNA transcripts corresponding to structural domains predicted to enclose stable dsRNA regions in the 5´and 3´NCRs of the viral genome were able to trigger an IFN-α/β response in epithelial porcine kidney cultured cells and induce an antiviral state [48] (Figure 3). A direct link between antiviral activity induced by FMDV NCR transcripts and IFN could be established in cultured cells, as treatment with monoclonal antibodies against IFN-α/β effectively blocked the antiviral activity induced by the RNAs [48]. Different levels of IFN-β mRNA induction were observed for the different RNAs assayed, being the one mimicking the complete 3´NCR, enclosing two predicted stem-loop structures, the best inducer. The in vitro RNA transcripts corresponding to the complete 5´ NCR, the IRES and the S hairpin (Figure 3), were also able to induce IFNβ transcription, though at lower levels than the 3´NCR transcript. The removal of the poly A tail within the 3´NCR RNA had a detrimental effect on IFN-β induction, but milder than removal of the 5´-ppp by treatment with alkaline phosphatase, which strongly reduced but did not completely abolish induction. However, deletion of any of the 3´NCR stem-loop (SL) structures rendered RNAs minimally active for IFN-β signaling, suggesting a relevant role for RNA structure in this region for its recognition as a PAMP. Unlike the FMDV NCR transcripts, the 5´-end of the viral genome is linked to the viral protein VPg lacking a 5´-ppp, making difficult to draw conclusions on the putative role of these structural regions in viral patho‐

Use of RNA Domains in the Viral Genome as Innate Immunity Inducers for Antiviral Strategies and Vaccine

Improvement

199

http://dx.doi.org/10.5772/56099

Encouraged by the results of IFN-β induction in swine cultured cells transfected with the NCR transcripts, we aimed to address the potential of such molecules as type-I IFN inducers in vivo. For that, the FMDV NCRs were inoculated intraperitoneally into Swiss suckling mice and the levels of IFN-α/β proteins and antiviral activity in sera were measured [48]. Newborn mice are a suitable model for innate immune responses while their adaptive immunity is still immature. All the FMDV NCRs were able to induce a peak of IFN-α/β in sera of the inoculated animals at 8 h after injection, remarkably higher than those observed for poly I:C-transfected mice. This peak was maintained up to 24 h in the case of the S RNA. The presence of antiviral activity in sera from NCR-transfected mice was also detected and measured, and a good correlation with IFN-α/β levels tested by ELISA was found. Interestingly, even those transcripts showing a lower capacity for IFN-β induction in porcine cultured cells were able to induce an innate immune response in mice. On one hand, this suggests that the effect of low level-inductions of type-I IFN observed in cultured cells can be magnified in vivo. On the other hand, the action of other viral sensors in vivo, mainly TLRs, may account for the enhancing effect observed. Thus, the specific immunostimulatory activity of each NCR RNA may be different depending on the host cell context assayed. This was the case for the IRES: despite of its complex structure, it was a poor inducer in cultured cells. However, the IRES acted as a strong IFN inducer in suckling mice. We further showed that the innate immune responses triggered by the NCRs in suckling mice resulted in a reduced susceptibility to FMDV infection in all cases, being

remarkable the antiviral effect of inoculation with the IRES RNA [48, 49].

genesis.

In addition to the direct antiviral function of RNase L degrading ssRNA, RNase L can generate viral or host-derived small RNAs that amplify the IFN response by generating PAMPs that activate the RLR pathway. RNase L mediated cleavage of HCV RNA generates svRNA that activates RIG-I, thus propagating innate immune signaling to the IFN-β gene [43, 44].

Given the ability of RLRs to sense viral RNAs and activate IFN signaling cascades that eliminate viral infections, many viruses have developed immune evasion strategies to overcome detection by RLRs. This is carried out through RNA modification of viral RNA genomes to prevent host detection [24]. For example, some viruses engage cap snatching (e.g. influenza virus), modification of 5´-ppp to monophosphate through virus encoded enzymes (e.g. Borna disease virus, Lassa virus), 2´O-methylation of viral mRNA cap structure by virus encoded methyltransferases, exploiting nucleotide modifications found at higher frequency in eukaryotic versus prokaryotic/viral RNA, and the use of proteins to protect the 5´ends (picornavirus have a virus encoded protein, VPg, covalently linked to the 5´end of their genome) or overhangs (e.g. arenavirus) [24].

In 2008, Saito et al. showed that the 3´ non-coding region (NCR) of HCV (a flavivirus) encoded PAMP motifs triggering innate immune signaling in the host cell. Thus, the 100 nt-polyuridine motif (poly U/UC) within the 3´ NCR was identified as a potent PAMP, substrate of RIG-I recognition and immune triggering in human and murine cells [45]. In contrast, the structured 3´-terminal X region failed to trigger signaling. The entire HCV 5´NCR, containing four major stem-loop structures including the internal ribosome entry site (IRES), was a weak inducer of IFN promoter signaling. However, prior treatment of cells with IFN-β to increase RIG-I levels rendered them responsive to signaling induced by the 5´NCR or the X region [45], suggesting that dsRNA regions of the HCV genome are not potent PAMPs but may confer signaling during the IFN response. Some studies on the 5´- and 3´-NCRs of other flaviviruses show remarkable differences in their IFN-inducing capacity. The 5´and 3´NCRs of dengue virus (DEN) elicited low but measurable stimulation of innate immune signaling, while the smaller highly structurally conserved 3´-terminal stem-loop RNAs of DEN, West Nile virus (WNV) and yellow fever viruses were minimally active [46]. Additionally, the base-pairing extent of the 5´-triphosphate of the RNAs may have an enhancing effect on RLR recognition and signaling [38, 47]. Therefore, the ability of different RNAs as IFN inducers must be tested independently, being difficult to predict their behaviour/potency by their sequence, secondary structure or homology with analog molecules. In this sense, we have recently shown that FMDV (a picornavirus) full-length transcripts with the 3´NCR deleted induce lower levels of IFN-β than complete RNA transcripts in cell culture [48]. These results are equivalent to those reported for HCV transcripts lacking the PAMP motif poly-U/UC. In this case, it has been proposed that, as the virus must maintain this motif in the 3´NCR for its viability, the host takes advantage of this and targets this region as a discriminator of PAMP RNA through RIG-I interaction [45]. Thus, HCV infection seems to be regulated by hepatic immune defenses triggered by the cellular RIG-I helicase. For FMDV, we also found that RNA transcripts corresponding to structural domains predicted to enclose stable dsRNA regions in the 5´and 3´NCRs of the viral genome were able to trigger an IFN-α/β response in epithelial porcine kidney cultured cells and induce an antiviral state [48] (Figure 3). A direct link between antiviral activity induced by FMDV NCR transcripts and IFN could be established in cultured cells, as treatment with monoclonal antibodies against IFN-α/β effectively blocked the antiviral activity induced by the RNAs [48]. Different levels of IFN-β mRNA induction were observed for the different RNAs assayed, being the one mimicking the complete 3´NCR, enclosing two predicted stem-loop structures, the best inducer. The in vitro RNA transcripts corresponding to the complete 5´ NCR, the IRES and the S hairpin (Figure 3), were also able to induce IFNβ transcription, though at lower levels than the 3´NCR transcript. The removal of the poly A tail within the 3´NCR RNA had a detrimental effect on IFN-β induction, but milder than removal of the 5´-ppp by treatment with alkaline phosphatase, which strongly reduced but did not completely abolish induction. However, deletion of any of the 3´NCR stem-loop (SL) structures rendered RNAs minimally active for IFN-β signaling, suggesting a relevant role for RNA structure in this region for its recognition as a PAMP. Unlike the FMDV NCR transcripts, the 5´-end of the viral genome is linked to the viral protein VPg lacking a 5´-ppp, making difficult to draw conclusions on the putative role of these structural regions in viral patho‐ genesis.

genomes or self-complementary influenza virus genome have been described as potent PAMPs sensed by RIG-I [39]. Data on MDA5 ligands are scarce. MDA5 seems to sense dsRNA analog poly I:C in mice [40] and higher-order RNA structures present in infected cells have been found to activate MDA5 [41]. A recent report shows the direct interaction of MDA5 with dsRNA replicative intermediate forms of positive strand RNA viruses [14]. RLRs have evolved to sense the presence of largely different sets of viruses but not always acting in a mutually

198 Current Issues in Molecular Virology - Viral Genetics and Biotechnological Applications

In addition to the direct antiviral function of RNase L degrading ssRNA, RNase L can generate viral or host-derived small RNAs that amplify the IFN response by generating PAMPs that activate the RLR pathway. RNase L mediated cleavage of HCV RNA generates svRNA that

Given the ability of RLRs to sense viral RNAs and activate IFN signaling cascades that eliminate viral infections, many viruses have developed immune evasion strategies to overcome detection by RLRs. This is carried out through RNA modification of viral RNA genomes to prevent host detection [24]. For example, some viruses engage cap snatching (e.g. influenza virus), modification of 5´-ppp to monophosphate through virus encoded enzymes (e.g. Borna disease virus, Lassa virus), 2´O-methylation of viral mRNA cap structure by virus encoded methyltransferases, exploiting nucleotide modifications found at higher frequency in eukaryotic versus prokaryotic/viral RNA, and the use of proteins to protect the 5´ends (picornavirus have a virus encoded protein, VPg, covalently linked to the 5´end of their

In 2008, Saito et al. showed that the 3´ non-coding region (NCR) of HCV (a flavivirus) encoded PAMP motifs triggering innate immune signaling in the host cell. Thus, the 100 nt-polyuridine motif (poly U/UC) within the 3´ NCR was identified as a potent PAMP, substrate of RIG-I recognition and immune triggering in human and murine cells [45]. In contrast, the structured 3´-terminal X region failed to trigger signaling. The entire HCV 5´NCR, containing four major stem-loop structures including the internal ribosome entry site (IRES), was a weak inducer of IFN promoter signaling. However, prior treatment of cells with IFN-β to increase RIG-I levels rendered them responsive to signaling induced by the 5´NCR or the X region [45], suggesting that dsRNA regions of the HCV genome are not potent PAMPs but may confer signaling during the IFN response. Some studies on the 5´- and 3´-NCRs of other flaviviruses show remarkable differences in their IFN-inducing capacity. The 5´and 3´NCRs of dengue virus (DEN) elicited low but measurable stimulation of innate immune signaling, while the smaller highly structurally conserved 3´-terminal stem-loop RNAs of DEN, West Nile virus (WNV) and yellow fever viruses were minimally active [46]. Additionally, the base-pairing extent of the 5´-triphosphate of the RNAs may have an enhancing effect on RLR recognition and signaling [38, 47]. Therefore, the ability of different RNAs as IFN inducers must be tested independently, being difficult to predict their behaviour/potency by their sequence, secondary structure or homology with analog molecules. In this sense, we have recently shown that FMDV (a picornavirus) full-length transcripts with the 3´NCR deleted induce lower levels of IFN-β than complete RNA transcripts in cell culture [48]. These results are equivalent to those reported for HCV transcripts lacking the PAMP motif poly-U/UC. In this case, it has been

activates RIG-I, thus propagating innate immune signaling to the IFN-β gene [43, 44].

exclusive way [13, 42].

genome) or overhangs (e.g. arenavirus) [24].

Encouraged by the results of IFN-β induction in swine cultured cells transfected with the NCR transcripts, we aimed to address the potential of such molecules as type-I IFN inducers in vivo. For that, the FMDV NCRs were inoculated intraperitoneally into Swiss suckling mice and the levels of IFN-α/β proteins and antiviral activity in sera were measured [48]. Newborn mice are a suitable model for innate immune responses while their adaptive immunity is still immature. All the FMDV NCRs were able to induce a peak of IFN-α/β in sera of the inoculated animals at 8 h after injection, remarkably higher than those observed for poly I:C-transfected mice. This peak was maintained up to 24 h in the case of the S RNA. The presence of antiviral activity in sera from NCR-transfected mice was also detected and measured, and a good correlation with IFN-α/β levels tested by ELISA was found. Interestingly, even those transcripts showing a lower capacity for IFN-β induction in porcine cultured cells were able to induce an innate immune response in mice. On one hand, this suggests that the effect of low level-inductions of type-I IFN observed in cultured cells can be magnified in vivo. On the other hand, the action of other viral sensors in vivo, mainly TLRs, may account for the enhancing effect observed. Thus, the specific immunostimulatory activity of each NCR RNA may be different depending on the host cell context assayed. This was the case for the IRES: despite of its complex structure, it was a poor inducer in cultured cells. However, the IRES acted as a strong IFN inducer in suckling mice. We further showed that the innate immune responses triggered by the NCRs in suckling mice resulted in a reduced susceptibility to FMDV infection in all cases, being remarkable the antiviral effect of inoculation with the IRES RNA [48, 49].

The antiviral effect exerted in vivo by these small synthetic non-infectious RNA molecules was analyzed extensively, using a wide range of viral doses and different serotype isolates [49]. The time course of resistance to FMDV of the RNA-transfected mice was also studied. Inoculation with all RNAs remarkably increased the 50% lethal dose (LD50) of the virus, determined for the control group. Mice inoculated with IRES or S transcripts 24 h before challenge became at least 10000-foldless susceptible tothevirus thanPBS-inoculatedmice.Interestingly, 90%ofthe IREStransfected mice survived after infection with a viral dose of 7 x 106 plaque forming units (PFU) (undiluted viral stock), showing the outstanding protective effect of these RNA molecules. The level of protection against viral infection was dose-dependent. Complete or very high protec‐ tion was achieved when IRES RNA was inoculated 8 or 24 h prior to FMDV infection with 7 x 104 PFU, with 100 and 86% survival, respectively. Inoculation of the transcripts at longer times pre-infectionstronglydecreasedtheirprotective effect against viralinfection. Co-inoculationof S or IRES transcripts and the virus induced high levels of protection (about 90%), and the IRES RNA had a higher protective effect inoculated at 8 h than at 4 h before infection, suggesting that a fine balance between the routes activating the innate immune response by the RNAs and the viral replication kinetics or antagonistic mechanisms triggered by the virus, might determine eithertheoutcomeofdiseaseortheviralclearance.Additionally,highsurvivalpercentageswere observed for those groups inoculated with the RNAs at short times after infection (89 and 87% ofmiceinoculatedwiththeIRESat4hand8hpost-infectionsurvived,respectively),andcomplete protection (100% survival) was achieved when mice were inoculated with the S transcripts at 4 or 8 h post-infection [49]. No protective effect was observed for mice inoculated with the RNAs 24 h after viral infection. These results suggest that the antiviral response induced by the RNAs is rapidly established and effective to counteract the viral replication if administered shortly afterinfection,while24hlateritwastoolatetorestraintheprogressofinfection.Ourdatasupport the potential use of this RNAs as both prophylactic as well as therapeutic molecules in a certain time window. These small non-infectious RNAs could be useful to induce a rapid antiviral state in combination with effective FMD vaccines to overcome the problem of the susceptibility window until protective levels of antibodies are produced by vaccinated animals. These results provide, as well, a new insightinto broad-spectrumantiviraldevelopment strategies (Figure 3).

have been found to play a role in other disease conditions like type-I diabetes and other autoimmune diseases like psoriasis or selective IgA-deficiency [53-55]. RLRs ligands have been shown to have a therapeutic effect for autoimmune inflammatory disease of the central

Use of RNA Domains in the Viral Genome as Innate Immunity Inducers for Antiviral Strategies and Vaccine

**3'NCR**

*FMDV genome*

Improvement

201

http://dx.doi.org/10.5772/56099

**SL1**

**ORF**

**IRES 3'NCR**

**Activation of Innate immunity**

**Antiviral effect Vaccine adjuvation**

**Figure 3.** Synthetic RNA transcripts derived from structural domains present in the 5' and 3'NCRs of the FMDV ge‐

Other than its role in driving innate immune defenses, IFN plays a major role in modulating the adaptive immune response [57]. IFN is required to promote T cell survival and clonal expansion after antigen presentation. IFN also induces the cytolytic activity of NK cells and cytotoxic lymphocytes and promotes B cell differentiation and antibody production, as well as expression of MHC class I molecules [58, 59]. The specific role of RLR signaling in regulating IFN production and its regulation of the adaptive immune response is less clear and appears

Luke et al. showed that the potency of a DNA vaccine against influenza virus could be augmented by the incorporation of a RIG-I activating immunostimulatory RNA into the vector

**Figure 3.** Synthetic RNA transcripts derived from structural domains present in the 5´and 3´NCRs of the FMDV genome as innate immunity elicitors

**5´ppp**

**An** **SL2**

*FMDV NCR RNAs*

**An5´ppp**

nervous system in mice [56].

**VPg**

**S**

**S**

nome as innate immunity elicitors

to vary from virus to virus [60].

**5´ppp**

**5'NCR**

*cre*

**IRES**
