Interferons (IFN) and cytokines are important mediators of immune defense against infections. An effective IFN response involves two phases: an early phase of IFN production triggered by pattern recognition receptors (PRRs) that detect pathogens, and a late phase of IFN signaling mediated by the type I IFN receptor (IFNR), which activates the JAK-STAT signaling pathway to induce hundreds of interferon stimulated genes (ISGs). ISGs collectively suppress viral infection, replication and assembly, but the mechanisms underlying the antiviral functions of most ISGs remain largely unknown. In this issue, Saitoh et al. (2011)
reveal a surprising role of an ISG, Viperin (virus inhibitory protein, endoplasmic reticulum-associated, IFN-inducible), in the first phase of IFN production mediated by toll-like receptor 7 (TLR7) and TLR9 in plasmacytoid dendritic cells (pDCs). Importantly, through investigating the mechanism of IFN regulation by Viperin, the authors uncover an interesting connection between lipid storage organelles called lipid bodies (also known as lipid droplets) and antiviral immune defense.
Most mammalian cells have the ability to secrete and respond to IFN. However, pDCs are specialized to produce copious amounts of IFNα, in part through constitutive expression of endosomal TLRs (TLR7 and TLR9) and IRF7, a master transcription factor for IFNα. In addition, TLR ligands (e.g, CpG DNA) capable of triggering IFN production in pDC are retained in specialized endosomal vesicles and TLRs on the cytosolic surface of these vesicles recruit signaling proteins including MyD88, TRAF6, TRAF3, IRAK4 and IRAK1 (Honda et al., 2005
)(). A signal transduction cascade through these proteins leads to the recruitment and activation of IRF7, which then enters the nucleus to induce IFNα. Recent studies identified AP-3 (adaptor protein 3) as a key protein required for TLR9 trafficking to specialized vesicles called lysosome related organelles (LRO) (Blasius et al., 2010
; Sasai et al., 2010
). LRO recruits TRAF3 and IRF7, thereby triggering IFNα production. Loss of AP-3 blocks TLR9 trafficking to LRO such that TLR9 is retained in another endosomal compartment capable of inducing proinflammatory cytokines, but not IFNα, through activation of NF-κB. Thus, TLRs in distinct intracellular compartments, termed NF-κB endosomes and IRF7 endosomes (or LRO), trigger production of inflammatory cytokines and IFN, respectively.
Type-I interferon induction by endosomal TLRs in plasmacytoid dendritic cells
The study of Viperin by Saitoh et al. adds another layer of regulation of IFN induction by endosomal TLRs. Viperin is strongly induced by IFN, as well as by viral and bacterial infection (Fitzgerald, 2011
). The induction of Viperin by viruses and bacteria is mediated by distinct PRRs, including TLRs, and requires the transcription factors IRF3 and IRF7. To study the role of Viperin in vivo, Saitoh et al. generated Viperin-deficient mice and found that Viperin deficiency impaired TLR7 and TLR9-mediated type I IFN production in pDCs, but not in other cell types including conventional DCs, macrophages and fibroblasts. The induction of IFN by other PRRs, such as TLR3, TLR4, RLR (RIG-I like receptors), was not affected by the loss of Viperin. Viperin was also dispensable for induction of inflammatory cytokines, including IL-12, TNFα, and IL-1β, in all cell types examined, including pDCs. Thus, Viperin is specifically required for IFN induction by TLR7 and TLR9 in pDCs. Mechanistically, the authors found that Viperin associates with TRAF6 and IRAK1 and that Viperin is required for lysine-63 (K63) polyubiquitination of IRAK1, which is known to mediate the activation of NF-κB and IRF7. The kinase IKKα is required for IFN induction through phosphorylation of IRF7. However, phosphorylation of IKKα is normal in the absence of Viperin, suggesting that IKKα phosphorylation is insufficient to activate IRF7. This result also implies that Viperin and K63 polyubiquitination of IRAK1 are not required for IKKα activation, but may facilitate IRF7 activation through a mechanism that remains to be elucidated.
Interestingly, the N-terminus of Viperin contains amphipathic alpha helix that targets the protein to the cytoplasmic face of ER-derived lipid bodies (Hinson and Cresswell, 2009
). The localization of Viperin on the lipid bodies may be important for its antiviral effect, as some viruses such as hepatitis C virus replicate in the lipid bodies (Miyanari et al., 2007
). Saitoh et al. discovered another important function of lipid bodies. They found that TRAF6 and IRAK1, and to a lesser extent, MyD88, are recruited to lipid bodies in a Viperin-dependent manner upon TLR9 stimulation. Viperin without its N-terminal membrane association domain is still able to associate with TRAF6 and IRAK1, but cannot rescue IFN induction in Viperin-deficient pDCs. A chemical inhibitor (U18666A) that disrupts lipid body formation also prevents the induction of IFN, but not IL-12, by a TLR9 ligand in pDCs. These results strongly suggest that the localization of Viperin on the surface of lipid bodies is important for IFN induction. Immunocytochemistry studies detect IRF7, but not TLR9, on Viperin-positive lipid bodies. Thus, lipid bodies may function at a step downstream of LRO, which contains TLR9, and the role of Viperin and lipid bodies may be serving as a platform to assemble a signaling complex including MyD88, TRAF6, IRAK1 and IRF7 (). Future studies should determine whether and how the formation of this complex on the lipid bodies facilitates IRF7 phosphorylation and/or nuclear translocation.
The finding that Viperin and its localization in lipid bodies are important for IFN induction provides a mechanism for positive feedback control, as Viperin expression is highly induced by IFN. If Viperin is essential for IFN induction, pDCs must express certain amounts of Viperin even in the absence of IFN. Viperin may also be induced through IFN-independent mechanisms, and the regulation of viperin by IFN-dependent and independent pathways allows for crosstalk between pDCs and other cells. In addition to Viperin, several other proteins that regulate IFN production are themselves induced by IFN. Examples include IRF7 and the viral RNA sensors RIG-I and MDA5. Through positive feedback loops, these proteins induce and amplify IFN, which is important for a rigorous antiviral response.
It is now clear that pDCs have the unique ability to engage proteins such as AP-3 and vipirin to assemble signaling complexes on specialized endosomes and lipid bodies to activate IRF7, which then induces large amounts of IFNα. However, AP-3 and vipirin expression is not restricted to pDCs, begging the question of what makes pDCs uniquely capable of forming membrane compartments optimized for IFNα production.
The discovery of the role of lipid bodies in IFN production reinforces the recurring theme that cell signaling, like real estate, is all about location, location, location. Multiple signaling pathways can lead to the production of IFN, and remarkably, all known pathways signal from distinct intracellular membrane compartments. In addition to intracellular TLRs, which detect DNA and RNA in the lumen of endosomes, other PRRs can detect cytosolic nucleic acids and activate signaling cascades leading to the production of IFN (Rehwinkel and Reis e Sousa, 2010). RIG-I-like receptors bind to viral RNA in the cytosol and transduce the signal to the adaptor protein MAVS (also known as IPS-1, CARDIF or VISA), which contains a C-terminal transmembrane domain that anchors the protein to the mitochondrial outer membrane. The mitochondrial localization of MAVS is essential for IFN induction (Seth et al., 2005
). Cytosolic DNA can also induce IFN through one or more DNA sensors and the adaptor protein STING (also known as MITA or MPYS). STING contains multiple transmembrane domains that target the protein to the ER membrane (Ishikawa and Barber, 2008
). In several cases, it has been demonstrated that mislocalization of the adaptor proteins to different membrane compartments abrogates their ability to induce IFN. Thus, signals emanating from the surface of specific intracellular organelles are endowed with the special ability to induce IFN. An exciting avenue of future research is to delineate the signals and signaling complexes on these intracellular membranes, including the fascinating lipid bodies.