The compartmentalization of immune responses into innate and adaptive components was a conceptual breakthrough that revolutionized our understanding of immunology. In particular, the idea that activated dendritic cells (DCs) instructed T cell differentiation rapidly gained traction in the 1990s (1, 2). This was driven in part by the discovery of TLRs, including TLR4 and its ligand LPS (3, 4). These exciting developments ushered in an era of mechanistic studies investigating how LPS and other TLR ligands worked as adjuvants, boosting the development of adaptive immune responses to otherwise weakly immunogenic Ags (5).
The discovery that Th1 differentiation was promoted by activated DCs through their production of IL-12 was an important advance (6), but the signals required to initiate Th2 immune responses remained somewhat mysterious. In 2001, Schnare et al. (7) reported that the development of Ag-specific T cell responses to OVA in CFA was markedly attenuated in mice deficient for the TLR adaptor protein MyD88. In these experiments, OVA-driven recall IFN-γ production was abolished in MyD88-deficient mice, but IL-4 production was unaffected. These results suggested that Th2 immunity might develop in a TLR/MyD88-independent manner.
With this background, the Pillars of Immunology article by Eisenbarth et al. (8) convincingly demonstrated that inhaled LPS could, in fact, act as a Th2-promoting adjuvant, but this was critically dependent on the dose of LPS. The authors used a straightforward experimental design in which inhalation of OVA alone, without adjuvant, induces a state of tolerance (9). Using endotoxin-free OVA, Eisenbarth et al. (8) added different doses of inhaled LPS to OVA during sensitization and reported that OVA-dependent Th2 recall responses preferentially developed when low-dose LPS (100 ng) was used as adjuvant. A higher LPS dose (100 µg) instead led to the production IFN-γ and not Th2 cytokines in lymphocytes from lung-draining lymph nodes. This cytokine profile was closely mirrored by the nature of the inflammatory response that developed following OVA inhalation challenge. Mice sensitized with OVA plus low-dose LPS developed airway eosinophilia, mucus production, and elevated serum IgE levels after subsequent OVA challenge, whereas mice sensitized with OVA plus high-dose LPS did not. These asthma-like features were markedly attenuated in TLR4-deficient mice sensitized with inhaled OVA plus low-dose LPS, but not in mice sensitized by i.p. injection of OVA with aluminum hydroxide.
The observation that inhalation of low-dose LPS primes for Th2 responses has been widely replicated and suggested a potential mechanism by which human subjects might become sensitized to inhaled allergens. Since its initial publication in 2002, this article has influenced subsequent research in different areas of immunology. Importantly, it laid the groundwork for mechanistic experiments investigating how innate immune cells instruct Ag-specific Th2 adaptive immunity during allergen sensitization. Eisenbarth et al. (8) proposed that low-dose LPS resulted in activation of DCs without inducing IL-12 production, which then induced Th2 differentiation in responding naive T cells. In a follow-up study, Piggott et al. (10) showed that the effects of low-dose inhaled LPS were attenuated in MyD88-deficient mice, in association with reduced MHC class II and CD86 expression in lung CD11c+ DCs. Adoptive transfer of wild-type OVA-pulsed bone marrow–derived DCs reconstituted eosinophilic airway inflammation in MyD88-deficient mice, as did the administration of inhaled TNF-α, instead of LPS, during OVA sensitization (10). In a later article, Tan et al. (11) demonstrated that lung DCs from mice sensitized with low-dose LPS plus OVA indeed induced Th2 differentiation when cocultured with naive OVA TCR-specific T cells in vitro. This rigorous approach definitively established that Th2 promoting DC develop in the lung in response to low-dose inhaled LPS as adjuvant. Although an IL-12lo phenotype is still the hallmark of a Th2-promoting DC, we now recognize that these DCs are heterogeneous (12) and can develop in response to signaling by different TLR and non-TLR ligands (13). Despite decades of subsequent study, it still remains challenging to predict how exactly a given DC subset will influence the differentiation of naive T cells.
Using variations of the mouse model reported by Eisenbarth et al. (8), several groups refined our understanding of the how inhaled LPS induces Th2 responses in the lung (14–18). It seemed plausible that airway DCs themselves would be directly activated by LPS, because they express TLR4, are often interdigitated in the airway epithelium, and seem likely to sample airway luminal contents. However, subsequent studies revealed that the picture was more complicated and that DCs are likely to be activated in an indirect manner. Hammad et al. (14) performed reciprocal bone marrow chimeras between wild-type and TLR4-deficient mice and discovered that TLR4 expressed by nonhematopoietic cells played a critical role during sensitization with inhaled LPS or house dust mite (HDM) extracts (which are known to be contaminated with LPS). Because this was accompanied by significant attenuation of LPS-induced lung DC migration, the authors concluded that inhaled LPS (and HDM) activates lung DC indirectly, via the release of soluble mediators from structural lung cells, such as airway epithelial cells. McAlees et al. (18) confirmed that TLR4 expression by nonhematopoietic cells was required for eosinophilic airway inflammation driven by HDM or inhaled LPS as adjuvant and also used conditional deletion to further refine our understanding of the key TLR4-expressing cell types. Using Shh-Cre to delete TLR4 deletion specifically in epithelial cells, HDM-driven airway eosinophilia was significantly attenuated, in conjunction with reduced levels of the epithelial-derived Th2 promoting cytokines IL-25 and IL-33 in lung lavage fluids. These and other studies helped underscore the importance of airway epithelial cells as key sensors of inhaled allergens and adjuvants (19), which then drive Th2 mucosal immune responses by releasing mediators that recruit and activate other cells in the airway epithelium (20).
Eisenbarth et al. (8) showed that higher doses of inhaled LPS during sensitization resulted in neutrophilic airway inflammation after OVA inhalation challenge and proposed that this reflected a switch from Th2 to Th1 immune responses in the lung. Subsequent studies by Wilson et al. (16) and Whitehead et al. (17) clearly demonstrated that inhaled LPS also primes for airway Th17 differentiation and neutrophilic airway inflammation in a dose-dependent manner. In contrast to the requirement for TLR4 expression in epithelial cells during the initiation of eosinophilic airway inflammation, deleting TLR4 in hematopoietic cells using Vav-Cre selectively attenuated both LPS and HDM-induced neutrophilic airway inflammation (18). Thus, expression of TLR4 in different cell types regulates qualitatively different types of inflammatory responses in the lung. In addition to dose, the timing of LPS administration also matters. Sustained administration of even low-dose LPS can attenuate Th2 responses and airway inflammation, reflecting inhibitory responses that are driven by epithelial cells as well as regulatory T cells (17, 21–23).
Eisenbarth et al. (8) also showed that inhaled TNF-α could substitute for LPS as a Th2-promoting adjuvant in both TLR4- and MyD88-deficient mice (8, 10). A key adjuvant role for TNF-α in models of allergen Th2 sensitization was confirmed by other groups (24, 25). In an elegant study, Whitehead et al. (25) showed that LPS-dependent allergen sensitization involved TNF-α produced from CD11c+ lung macrophages, which reprogrammed airway epithelial cells in a TNFR1-dependent manner. Similar to LPS, the ability of inhaled TNF-α to prime for Th2 responses was also dose-dependent, being more pronounced at lower doses. The exact mechanisms by which TNF-α–conditioned epithelial cells induced Th2 responses require further study, but interestingly did not seem to involve IL-25, IL-33, or thymic stromal lymphopoietin (25).
In summary, using a simple experimental design, Eisenbarth et al. (8) set the stage for decades of research investigating the initiation of Th2 allergic immune responses in the lung. Although a little bit of LPS is a Th2-promoting adjuvant, it is now clear that relatively small differences in the dose and timing of inhaled LPS during initial Ag exposure can markedly impact the subsequent Ag-specific adaptive immune response, both quantitatively and qualitatively. The ability of inhaled LPS to act as a Th2-promoting adjuvant is also dependent on exposure history, because mice previously tolerized with inhaled OVA react differently than naive hosts (26). Although great progress has been made in deciphering cross-talk between the innate and adaptive arms of the immune system during allergen sensitization in the airway, we still have more to learn. For example, it is not clear where exactly inhaled allergens deposit in the respiratory tract and how they traverse the epithelial barrier. The role of innate lymphoid cells and other mucosal immune cells during allergen sensitization is an exciting area for future research (27). Because humans are continually exposed to inhaled LPS, allergens, and other pattern recognition receptor ligands, it will be important to try and mimic real-world exposure conditions wherever possible in models of allergen sensitization.
The author has no financial conflicts of interest.