TREM‐1 induces pyroptosis in cardiomyocytes by activating NLRP3 inflammasome through the SMC4/NEMO pathway

Sepsis often causes cell death via pyroptosis and hence results in septic cardiomyopathy. Triggering receptors expressed in myeloid cells‐1 (TREM‐1) may initiate cellular cascade pathways and, in turn, induce cell death and vital organ dysfunction in sepsis, but the evidence is limited. We set to investigate the role of TREM‐1 on nucleotide‐binding oligomerization domain‐like receptors with pyrin domain‐3 (NLRP3) inflammasome activation and cardiomyocyte pyroptosis in sepsis models using cardiac cell line (HL‐1) and mice. In this study, TREM‐1 was found to be significantly increased in HL‐1 cells challenged with lipopolysaccharide (LPS). Pyroptosis was also significantly increased in the HL‐1 cells challenged with lipopolysaccharide and an NLRP3 inflammasome activator, nigericin. The close interaction between TREM‐1 and structural maintenance of chromosome 4 (SMC4) was also identified. Furthermore, inhibition of TREM‐1 or SMC4 prevented the upregulation of NLRP3 and decreased Gasdermin‐D, IL‐1β and caspase‐1 cleavage. In mice subjected to caecal ligation and puncture, the TREM‐1 inhibitor LR12 decreased the expression of NLRP3 and attenuated cardiomyocyte pyroptosis, leading to improved cardiac function and prolonged survival of septic mice. Our work demonstrates that, under septic conditions, TREM‐1 plays a critical role in cardiomyocyte pyroptosis. Targeting TREM‐1 and its associated molecules may therefore lead to novel therapeutic treatments for septic cardiomyopathy.


Introduction
Sepsis results in a high mortality which is primarily due to multiple organ failures [1,2]. The heart is often affected by multiple organ failures, namely septic cardiomyopathy. The incidence of septic cardiomyopathy in septic patients can be as high as 60% [3] and is associated with an increased risk of death [4]. The underlying mechanisms of septic cardiomyopathy, however, still remain elusive [5]. The nucleotidebinding oligomerization domain-like receptor with pyrin domain-3 (NLRP3) inflammasome is an important mediator in the innate immune system [6]. It contains three components of NLRP3, caspase-1 and apoptosis-associated speck-like protein containing a caspase recruitment domain. Inflammasome activation by LPS (primer), followed by nigericin (activator), results in caspase-1 activation, which causes cleavage of precursor forms of IL-1b and IL-18 into mature (cleaved) and active cytokines that are then secreted [6,7]. While the main products of the NLRP3 inflammasome, IL-1b and IL-18, are strongly proinflammatory, they have been shown to have a prominent suppressive effect on cardiac myocytes [8,9], which has been considered the main mechanism for septic cardiomyopathy. It is shown that neutralization of IL-18 attenuates LPS-induced myocardial injuries [9]; however, the upstream pathway of NLRP3 inflammasome activation remains unknown.
Structural maintenance of chromosome protein complex-4 (SMC4) epigenetically enhanced the transcription of NEMO by recruiting H4K5ac to the nemo promoter and promoted innate activation of NF-jB and IRF3 [21]. Herein, we set to investigate the relationship between TREM-1 and SMC4, and the role of TREM-1 on cardiomyocyte pyroptosis and its association with NLRP3 inflammasome activation in sepsis models using cardiac cell lines (HL-1) and mice.

TREM-1 was expressed in cardiomyocyte HL-1 after stimulation by LPS
TREM-1 was thought to be expressed exclusively in myeloid cells; however, there are reports of TREM-1 also being expressed in other cell types such as gastric epithelial cells and hepatic endothelial cells [22,23]. Therefore, to assess whether TREM-1 was expressed in cardiomyocytes and if it played an important role in the septic heart, we first examined its expression in cardiomyocyte HL-1. By using a confocal laser scanning microscope and qRT-PCR, we found that TREM-1 was heavily expressed in cardiomyocytes after stimulation by LPS when compared with a negative control (Fig. 1A,B). Furthermore, this upregulated expression was maximized following an LPS challenge at 20 lgÁmL À1 for 12 h (Fig. 1C,D).

Stimulation of TREM-1 exacerbated pyroptosis of cardiomyocytes induced by LPS
Pyroptosis is a common cause of cell death and contributes to septic cardiomyopathy as TREM-1 can act synergically with pattern recognition receptors to intensify inflammation [16]. We utilized both LR12 (an inhibitory peptide of TREM-1) and monoclonal antibody of TREM-1 (mABTREM-1) to either block or amplify the effect of TREM-1 respectively. As demonstrated in Fig. 2A, the LPS + NI + mABTREM-1 group showed more cell death than the others, while the cardiomyocyte in the LPS + NI + LR12 group was less injured than in the LPS + NI group. Similarly, the results from flow cytometry showed that 22.1% of cardiomyocytes underwent pyroptosis after stimulation of LPS + NI, compared with 29.4% after stimulation of LPS + NI plus mABTREM-1, this was reduced to 16.4% following the addition of LR12 (Fig. 2B). In western blot analysis (Fig. 2C), pyroptosis-associated proteins, such as inflammasome protein NLRP3 and CASPASE-1 p10, were upregulated after stimulation of LPS, this effect was further enhanced by adding the TREM-1-specific stimulant mABTREM-1, and was reversed by LR12. Furthermore, downstream proteins GSDMD and CASPASE-1 were cleaved to a greater extent after stimulation of mABTREM-1 plus LPS and NI (Fig. 2C). The mRNA levels of TREM-1, NLRP3, IL-18 and IL-1b were also significantly increased after LPS + NI + mABTREM-1 treatment, while LR12 suppressed these effects (Fig. 2D).

SMC4 interacted physically with TREM-1
In order to investigate the exact mechanism and downstream effects of TREM-1, we used immunoprecipitation to examine the interaction between TREM-1 and SMC4. LPS increased the expression of TREM-1 (green), while TREM-1 and SMC4 were colocalized in HL-1 cell line (Fig. 3A). For immunoprecipitation, the amount of TREM-1 pulled down by SMC4 in the LPS group was much more than that in the control group, indicating the physical interaction between SMC4 and TREM-1 (Fig. 3B).
Knockdown of TREM-1 or SMC4 attenuated cardiomyocyte pyroptosis The expression of NEMO, P-p65, NLRP3, GSDMD-N and mature IL-1b was increased by LPS and NI stimulation, while RNA interfering with TREM-1 or SMC4 reversed the effects of LPS and NI (Fig. 5A,B). Through qRT-PCR analysis, we demonstrated similar changes in the mRNA level of NEMO, NLRP3 and IL-1b ( Fig. 5C) in these different groups.

Inhibition of TREM-1 attenuated cardiomyocyte pyroptosis in mice
The expression of proteins TREM-1, NEMO, P-p65, NLRP3, N-GSDMD and mature IL-1b (Fig. 6A,B) and the level of mRNAs of TREM-1, SMC4, NEMO, NLRP3 and IL-1b were significantly increased in the CLP-treated mice compared to the control group (Fig. 6C). The upregulated proteins and mRNAs were, however, inhibited in mice that were given LR12. The change in NLRP3 expression in cardiac tissue was further confirmed by immunohistochemistry (Fig. 6D). Furthermore, the cardiac tissue in CLP mice appeared to be oedematous and injured compared to the control and LR12 groups, and there was a higher degree of cell infiltration observed in the CLP mice (Fig. 6E).
Inhibition of TREM-1 improved cardiac function and increased the 7-day survival of septic mice Cardiac function, measured using ejection fraction (EF), fractional shortening (FS) and ratio of early (E)- , and the Student's t-test was used to determine the significance. (C, D) Western blot analysis of the dose responses for expression of TREM-1 in HL-1 exposed to LPS (0-20 lgÁmL À1 ) for 12 h at 37°C, and time course for expression of TREM-1 in HL-1 exposed to 20 lgÁmL À1 LPS. Data are mean AE SD (n = 5) and one-way ANOVA followed by Dunnett's multiple-comparison test was used to determine the significance. *P < 0.05, **P < 0.01 vs. control group.

1551
The to-late (A) ventricular filling velocities (E/A ratio), was severely impaired in the CLP mice. Although not significant, the cardiac function of mice in the LR12 group had improved (Fig. 7A). Furthermore, LR12 significantly improved the 7-day survival rate of CLP mice (Fig. 7B).

Discussion
This study demonstrated that TREM-1 was upregulated in both cardiomyocytes of septic mice and cardiomyocyte cell lines following an LPS challenge. TREM-1 activation caused cardiomyocyte pyroptosis and, hence, septic cardiomyopathy while inhibition of TREM-1 decreased cardiomyocyte pyroptosis, improved cardiac function and prolonged survival of septic mice. Our data further suggested that under septic conditions, TREM-1 activation led to a series of alterations in downstream cellular signalling. This included changes in SMC4, NEMO, NF-kB and NLRP3 release causing pyroptosis via caspase-1 and GSDMD cleavage. These cascade changes precipitated the release of cytokines (e.g. TNF-a, IL-1b and IL-18), which, in turn, led to further organ dysfunction and even death (Fig. 8).
TREM-1 is a type I transmembrane protein consisting of a single extracellular immunoglobulin (Ig)-like domain, a transmembrane region with a positively charged lysine residue and a short cytoplasmic tail lacking any signalling motifs [24]. TREM-1 can be homooligomerized through its ectodomain [25], while its transmembrane domain mediates binding with the signalling adaptor DNAX activation protein 12 [13]. Several cellular signallings including CARD9 [13], PLCc and ERK1/2 [14,15] interact with TREM-1, leading to the activation of transcription factors AP1, CREB and NF-jB [16]. This process results in the upregulation of proinflammatory cytokines, contributing to the development of inflammatory disease [10]. Previous studies have demonstrated that, by blocking TREM-1, organ dysfunction is mitigated and mortality of septic mice is reduced [26][27][28][29]. Our previous work found that TREM-1 was associated with the progression of septic cardiomyopathy, and that TREM-1 activation aggravated myocardial damage and LPS-induced left ventricular systolic dysfunction [30]. In a model of neonatal sepsis, the inhibition of TREM-1 by M3 (an inhibitor specified for the interaction of eCIRP and TREM-1) reduced the expression of TREM-1 on cardiomyocytes, attenuated cytokine release and improved cardiac function and the survival of septic mice [26]. Our current study further demonstrates that activation of TREM-1 triggers cardiomyocyte pyroptosis, causing septic cardiomyopathy, and that SMC4 is likely a downstream molecule of TREM-1.
The condensin protein complex, containing SMC4, SMC2 and non-SMC subunits, plays a key role in the segregation of replicated genomes during prokaryotic and eukaryotic cell division [31]. It contributes to DNA damage repair, recombination and replication [32], while also determining the three-dimensional landscape of interphase nuclei to regulate gene expression [33]. SMC4 (and partly SMC2) promotes innate activation of NF-jB and IRF3 by recruiting H4K5ac to the nemo promoter and epigenetically enhancing the transcription of NEMO [21]. Subsequently, the interaction of TREM-1 and SMC4 promotes NF-jB activation resulting in the upregulation of inflammatory genes and proteins, e.g. TNF-a and IL-1b, as shown in our study.
The NLRP3 inflammasome is an intracellular multiprotein complex, and also the source of inflammatory cytokines of IL-1b and IL-18 [6,7]. Activation of the NLRP3 inflammasome first requires priming, followed by a second signal [34]. The priming signal process is triggered by pattern recognition receptors, like TLR4, TNFR and IL-1R. These signals lead to the activation of nuclear factor kappa B (NF-kB), which ultimately increases pro-IL-1b and NLRP3 synthesis [18]. The second signal involved in NLRP3 inflammasome activation is triggered by a vast number of structurally dissimilar agonists, including silica, nigericin and ATP [35,36]. All these changes contribute to vital organ dysfunction including cardiac injury in sepsis. This is evidenced by the fact that astrocyte pyroptosis promotes neuronal injury in sepsis [37], while the genetic deletion of NLRP3 reverses CLP-induced acute kidney injury [38]. Moreover, LPS administration intratracheally promotes the development of acute lung injury, while the genetic deletion of NLRP3 or CASPASE-1 reverses lung dysfunction [39]. Likewise, NLRP3 inflammasome activity was found to be increased in atrial cardiomyocytes leading to paroxysmal atrial fibrillation but the genetic inhibition of NLRP3 prevented AF development in transgenic mice [40]. Additionally, activation of the NLRP3 inflammasome played a critical role in calcineurin transgene-induced structural heart disease [41], whereas knockout of NLRP3 improved the cardiac function of septic mice induced by CLP [42]. Activation of TREM-1 increases the synthesis of pro-IL-1b and NLRP3 by enhancing the phosphorylation of NF-kB [17,43]. In contrast, blocking TREM-1 by LR12 reduces the expression of NLRP3 and suppresses the progression of acute lung injury [18]. Furthermore, blocking TREM-1 by LP17 reduced the expression of NLRP3 and mitigated pyroptosis in an experimental subarachnoid haemorrhage model [19]. Consistent with this work, we found , and one-way ANOVA followed by Dunnett's multiple-comparison test was used to determine the significance. (C) qRT-PCR analysis of TREM-1, SMC4, NLRP3, NEMO and IL-1b in HL-1 cells after stimulated by LPS (20 lgÁmL À1 ) and NI (10 lM). HL-1 was pre-treated with siSMC4 or siNC according to the group setting to knockdown the expression of TREM-1 or SMC4. Data are mean AE SD (n = 6), and one-way ANOVA followed by Dunnett's multiple-comparison test was used to determine the significance. *P < 0.05 and **P < 0.01 vs. control group; # P < 0.05 and ## P < 0.01 vs. LPS + NI group.

1555
The that NLRP3 contributed to CLP-induced cardiac dysfunction, and that downregulation of NLRP3 (by TREM-1 inhibition) attenuated cardiac dysfunction. Interestingly, our data indicate that cardiomyocytic pyroptosis is very likely due to the caspase-1dependent canonical NLRP3 inflammasome pathway but not the caspase-11-dependent non-canonical inflammasome pathway as it was not changed under our experimental conditions (data not shown). Sepsis is the most important cause of morbidity and mortality in intensive care units [44], and myocardial dysfunction is often concomitant with severe sepsis and septic shock [1]. Effective treatments for sepsisinduced cardiomyopathy are limited and novel approaches targeting cellular signalling, e.g. TREM-1, related to pyroptosis, may provide hope for future therapeutic developments. Along with our previous study [30], current data suggest that inhibition of TREM-1 by LR12 preserves left ventricular systolic function and improves the survival of septic mice.
We realized that most work was done in the cell line while sepsis is a systemic disease that multi-cell processes are involved. However, our in vivo data also corroborated with our in vitro data and hence the conclusions are likely valid. In addition, the relationship between TREM-1 and NLRP3 inflammasome in sepsis remains unknown and would need to be studied further. TREM-1 À/À and NlLRP3 À/À mice are being used to answer this question in our laboratory currently. Furthermore, an interaction between TREM-1 and SMC4 was not investigated in in vivo setting and warrants further study in the future. In addition to inconsistency with the in vivo results, we found that in an HL-1 cell line, TREM-1 was decreased after treated with monoclonal antibodies (agonists) and increased after treated with LR12 (inhibitor) (Fig. 4). It was reported that TREM-1 may be shed from the membrane by metalloproteinases, which is called sTREM-1 [46]. It is possible that LR12 prevented TREM-1 from shedding. However, in in vivo study, LR12 may stop the escalated inflammatory response, which made the TREM-1 less expressed, and the exact mechanism also needs to be further studied.
In conclusion, our study demonstrated the central role of TREM-1/SMC4/NEMO signalling in triggering the development of cardiomyocyte pyroptosis and hence septic cardiomyopathy. Inhibition of TREM-1 decreased cardiomyocyte pyroptosis, improved cardiac function and prolonged the survival of septic mice. Treatment of cardiac inflammation through pharmacological inhibition of TREM-1 or other targets in this pathway may provide significant benefits in treating or preventing the progression of septic cardiomyopathy and hence improving the outcome of septic patients.

Animal model of sepsis
The study was approved by the University Committee on Use and Care of Animals University Committee, Guangdong Medical University, Zhanjiang, Guangdong, China. Specific pathogen-free male C57BL/6 mice (6-10 weeks old, weight 25-30 g; Dien Corporation, Guangzhou, China) were used in the experiments. Sepsis was induced by the caecal ligation puncture (CLP) procedure [45] under ketamine anaesthesia (80 mgÁkg À1 body weight) (Gutian Corporation, Fujian, China). The caecum was exposed, ligated at half the distance between the distal pole and the base of the caecum and punctured, and a small amount of faeces was expressed. The wound was closed in layers. The TREM-1 inhibitor LR12 (5 mgÁkg À1 , i.p.) or 0.9% normal saline was given once to the mice intraperitoneally for all experiments except for the 7-day survival analysis in which the LR12 (5 mgÁkg À1 , i.p.) was given right after CLP was completed and then repeated once every 12 h for 7 days.
The sham controls were subjected to the same surgical procedure without ligation and puncture. The animals were euthanized by cervical dislocation 12 h after CLP and then they were cardiac perfused with 5% formalin. Their hearts were harvested and embedded in wax for further analyses.

Echocardiography
At 12 h after CLP surgery, echocardiography was performed under inhaled isoflurane anaesthesia as reported previously [30] with a Vevo 2100 System (Fujifilm Visual Sonics, Shanghai, China).

Flow cytometry
The single-cell suspension of the cultured HL-1 was washed twice and 1 9 106 cells were suspended in 50 lL PBS, supplemented with 1% FBS. The cells were stained with FLICA (FAM-YVAD-FMK, ImmunoChemistry Technologies, Cat: #9145) for 40 min at 37°C and then counted with a BD FACS Celesta flow cytometer, and data were analysed with FlowJo software.

Immunoprecipitation
To detect the TREM-1-SMC4 interaction by immunoprecipitation, LPS-primed HL-1 was treated either with or without 10 mM nigericin for 6 h. Cells were lysed, sonicated and centrifuged at 12 000 g at 4°C for 30 min. Supernatants were then incubated overnight at 4°C with either anti-SMC4 (ab250130, Abcam, Shanghai, China) or anti-IgG antibody (ab172730, Abcam). The sample antigen-antibody mixture was added to a tube containing pre-washed magnetic beads and incubated at room temperature for 1 h. The beads were collected using a magnetic stand and the samples were heated with a Lane Marker Sample Buffer at 96-100°C for 10 min and then used for western blot.

Quantitative real-time PCR
Total RNA was extracted from cultured HL-1 cell lines or mouse heart tissues using TRIzol reagent (Sigma-Aldrich, USA) for quantitative real-time PCR with the following specific mouse primers, Β-actin:

ELISA
Cells were cultured and treated as indicated. The culture medium was then analysed for mouse TNF-a and IL-1b with an ELISA kit (EK0527, EK0394, BOSTER).

Statistical analysis
All values were expressed as means AE SD and then analysed using one-way analysis of variance (ANOVA) followed by Dunnett's multiple-comparison test. For comparison of two groups, the Student's t-test was used. The survival rate was evaluated by the Kaplan-Meier test. All data were analysed with PRISM (GraphPad, La Jolla, CA, USA). A P-value of < 0.05 was considered to be of statistical significance.