RIPK3 is a well-known mediator of the necroptosis cell death pathway, which is an important antiviral defence mechanism. In an unexpected twist, RIPK3 has now been shown to also drive neuroprotective inflammation in the central nervous system during West Nile virus infection in a cell-death-independent manner.
Programmed cell death (PCD) is a crucial component of the host antiviral response that restricts viral replication and spread. Necroptosis is a form of PCD that has received significant attention for its role during infection by numerous viral pathogens1. The interaction of death receptors or pathogen-recognition receptors (PRRs) with viral ligands results in the oligomerization and activation of receptor-interacting protein kinase 3 (RIPK3), a central regulator of this death pathway. Upon activation, RIPK3 phosphorylates downstream effectors that mediate membrane rupture and the execution of necroptotic cell death. The identification of viral inhibitors of RIPK3 encoded by murine cytomegalovirus and herpes simplex virus underscore the importance of RIPK3-dependent antiviral necroptosis in host defence2. A recent report by Daniels et al. in Cell3 reveals an unexpected, cell-death-independent function for RIPK3 in host defence against West Nile virus (WNV), expanding the role of this important host antiviral effector.
WNV is a vector-borne flavivirus that infects peripheral organs, lymphoid tissues and the central nervous system (CNS), often resulting in fatal encephalitis. CNS control of WNV involves both the innate immune response and the coordinated influx of inflammatory leukocytes4. However, the CNS is particularly susceptible to immunological injury and has a limited ability to repair damage: thus, immune responses in the CNS must be carefully regulated. Daniels and colleagues initially observed that Ripk3−/− mice succumb to WNV with accelerated progression of symptoms and increased mortality in comparison with control animals. Surprisingly, they also found that mice lacking critical downstream effectors of necroptosis and apoptosis have normal susceptibility to WNV infection, suggesting that neither necroptosis nor extrinsic apoptosis contribute to RIPK3-mediated protection. Although viral loads are transiently elevated at early times post infection, Ripk3−/− mice clear WNV from peripheral tissues as effectively as wildtype mice, concomitant with normal T cell and humoral responses. In contrast, Ripk3−/− animals have significantly higher viral loads in CNS tissues, indicating a tissue- or compartment-specific role for RIPK3 in controlling WNV. Viral replication is normal in Ripk3−/− neuronal cell cultures, suggesting that the role of RIPK3 is not intrinsic to infected neuronal cells, but is potentially involved in host immune or inflammatory responses.
Neuronal cultures derived from Ripk3−/− animals produce normal levels of tumour necrosis factor (TNF) and interleukin-1β (IL-1β) in response to WNV infection, whereas a number of chemokines implicated in WNV-induced neuroinflammation, including chemokine ligand 2 (CCL2) and C-X-C motif chemokine 10 (CXCL10) (ref. 5), are dramatically reduced. This phenotype is also seen in the brains of infected mice, indicating that RIPK3 controls the expression and secretion of a subset of chemokines in neuronal cells of the CNS. Decreased CCL2 and CXCL10 levels correlate with decreased CD4 and WNV-specific CD8 T cells in the brains of Ripk3−/− animals, as well as impaired recruitment of infiltrating inflammatory cells (Fig. 1a). Chemokine expression requires the kinase activities of both RIPK3 and RIPK1, a known RIPK3 interacting partner, and can be elicited by a variety of ligands, showing that PRR-mediated activation of RIPK3 activity drives neuroprotective inflammation during WNV infection.
To directly address the role of RIPK3 activity in driving chemokine expression in neurons without the pleiotropic effects of viral infection, Daniels et al. generated a transgenic mouse in which oligomerization and activation of RIPK3 is controlled by administration of a synthetic ligand6. As expected, addition of the ligand induced necroptosis in murine embryonic fibroblasts. However, synthetic activation of RIPK3 in primary neuronal cultures did not induce cell death but robustly induced expression of CCL2 and CXCL10. Thus, neuron-specific RIPK3 activation is sufficient to induce cell-death-independent chemokine production, highlighting the cell-type specificity of RIPK3 function (Fig. 1b).
The specific molecular mechanisms by which RIPK3 induces pro-inflammatory chemokines in neuronal cells remain unclear. RIPK3 has been previously shown to be crucial for cytokine induction in a dextran sulfate sodium (DSS)-induced colitis model via activation of the NF-κB subunits RelB and p50 (refs 7 and 8). However, induction of IL-1β or TNF-α by WNV infection is largely unaffected in Ripk3−/− mice or cells, suggesting the difference does not derive from a widespread defect in NF-κB signalling. DSS-induced cytokine expression is dependent on the RIP-homotypic interaction motif (RHIM) of RIPK3 for activation and translocation of both RelB and p50, but does not require kinase activity7. It is thus possible that the protein–protein interactions that mediate RIPK3 activation vary in different cell types. RIPK1 and RIPK3 form a complex through RHIM-mediated interactions. Whether they drive chemokine expression in neurons as a complex or as part of coordinate signalling pathways in neurons remains an open question. Contrary to the DSS model, WNV-induced chemokine expression in neurons requires the kinase activity of both RIPK3 and RIPK1, raising the additional possibility of a tissue-specific target for these kinases. RIPK3 kinase activity has also been implicated in control of cytokine expression in bone-marrow-derived cells in response to lipopolysaccharide through an ERK1/2-dependent pathway9. Whether similar mechanisms are at play in the brain is still unknown. A more complicated question is why neuronal cells in which RIPK3 is activated produce chemokines, rather than undergoing necroptosis, especially considering that RIPK3 has been implicated in neuronal necroptosis in a number of neurodegenerative diseases10. The necroptotic pathway is thus clearly intact in neuronal cells, so why WNV infection, toll-like receptor (TLR) activation (below), or forced activation of RIPK3 results in cytokine secretion, while other insults result in death, remains an important question to answer.
Another outstanding question is how infected neurons sense WNV to activate RIPK3. Daniels et al. show that stimulating neurons with a variety of synthetic TLR ligands results in RIPK1 and RIPK3 kinase-dependent induction of CCL2 and CXCL10, implicating both TRIF- and Myd88-dependent signalling pathways. What role, if any, do specific TLRs play in neuronal RIPK3-dependent production of cytokines during WNV infection? TLR3 and TLR4 are both known activators of RIPK3-dependent cell death1. Patients with TLR3 genetic deficiencies are more susceptible to herpes simplex virus encephalitis11, and TLR signalling clearly plays a critical role in controlling many other viral infections in the CNS12. However, it remains possible that additional or alternative PRRs, including cytoplasmic nucleic acid sensors such as RIG-I or DAI, may contribute to RIPK3 and RIPK1 activation during WNV infection. Understanding the neuronal-specific WNV sensing mechanism may provide important insights into other neurotropic infections and shed light on the dual role of RIPK3 in neuroprotection and neurodegeneration. Ultimately, the identification of the cell-death-independent neuroprotective role for RIP kinases is an important first step in understanding and ultimately treating viral infections in the CNS.
Sridharan, H. & Upton, J. W. Trends Microbiol. 22, 199–207 (2014).
Guo, H., Kaiser, W. J. & Mocarski, E. S. Med. Microbiol. Immunol. 204, 439–448 (2015).
Daniels, B. P. et al. Cell 169 301–313 (2017).
Suthar, M. S., Diamond, M. S. & Gale, M. Jr. Nat. Rev. Microbiol. 11, 115–128 (2013).
Lim, J. K. & Murphy, P. M. Exp. Cell. Res. 317, 569–574 (2011).
Orozco, S. et al. Cell Death Differ. 21, 1511–1521 (2014).
Moriwaki, K., Balaji, S., Bertin, J., Gough, P. J. & Chan, F. K. Cell Rep. 18, 2441–2451 (2017).
Moriwaki, K. et al. Immunity 41, 567–578 (2014).
Najjar, M. et al. Immunity 45, 46–59 (2016).
Zhao, H. et al. Cell Death Dis. 6 e1975 (2015).
Zhang, S. Y. et al. Science 317, 1522–1527 (2007).
Suh, H. S., Brosnan, C. F. & Lee, S. C. Curr. Top. Microbiol. Immunol. 336, 63–81 (2009).
The authors declare no competing financial interests.
About this article
Cite this article
Ragan, K., Upton, J. Host response: Neurons loosen the gRIP of death. Nat Microbiol 2, 17090 (2017). https://doi.org/10.1038/nmicrobiol.2017.90