NINJ1 mediates inflammatory cell death, PANoptosis, and lethality during infection conditions and heat stress

Innate immunity provides the first line of defense through multiple mechanisms, including pyrogen production and cell death. While elevated body temperature during infection is beneficial to clear pathogens, heat stress (HS) can lead to inflammation and pathology. Links between pathogen exposure, HS, cytokine release, and inflammation have been observed, but fundamental innate immune mechanisms driving pathology during pathogen exposure and HS remain unclear. Here, we use multiple genetic approaches to elucidate innate immune pathways in infection or LPS and HS models. Our results show that bacteria and LPS robustly increase inflammatory cell death during HS that is dependent on caspase-1, caspase-11, caspase-8, and RIPK3 through the PANoptosis pathway. Caspase-7 also contributes to PANoptosis in this context. Furthermore, NINJ1 is an important executioner of this cell death to release inflammatory molecules, independent of other pore-forming executioner proteins, gasdermin D, gasdermin E, and MLKL. In an in vivo HS model, mortality is reduced by deleting NINJ1 and fully rescued by deleting key PANoptosis molecules. Our findings suggest that therapeutic strategies blocking NINJ1 or its upstream regulators to prevent PANoptosis may reduce the release of inflammatory mediators and benefit patients.

Editorial Note: This manuscript has been previously reviewed at another journal that is not operating a transparent peer review scheme.This document only contains reviewer comments and rebuttal letters for versions considered at Nature Communications.Mentions of the other journal have been redacted.
Parts of this Peer Review File have been redacted as indicated to maintain the confidentiality of unpublished data.

REVIEWER COMMENTS
Reviewer #1 (Remarks to the Author): I reviewed this manuscript for [redacted] earlier and disagreed with other reviewers on their pointed out weaknesses and perceived lack of novelty.This manuscript went through a high quality revision at [redacted], where the authors provided a large body of new experiments to further support their conclusion.
In my opinion, this is a very compelling and high-quality manuscript about the mechanism of heat stress induced cell death.The authors use genetic deletion or siRNA silencing of all possible key players involved in pyroptosis, apoptosis and necroptosis to dissect their respective contributions to-and to dissect the pathway of heat stress induced cell death revealing PANoptosis as the involved mechanism.
I have no additional questions and consider this manuscript to be of high quality, high importance to the field and an essential complementation and extension of earlier studies.
Reviewer #4 (Remarks to the Author): The authors demonstrate a cell death response to heat stress alongside TLR signaling through TRIF, inducing cell death through Caspase8-Ripk3, Caspase-1/11, and Ninjurin1.
However, these findings seem to conflict with previous published findings of heat stress induced cell death.No resolution or serious discussion is given on why that may be.Yet, I am missing mechanistic depth on how this pathway is triggered.Instead, the authors focus on searching for executioners of this cell death.The complex genetic analysis using a variety of mouse KO is appreciated, and the progression from individual regulators to triple KO and quadruple KO, providing more and more complete cell death reduction.Highlighting the complexity and redundancy of this form of cell death is interesting, but this cannot cover the shortcomings in explanation of how the signals are triggered.In addition, the authors make some conclusions from their cell death-Crispr screen that I don't really find well supported.Overall, this publication has substantial shortcomings.It is missing some essential controls, some necessary additional analyses, and shows a lack of depth on search for mechanism.An earlier publication by Yuan et al Science 2022 also discusses heat stress and this takes some novelty away, in addition to providing different results about signaling molecules -this is also difficult to comprehend.

Specific comments:
It is unsatisfying that little effort has been made to identify which mechanisms may trigger the heat related stress signaling.I cannot consolidate Yuan, F. et al.'s finding that ZBP1 is key to HS cell death, which did not involve TRIF.With the findings here that are dependent on TRIF but do not involve ZBP1.
Especially in Extended figure 4, when performing the 1hr heat stress (HS) without LPS, which appears identical to Yuan, F. et al 's HS method.Could the authors highlight differences between your methods that may explain this? Figure 2A: HS of 30min alone failed to activate GSDMD, caspase-1/11, or GSDME, but earlier was shown that heat stress of 30min did not cause robust cell death.So this is not surprising, it should serve as a negative control.Are the authors trying to claim that HS death without LPS is not pyroptotically driven?For that claim it would be more telling if HS that does cause robust death was used, like HS induced on a longer period of time.Then it could be analyzed for what executor molecules are responsible for cell death without LPS.
Ext. Fig. 5 G-H: The authors claim this LPS/HS condition forms one complex with NLRP3, RIPK3, and Casp8.However, NLRP3 is not probed in the co-immunoprecipitation blots and only one image of one NLRP3 speck colocalizing with Ripk3 and Caspase-8 was provided.
Furthermore, in the one image provided, there are other NLRP3 specks present that do not colocalize with Ripk3 and Casp8.I'm not certain this has to be one big speck complex, as different complexes could be formed in one cell at the same time.but then in these figures, none of them contribute to PI uptake at all.When these molecules are activated, they make pores on the membrane, making the membrane permeable to dyes such as PI.Thus, the authors conclude that another pore must be made that can supersede all of these.However, an alternative explanation could be that these activated molecules are unable to aggregate on the cell membrane and form pores in HS conditions.
Has gasdermin/MLKL oligomerization and pore formation been analyzed?This could be examined.show up as it regulates larger cell ruptures and not PI permeability.Ninj1 is not a topmost enriched gene as it's ranked 5000+ in fold-change and 300+ by p-value.GsdmC isoforms are also ranked 300+ by p-value and in the thousands by fold-change.So, it doesn't appear that this whole genome CRISPR screen supports the claims made.It is then not surprising that silencing all isoforms of GsdmC failed to provide protection.I would say that the interpretations from the authors on this aspect are not supported well.I don't think it is possible to identify Ninj1 from the screen alone.Showing equivalent YOYO-1 uptake as WT cells when undergoing pyroptosis by LPS electroporation.Thus, it is surprising that in LPS+HS and HS conditions, Ninj1 deletion resulted in a significant reduction in PI uptake.I cannot understand this discrepancy and it should be carefully discussed.It would be highly informative then to have a negative control in these PI uptake experiments, and at the very least a no treatment condition.At minimum Furthermore, it appears that Ninj1 deletion did not fully suppress cell death, though it is hard to tell without comparing to no treatment.This may imply that Ninj1 is not solely responsible for executing cell death by HS, but plays a large significant role in its membrane permeability.The discussion proposes cytokine storms as a cause for the pathological effects of HS.
However, the paper does little to no investigation of cytokine release and regulation.Only in Extended Fig. 7 are cytokines even measured, and only IL-1b and IL-18 in one in vitro experiment.More should be done on measuring cytokines throughout, such as by ELISA.
Extended figure 2 demonstrates that TLR3/TLR34 responses are the most robust in causing cell death under heat stress.Usually Poly(I:C) and LPS provide priming for secondary signals, but provide little cell death on their own, yet these primed, heat stressed cells are directly going into programmed death.Heat stress can cause a wide range of effects on the cell.So it is curious why TLR3 and TLR4 respond so robustly to this heat stress induced death, but not the other TLRs tested.TRIF signaling of course is the easiest explanation, and the KOs confirm the expectation that TRIF is important for LPS+HS.Yet TRIF is early in TLR3/TLR4 signaling cascade, and it is hard to dissociate what molecules are affected by HS, and if it is signaling through the expected TRIF pathway.Could the authors address this?
Minor comments: Perhaps testing of other TRAF related responses independent of TLRs, such as TNFR1, ILR1, or RigI, would be informative.As well as looking at TRAF3, other TRAFs, TBK1, or IRF3.
Later the authors mention 1hr of HS generates similar cell death in BMDMs to that seen in LPS primed cells.Why not test if this cell death is also TRIF dependent?
Reviewer #5 (Remarks to the Author): I was asked to review whether Rev #3's questions have been addressed.Upon investigating questions raised by both Rev #3 and #2, and the answers by the authors, I would like to present my analysis below.
The reviewers' questions were largely about the novelty of the study and whether the study provided sufficient advance over previously published papers to warrant its publication in [redacted].I imagine that the authors then transferred their study to Nat Comm by addressing the comments.In my opinion, the study did have some overlap with previous papers, but because NINJ1 is still quite a new molecule implicated in membrane rupture post cell death, the new insights would still serve the field well.The publication in a lower tier, but still very good, journal of this study would further expand the interest of the field in this molecule.In particular, the conclusions that heat stress activates panoptosis in addition to apoptosis and necroptosis and that NINJ1 is involved in this process are useful to the field.I also think that the in vivo data on NINJ1 KO adds additional insights to the current understanding.Thus, I support the publication of the data.Reviewer #4 (Remarks to the Author): The manuscript is improved upon revision and I appreciate that teh authors have added new data and text.A few remaining points: 1-Trigger of heat related stress signaling.
All of these comments in the rebuttal are about how LPS triggers Trif, but not about the mechanism of how heat stress itself initiates signaling.Adding information on this would help the publication.The methods section for this is a necessary addition.Still it is unsatisfactory to claim Ninj1 and GsdmC are hits from this screen, however.These "hits" could have easily been picked out of the literature without the screen, which only partially supports them, being far down on the list.

-Fig 6 Mouse model evaluation.
The paper would benefit from more data evaluating the model overall, to my knowledge this is a new mouse model, and therefore should be given a proper evaluation.Cytokine data gives more information about the immune response but doesn't necessarily evaluate how the heat stress is affecting the mouse physiology.

4-Ext Fig 2
There is still a lack of clarity on what is the secondary signal induced by heat in this cell death pathway.This is linked back to comment 1.
We greatly appreciate the time and expertise the reviewers have contributed to the peer review process.We are grateful for the comments from Reviewer 4 noting that "The manuscript is improved upon revision and I appreciate that the authors have added new data and text."In our revision, we have carefully considered the remaining comment s from Reviewer 4 and followed the expert guidance of the Reviewer to improve the clarity of our manuscript.
We are excited to share this study with the broader scientific community.Overall, this study is the culmination of a decades-old question about innate immune mechanisms of heat stress-mediated pathology and combines an analysis of 40 genetically distinct models to define the regulation of the innate immune response to heat stress.We have included a point-by-point response to the specific comments in the following pages.
Reviewer #4: The manuscript is improved upon revision and I appreciate that teh authors have added new data and text.
We greatly appreciate the time and expertise the reviewer has contributed to reviewing our study.
A few remaining points: 1-Trigger of heat related stress signaling.All of these comments in the rebuttal are about how LPS triggers Trif, but not about the mechanism of how heat stress itself initiates signaling.Adding information on this would help the publication.

-Fig 4A
Crispr cell death screen.The methods section for this is a necessary addition.Still it is unsatisfactory to claim Ninj1 and GsdmC are hits from this screen, however.These "hits" could have easily been picked out of the literature without the screen, which only partially supports them, being far down on the list.

FIGURE REDACTED text redacted
Based on these in vitro results, we also performed an in vivo assessment to conclusively test whether NINJ1 plays a role in the pathophysiology of heat stress.We found that we found that NINJ1-mediated cell death significantly contributes to heat stress-induced mortality (shown below).
Overall, these findings show a clear role for NINJ1 in the inflammatory cell death and pathophysiology induced by heat stress plus PAMPs.Therefore, our CRISPR analysis laid a solid foundation for these subsequent validations to provide multiple lines of evidence supporting the role of NINJ1, making this analysis an essential component of our study.Furthermore, this CRISPR screen is a powerful tool to identify molecules at every stage of the cell death process, including upstream regulators, effectors, and downstream executioners.We feel that the inclusion of these data also allows other investigators to identify their molecules of interest and probe further into their functions in follow up studies.
In the revised manuscript, we have not referred to these molecules as "hits" and instead refer to them as "the topmost enriched genes based on P value" among the cell death executioners for clarity.As suggested by the reviewer, to confirm that the heat stress was inducing a physiological response, we also performed additional analyses from our in vivo heat shock model.We observed a significant increase in serum levels of alanine aminotransferase (ALT) and aspartate aminotransferase (AST) in mice subjected to LPS plus heat stress, which is indicative of liver injury.The levels of ALT and AST were reduced in the serum of Casp1 -/- Casp8 -/-Ripk3 -/-(TKO) and Casp1 -/-Casp11 -/-Casp8 -/-Ripk3 -/-(QKO) mice, indicating that the liver injury rate induced by LPS plus heat stress was decreased in these animals (shown below).However, due to space limitations, these data have not been included in the revised manuscript.

Fig 4A :
Fig 4A: Crispr cell death screen: Much of these data are difficult to understand, including the evaluation put forward in the paper.First, I am trying to find out how the screen was done, and the various descriptions appear incomplete.Information about this needs to be clearly stated.Importantly -what is the read-out in the screen?If it is PI, then Ninj1 should not

Fig
Fig 4E-J: As for Ninj1, Ninj1 is necessary for LDH and DAMP release, as well as lysis of the cell.But in Kayagaki et al. in Nature 2021 where this was discovered, Ninj1 was not responsible for membrane permeability in uptake of YOYO-1, or the release of IL-1b.

Fig 5D :
Fig 5D: Why is there no ladder provided?