In the course of the GENESIS project, several important findings pertaining to the role of nucleic acid-sensing and signaling cascades were made. In the following, the three most significant studies are briefly highlighted:
1. The human ‘DNA inflammasome’
Studying mechanisms of DNA recognition within human immune immune cells, we were able to establish a critical role for the cGAS-STING axis in NLRP3 inflammasome activation. As such, re-evaluating the role of AIM2 in human cells, we made the surprising finding that AIM2 is dispensable for DNA-mediated inflammasome activation in human myeloid cells (Gaidt et al. Cell 2017). Instead, in these cells, the ‘DNA inflammasome’ was dependent on NLRP3. We identified a cGAS-STING-dependent signaling cascade that resulted in a lytic form of cell death upstream of NLRP3 activation. Conducting a forward-genetic screen, our attention was directed towards lysosomes as potential modulators of this cGAS-STING-dependent cell death cascade leading to NLRP3 activation. Indeed, following up on these results we found that STING was recruited to lysosomes upon activation, resulting in membrane permeabilization and thus the induction of lysosomal cell death (LCD). This type of cell death triggered potassium efflux and thus NLRP3 activation. All in all, these results uncovered a new type of signaling cascade in human myeloid cells that functions to connect the predominant DNA recognition module (cGAS-STING axis) to its membrane perturbation sensing module (NLRP3 inflammasome). The connecting link between these two modules is the induction of lysosomal cell death, which we speculate to be an ancient functionality of STING. From an evolutionary perspective, these results would also explain why AIM2 is a rather ‘novel invention’ in the repertoire of innate immune defense mechanisms that is not stringently conserved across different mammalian lineages. Indeed, some mammalian species have completely lost AIM2, although they are facing the same microbial pathogens as AIM2-sufficient species.
2. RNase T2 functions upstream of TLR8
TLRs detect a diverse set of molecular patterns, ranging from bacterial cell wall components to nucleic acids. One special subgroup of TLRs, comprising of TLR7,8 and 9, resides in the endolysosomal compartment where it senses RNA and DNA degradation products. Among these TLRs, TLR8 plays a particularly important role in the human system, in that it is broadly expressed in myeloid cells, while functional studies have indicated that it is especially important for the detection of bacterial RNA. While TLR8 is also expressed in the murine system, it does not share the same functionality with human TLR8. Biochemical and structural studies had indicated that TLR8 has two binding sites for RNA degradation products, while it had remained elusive how these RNA fragments are generated. We went on to address this question by systematically knocking out RNA degrading enzymes annotated to be localized to the endolysosomal compartment. Doing so, we found that RNase T2 played a non-redundant role in TLR8 activation following its stimulation with complex RNA molecules. Mechanistically, RNase T2 was identified to cleave RNA molecules between purine and uridine residues, thereby contributing to the supply of catabolic uridine and generating fragments that are terminated with a purine-2',3'-cyclophosphate (Greulich, Wagner et al. Cell 2019). While the former engages the 1st binding pocket of TLR8, the latter has been found to occupy the 2nd binding pocket of TLR8. Functional studies further revealed that the recognition of the bacterial pathogen Staphylococcus aureus was severely compromised in the absence of this enzyme – receptor axis.
3. NLRP1 is a sensor for dsRNA
NLRP1 was one of the first NLR proteins to be identified as an inflammasome forming sensor. However, ever since its description, its exact function in non-self recognition has remained elusive. More recently, it was realized that NLRP1 functions as a sensor for proteolytic activity. Here, it was shown that pathogen-encoded proteases lead to the cleavage of NLRP1 and thereby its activation. This process also called ‘functional degradation’ can be explained by its unique domain architecture that contains an auto-cleaving FIIND domain. As a function of its autoprocessing, NLRP1’s C-terminal region containing the CARD, which transduces the signal towards inflammasome formation, is non-covalently associated with the ‘NLR-part’ of the protein. Upon degradation of the N-terminus, the C-terminal fragment can be released to initiate inflammasome formation. Human NLRP1 is highly expressed in keratinocytes so that we went on to screen a panel of different viruses for their potential NLRP1 inflammasome activation. Doing so, we identified the positive-strand RNA virus (+ssRNA virus) Semliki Forest virus (SFV) as a potent activator of the NLRP1 inflammasome. As a +ssRNA virus SFV generates ample amounts of dsRNA during its life cycle and the formation of this special nucleic acid intermediate coincided with NLRP1 inflammasome activation. Indeed, transfecting keratinocytes with long dsRNA recapitulated the effect of SFV infection, indicating that NLRP1 detects dsRNA. Employing recombinantly expressed NLRP1 constructs, we found that NLRP1 directly bound dsRNA with high affinity, while this binding was mainly due to its LRR domain. Moreover, dsRNA-binding triggered NLRP1 to exert ATPase activity, a typical hallmark of NLR proteins being activated. In summary, these studies identified human NLRP1 as a direct sensor for dsRNA (Bauernfried et al. Science 2020).
