Innate immunity, the first line of defense against infection, has revealed surprising complexity for what is considered a relatively primitive and conserved function. A diverse array of Toll-like receptors (TLRs) serves to recognize specific components of foreign invaders. Each TLR contains a Toll interleukin-1 receptor (TIR) domain that recruits IL-1R-associated kinases via adaptor molecules. These adaptors induce nuclear translocation of transcription factors like NF-κB or IRF-3, which turn on a variety of cytokines, including IFN-β and IFN-α, and subsequently a large number of interferon-stimulated genes. Most TLRs have been shown to signal through the adaptor protein myeloid differentiation factor 88 (MyD88) or the MyD88 adaptor-like (MAL) protein, but evidence soon suggested that more adaptors were involved in translating the response to pathogens into cytokine production.

In 2001, evidence mounted that neither TLR3 nor TLR4 signals through MyD88 or MAL: Shizuo Akira of Osaka University, Japan, and colleagues generated MyD88 knockout mice and found that those mice could still respond to TLR3 and 4 ligands, including the endotoxin lipopolysaccharide (LPS).[1] "The big question was, 'What's the missing adaptor?"' says Luke O'Neill of Trinity College, Dublin, Ireland.

This issue's Hot Papers identified and characterized a novel adaptor molecule, TRIF, for TIR domain-containing adaptor inducing IFN-β, (or TICAM-1, for TIR-containing adapter molecule). Akira's group identified TRIF by screening expressed sequence tag databases for the TIR domain motif, and showed that TRIF activates the promoter of IFN-β and the NF-κB-responsive promoter of the ELAM-1 gene. A dominant-negative form of TRIF abolishes TLR3 signaling. The team later generated TRIF knockout mice and conclusively demonstrated that TRIF is required for TLR3- and TLR4-mediated antiviral responses.[2,3]

Tsukasa Seya, of the Hokkaido University Graduate School of Medicine, and colleagues, used yeast two-hybrid experiments to show that TLR3 binds to TRIF, and demonstrated in vitro that TLR3 recruitment of TRIF leads to activation of IFN-β.[4] Bruce Beutler and colleagues at the Scripps Research Institute, La Jolla, Calif., characterized the MyD88-independent signaling pathway and revealed that it plays a role in the innate immune response to both bacterial and viral invaders.

The findings offered another clue to how TLRs discriminate between different types of pathogens. Previously, researchers had assumed that TLRs signaled through similar mechanisms, says Katherine Fitzgerald of the University of Massachusetts Medical School, who codiscovered MAL. "So I think the discovery of TRIF was really one of the key findings in the field because it allowed us to understand how specific downstream responses were coordinated by different TLRs upon ligand differentiation."

Before the TRIF gene was identified or named, Bruce Beutler's group had created a TRIF-deficient mouse using random germline mutagenesis, and picked the mutant animal from among many thousands of other mice by phenotypic screening. They called the mutation Lps2, and by positional cloning it, established that a single TIR adaptor protein was required for both TLR3 and TLR4 signaling, as well as responses to viral infections in vivo, Beutler notes. At about the same time, Akira's team had speculated about the existence of a pathway mediated by a yet-unknown adaptor. Beutler says, "The basis of this pathway wasn't understood, [but] when we observed the phenotype that we did, we thought this probably explained the MyD88-independent pathway or was a component of it." His team used positional cloning to track down the gene responsible for this phenotype, and discovered that the mutation occurred on the same gene previously identified by Seya as encoding TRIF. "But the function of TRIF and TICAM-1 wasn't known yet at the time that we discovered the mutation," says Beutler, who notes that Akira's TRIF knockout experiments independently determined TRIF's function.

Subsequently, Fitzgerald and coworkers identified a fourth adaptor, which they named TRIF-related adaptor molecule (TRAM or TICAM-2).[6] "We used an siRNA approach to show that TRIF was involved in both the Toll3 and the Toll4 pathways while the fourth adapter, TRAM, was involved only in Toll4 and not Toll3 signaling," says Fitzgerald. Her team, together with a group, from Harvard, later demonstrated that the kinases TBK1 and IKKε is responsible for the phosphorylation of IRF-3.[7] "It turns out that TRIF actually recruits in TBK1 and causes activation of the kinase, which then ultimately leads to the phosphorylation of IRF-3, and then the transcriptional induction of type 1 interferon," she adds.

Beutler's group later looked at downstream TRIF signaling events and showed that TRIF is required for the adjuvant effect of endotoxin. Macrophage and dendritic-cell receptors recognize endotoxin, activating TRIF and upregulating costimulatory molecules to enable T-cell activation.[8] "Essentially what we identified is that half of the mediators downstream of TLR4 are depending on the activation of TRIF, and showed that TRIF is required for the development of an adaptive immune response. This is an important consequence of activating the TLR4 pathway," says coauthor Kasper Hoebe, a senior research associate in the Beutler lab.

In more recent work using a forward genetic approach, Beutler's team showed that CD14 is needed for activation of the MyD88-independent signaling pathway by LPS.[9] "CD14 is a molecule that is tethered to the surface of cells, and it binds to LPS. Nonglycosylated LPS can signal without CD14 by way of the MyD88/MAL adapter pathway. But if CD14 is present, it permits TLR4 to assume a different conformation, so it can also recruit TRIF and TRAM. Then both pathways can be activated," Beutler says.

Lena Alexopoulou of the Centre d'Immunologie de Marseille-Luminy, in France, notes that detailed understanding of TLR signaling pathways may ultimately provide new therapeutic approaches for treating viral and bacterial infections, "because we will be able to block or to enhance a specific signaling pathway without knocking out the whole immune system of the organism."

Some of these pathways are currently being studied in specific infectious models. For instance, the hepatitis C virus (HCV) expresses a protease called NS3/4A that blocks phosphorylation of IRF-3 and activation of IFN-β. While examining this pathway, Stanley Lemon's team discovered that TRIF contains a domain with high homology to protease cleavage sites within HCV proteins required for viral replication. "We demonstrated and reported ... that the protease actually cleaves TRIF," says Lemon, director of the Institute of Human Infections and Immunity at the University of Texas Medical Branch at Galveston.[10]

Lemon notes that relative to the other adaptors, TRIF is quite large, suggesting that it may have functions in addition to TLR signal transduction. He adds that the function of SARM, another large molecule with crude domain homology to TRIF, is still unknown. "We'll probably learn a lot more about these adaptors in the next few years, and how the toll-like receptors ... are able to control a huge diversity of pretty specific innate immune responses, depending on whether it's a virus or a fungus that's infecting, whether it's LPS or dsRNA that's activating. There are some very interesting questions there."