Adjuvants enhance the immune response to an antigen with which they are mixed. Thus, they often form an essential part of vaccines. The identification of new adjuvants is therefore an important challenge for vaccine development in the fight against the world's most common infectious killers, including malaria and tuberculosis, for which even natural infection itself is insufficient to induce protective immunity. The development of adjuvants has traditionally been focused on both modified microbial products and synthetic mimics of these molecules. An exceptionally large percentage of these immunologically active microbial molecules are ligands for the family of Toll-like receptors (TLRs), archetypical activators of innate immunity1. The most studied TLR ligand is bacterial lipopolysaccharide (LPS; endotoxin), the main component of the outer membrane of Gram-negative bacteria. LPS activates animal cells through TLR4 in cooperation with its coreceptor, MD-2 (ref. 1). Another TLR ligand that has been studied intensively is poly(I:C), considered to be a surrogate of viral double-stranded RNA, which uses TLR3 (refs. 1,2). In this issue of Nature Immunology, Hoebe and colleagues3 address the mechanism behind the adjuvant effect induced by these two compounds, both ex vivo and in vivo.

This study follows earlier reports4,5 of mice with a mutation in Trif (also known as TICAM-1), one of five known adaptor proteins containing a Toll–interleukin 1 receptor–resistance domain6. Trif functions as an adaptor protein for TLR3 and TLR4 signaling. Trif-deficient mice have a profound defect in activation of the transcription factor NF-κB and the related release of proinflammatory cytokines after exposure to poly(I:C) and LPS4,5. Complicating this picture is the finding that Trif is not the sole adaptor associated with TLR3 and TLR4 (Fig. 1). MyD88, the first adaptor to be discovered, has been believed to associate with all the known TLRs to induce proinflammatory cytokines6. MAL (also known as Tirap), associates with MyD88 to mediate TLR2 and TLR4 signaling6. Finally, the most recently discovered adaptor protein, called TRAM, mediates TLR4 signaling. Engagement of TRAM leads to activation of NF-κB, and the subsequent release of inflammatory cytokines such as tumor necrosis factor (TNF) and interleukin 6; it also induces cytokines regulated by the transcription factors IRF-3 and IRF-7, such as interferon-β (IFN-β)7,8. It is becoming increasingly apparent that the specificity and fine tuning of responses to TLR ligands are determined by both the relevant receptor and its ability to interact with specific combinations of downstream adaptor molecules.

Figure 1: Innate immune signaling pathways that enhance antigen presentation.
figure 1

C.C.

After LPS or poly(I:C) (double-stranded RNA; dsRNA) challenge, both TLR-dependent and TLR–independent signaling pathways lead to upregulation of costimulatory molecules (such as CD80 and CD86) and adjuvant effects, as suggested by Hoebe3 and others9. In this issue, Hoebe et al.3 suggest Trif (TICAM-1) as the common adapter molecule in TLR3- and TLR4-mediated responses leading to upregulation of CD80 and CD86, whereas a new locus (which they call dsRNA1) may be responsible for a non-TLR response to poly(I:C). Furthermore, IFN-RI signaling apparently is central in a feedback loop leading to upregulation of CD80 and CD86. Of the other TLR adapters, MyD88 interacts with MAL (Tirap), mediating the release of proinflammatory cytokines through TLR4, whereas it is unclear whether MyD88 interacts with TLR3 at all. A new adapter called TRAM7 seems to pair with Trif, mediating MyD88-independent signals through TLR4. Some TLRs may function as intracellular receptors.

Full size image

Hoebe et al. delve into the mechanisms mediating upregulation of the costimulatory molecules CD80, CD86 and CD40 after poly(I:C) and LPS stimulation. Their study focuses on the phenotype of Trif mutant mice (which express a dominant negative version of Trif known as Lps2), as well as mice deficient in TLR3, MyD88, interferon receptor type I (IFN-RI) and the double-stranded RNA–interacting protein kinase (PKR). Hoebe et al. show type I interferon, most likely IFN-β, is crucial in the pathway leading to costimulatory molecule upregulation induced by both LPS and poly(I:C). Distinct differences in mechanisms between the adjuvant activity induced by LPS and poly(I:C) emerged. Most startling was the finding that poly(I:C)-induced upregulation of the costimulatory molecules was mediated in part by a TLR3-Trif-independent pathway. Indeed, Hoebe et al. identified a genetic locus on mouse chromosome 7 that is responsible for this. This locus does not correspond to any of the known TLRs. An earlier study9 supports the idea of a central function for IFN-RI adjuvant effects and agrees with the existence of a TLR3-independent pathway for poly(I:C). However, that earlier study9 suggests only limited involvement of type I interferon in LPS adjuvant activity.

Despite the clear-cut demonstration of involvement of type I interferons ex vivo, the question remains as to whether IFN-RI signaling is also essential for the in vivo adjuvant effects of LPS, poly(I:C) and their mimetics. In addition, certain issues will take time to be resolved. For example, the Trif mutant mice are clearly defective in the adjuvant effect mediated by LPS, but so are MyD88-deficient mice, which show normal upregulation of costimulatory molecules. The accessibility of different TLR adaptors to their receptors in different cell types and tissues, and the ability of these adaptors to generate adjuvant signals by TLR ligands, may prove a critical determinant of the subsequent acquired immune response. As four adaptor proteins seem to cooperate to mediate the full response to LPS and other TLR4 agonists7, additional studies of adjuvanticity in mice with targeted deletions in one or more of each of these adaptor molecules are needed.

As with all important findings, Hoebe et al.3 cover exciting new territory but leave certain issues unresolved. How important is the TLR3-Trif-independent pathway that is suggested? This pathway is not sensitive to inhibition by the drug 2-aminopurine, in contrast to the Trif pathway. Although this drug is thought to be an inhibitor of PKR, PKR-null mice upregulated CD80 and CD86 indistinguishably from wild-type mice, indicating that 2-aminopurine works by an alternative mechanism to suppress Trif signaling. In vivo testing of mice that seem to lack the TLR–independent pathway of adjuvanticity should answer questions related to this pathway. Earlier studies showed a severe impairment of poly(I:C)-induced cytokine release in cells from MyD88-deficient mice2, whereas Hoebe et al. provide unequivocal data against this conclusion. Although several lines of evidence indicate that MyD88 has limited, if any, interaction with TLR3, this question needs to be resolved definitively. If MyD88 is dispensable for TLR3 signaling, it will shake the common belief that it is a 'core' adapter for all TLR signaling pathways. It has also been suggested that a TLR3-independent, PKR-dependent pathway induces IFN-α by intracellular (transfected) poly(I:C)10. Finally, a mystery well worth addressing is how different TLRs induce similar effects with different sets of adaptor molecules. For example, CpG oligodeoxynucleotides apparently induce all of their known effects through TLR9 and MyD88 alone.

The experiments of Hoebe et al. emphasize the potential for TLR ligands as future constituents of human vaccines, but simultaneously warn against simplistic conclusions. Few compounds have been licensed as adjuvant components of human vaccines, and improvements in the pharmacopoeia of adjuvants can be anticipated in the near future. Nevertheless, a key feature for a successful adjuvant is its safety for widespread use. Development of TLR ligands with a reasonable therapeutic index may prove challenging; after all, these are the very same molecules that cause fever and end-organ damage in septic patients.

What is the future for TLR ligands as vaccine adjuvants? Deciphering the complicated signaling pathways for the TLRs will undoubtedly lead to new ideas for how to manipulate this family of recep- tors for a maximal beneficial response. The achievement of additional insights into signaling events that are mediated by each of the TLRs and their adaptors, and the delineation of non-TLR-mediated pathways with similar effects, seem a certain road to travel to achieve our goal of developing safe and powerful vaccine adjuvants.