The innate immune system enables animals to deal quickly and efficiently with invading microorganisms. A key component of this rapid-reaction force is a finely honed detection system, which consists of a set of receptors, found on the surface of immune cells, that recognize particular molecular patterns associated with pathogens. When they sense the presence of intruders, these so-called Toll-like receptors trigger an intracellular signalling pathway that culminates in various immune responses1. This pathway has been well characterized. But it has become clear that it does not elicit all of the responses to bacteria and viruses — in particular the production of interferon-
protein2. So a second signalling pathway with this specific function has been postulated. Writing on page 743 of this issue and in Science respectively, Hoebe et al.3 and Yamamoto et al.4 report that a protein variously named Trif or Ticam-1 is a key part of that second pathway. Animals with defective Trif protein, or lacking Trif altogether, are severely defective in their ability to mount inflammatory responses against viral and bacterial components.
Toll-like receptors (TLRs) come in various subtypes, which detect different components of pathogens. When TLRs on the surface of an immune cell are engaged, two major events are triggered downstream. First, all TLRs trigger the secretion of intercellular messenger proteins called cytokines, which initiate and perpetuate inflammation. Second, the stimulation of the TLR3 and TLR4 receptors also triggers unique signals that lead to the secretion of interferon-. This protein subsequently causes the activation of a series of genes whose products help to eliminate the pathogen. TLR3 detects the presence of double-stranded RNAs (such as some viral genomes) and similar structures, and is therefore crucial in warding off viruses5. TLR4 is stimulated by lipopolysaccharide, a major component of the cell wall of certain bacteria6.
In most cases, the production of inflammatory cytokines in response to TLR stimulation depends on a common molecular pathway that is anchored by an 'adaptor' protein called MyD88. This protein contains a structural region called a TIR domain, which allows it to associate with the intracellular tail of a TLR and to connect the receptor to molecules that propagate a signal further inside the cell. The pathway culminates in the nucleus with the activation of appropriate genes, such as cytokine genes. But MyD88 is not involved in some responses to TLR3 and TLR4 activation, including interferon- production2. So how do such events come about?
A clue lies in the report by Hoebe and colleagues3 that mice with a chemically induced mutation called Lps2 show severe impairment of immune responses mediated by TLR3 and TLR4, and are highly susceptible to viral infections. The Lps2 mutation turns out to occur in the Trif protein — which has a similar receptor-interacting domain to MyD88 — and may inactivate it. The mutation thereby abolishes the activation of the gene-transcription factor IRF-3. Stimulation of this factor was known to be triggered by TLR3 and TLR4, but to be independent of MyD88; so these data distinguish the function of Trif from that of MyD88.
What happens when Trif is missing completely? Yamamoto et al.4 used conventional gene targeting to generate a Trif-deficient mouse strain, and report similar results to those of Hoebe and colleagues. Specifically, they found that Trif-deficient cells from several tissues show impaired TLR3- and TLR4-mediated responses, including all of the responses known to be independent of MyD88. In particular, the almost complete absence of TLR3-mediated signalling in Trif-null and Lps2-mutant cells suggests that Trif is a major adaptor protein for TLR3 signalling3, 4.
It is worth noting that Trif was initially identified7, 8 as a protein that transduces signals from activated TLR3; the new findings confirm this role. What wasn't known, however, is that Trif is also essential in transducing signals from bacterial lipopolysaccharide, via TLR4, for interferon- production (Fig. 1). Curiously, the two groups3, 4 also show that mice with defective or missing Trif protein exhibit impaired production of inflammatory cytokines in response to lipopolysaccharide. As mentioned above, the production of such cytokines proceeds via MyD88 — so it seems that Trif can also collaborate with this protein following TLR4 activation (Fig. 1). The major signalling pathway leading to production of inflammatory cytokines involves the transcription factor NF-
B and various mitogen-activated protein kinase enzymes. Hoebe et al. and Yamamoto et al. found that the activation of these proteins by lipopolysaccharide was only slightly impaired in mice with a mutation in either MyD88 or Trif. But there was no activation at all in mice lacking both adaptor proteins.
Figure 1: Fighting bacteria.
A major component of the cell wall of certain bacteria is lipopolysaccharide, which is recognized by Toll-like receptor 4 (TLR4) on animals' immune cells. TLR4 then initiates the well-characterized MyD88-dependent signalling pathway, which contributes primarily to the activation of the transcription factor NF-B and the production of inflammatory cytokines. Hoebe et al.3 and Yamamoto et al.4 have identified a second pathway, mediated by a MyD88-related molecule called Trif (or Ticam-1). This pathway leads to activation of the transcription factor IRF-3, which controls the production of interferon-. Interaction of interferon- with its receptor initiates signalling through STAT1, which helps to activate further genes. Trif also collaborates with the MyD88 pathway by activating NF-B and various mitogen-activated protein kinases (MAPKs), and promoting cytokine production. Mal is another MyD88-related protein. Thicker arrows represent major pathways.
One advantage to the cell of using multiple adaptors and signalling pathways may be to provide response specificity: different pathogen components engage different TLRs, which recruit different combinations of adaptors, yielding different outcomes appropriate for eliminating those pathogens. Yet another TIR-domain-containing adaptor is Mal, which collaborates with MyD88 in pathways triggered by engagement of TLR2 or TLR4 (refs 9, 10). And at least two more members of this family have been identified, although their physiological functions are unknown11.
The fact that MyD88, Mal and Trif can potentially interact with TLR4 in response to lipopolysaccharide, and that all contribute to downstream signal transduction, raises several questions. Does TLR4 interact with these adaptors directly? Does it bind to them simultaneously or successively? Are the effects of the interactions synergistic, or are they mutually exclusive in a way that provides temporal regulation of different signalling pathways? Hoebe et al. observed that Lps2-mutant mice have two populations of Trif-defective macrophages (a major type of immune cell): one population is severely impaired in lipopolysaccharide-induced cytokine production whereas the other is virtually unaffected. (In contrast, all MyD88-deficient macrophages exhibit impaired lipopolysaccharide-induced cytokine production.) It remains to be seen whether the 'Trif-independent' Lps2 macrophages naturally use an alternative adaptor in place of Trif, or whether they transduce an inherently different type of MyD88-dependent signal.
The bipartite nature of the lipopolysaccharide-induced signalling pathways that lead to cytokine production might prove useful clinically. Inflammatory responses initiated by TLRs are crucial for defence against pathogens — but they can be harmful if uncontrolled. For example, a severe bacterial infection can lead to death from septic shock, which is caused not directly by the pathogen but rather by excessive inflammation. Modulating one branch of the lipopolysaccharide-induced pathways in vivo might eliminate the risk of septic shock while preserving an adequate immune response. Hoebe et al. and Yamamoto et al. have shown that disrupting the Trif-dependent pathway can reduce the toxicity of lipopolysaccharide in vivo. It will be interesting to see how well Trif-deficient animals fare against bacteria.
One final question is which other molecules contribute to the Trif pathway: signal transduction downstream of a given molecule depends on the transducers it can recruit. Apart from certain conserved structural features in the TIR domain, the amino-acid sequence of Trif diverges sharply from that of MyD88. So Trif might recruit a different set of transducers to relay signals downstream. Identifying this molecular pathway12, 13 will be crucial in understanding antiviral defences, and may prove useful in modulating inflammatory responses to bacteria.