Special Reviews

Type 1 Interferons and the Virus-Host Relationship: A Lesson in Détente

See allHide authors and affiliations

Science  12 May 2006:
Vol. 312, Issue 5775, pp. 879-882
DOI: 10.1126/science.1125676


The interface between an infectious agent and its host represents the ultimate battleground for survival: The microbe must secure a niche for replication, whereas the host must limit the pathogen's advance. Among the host's arsenal of antimicrobial factors, the type 1 interferons (IFNs) induce potent defense mechanisms against viruses and are key in the host-virus standoff. Viruses have evolved multiple tricks to avoid the immediate antiviral effects of IFNs and, in turn, hosts have adapted use of this innate cytokine system to galvanize multiple additional layers of immune defense. The plasticity that exists in these interactions provides us with a lesson in détente.

Viruses have adapted strategies to evade or even inhibit key elements of host immune responses. Because particular arms of the responses are susceptible to neutralization, the host has evolved means of activating a broad range of defense mechanisms to ensure effective protection. In the midst of this power struggle, a remarkable picture is emerging of the importance of type 1 IFNs for antiviral defense and immune regulation. This family of innate cytokines has many members, with the best characterized being a single β and multiple α gene products. The host has a variety of pathways to elicit the expression of the IFN-α/β cytokines, including pathways protected from virus-mediated inhibition in infected cells. Viruses have countered by incorporating machinery that can dampen specific defense mechanisms that type 1 IFNs induce. Host organisms have taken advantage of the cytokines to shape and enhance its panoply of independent immune effector mechanisms so that any agent overcoming the direct antiviral effects activated by type 1 IFNs will have to contend with other responses. At their best, the tools used by the virus and host achieve conditions for coexistence.

Induction of Type 1 IFNs by Viruses

Mammalian hosts have evolved a variety of cellular sensors for viral infection, and it is the engagement of these protein receptors that ultimately leads, through complex and redundant pathways, to the production of type 1 IFN. Sensors fall into two functional classes that differ fundamentally with respect to localization, associated with either the cell membrane or the cytoplasm (Fig. 1). An important biological consequence of this differential localization is the flexibility it affords the host in triggering type 1 IFN production, either in infected cells and/or before cells are exposed to viral products capable of blocking induction.

Fig. 1.

The pathways to and from type 1 IFN are flexible. Families of sensors are available to detect viral products and induce expression of the cytokines. One set senses components in the cytoplasm. Another set is localized in cell membranes. There is also flexibility in the signaling pathways used by type 1 IFNs, with the potential to induce the activation of multiple STAT molecules and their downstream targets of transcription.

Extracytoplasmic pathways for pathogen sensing. The Toll-like receptor (TLR) family is composed of membrane proteins with domains designed to sample the environment for pathogen-associated molecular patterns (PAMPs) (1). Subsequently, TLRs become activated and transmit signals through their cytoplasmic Toll/interleukin-1 receptor (TIR) domains, resulting in the transcriptional induction of multiple genes involved in innate and adaptive immunity, including type 1 IFN. Different TLR molecules recognize specific PAMPs and, among these, TLR3, TLR7, TLR8, and TLR9 appear to play important roles in identifying viral products. TLR7 and TLR8 become activated by recognition of single-stranded RNA (24), whereas DNA activates TLR9 (5). Accordingly, TLR9 becomes activated in response to infection with DNA viruses such as herpesviruses, and TLR7 and TLR8 respond to RNA viruses such as influenza viruses and HIV. In contrast, TLR3 appears to represent a more general sensor of viral infection through detection of double-stranded RNA (dsRNA) (6), a by-product of viral replication and transcription for both RNA and DNA viruses. These four TLR molecules localize mainly in endosomal compartments of the cell, and it is upon endosome-mediated internalization of viruses (in the case of TLR7/8/9) or products of viral replication from lysed and/or apoptotic virus-infected cells (in the case of TLR3) that they are believed to come into contact with their respective activating ligands. To expose the viral genome to the corresponding TLR, this process most likely involves degradation of a subset of virus particles in the endosome. In addition, some cell surface–expressed TLRs, such as TLR4, have been shown to bind to specific viral glycoproteins and induce IFN production in a different range of cells.

TLR3 induction of type 1 IFN is mediated through the TIR domain–containing adaptor-inducing IFN-β (TRIF). TRIF mediates the activation of IκB kinase ϵ (IKKϵ) and tank-binding kinase 1 (TBK1), which phosphorylate IFN regulatory factor 3 (IRF3), resulting in its dimerization and nuclear translocation where it promotes gene transcription. TRIF also mediates the activation of neural factor κB (NF-κB) and activating protein 1 (AP1) through the kinase complex IKKα/β/γ and the mitogen-activated protein kinase (MAPK) cascade, respectively (7). These three transcription factors (IRF3, NF-κB, and AP1) coordinate the transcriptional regulation of the IFN-β gene (8).

Induction of type 1 IFN by TLR7/8/9 is mediated by the adaptor molecule myeloid differentiation primary response protein 88 (MyD88), which associates with the TIR domain of the TLRs, the interleukin-1 receptor–associated kinases IRAK1 and IRAK4, and the tumor necrosis factor receptor–associated factor 6 (TRAF6). This results in downstream activation of IRF7, and of the IKKα/β/γ and the MAPK cascades, leading to NF-κB and AP-1 activation (7). IRF7 is functionally similar to IRF3 and mediates the induction of IFN-β but, unlike IRF3, it also initiates the general induction of the IFN-α genes (9). TLR7/8/9 and IRF7 appear to be constitutively expressed in only a subset of cells, the plasmacytoid dendritic cells (PDCs), which are characterized by high IFN production and can spearhead the early IFN response (10).

Cytoplasmic pathways for sensing. TLR-independent pathways of sensing viral infections and triggering IFN production appear to exist as well. At least one such pathway is mediated by members of a family of DExD/H box RNA helicase proteins that contain caspase-recruiting domains (CARDs) (11). In contrast to TLRs, these sensors are all cytoplasmic and, thus, exclusively mediate intracellular recognition of viruses. Two such sensors have so far been described: the retinoic acid–inducible gene I (RIG-I) and the melanoma differentiation–associated gene 5 (mda5). These RNA helicases, upon binding to dsRNA, interact with a downstream molecule named independently by four different groups as mitochondrial antiviral signaling protein (MAVS) (12), IFN-β promoter stimulator 1 (IPS-1) (13), virus-induced signaling adaptor (VISA) (14), and CARD adaptor–inducing IFN-β (CARDIF) (15). This mitochondrial-resident protein interacts with the CARD of RIG-I and recruits and activates, by not yet well-defined mechanisms, IKKϵ/TBK1, the IKKα/β/γ complex, and MAPK, resulting in activation of IRF3, NF-κB, and AP-1 and in IFN-β induction. Evidence for a TLR and RIG-I independent pathway in recognition of cytoplasmic DNA has recently been obtained (16, 17), and additional sensors of other viral products are likely to exist.

IFN-Mediated Effects On Defense

As secreted factors, the type 1 IFNs regulate a range of immune responses through the type 1 IFN receptor, a cell surface transmembrane receptor composed of two subunits, IFN-α receptor 1 (IFNAR1) and IFNAR2 (1820). Binding of the IFNAR results in receptor subunit dimerization and activation of kinases that associate with their cytoplasmic tails: the Janus-activated kinase 1 (JAK1) and tyrosine kinase 2 (TYK2). In turn, tyrosine phosphorylation activates the signal transducers and activators of transcription 1 and 2 (STAT1 and STAT2), to form a trimeric STAT1-STAT2-IRF9 complex, also known as IFN-stimulated gene factor 3 (ISGF3), as well as a STAT1 homodimer complex, known as the IFN-γ–activated factor (GAF). Both complexes translocate to the nucleus and bind to DNA regulatory sequences containing IFN-stimulated response elements (ISREs) and IFN-γ–activated sites (GAS), respectively. The ensuing stimulation leads to the transcription of more than 100 IFN-stimulated genes (ISGs), whose concerted action leads to the generation of an “antiviral state.” Although signaling through STAT1/STAT2 and STAT1/STAT1 dimers are the best characterized of the type 1 IFN-induced intracellular pathways to gene expression, cells can differ greatly in their signaling response to type 1 IFN, and a variety of other signaling pathways can also be activated (18, 19). Some of these are revealed in the absence of STAT1 (21), depend on STAT4 (22, 23), may be STAT2 dependent but STAT1 independent (24), and/or are a result of STAT 3 or 5 activation (18). The flexibility that receptors for type 1 IFNs have for accessing downstream signaling molecules may be regulated by the relative abundance of STATs themselves (22).

Antiviral activities of ISGs. The myxovirus resistance gene Mx, the protein kinase stimulated by dsRNA PKR, and the 2′-5′ oligoadenylate synthetases (OAS) are among the best characterized ISGs with antiviral activity (20). Mx is a guanosine triphosphatase (GTPase) belonging to the dynamin family that sequesters viral ribonucleoproteins to specific subcellular compartments. PKR is a serine-threonine kinase that phosphorylates downstream substrates upon recognition of dsRNA, including the elongation initiation factor eIF2α, resulting in the inhibition of protein translation. The OAS proteins are also activated by dsRNA, leading to the generation of 2′-5′ oligoadenylates, which activate ribonuclease L (Rnase L) that degrades cellular and viral RNA. Both PKR and the OAS/RNaseL systems have profound inhibitory effects on basal cellular processes that eliminate virus-infected cells by suicide (25). Even in the absence of these IFN-regulated antiviral pathways, however, IFN can still induce an effective antiviral response (26). The existence of other multiple ISGs with antiviral activity, multiple IFN genes, and multiple pathways leading to the production of IFN raises the intriguing question of whether hosts have evolved redundant pathways to make it generally difficult for viruses to use any single mechanisms to inhibit the IFN antiviral response or whether different factors of the IFN system mediate inhibition of specific virus families.

Regulation of immune responses. Type 1 IFNs extend their antiviral defense functions to a number of other immune response components. They amplify their own expression through two independent mechanisms: the induction of IRF7 to extend IFN gene expression to a broader range of IFN-αs (9) and the accumulation of PDCs, major contributors to IFN-α/β responses (27). The cytokines also activate natural killer (NK) cells to mediate elevated cytotoxicity (28, 29) and induce interleukin-15 (IL-15) to promote NK cell proliferation (29). At high concentrations, type 1 IFNs inhibit IL-12 (30) and NK cell responsiveness for IFN-γ expression (21). All of these effects on innate immune responses depend on STAT1. Type 1 IFNs also regulate adaptive immunity. The early STAT1-dependent induction of IL-15 contributes to short-term proliferation of memory CD8 T cells (31). However, only certain subsets of cells are equipped to respond to IL-15 and, consistent with the antiproliferative effects of type 1 IFNs (32), STAT1 acts to limit nonspecific CD8 T cell expansion, at times overlapping with the induction of innate responses (33). Hence, early during infections, the type 1 IFNs work to enhance proliferation of certain cell subsets and inhibit others through STAT1-dependent mechanisms.

There are many other paradoxical effects assigned to type 1 IFNs, and some of these may be explained by regulation of accessibility to different signaling pathways (22, 3335). STAT1 is induced by the cytokines, and the protein concentrations of this molecule are elevated at times of type 1 IFN induction during viral infection (22). These conditions may steer responses to STAT1/STAT2 and STAT1/STAT1 with induction of the antiviral defenses and general inhibition of proliferation as described above, but they also limit the ability of type 1 IFNs to activate STAT4 (22). To be effective in defense, antigen-specific CD8 T cells must expand in the presence of type 1 IFNs, and their IFN-γ production is aided by type 1 IFN- and STAT4-dependent events (22). Thus, the elevated STAT1 levels induced early during infection and the consequences this has for proliferation and STAT4 accessibility present inherent obstacles for generating protective T cell–mediated immunity. To overcome this, antigen-specific CD8 T cells with lower relative levels of STAT1 are induced and preferentially undergo proliferation at times when STAT1 levels are elevated in most other cells (33). Thus, the critical role for STAT1 in controlling the effects of type 1 IFNs may be a result of its own differential expression in different cell subsets. Linking cytokine signaling to different pathways, depending on transcription factor levels, is a sophisticated means of providing cells with a variety of downstream options using only a limited number of genes.

In addition to these effects during acute responses, the type 1 IFN receptor helps with long-term maintenance of the CD8 T cell pool (36). Although the role for STAT1 in this effect has yet to be defined, its ability to inhibit proliferation may serve to protect CD8 T cells from chronic stimulation and eventual depletion through clonal “exhaustion” (37). In the absence of STAT1, type 1 IFNs can induce growth as well as have antiapoptotic effects on T cells through possible STAT3- and/or STAT5-dependent events (3840).

Viral Evasion of IFN Responses

Many viruses dedicate a substantial part of their genomes to down-modulating the IFN pathways. A general mechanism used by several viruses is inhibition of cellular gene expression by inhibiting transcription, RNA processing, and/or translation (20). Virus-induced shutoff of cellular protein expression not only favors the diversion of cellular resources for viral protein expression but also prevents the synthesis of IFN and of ISG products. Most viruses also encode viral products that specifically target pathways involved in the response to IFN, and such products are generically known as viral IFN antagonists. Typically, different virus families are characterized by the presence of specific viral IFN antagonists lacking homology with those from other families. Nevertheless, viral IFN antagonists focus inhibition on at least one of three key pathways: the IRF3, the JAK-STAT, and the PKR pathways.

Antagonism of type 1 IFN induction. Viral inhibition of IRF3, resulting in decreased type 1 IFN production, has been documented to occur at multiple levels. For example, influenza and poxviruses encode dsRNA binding proteins NS1 and E3L that prevent IRF3 activation (41, 42), at least in part by sequestering viral dsRNA and preventing stimulation of cellular sensors of dsRNA, such as RIG-I and mda5. Direct binding to mda5 of a viral IFN antagonist resulting in mda5 inhibition has been shown in the case of the V protein of several paramyxoviruses (43). MAVS/IPS-1/VISA/CARDIF is the target for cleavage by the NS3/4A protease of hepatitis C virus (15). This protease also cleaves TRIF (44) and therefore blocks both TLR3- and RIG-I–mediated activation of IFN. In addition, the P proteins of some negative RNA viruses appear not only to be essential components of the viral RNA polymerase but also to inhibit the action of TBK1, preventing IRF3 phosphorylation (45). The human herpesvirus 8 encodes several analogs of IRF, known as viral IRFs (20), some of which act as dominant negative mutants of IRF3 action.

Antagonism of type I IFN signaling. The JAK/STAT pathway is also targeted at multiple levels by viral IFN antagonists. Poxviruses secrete a soluble form of the IFNAR that sequesters type 1 IFN before it can bind to the natural IFNAR (46). Inhibition of the JAK kinases has been documented for several viruses (20). STATs appear to be a preferred target for many paramyxoviruses, whose accessory V and W proteins bind to these factors and prevent their activation in response to IFN. Different paramyxoviruses have different specificities for STATs, with some of them inhibiting STAT1, STAT2, STAT3, or a combination of these factors. In addition, targeted degradation of the STATs is seen with a subset of paramyxoviruses (47). Intriguingly, paramyxovirus IFN antagonists, including the V proteins of several paramyxoviruses, and the NS1 and NS2 proteins of respiratory syncytial virus, have the remarkable property of inhibiting both IRF3 and STAT activation (48).

Antagonism of type 1 IFN-inducible genes and their products. It is intriguing that among more than 100 ISGs, the PKR product appears to be a common target for many viral IFN antagonists. The ways viruses inhibit the PKR pathway are again very diverse and rank from sequestration of the PKR-activating dsRNA, expression of dsRNA mimics, binding to PKR preventing its dimerization and activation, and substrate competitive inhibition (49). Finally, activation of cellular proteins involved in negative regulation of many aspects of the IFN response is also a strategy used by several viruses. Induction of cellular inhibitors of STATs, such as suppressors of cytokine signaling (SOCS), has been reported to occur during herpesvirus infections. Influenza virus infection activates a cellular inhibitor of PKR, the p58IPK, and herpesviruses encode a protein, γ34.5, that recruits a cellular phosphatase for the dephosphorylation of eIF2α, reverting the PKR-mediated translational block (49).

Smart Virus, Smart Host

As we've described, multiple offensive/defensive mechanisms have evolved to result in a balance for coexistence of hosts and viruses, and type 1 IFNs have emerged as pivotal in this conflict (Fig. 2). Antagonism of these cytokines appears to be a common feature shared by all viruses, and the variety of mechanisms for “hiding” from the direct antiviral effects of type 1 IFNs provides the virus with opportunities to replicate and infect a new host before being eliminated by secondary innate and adaptive immune responses in the case of acute viral infections. In chronic infections, this can emerge as a means to establish viral persistence in the host. The diverse and multiple viral approaches to avoiding IFN induction and function provide persuasive evidence for the potency of these mediators in early defense. A direct consequence of disrupting the function of viral IFN antagonists is a decrease in viral replication and pathogenicity in the host, apparent, for example, in poxviruses, herpesviruses, and influenza viruses. On the far side of the spectrum, however, viruses equipped with highly refined means of disrupting innate mechanisms activated by type 1 IFN are likely to be in the highest order of pathogenicity, a notorious example being the virus strain responsible for the pandemic influenza A virus of 1918 (50, 51). High pathogenicity, however, is not necessarily advantageous for the virus: Conditions that result in too rapid a destruction of the host may have deleterious effects on long-term survival of an infectious agent that requires the host for further propagation.

Fig. 2.

The host uses type 1 IFNs to its advantage despite viral evasion mechanisms. The sensors for detecting viral products include components in the cytoplasm that are particularly sensitive to viral blocks. The set localized in cell membranes is available for sensing viral products before cells are infected. Once induced, the cytokines enhance innate and adaptive antiviral defense mechanisms as well as direct antiviral pathways. The intracellular signaling pathways used by type 1 IFNs appear to be modified to access a variety of downstream target effects in different immune cell subsets. The concentration of the various STAT molecules may act to shape accessibility to different signaling pathways. As a result of these events, an infectious agent overcoming the direct antiviral effects of type 1 IFNs has to deal with additional immune mechanisms of defense.

Although viruses evade the direct antiviral defense mechanisms activated by type 1 IFNs, hosts take advantage of these cytokines to elicit a wide range of responses. The classical signaling pathways used to induce direct antiviral defenses also activate some of these, particularly the ones elicited during innate periods of responses. The importance for protection is clearly shown by the fact that deficiencies in STAT1, and therefore in innate responses activated by type 1 IFNs, result in extreme sensitivities to viral infections (5255). The host, however, has adapted to also use other signaling pathways to activate additional defense mechanisms, most clearly demonstrated for STAT4-dependent induction of T cell IFN-γ production. Thus, new complex and important defense responses have been attached to the critical early antiviral cytokines. One consequence of using these factors to promote downstream innate and adaptive responses is that, even if viruses have escaped their direct antiviral effects, the host can limit the window of opportunity for pathogen advancement.

Much progress has been made, but there is still much to be learned about the pathways regulating IFN induction and function during viral infections. Nevertheless, there is evidence to conclude that, as in many difficult relationships, viruses and their hosts are learning to live together and that type 1 IFNs are important players in this détente.

References and Notes

Stay Connected to Science

Navigate This Article