HLJ1 amplifies endotoxin-induced sepsis severity by promoting IL-12 heterodimerization in macrophages

Article activity feed

1. 

   Version published to 10.7554/elife.76094 on eLife

   Aug 19, 2022
2. 

   eLife

   Jul 11, 2022

   Author Response

   Reviewer #1 (Public Review):

   This work by Wei-Jia Luo and colleagues elegantly employs in vitro and in vivo models to demonstrate that within the mouse liver, macrophages respond to lipopolysaccharide (LPS) by releasing active IL-12 (IL-12p70), which is a heterodimer of IL-12p35 and IL-12p40. They observed that the availability of "free" IL-12p35 to this heterodimerization process is governed by the molecular chaperone HLJ1. In response to LPS, HLJ1 separates homodimerized IL-12p35 into monomers, which then can heterodimerize with IL-12p40 to form active IL-12p70. This active IL-12 is released from macrophages in the liver, which then act on neighboring natural killer T cells to release interferon gamma. This interferon gamma circulates systemically and is responsible for mortality in a mouse model of endotoxic shock.

   Author Response

   Reviewer #1 (Public Review):

   This work by Wei-Jia Luo and colleagues elegantly employs in vitro and in vivo models to demonstrate that within the mouse liver, macrophages respond to lipopolysaccharide (LPS) by releasing active IL-12 (IL-12p70), which is a heterodimer of IL-12p35 and IL-12p40. They observed that the availability of "free" IL-12p35 to this heterodimerization process is governed by the molecular chaperone HLJ1. In response to LPS, HLJ1 separates homodimerized IL-12p35 into monomers, which then can heterodimerize with IL-12p40 to form active IL-12p70. This active IL-12 is released from macrophages in the liver, which then act on neighboring natural killer T cells to release interferon gamma. This interferon gamma circulates systemically and is responsible for mortality in a mouse model of endotoxic shock.

   Overall, this work is mechanistically compelling and demonstrates a novel multicellular inflammatory pathway that contributes to death in a murine model of endotoxic shock. However, it is unclear if the observed pathway is limited to this highly reductionist model, or if it applies to models that better approximate the complexity of human sepsis. Indeed, the long-standing concept of "cytokine storm" as the major mediator of sepsis has largely failed to yield benefits in clinical trials. These numerous and repeated translational failures cast doubt on the translational validity of reductionist in vivo animal models of sepsis.

   Thank the reviewer’s affirmation. One of the major aims of our work is to identify a novel multicellular inflammatory pathway mediated by HLJ1 that contributes to endotoxic shock. We agree that although our understanding of cytokine storm as the major mediator of sepsis had made dramatic progress over the past decade, these findings could not translate yet into effective treatments. As the reviewer mentioned, almost all clinical trials targeting cytokine effects failed, especially in the context of sepsis. We also know that among several explanations, the appropriateness of in vivo animal models should be concerned (Chousterman et al., 2017). Some approaches to treat cytokine storm were aimed to target the direct tissue consequences of inflammation cascade such as the blood vessel (London et al., 2010). Another possible strategy to treat cytokine storm was to target signaling that promotes cytokine synthesis and secretion (Maceyka et al., 2012). It may be feasible to quell the cytokine storm after infection by targeting upstream signaling, and reducing cytokine synthesis as well as secretion is a valid alternative to direct cytokine antagonism (Chousterman et al., 2017). Furthermore, in this study we found Hlj1−/− mice showed reduced IFN-g and improved survival when treated with daily systemic antibiotics after CLP surgery (Figure 6), indicating that targeting cytokine storm in combination with antibiotics provides a promising therapeutic strategy to treat sepsis. Combined, we think HLJ1-targeting strategy might be a potential therapy to treat cytokine storm-associated sepsis. We emphasized and discussed the concept in the Discussion of our revised manuscript (Page 19, line 441-453).

   We highly appreciated the reviewer #1 and other reviewers raised the same issue. We worked hard and attentively to response comments point-by-point below.

   This raises several specific concerns with regard to the model used by the investigators:

   (1) The authors use a massive dose of LPS that rapidly leads to the death of mice in 24 hours. This massive and rapid mortality is not consistent with human sepsis, which is a more crescendo course with a mortality of ~30%. Indeed, when the authors used a more clinically-relevant model of mild endotoxemia, HLJ1 appeared to have no impact on mortality (Figure 1A).

   Thank for the comment. Indeed, since we observed HLJ1 knockout mice could survive from high dose of LPS, we use 20 mg/kg LPS to perform the subsequent experiments based on these obvious and significant phenomena. We also recognized the importance of administration of low dosages of LPS. To address this issue, we performed additional experiments and made some revisions point-by-point.

   i. Because 4 mg/kg is a common non-lethal dosage to induce TLR4 and IFN-γ signaling (Kunze et al., 2019; Malgorzata-Miller et al., 2016), we performed additional experiment with 4 mg/kg LPS according to the editor’s suggestion. As a result, Hlj1−/− mice showed lower serum levels of BUN, creatinine and ALT and thus less severe organ damage than Hlj1+/+ mice after 4mg/kg LPS treatment. The data are showed in Figure 1C and D of revised Figure 1 (Figure 1).

   ii. We also performed ELISA test and found that serum levels of IFN-γ were lower in Hlj1−/− mice than in Hlj1+/+ mice after 4 mg/kg LPS injection. The result is in Figure 2C of revised Figure 2 (Figure 2).

   iii. Combined, this result indicated the effect of HLJ1 deletion on reducing IFN-γ and alleviating organ injury can also be found during moderate endotoxemia. We described and discussed the result in the revised manuscript (Page 6, line 134-141; Page 18, line 423-437)

   (2) LPS is a model of endotoxemia, not a model of sepsis. Accordingly, it is unclear if the protective benefit of blocking IL-12 will similarly be seen as a live-infection model of sepsis, in which inflammatory signaling may be necessary for pathogen clearance.

   Thank the reviewer for raising these critical issues and providing valuable suggestions. This issue was also mentioned by other reviewers. Although the LPS-induced endotoxemia is a simple model with higher reproducibility and reliability comparing to other sepsis models, it indeed cannot represent actual sepsis and is based on the notion that it is the host’s response to bacteria but not the pathogen itself, that leads to mortality and organ failure (Deitch, 2005). Therefore, according to the reviewers’ suggestion, we performed additional live-infection model of sepsis including cecal ligation and puncture (CLP) which resembles clinical disease and septic shock (Deitch, 2005) to reassure the importance of HLJ1 on sepsis. As a consequence, we found IFN-γ expression was lower in liver and spleen of Hlj1−/− mice comparing to Hlj1+/+ mice (Figure 6A and B). We analyzed serum markers of organ dysfunction and Hlj1−/− mice showed lower serum levels of BUN, creatinine and AST (Figure 6C). H&E staining showed kidney injury at the histology level after CLP surgery, while Hlj1−/− mice showed less severe kidney injury than Hlj1+/+ (Figure 6D). We further found Hlj1−/− mice showed significantly improved survival compared to Hlj1+/+ mice when mice were treated with systemic antibiotics (Figure 6E). Combined, we demonstrated the effect of HLJ1 deletion on attenuation of CLP-induced sepsis with down-regulated IFN-γ, and concluded that the benefit of blocking IL-12 and HLJ1 can similarly be seen as a live-infection model of sepsis. The result is showed as below (revised Figure 6). The corresponding result was also added in the revised manuscript (Page 11-12, line 268-286). Please check it as well as the above responses to other reviewers.

   Page 11-12, line 268-286 "HLJ1 deletion protect mice from CLP-induced organ dysfunction and septic death To address the question whether HLJ1 also regulates IFN-γ-dependent septic shock in live infection model, we performed CLP (cecal ligation and puncture) surgery which more resembles clinical disease and human sepsis. CLP significantly induced transcriptional levels of IFN-γ in the liver of Hlj1+/+ mice comparing to mice receiving sham surgery while Hlj1−/− mice showed significantly lower IFN-γ mRNA than Hlj1+/+ mice (Figure 6A). This phenomenon was not restricted to the liver since lower expression of splenic IFN-γ was also found in Hlj1−/− mice (Figure 6B). The CLP surgery resulted in serious renal and liver damage while Hlj1−/− mice showed alleviated organ dysfunction with significantly lower serum levels of BUN, creatinine and AST (Figure 6C). H&E staining showed kidney injury at the histology level after CLP, while Hlj1−/− mice showed less severe kidney injury than Hlj1+/+ mice (Figure 6D). However, there was no significant difference in survival when comparing Hlj1+/+ and Hlj1−/− mice (Figure 6E). We hypothesized that severe bacteremia contributed to mortality in mice that did not receive any treatment, so we treat mice with systemic antibiotics. As a result, Hlj1−/− mice displayed significantly improved survival compared with Hlj1+/+ mice when mice received daily systemic antibiotics after CLP (Figure 6E). These results implied the agent responsible for bacteria clearance can be combined with immune modulation such as HLJ1 targeting to improve the outcome of sepsis."

   (3) Finally, it is unclear if the findings are only relevant to mice, or if they also have relevance to humans.

   We admit human studies is important, while there are some objective difficulties need to be overcame; for example, cohort identification, individual variation, and clinical considerations. This is our limitation since our findings were only based on animal models and human cell lines. We further performed CLP experiments which is more relevant to human sepsis, while it is not a true human study. That had been added as Figure 6 of our revised manuscript (Figure 6). Actually, based on the present result, we plan to initiate some specific clinical human studies. For example, we plan to collect blood monocytes from critically ill patients from ICU to see whether HLJ1 expression levels in monocytes is higher in patients with sepsis than in patients without sepsis. On the other hand, we also want to know whether HLJ1 expression levels in monocytes or in serum are correlated to inflammatory markers such as C-reactive protein, procalcitonin, and lactate in sepsis patients, because we found serum levels of HLJ1 correlated to IL-12 in mouse. In our unpublished preliminary result, HLJ1 can be detected in serum of patients with sepsis. This inspires us to investigate whether HLJ1 can be a diagnostic or prognostic marker in the future. We anticipate these results can be in our future publications. Thank you very much for your understanding.

   Reviewer #2 (Public Review):

   The authors show that HLJ1 converts misfolded IL-12p35 homodimers to monomers, which maintains bioactive IL-12p70 heterodimerization and secretion. In turn, this contributes to increased IL-12 activity, leading to enhanced IFN-gamma production and lethality in mice challenged with LPS to model sepsis.

   Strengths:

   • Huge and diverse dataset (e.g. in vivo, in vitro, single cell RNAseq, adoptive transfer etc.) with interesting findings that could be of relevance to the field.

   We deeply thank the reviewer for the affirmation. We hope our comprehensive dataset can provide a novel insight of relevance to the field. With this information, we also keep investigating the underlying molecular alteration resulting from endotoxin-induced immune responses. Thank you very much. At the mention of our weaknesses raised by the reviewer, we totally agreed on it and take it very seriously and revised point-by-point. Thank you very much.

   Weaknesses:

   • The flow/narrative of the paper is very hard to follow. This may result from the fact that the order of presented results is a bit puzzling. Normally, one would add-in the cytokine results (now figure 3), after the survival curves in Figure 1. Furthermore, the flow cytometry data presented in Figure 4 is more or less a validation of the scRNAseq data presented in Figure 2 in another organ. Likewise, Figure 5 is sort of a validation of Figure 3 in another organ. The authors seem to jump from organ to organ, from in vivo to in vitro and vice-versa all the time which makes the paper extremely difficult to follow.

   Thank the reviewer for the valuable suggestion. Actually, we were also hesitant to this arrangement in our first submission. We rearranged our results so that the flow/narrative of the paper can be easier to follow:

   1. We moved the result of figure 3 to become figure 2 so that the cytokine array results would after the survival curve results.

   2. The flow cytometry result presented in Figure 4 was moved to Figure 5 so that it would after the result of sc-RNA sequencing.

   3. The qPCR result of pro-proinflammatory cytokines presented in figure 5 was moved to Figure 2-figure supplement 1 so that it would be a validation of cytokine array in another organ.

   In addition, along with other suggestions from reviewers, we have rewritten the introduction and the discussion sections and reorganized whole manuscript so that we can focus more on important issues. All the modification and rearrangement can be checked in the revised manuscript with changes tracked. Please check our revised manuscript. Thank you for your kind suggestions.

   • Use of extremely high dosages of LPS.

   Thank for the comment. This issue had been raised by several reviewers and the editor. Indeed, since we observed HLJ1 knockout mice could survive from high dose of LPS, we use 20 mg/kg LPS to perform the subsequent experiments based on this obvious and significant phenomenon. We also recognized the importance of administration of low dosages of LPS. To address this issue, we performed additional experiments and made some revisions point-by-point.

   i. Because 4 mg/kg is a common non-lethal dosage to induce TLR4 and IFN-γ signaling (Kunze et al., 2019; Malgorzata-Miller et al., 2016), we performed additional experiment with 4 mg/kg LPS according to the editor’s suggestion. As a result, Hlj1−/− mice showed lower serum levels of BUN, Creatinine and ALT and thus less severe organ damage than Hlj1+/+ mice after 4mg/kg LPS injection (Figure 1C). H&E staining showed kidney injury at the histology level after LPS treatment, while Hlj1−/− mice showed less severe kidney injury than Hlj1+/+ mice (Figure 1D). The data are showed in Figure 1C and D (in below) of revised Figure 1 (Figure 1).

   ii. We also performed ELISA test and found that serum levels of IFN-γ were lower in Hlj1−/− mice than in Hlj1+/+ mice after 4 mg/kg LPS injection. The result is in Figure 2C (in below) of revised Figure 2 (Figure 2).

   iii. Combined, this result indicated the effect of HLJ1 deletion on reducing IFN-γ and alleviating organ injury can also be found during moderate endotoxemia. We described and discussed the result in the revised manuscript (Page 6, line 134-141; Page 18, line 423-437)

   • Much of the presented data is replication of previous work. For instance, neutralization of IFN-γ (e.g. Billiau et al., Eur. J. Immunol. 1987; Car et al. J. Exp. Med. 1994) and anti-IL-12 (e.g. Zisman et al., Shock 1997) has been shown to lower mortality in LPS models in mice.

   Thank reviewer for the reminding. We apologized for our unclear description leading to misunderstanding. To carefully and firstly identify the novel role of HLJ1 in sepsis, we actually investigated it on several well-known bases. Indeed, the role of IFN-γ and IL-12 has been recognized in previous studies and their neutralization attenuating LPS-induced endotoxic shock have been reported. However, our study focused on the effect of HLJ1 deletion on IL-12/IFN-γ-axis and septic death. Firstly, we observed IFN-γ and IL-12 decreased after HLJ1 deletion during sepsis. On the one hand, we use IL-12/IFN-γ neutralization and found it could improve survival in wild-type mice rather than in Hlj1 knockout mice, suggesting the importance of HLJ1 in IL-12/IFN-γ-mediated mortality. On the other hand, if the difference of mortality rate across genotypes could become no difference after IL-12 or IFN-γ neutralization, then we can infer that HLJ1 contributes to mortality mainly through IL-12 and IFN-γ signaling. These ideals came from a previous study published in Cell (Ponzetta et al., 2019). The authors elegantly proved the role of Csf3r in IL-12/IFN-γ-axis and subsequent tumor incidence by showing that IFN-γ neutralization can alter the phenotype in wildtype mice rather than in knockout mice. This rationale inspired and prompted us to perform the similar neutralization experiment for understanding the precise role of HLJ1 in sepsis.

   • No true sepsis model is used, only LPS. This is important, as for instance neutralization of IFN-γ and IL-12 has been shown to improve outcome in endotoxemia before (see above), but had no effect on survival in more relevant sepsis models such as cecal ligation and puncture (e.g. see Romero et al., Journal of Leukocyte Biology 2010; Zisman et al., Shock 1997). Furthermore, IFN-γ is even proposed (and used on a small scale) as therapy in sepsis patients to reverse immunosuppression.

   Thank the reviewer raised these critical issues and provided valuable suggestions. It was also mentioned by other reviewers. Although the LPS-induced endotoxemia is a simple model with higher reproducibility and reliability comparing to other sepsis models, it indeed cannot represent actual sepsis and is based on the notion that it is the host’s response to bacteria but not the pathogen itself, that leads to mortality and organ failure (Deitch, 2005). Therefore, we performed additional model including cecal ligation and puncture (CLP) which resembles clinical disease and septic shock (Deitch, 2005) to reassure the importance of HLJ1 to human sepsis. Please see our revised Figure 6 (Figure 6) and responses to other reviewers above.

   In accordance with the previous result from Romero et al showing that IFN-γ neutralization did not improve survival rate, we observed similar survival rate between Hlj1+/+ and Hlj1−/− mice after CLP. However, when they treated mice with systemic antibiotics, IFN-γ knockout mice survived significantly better than wild-type mice (Romero et al., 2010). In CLP model, it is possible that severe bacteremia contributed to mortality in mice that did not receive antibiotics in an IFN-γ-independent manner, so we treated mice with systemic antibiotics immediately after CLP. As a result, we further found Hlj1−/− mice showed significantly improved survival compared to Hlj1+/+ mice when mice were treated with systemic antibiotics after CLP surgery (Figure 6E), indicating that targeting cytokine storm in combination with antibiotics provides a promising therapeutic strategy to treat sepsis. The result is showed in Figure 6E (in below) of revised Figure 6 (Figure 6). This suggests that HLJ1-targeting strategy can be combined with antibiotics to become combined therapy for future clinical applications. We emphasized and discussed the concept in the Discussion of the revised manuscript (Page 18-19, line 441-453).

   Read the original source
3. 

   eLife

   Jul 11, 2022

   Evaluation Summary:

   Wei- Jai et al., demonstrate the protein Human liver DnaJ-like protein (HLJ1) converts misfolded IL-12p35 homodimers to monomers in sepsis. This contributes to increased IL-12 activity, in macrophages which in turn leads to an increased production of IFN-gamma production and lethality in mice. This study suggests that HLJ1 plays an role in regulating IFN dependent mortality and future studies may determine a therapeutic role in blocking this protein.

   (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. The reviewers remained anonymous to the authors.)

   Read the original source
4. 

   eLife

   Jul 11, 2022

   Reviewer #1 (Public Review):

   This work by Wei-Jia Luo and colleagues elegantly employs in vitro and in vivo models to demonstrate that within the mouse liver, macrophages respond to lipopolysaccharide (LPS) by releasing active IL-12 (IL-12p70), which is a heterodimer of IL-12p35 and IL-12p40. They observed that the availability of "free" IL-12p35 to this heterodimerization process is governed by the molecular chaperone HLJ1. In response to LPS, HLJ1 separates homodimerized IL-12p35 into monomers, which then can heterodimerize with IL-12p40 to form active IL-12p70. This active IL-12 is released from macrophages in the liver, which then act on neighboring natural killer T cells to release interferon gamma. This interferon gamma circulates systemically and is responsible for mortality in a mouse model of endotoxemic shock.

   Overall, this …

   Reviewer #1 (Public Review):

   This work by Wei-Jia Luo and colleagues elegantly employs in vitro and in vivo models to demonstrate that within the mouse liver, macrophages respond to lipopolysaccharide (LPS) by releasing active IL-12 (IL-12p70), which is a heterodimer of IL-12p35 and IL-12p40. They observed that the availability of "free" IL-12p35 to this heterodimerization process is governed by the molecular chaperone HLJ1. In response to LPS, HLJ1 separates homodimerized IL-12p35 into monomers, which then can heterodimerize with IL-12p40 to form active IL-12p70. This active IL-12 is released from macrophages in the liver, which then act on neighboring natural killer T cells to release interferon gamma. This interferon gamma circulates systemically and is responsible for mortality in a mouse model of endotoxemic shock.

   Overall, this work is mechanistically compelling and demonstrates a novel multicellular inflammatory pathway that contributes to death in a murine model of endotoxemic shock. However, it is unclear if the observed pathway is limited to this highly reductionist model, or if it applies to models that better approximate the complexity of human sepsis. Indeed, the long-standing concept of "cytokine storm" as the major mediator of sepsis has largely failed to yield benefits in clinical trials. These numerous and repeated translational failures cast doubt on the translational validity of reductionist in vivo animal models of sepsis. This raises several specific concerns with regard to the model used by the investigators:

   (1) The authors use a massive dose of LPS that rapidly leads to the death of mice in 24 hours. This massive and rapid mortality is not consistent with human sepsis, which is a more crescendo course with a mortality of ~30%. Indeed, when the authors used a more clinically-relevant model of mild endotoxemia, HLJ1 appeared to have no impact on mortality (Figure 1A).
   (2) LPS is a model of endotoxemia, not a model of sepsis. Accordingly, it is unclear if the protective benefit of blocking IL-12 will similarly be seen as a live-infection model of sepsis, in which inflammatory signaling may be necessary for pathogen clearance.
   (3) Finally, it is unclear if the findings are only relevant to mice, or if they also have relevance to humans.

   Read the original source
5. 

   eLife

   Jul 11, 2022

   Reviewer #2 (Public Review):

   The authors show that HLJ1 converts misfolded IL-12p35 homodimers to monomers, which maintains bioactive IL-12p70 heterodimerization and secretion. In turn, this contributes to increased IL-12 activity, leading to enhanced IFN-gamma production and lethality in mice challenged with LPS to model sepsis.

   Strengths:
   - Huge and diverse dataset (e.g. in vivo, in vitro, single cell RNAseq, adoptive transfer etc.) with interesting findings that could be of relevance to the field.

   Weaknesses:
   - The flow/narrative of the paper is very hard to follow. This may result from the fact that the order of presented results is a bit puzzling. Normally, one would add-in the cytokine results (now figure 3), after the survival curves in Figure 1. Furthermore, the flow cytometry data presented in Figure 4 is more or less a …

   Reviewer #2 (Public Review):

   The authors show that HLJ1 converts misfolded IL-12p35 homodimers to monomers, which maintains bioactive IL-12p70 heterodimerization and secretion. In turn, this contributes to increased IL-12 activity, leading to enhanced IFN-gamma production and lethality in mice challenged with LPS to model sepsis.

   Strengths:
   - Huge and diverse dataset (e.g. in vivo, in vitro, single cell RNAseq, adoptive transfer etc.) with interesting findings that could be of relevance to the field.

   Weaknesses:
   - The flow/narrative of the paper is very hard to follow. This may result from the fact that the order of presented results is a bit puzzling. Normally, one would add-in the cytokine results (now figure 3), after the survival curves in Figure 1. Furthermore, the flow cytometry data presented in Figure 4 is more or less a validation of the scRNAseq data presented in Figure 2 in another organ. Likewise, Figure 5 is sort of a validation of Figure 3 in another organ. The authors seem to jump from organ to organ, from in vivo to in vitro and vice-versa all the time which makes the paper extremely difficult to follow.
   - Use of extremely high dosages of LPS.
   - Much of the presented data is replication of previous work. For instance, neutralization of IFNg (e.g. Billiau et al., Eur. J. Immunol. 1987; Car et al. J. Exp. Med. 1994) and anti-IL-12 (e.g. Zisman et al., Shock 1997) has been shown to lower mortality in LPS models in mice.
   - No true sepsis model is used, only LPS. This is important, as for instance neutralization of IFNg and IL-12 has been shown to improve outcome in endotoxemia before (see above), but had no effect on survival in more relevant sepsis models such as cecal ligation and puncture (e.g. see Romero et al., Journal of Leukocyte Biology 2010; Zisman et al., Shock 1997). Furthermore, IFNg is even proposed (and used on a small scale) as therapy in sepsis patients to reverse immunosuppression.

   Read the original source
6. 

   Version published to 10.1101/2022.01.05.475083v1 on bioRxiv

   Jan 6, 2022