The intracellular NOD1 and NOD2 receptors have been found to activate innate inflammation when a condition known as endoplasmic reticulum stress is induced by bacterial infection. See Letter p.394
The ability of mammalian cells to recognize and respond to pathogenic invaders at the earliest stages of infection is crucial for immunity. Many host-cell receptors have been identified that recognize evolutionarily conserved microbial components, such as cell-wall peptidoglycan, to rapidly activate the innate immune system1. In this issue, Keestra-Gounder et al.2 (page 394) describe a curious connection whereby stress in the endoplasmic reticulum (ER), a membrane-bound compartment responsible for the synthesis of secreted cytokine proteins, triggers a cytoplasmic 'danger' signal leading to the release of inflammatory proteins that activate an innate immune response. Even more surprisingly, this ER stress signal is transduced through NOD1 and NOD2 — sensors that respond to fragments of bacterial peptidoglycans — yet takes place in the absence of these molecules.
A wide variety of pathogens hijack and/or manipulate the ER of host cells3. Such infection often activates the stress-responsive control system of this organelle, termed the unfolded protein response (UPR)4,5. Previous work indicates that activation of IRE1α, a transmembrane kinase enzyme required for the UPR, leads to inflammation, although the mechanisms involved remain unclear6,7. To avoid triggering the immune response, several bacterial pathogens actively inhibit the UPR3. By contrast, Brucella abortus, a bacterial pathogen that causes septic abortion in livestock, and which occasionally infects humans through contaminated dairy products, actively promotes UPR-induced inflammation through its secreted effector protein VceC (ref. 8).
Keestra-Gounder et al. demonstrate that blocking either ER stress or the activity of the stress sensor IRE1α during infection with B. abortus in mice markedly lowered production of the cytokine IL-6 and other mediators of inflammation. Conversely, triggering the UPR in mouse macrophage cells, either with chemical inducers or by expressing the B. abortus VceC protein, resulted in IL-6 release. However, a major twist came when the researchers investigated the mechanism by which the UPR triggered IL-6 release and identified the requirement for NOD1 and NOD2. Keestra-Gounder and colleagues found that genetic inactivation of both receptors, or deletion of the downstream enzyme RIP2, greatly reduced cytokine signalling in response to ER stress. Importantly, these sensors were activated in the absence of bacterial ligands, implying that NOD receptors can participate in both bacterial-ligand-dependent and -independent inflammatory responses (Fig. 1).
The authors then investigated UPR-mediated inflammation in B. abortus infection in vivo. When they treated pregnant mice infected with B. abortus with the bile salt TUDCA, a chemical used to mitigate the effects of ER stress, they observed dramatically damped inflammation in the placenta, and increased survival of newborn pups, without a change in the bacterial burden. B. abortus is usually transmitted between animals when a pregnant female becomes infected, develops placental inflammation and has a spontaneous abortion, releasing the infected uterine contents. Animals grazing nearby subsequently eat these infected tissues and become infected. The finding that B. abortus secretes a protein that actively triggers placental inflammation suggests that it may be hijacking a host inflammatory pathway to promote its transmission to the next host.
The mechanism by which IRE1α promotes inflammatory signalling remains unresolved. The authors propose a simple model in which IRE1α stimulation leads to the formation of a complex between TRAF2 (a ubiquitin ligase enzyme that binds to IRE1α), NOD1 and NOD2 (Fig. 1). This then leads to RIP2 activation, presumably by promoting interactions between these proteins. Although this is consistent with the requirement for TRAF2, NODs and RIP2 for inflammatory responses, the question of how the NOD receptors would be activated in the absence of bacterium-derived ligands remains untested. Although NODs have long been known to have a role in peptidoglycan sensing, only more recently have data emerged showing that peptidoglycan can directly bind NOD proteins9. It is possible that the peptidoglycan-independent signalling triggered by ER stress activates NODs through a distinct biochemical mechanism. An intriguing alternative possibility is that host-derived ligands are generated in response to ER stress, and activate NOD receptors in the cytoplasm.
Despite these uncertainties, Keestra-Gounder and colleagues' work compels researchers in the field to expand their thinking about the role of NODs in innate immunity beyond simply the sensing of bacterial ligands. It has long been recognized that the immune response to pathogens in plants relies in part on the monitoring of several normal host-cell physiological processes, and that disruption of these processes by a pathogen triggers an immune response10. It is now becoming increasingly evident that mammalian innate immunity also uses this strategy to 'guard' multiple cellular nodes commonly perturbed by pathogens; these include translation11,12, membrane integrity13, mitochondrial function14, ER homeostasis3 and cytoskeletal integrity15. Strikingly, both the ER and cytoskeletal danger signals are relayed by the NODs in a ligand-independent manner, allowing NODs to participate in sensing a broad range of pathogens2,16.
This ligand-independent NOD signalling also has potential implications for human health beyond the realm of infectious diseases. Several common chronic diseases, including type 2 diabetes and inflammatory bowel disease, are associated with ER stress and inflammation5, and might be therapeutically targeted through this newly identified ER–NOD pathway.Footnote 1
Kagan, J. C. & Barton, G. M. Cold Spring Harb. Perspect. Biol. 7, a016253 (2015).
Keestra-Gounder, A. M. et al. Nature 532, 394–397 (2016).
Janssens, S., Pulendran, B. & Lambrecht, B. N. Nature Immunol. 15, 910–919 (2014).
Walter, P. & Ron, D. Science 334, 1081–1086 (2011).
Garg, A. D. et al. Trends Mol. Med. 18, 589–598 (2012).
Cho, J. A. et al. Cell Host Microbe 13, 558–569 (2013).
Urano, F. et al. Science 287, 664–666 (2000).
de Jong, M. F. et al. mBio 4, e00418–12 (2013).
Grimes, C. L., Ariyananda, L. De Z., Melnyk, J. E. & O'Shea, E. K. J. Am. Chem. Soc. 134, 13535–13537 (2012).
Jones, J. D. G. & Dangl, J. L. Nature 444, 323–329 (2006).
Fontana, M. F. et al. PLoS Pathog. 7, e1001289 (2011).
Vance, R. E., Isberg, R. R. & Portnoy, D. A. Cell Host Microbe 6, 10–21 (2009).
Thurston, T. L. M., Wandel, M. P., von Muhlinen, N., Foeglein, A. & Randow, F. Nature 482, 414–418 (2012).
West, A. P. et al. Nature 520, 553–557 (2015).
Kustermans, G. et al. Biochem. Pharmacol. 76, 1214–1228 (2008).
Legrand-Poels, S. et al. J. Cell Sci. 120, 1299–1310 (2007).
Keestra, A. M. et al. Nature 496, 233–237 (2013).
Related links in Nature Research
About this article
Cite this article
Penn, B., Cox, J. Organelle stress triggers inflammation. Nature 532, 321–322 (2016). https://doi.org/10.1038/nature17882