Kaposi’s sarcoma-associated herpes virus (KSHV) is a member of the γ herpes virus family and is genetically similar to EBV and monkey Herpes Virus Saimiri (HVS) [²⁸]. Sequence analysis of the KSHV genome revealed the presence of about 80 open reading frames (ORFs) and a number of ORFs showing homology to cellular genes including a cluster of four ORFs with homology to the cellular IRF family transcription factors [²⁹], three of which have been cloned and characterized. The open reading frame 9 of KSHV genome (ORF K9)-encoded vIRF-1, has been studied most extensively and was shown to inhibit both virus-mediated induction of Type I Ifn genes and IFN-induced genes (ISG) [³⁰–³⁴].

vIRF-2 (ORF K11.1) encodes a small nuclear protein (163 aa) that unlike the cellular IRFs, binds dsRNA-activated protein kinase (PKR), inhibits its kinase activity and blocks the phosphorylation of the PKR substrate, eukaryotic translation initiation factor 2α [³⁴,³⁵].

vIRF-3/LANA2, encoded by ORFs K10.5 and K10.6 [³⁶,³⁷], is a multifunctional nuclear protein constitutively expressed in KSHV-positive B cell lymphoma cells (PEL) and Castleman’s disease tumors [³⁸]. The vIRF-3 protein binds to IRF-3, IRF-7 and IRF-5 [³⁷,³⁸] vIRF-3 also interacts with two tumor suppressor genes, p53 and MM-1 [³⁹].

The observation that some IRFs also have an essential role in the differentiation and function of lymphoid cells emerged mainly from a large number of studies in genetically modified mice. Thus, IRF-1 participates not only in the induction of the IfnG gene, but also in several autocrine loops that induce and enhance the Th1 response [⁴⁰,⁴¹]. IRF-4 has been also associated with Th cell development and the differentiation of B cells to plasma cells [⁴²]; Irf4−/− mice show a complete absence of plasma cells [⁴³]. Irf8−/− mice show major defects in CD8+DC and plasmacytoid dendritic cells (PDC). These mice also display increased susceptibility to infection, which is due to a defect in the Th1 immune response and an inability to express IL-12 [⁴⁴,⁴⁵]. IRF-3 and IRF-7 play essential roles in the antiviral response, however IRF-7 also has an important role in the differentiation of monocytes to macrophages [⁴⁶]. The Th1 and Th2 responses of Irf3−/− or Irf7−/− mice have yet to be characterized, however a DNA-mediated Th1 immune response was enhanced by co-expression of IRF-3 and IRF-7 [⁴⁷]. IRF-5 plays an important role in B cell differentiation and function in vitro. Infected B cells over expressing IRF-5 show up-regulated expression of several immune response genes, including CD80, MyD88 and ISG [²¹], and IRF-5 promotes macrophage polarization and TH-1 response [²⁵,²⁷]. Irf5−/− mice show a defect in differentiation of B cells to plasma cells and an attenuated response to antigenic stimulation [²⁶]. IRF-9 is an important component of the ISGF-3 complex with an essential role in the type I IFN signaling pathway and induction of many ISGs including IRF-7 and IRF-5 [⁴⁸]. The Ifr9−/− mice show a significant defect in Type I IFN signaling and share many characteristics with the Stat1−/− and InfR1−/− mice. These mice are also predicted to be defective in expression of many ISGs including both IRF-7 and IRF-5. However, whether they show modulation of the adaptive immune response similar to Irf5−/− mice has yet to be evaluated.

Understanding the mechanism by which cells detect invading pathogens was significantly advanced by the observation that this immune response is initiated by Toll like receptors (TLR) [⁵²,⁵³]. The TLR are Type I transmembrane protein receptors expressed in endosomes that recognize conserved motifs unique to pathogens, denoted as pathogen-associated molecular patterns (PAMPS). The presence of dsRNA, which has been considered the common signature of virus infected cells, is recognized by TLR3, while ssRNA activates TLR7/8, and unmethylated dsDNA activates TLR9. The binding of viral nucleic acids to their respective TLR induces signaling pathways leading to the activation of IRF and NFkB. Some TLR, like TLR2 can form homo or heterodimers, which alters the specificity of the ligand binding. With the exception of TLR3, activation of all TLR family members is dependent on the adaptor MyD88 [¹⁷]. Upon ligand activation, the intracellular part of TLR-TIR binds a tetramer of MyD88, and the signaling pathway is initiated by recruitment of IRAK1 and IRAK4, which are autophosphorylated and form a complex with the E3 ubiqitin ligase TRAF6 [⁵⁴]. This can activate mitogen activated protein kinase and lead to the activation of NFkB, which binds to the promoters of inflammatory genes such as IL-6 and TNFα or leads to the activation of IRF [¹⁷,⁵⁵].

The expression of TLR is cell type specific. Thus the PDC, which represent the major source of Type I IFN, (largely IFNα), constitutively express TLR7/8 and TLR9 as well as relatively high levels of IRF-5 and IRF-7 [⁵⁶]. In contrast, TLR3 signaling is distinct, it does not associate with MyD88, and binding of dsRNA to the TLR3 receptor activates IRF-3 through the TRIF adaptor [⁵⁷]. Two non-canonical IKB kinases, IKKε and TBK-1, have been implicated in the phosphorylation and activation of IRF-3 and IRF-7 in cultured human cell lines in vitro [⁵⁸]. Activated IRF-3 dimerizes and binds to p300/CBP co activators [⁴⁹], translocates to the nucleus and activates transcription of the IfnB gene and some ISGs in IFN-independent manners [⁵⁹]. Both IKKε and TBK-1 kinases also phosphorylate IKBα, but only on serine 36, which does not lead to the degradation of IKBα by the ubiquitination pathway or to activation of NFkB [⁶⁰].

The TLR adaptor, TRIF, is a potent activator of the IfnB promoter. Mice that are Trif-deficient or have a mutation in the Trif gene [⁶¹] have a profound defect in IRF-3 activation and fail to produce IFNβ. In addition, activation of IRF-3 and stimulation of IfnB and Rantes in infected or LPS-treated cells does not occur in TBK-1-deficient cells. It is noteworthy that whereas TRIF was shown to be solely responsible for the TLR3-mediated activation of IRF-3, activation by TLR4 requires an additional adaptor protein MAL [⁶²]. Taken together, these results indicate that the initial signaling events leading to the activation of TBK-1 by TLR3 and TLR4 may not be identical. However, there is strong evidence that IRF-3 plays an essential role in the antiviral response. First, ubiquitous expression in most of the cell types, allows stimulation of the antiviral response and synthesis of IFNβ in a variety of infected cell types [¹⁴]. Second, even low levels of autocrine or paracrine IFNβ stimulate expression of Irf7 and Irf5 and triggers amplification of the antiviral response [¹²,⁶³]. Finally, the observation that many viruses prevent the induction of Type I IFN by targeting the function of IRF-3, underlines a central role for IRF-3 in the induction of the antiviral response [⁶⁴].

TLR7/8 and TLR9 are an evolutionarily related subgroup within the TLR family [⁶⁵]. Both of these receptors are expressed in PDC, and act intracellularly in association with endosomes in a MyD88 dependent fashion. The TLR7/8 and TLR9 signaling pathway activates IRF-7 and IRF-5, (but not IRF-3). While the unmethylated CpG DNA [⁶⁶] specifically stimulates TLR9, single stranded (ss) viral RNA and synthetic siRNA are recognized by TLR7 and TLR8 [⁶⁷]. In mice, TLR7 is activated by infections with ssRNA viruses, including influenza virus and vesicular stomatitis virus (VSV) [⁶⁷,⁶⁸]. The activation of MyD88 pathways is initiated by the assembly of a multicomponent complex containing MyD88 tetramers, IRAK1 and IRAK4, TRAF6 and IRF-7 and or IRF-5. Activation of these two IRFs requires ubiqiutination by TRAF6 and phosphorylation by a not yet well defined kinase. Thus while IRF-7 can be activated both by TRIF and MyD88 pathways, IRF-5 is activated only by TLR7 and TLR9, and the activation is dependent on MyD88 and K63 ubiquitination by TRAF6 [⁶⁹,⁷⁰].

dsRNA, generated as an intermediate during virus replication, can be also recognized by cytoplasmic sensors. Two closely related intracellular RNA helicases, Rig-I and MDA5, which are ubiquitously expressed in nearly all cell types, can recognize dsRNA and 5′triphosphorylated RNA. Both helicases contain two N’ terminal caspase-recruiting domains (CARD) that activate downstream signaling pathways and a C-terminal DExD/H helicase domain that binds RNA. Signaling by Rig-I is mediated through downstream interaction with the adaptor protein IPS-1/MAVS. Both sensors contain CARD domains, and conformational changes occurring upon RNA binding permit these domains to interact with the CARD domain of an adaptor IPS-1/MAVS. This protein is localized on the outer mitochondrial membrane and this association is critical for its activity. IPS-1 activates TBK-1 and IKKe that leads to the activation of IRF-3, IRF-7 and NFkB. The IPS-1-dependent signaling is an essential feature of host immunity to RNA virus infection [⁷¹,⁷²].

The analysis of TLR ligands clearly shows that the endosomal TLR9 senses hypo/unmethylated CpG DNA that is present in genomes of bacteria and viruses. TLR9 is also the major sensing mechanism for DNA viruses and TLR9 dependent recognition of virus infection results in production of Type I IFN [⁷³]. Rather unexpected was the observation that TLR2 plays a critical role in recognition of herpes viruses. In DC that expresses multiple TLR receptors, HSV-2 engaged both TLR2 and TLR9, but the TLR2 recognition was limited only to a distinct population of HSV-2 [⁷⁴]. The synergistic interaction between TLR2 and TLR9 was also found to control HSV-2 infection in brain [⁷⁵]. It has been assumed that TLR2 detects viral envelope glycoprotein [⁷⁶].

As shown later, detection of viral DNA by cytoplasmic DNA sensors also leads to a profound antiviral response, and multiple cytoplasmic DNA sensors have been identified [⁷⁷]. A common feature of the signaling pathway induced by these receptors is activation of a transmembrane protein present on the endoplasmic reticulum designated STING (stimulator of IFN genes), which activates TBK-1 kinase that phosphorylates and activates IRF-3 and IRF-7 [⁷⁸]. Redundancy may be another feature of these sensors. Thus DAI (DNA dependent activator of IRF) induces in vitro activation of Type I IFN in response to synthetic DNA and HSV infection by the TBK-1-IRF-3 axis, but DAI knockout mice did not show any attenuation of IFN induction [⁷⁹,⁸⁰]. Another DNA sensor, interferon induced protein 16 (IF16) also interacts with STING and activates the TBK-1-IRF-3 axis and Type I IFN production [⁸¹]. Two cytoplasmic sensors recognize AT-rich B form DNA. The first sensor is leucine rich repeat Fli-I interacting protein (LRRFIPI) that activates IFN production by association with β Catenin, followed by recruitment of p300 histone acetyltransferase and IRF-3 to the IfnB promoter [⁷⁷]. The second sensor is RNA polymerase III, which recognizes AT rich DNA and transcribes it to immunostimulatory RNA recognized by Rig I [⁸²]. The importance of these two sensors for sensing viral infection remains poorly characterized.

The response to DNA in the cytoplasm can also be mediated through inflammasome complexes, resulting in the production of inflammatory cytokines that are also an important part of the innate immune response [⁸³]. Among these mediators is the protein AIM2 (absent in melanoma), which is an interferon induced protein that recognizes viral DNA and triggers the assembly of the inflammasome complex, resulting in activation of inflammatory cytokines IL1β and IL18, but not Type I IFN [⁷⁷].

In vitro, HIV-1 infects a number of diverse cell types; however the characteristics of the infections in these cells are distinct. Thus while infection in activated CD4+ T cells is productive and cytopathic, in naïve T cells HIV-1 is present in latent state. In macrophages HIV-1 replicates at low levels and does not induce cytopathic effects. Thus the innate antiviral response to HIV-1 infection in different cell types may be distinct. Several cell types, including CD4+ T cells, primary macrophages, and PBMC, infected with HIV-1 in vitro induced interferon signature, while induction of either IFNα or IFNβ was generally difficult to detect. Analysis of this typical interferon signature transcription profile in HIV infected CD4+T cells showed significant up-regulation of expression of interferon stimulated genes with increasing viral load, including genes of the intrinsic antiretroviral defense. Upon successful antiretroviral treatment, the signature profile of previously viremic individuals reverted to a pattern comparable to that of non progressors and uninfected individuals [¹²⁵]. In contrast, when the transcription profile of lymph nodes from HIV-1 infected patients was analyzed, only about 5% of the identified genes were associated with the antiviral response and the majority of the genes were negatively associated with HIV-1 replication. Based on these data the authors concluded that the antiviral response is not sufficient to attenuate HIV-1 replication [¹²⁶]. Recent data have shown up-regulation of type I IFN induced ISG in activated HIV-1+CD4+ T cells [¹²⁷], indicating that only a subset of CD4+ cells can induce IFNα. However the IFN signature induced in PBMC can be also induced by exogenous Type I IFN produced by other types of cells such as PDC, which are high producers of IFNα. The interaction between HIV-1 and DC is complex and not yet fully understood. It was shown that PDC are recruited to the site of HIV-1 replication in HIV-positive lymph nodes and this recruitment has been facilitated by expression of increased levels of lymphocyte homing marker CCR7 on circulating DC of HIV-1 infected individuals [¹²⁸]. Activated DC from HIV-1 infected individuals expressed high levels of IFNα and other inflammatory cytokines and these could well contribute to increased HIV-1 replication through immune activation.

The mechanism and the signaling pathway by which HIV-1 activates IFN production in dendritic cells has been subject of intensive studies. The PDC are an essential part of the innate immune system and the major producer of IFNα It was shown that the co-culture of PBMC with HIV-1 infected lymphocytes or infected CD4+ cells resulted in production of relative high levels of Type I IFN and that the major producers were the DC [¹¹⁸]. Production of IFNα was observed in lymph nodes of AIDS patients [¹²⁸] and has been associated with disease progression [¹²⁹,¹³⁰]. Recent clinical data suggest that production of IFNα may drive the hyper activation of the immune system resulting in CD4 T cell depletion and AIDS [¹²⁷]. Presently, it is assumed that IFNα alone contributes to the progression of AIDS disease, however a decrease in CD4+T cells is also associated with a significant increase in serum levels of inflammatory cytokines including TNFα. IL-6, IL-18 and IL-7 which may contribute to the observed hyper activation of the immune system.

The DC are generally resistant to HIV-1 infection, and co-culture of HIV-1 particles with monocyte derived dendritic cells (MDDC) neither infects DC or stimulates their activation. However when the freshly isolated DCs were infected with various HIV-1 strains in vivo, they were permissive for productive infection with the macrophage tropic HIV-1 (R5 HIV-1) and the infected DCs transferred HIV-1 to CD4 T cells [¹³¹]. It was shown that DC and T cells can form conjugates that facilitate transfer of HIV-1 from DC to T cells [¹³²]. It was also shown that this transfer is bi-phasic [¹³³].

As described above, cell recognize RNA viruses either by the membrane bound receptors such as TOLL receptors or cytoplasmic receptors such as RNA helicase RigI and MDA5. Which of these receptors are engaged upon HIV-1 infection has not been fully explored. However it was shown that the peptide inhibitors of TLR7 ligands inhibited IFN production in PBMC co-cultivated with HIV-1 infected T cells, indicating that the HIV-1 recognition is at least partially mediated by TLR7 [¹³⁰,¹³⁴]. The TLR7 independent pathway, which depends on IRF-3, was also induced in the co-culture experiments of HIV-1 infected T cells with TLR negative cells. Interestingly the induction of Type I IFN did not depend on HIV-1 replication in the infected cells, but was facilitated by the fusion of the infected T cells with DC [¹³⁴]. It is possible that the fusion transfers large amounts of viral nucleic acids to the cytoplasm of DC, where the viral RNA may be recognized either by TLR7 or one of the cytoplasmic sensors such as Rig I or MDA5. Alternatively, the un-integrated proviral DNA could be recognized by one of the recently identified cytoplasmic DNA sensors. Further studies will certainly address these questions.

Another very interesting and novel aspect of the mechanism of IFNα induction in DC is the observation that the restriction of HIV-1 replication in DC can be overcome by the SIV encoded protein Vpx. Expression of this protein or co-infection with SIV (that encodes Vpx) allowed HIV-1 replication in DC, stimulated production of IFNα and activated DC [¹³⁵]. It was also shown recently that Vpx induces proteasomal degradation of SAMHD1, a negative regulator of the interferon response. Silencing of SAMHD1 in non-permissive cell lines alleviated HIV-1 restriction and overexpression of SAMHD1 in sensitive cells attenuated HIV-1 replication [¹³⁶]. Further experiments demonstrated requirement of an interaction of HIV-1 capsid with CYPA for IFN production [¹³⁵]. The induction of IFN was also dependent on nuclear localization of phosphorylated IRF-3. Whether the phosphorylation and activation of IRF-3 in these cells depends entirely on the capsid interaction with cellular proteins or whether the requirement for the capsid interaction is connected to the TLR3 or RIG I-IRF-3 axis needs yet to be clarified. However it is unlikely that the SAMHD1 is the only factor that can modulate HIV-1 replication in DC. The attenuation of HIV-1 replication by the intrinsic cellular antiviral proteins such as APOBECC3G, TRIM and others [⁹⁵] may also contribute to this inhibition.

In addition to the antiviral ISG described earlier, type I IFN induces expression of several ISGs that interfere with different stages of HIV-1 and retroviral replication. These proteins are constitutively expressed at low levels in a cells type specific manner, and unlike the antiviral enzymes PKR or 2′5′OAS (already described above), they do not seem to require virus mediated activation (reviewed in [¹³⁷,¹³⁸]. The first of these cellular factors, dubbed intrinsic immunity factors was APOBEC3G [¹³⁹]. APOBEC3G is a cytidine deaminase that represses reverse transcription of viral RNA, acts on CC dinucleotide motifs and induces deamination of cytidine to uridine during synthesis of negative strand proviral DNA transcripts. This change leads to a G to A transition in proviral DNA and decreased ability (fitness) in virus replication. The expression of APOBEC3G is stimulated by Type I IFN [¹⁴⁰], and depends on NFATc1/NFATc2 and IRF-4. When either NFATc1 or NFATc2 and IRF-4 were co-expressed, activity of the APOBEC3G promoter was observed in cells that normally lack APOBEC3G expression. In addition to its effect on virus fitness APOBEC3G editing results in misfolding of viral proteins and enhances their recognition by CTL cells, thereby stimulating the antiviral adaptive immune response [¹⁴¹].

Type I IFN effectively induces the family of tripartite motif (TRIM) proteins that also restrict HIV-1 replication. TRIM5a from rhesus macaque (RhTRIM5a) blocks the early replication steps of HIV-1 and other retroviruses [¹⁴²] by targeting the incoming viral capsid (CA) protein for degradation. Anti-HIV-1 activity has also been reported for TRIM1 that also targets the CA protein at an early stage of pre-reverse transcription. TRIM19 was suggested to affect trafficking of viral proteins [¹⁴³], and TRIM22 blocked the release of HIV-1 Gag particles [¹⁴⁴].

It was shown that Type I IFN inhibits both acute and chronic retroviral (MuLV) infection and that the inhibition was at the posttranslational level affecting virus assembly and release [¹⁴⁵]. HIV-1 replication was found to be also sensitive to IFN inhibition, and the block was mapped to the late stages of HIV-1 replication [¹⁴⁶]. IFNα induced antiretroviral activities also include a decrease in proviral DNA accumulation [¹⁴⁷], restriction of viral entry and viral nucleic acid synthesis [¹⁴⁸], and in macrophages Type I IFN restricts HIV-1 replication during the initial stages of infection [¹⁴⁹].

At least three ISGs were shown to affect HIV-1 assembly and release. ISG15, a ubiquitin-like protein inhibits HIV-1 infection [¹⁵⁰] by interfering with the ubiquitination of HIV-1 Gag, blocking the interaction of TSG101 with Gag and consequently inhibiting the endosomal trafficking pathway, used by HIV-1 virions to exit cells [¹⁰³]. The ISG 15 block is not specific for HIV-1, but also affects release of the Ebola virus particles [¹⁰⁴]. TRIM22 blocks the intracellular trafficking of the viral structural protein Gag to the surface of the cell. The antiviral activity of TRIM22 is dependent on the E3 ligase activity of the RING-containing region of TRIM22 [¹⁴⁴]. Finally, the cellular protein BST-2, also known as tetherin, has been shown to be a very effective inhibitor of HIV-1 release [¹³⁸]. This protein is expressed constitutively in DC and B cells, but can be also effectively induced by Type I IFN in other cell types. Unlike the trapped retroviral virions in IFN treated cells, which have decreased infectivity, the HIV-1 virions retained on the cell surface by tetherin are infectious. The tetherin mediated inhibition is not specific but inhibits release of a large group of retroviruses, filoviruses and herpes viruses [¹⁵¹,¹⁵²].

Altogether these results indicate that HIV-1 can induce a strong innate immune response, associated with type I IFN mediated induction of several ISGs with a profound anti HIV-1 activity. Therefore the absence of the innate antiviral control of HIV-1 replication has been somewhat puzzling. One reason is that HIV-1, like the other RNA viruses, developed mechanisms by which it down-regulates the action of the antiviral ISG. HIV-1 encodes three small accessory proteins Vif, Vpr and Vpu, which circumvent the innate immune response by targeting the cellular antiviral proteins APOBEC3G and tetherin for ubiquitin dependent proteasome degradation. The expression of Virion infectivity factor, Vif, is essential for HIV-1 replication in primary T cells [¹³⁷]. In the absence of Vif, APOBEC3G is incorporated into HIV-1 virions; thus Vif prevents packaging of APOBEC3G and its availability during the very early stages of HV-1 infection [¹⁵³]. Virion associated viral protein R (Vpr) stimulates proviral transcription, and similar to SIV encoding Vpx, facilitates infection of macrophages. Both Vpr and Vif interact with cullin associated ubiquitin complex [¹³⁷]. Vpr and Vif also target IRF-3 for degradation in T cells, which results in inhibition of the antiviral response and IFNβ synthesis [¹¹³]. Finally, the membrane protein, Vpu, is required for efficient release of virus particles. Vpu interacts with the trans membrane domain of tetherin and targets it for proteasomal or lyosomal degradation [¹⁵⁴,¹⁵⁵]. Virions of HIV-1 mutants lacking the vpu gene accumulate on the plasma membrane of infected cells and are not released to the medium. Vpu also interacts with CD4 and targets it for degradation by the cullin associated ubiquitin complex [¹⁵⁶].

Other viral encoded functions may also counteract innate immunity. Interestingly, the negative factor Nef that has multiple functions, mostly in the modulation of adaptive immunity, can overlap Vpu and antagonize tetherin. In fact, primate SIV variants that lack Vpu use Nef to block the function of tetherin [¹⁵⁷] (Figure 1).

The role of HIV-1 encoded protease was recently shown to play a role in the inhibition of Rig I mediated signaling induced by in vitro synthesized HIV-1 RNA. Ectopic expression of HIV-1 protease resulted in inhibition of IRF-3 phosphorylation and induction of ISG [¹⁵⁸]. Since it is not clear whether ss HIV-1 RNA induces an interferon response in infected cells by the Rig I pathway, the in vivo relevance of this observation is yet to be confirmed.

In addition to virus encoded antiviral factors, several cellular proteins also contribute to down-regulation of the HIV-1 induced antiviral response. Thus the cellular DNAse Trex1 degrades the unintegrated proviral DNA in infected cells, thus preventing its detection by TLR9 or by the cytoplasmic DNA sensors. In TREX1 deficient cells HIV-1 DNA accumulates, which leads to induction of the IFN response [¹⁵⁹]. Interestingly induction of IFN required the cellular adaptor STING and IRF-3, suggesting that proviral DNA is recognized by one of the cytoplamic DNA sensors, rather than by TLR9 [⁷⁸]. Finally it was shown in SIV infected CNS macrophages, that IFNβ signaling is attenuated by induction of Suppressor of interferon signaling, SOCS3 [¹⁶⁰]. Also the virus induced miRNA may modulate the antiviral response. In vitro studies have shown that several IFNβ induced miRNA including miR-26a, -34a, -145 targeted IFNB mRNA and attenuated the antiviral response [¹⁶¹]. Whether these miRNA also regulate the Type I IFN system in HIV-1 infected cells in vivo is yet to be determined.