HMGB1-activatied NLRP3 inflammasome induces thrombocytopenia in heatstroke rat

Experimental animals

Because estrogen has certain effects on HS pathogenesis and organ injury (Chen et al., 2006), only adult male, pathogen-free Sprague-Dawley (SD) rats (Experimental Animal Center of the General Hospital of Southern Theatre Command, license number for animal experimentation: SCXK, Guangdong 2016-0041), weighing 220 to 250 g, were used in this study. The rats received humane care according to the criteria outlined in the Guide for the Care and Use of Laboratory Animals (National Institutes of Health publication 86-23, 1985 revision). We adhered to the guidelines for animal care of the General Hospital of Southern Theatre Command. The animal ethics committee of the General Hospital of Southern Theatre Command of PLA approved this study.

Rat model of HS

Rats housed for 6 h at ambient temperature (25 °C ± 0.5 °C) with a humidity of 35% ± 5%, 126 rats were randomly divided into 21 groups, with 6 rats per group. Animals have free access to standard food and water. To induce HS, rats were placed in a prewarmed incubator maintained at 39.5 °C ± 0.2 °C with a relative humidity of 60.0% ± 5.0%. The rectal temperature (Tr) was monitored at 10-min intervals using a thermocouple (BW-TH1101; Biowill, Shanghai, China), which was inserted 5.5 cm into the rectum of each rat. The time point at which the Tr reached 43 °C was considered a reference point of HS onset (Geng et al., 2015). The rats were removed from the incubator after HS onset, transferred to room temperature (22.0 °C ± 0.5 °C), and fed with adequate food and water. The rats in control group were sham-heated at a temperature of 25 °C ± 0.5 °C and a humidity level of 35% ± 5%. Ethyl pyruvate (EP, 50 mg/kg body weight, MCE, USA) was injected intraperitoneally (i.p.) to inhibit HMGB1 release, or the same volume phosphate buffered solution (PBS) were injected as a negative control before rats subjected to heat stress. A neutralizing antibody specific for rats against HMGB1 (3 mg/kg body weight, Santa Cruz Biotechnology, Santa Cruz, CA, USA) was immediately injected i.p. into the rats before heat stress. In some rats, TLR4 neutralizing antibody specific for rats (8 µg/kg body weight, Abcam, Cambridge, MA, USA), RAGE neutralizing antibody (80 µg/kg body weight, R&D, USA), or both were injected i.p. before rats subjected to heat stress. Other rats were injected i.p. with Ac-Tyr-Val-Ala-Asp-chloromethylketone (ac-YVAD-cmk, 0.3 mg/kg body weight, Enzo Biochem Inc, New York, USA) which is the caspase-1 inhibitor or PBS containing 2.8% dimethyl sulfoxide (DMSO) as a control before rats subjected to heat stress. Some rats were given MCC950 (20 mg/kg body weight, Selleckchem, Houston, TX, USA), a specific blocker of NLRP3, 1 h before heat stress. Some rats were injected i.p. with the antioxidant N-acetylcysteine (NAC, 300 mg/kg body weight, MCE) immediately before heat stress. At the end of the experiment, rats were anesthetized by inhalation of isoflurane, and sacrificed to collect blood from abdominal aorta.

Platelet count

50 µL of whole blood was collected into a sodium citrate anticoagulation tube to measure the platelet count using a blood cell analyzer (BC-3000; Mairui, Shenzhen, China).

Platelet isolation

Platelets were isolated as described previously (Qiao et al., 2018). Briefly, three mL of anticoagulant blood samples were centrifuged at 150 g for 15 min to obtain platelet-rich plasma (PRP), and about 3/4 of the PRP was carefully pipetted into two volumes of modified Tyrode buffer (12 mM NaHCO₃, 138 mM NaCl, 5.5 mM glucose, 2.9 mM KCl, 2 mM MgCl₂, 0.42 mM NaH₂PO₄, 10 mM HEPES, pH 7.4). The mixture was centrifuged at 37 °C at 500 g for 8 min in the presence of PGE1 (0.1 µg/mL; Cayman Chemical) and apyrase (1 U/mL, Sigma-Aldrich), washed twice with CGS buffer (120 mM sodium chloride, 12.9 mM trisodium citrate, 30 mM D-glucose, pH 6.5), and resuspended in modified Tyrode buffer.

Transmission electron microscopy examination

One volume of PRP prepared above was added into 45 volumes of 3% glutaraldehyde and to stand for 4 h. The samples were centrifuged at 800 g for 10 min and the supernatant was discarded. The pellet was washed with 0.1 M phosphate buffer for twice, postfixed with 1% osmium tetroxide at 4 °C for 1 h, and dehydrated in a graded ethanol series and embedded in Epon. Epon-embedded platelets were cut into sections (70–80 nm) with an Ultracut UCT ultramicrotome, transferred to copper grids, and stained with uranyl acetate and lead citrate. The sections were then observed with transmission electron microscope (TEM) (H-7560, Hitachi, Japan).

Flow cytometric analyses

Freshly isolated platelets (10⁶ to 10⁷) were resuspended in one mL of modified Tyrode buffer, and treated with 0.2% triton before to measure intracellular cytokines. A minimum of 10,000 events per gate was acquired using a flow cytometer (BD LSR Fortessa). P-selectin (CD62P) was determined using the anti-CD62P monoclonal antibody (1:1000; Biolegend, USA), and TLR2 (toll-like receptor 2), TLR4, and RAGE were assessed with antibodies targeted to TLR2 (1:500; Abcam), TLR4 (1:1000; Abcam), and RAGE (1:1000; Abcam). NLRP3 expression was evaluated by an NLRP3 antibody (1:1000; Abcam), and activation of caspase-1 was assessed by the fluorescent probe green fluorescent-labeled inhibitor of caspase-1 (FLICA, 1:100, Immunochemistry Technologies, Bloomington, MN, USA), which irreversibly binds to activated caspase-1. IL-1β was evaluated by the anti–IL-1β antibody (1:1000; Biorbyt). Mitochondrial-derived reactive oxygen species (ROS) were detected using the molecular probes Mito-SOX (for mitochondria O₂•-, 1:1000; Invitrogen), DHE (for cytoplasmic O₂•-, 1:1000; Invitrogen), and DCF-DA (for H₂O₂, 1:1000; Invitrogen). All indices were doubly labeled with a phycoerythrin- or fluorescein isothiocyanate–conjugated CD61 monoclonal antibody (1:1000; Biolegend) to label platelets. The results were analyzed using FlowJo software 7.6.

Confocal fluorescence microscopy

Immunofluorescence for NLRP3 and apoptosis-associated speck-like protein containing a caspase-recruitment domain (ASC) was performed according to previous research (Feng et al., 2014). Platelets attached on a poly-L-lysine-coated coverslip were fixed with pre-cooled methanol for 20 min, washed twice with PBS (135 mM NaCl, 4.7 mM KCl, 10 mM Na₂HPO₄, 2 mM NaH₂PO₄, pH 7.4), and blocked with 10% fetal calf serum (FCS) in PBS. Then, platelets were incubated with goat anti-rat NLRP3 antibody (1:1000; Abcam) or rabbit anti-rat ASC antibody (1:500; Santa Cruz) for 30 min, washed twice with PBS, and labeled with corresponding Dlight 488 and Dlight 594 conjugated secondary antibodies (1:1000; Invitrogen) for 30 min respectively. Preparations were analyzed under a laser scanning confocal microscope (Fluo View FV1000; Olympus), and FV1000 operations software was used for recording.

Plasma HMGB1 and IL-1β levels measurement

According to the manufacturer’s instructions, plasma HMGB1 levels were measured by a commercially available HMGB1 enzyme-linked immunosorbent assay kit (Shino-Test Corporation, Sagamihara, Japan), and IL-1β levels were measured by a commercially available IL-1β enzyme-linked immunosorbent assay kit (Invitrogen, Waltham, MA, USA).

Statistical analysis

All data were presented as the mean ± SD unless stated otherwise. All data followed a normal distribution except for one group, which was analyzed with Kruskall-Wallis. For the data followed a normal distribution, statistical significance was determined with the least significant difference t-test or one-way analysis of variance (ANOVA). IBM SPSS Statistics 26 software was used for data. P < 0.05 was considered statistically significant.

HS induces platelet activation and thrombocytopenia in HS rats

The platelet count began to decline at 3 h after HS onset (Fig. 1A), whereas the expressions of IL-1β (Fig. 1D) and CD62P (Fig. 1E) in platelets increased progressively during the same period after HS onset. All changes were peaked at 9 h after HS onset. The platelet parameters such as mean platelet volume (MPV) and platelet distribution width (PDW) were assessed. MPV (Fig. 1B) and PDW (Fig. 1C) at 3 h and 6 h after HS maintained in normal ranges, with no statistical differences comparing to the sham group. However, both MPV and PDW at 9 h after HS increased significantly compared to those in the control group after 9 h of HS (P < 0.01). Therefore, for the remaining experiments, we observed heat-stressed animals only at the time point of 9 h after HS onset.

NLRP3 inflammasome activation in platelets in HS rats

The association of NLRP3 and ASC as well as caspase-1 represents activation of the inflammasome. The colocalization of NLRP3 (green) and ASC (red) in platelets was observed under a confocal microscopy (Fig. 2C). In addition, flow cytometric analyses showed that NLRP3 (Fig. 2A) and cleaved caspase-1 (Fig. 2B) increased at 3 h and reached the peak levels at 9 h after HS onset. All results indicate the HS-inducing NLRP3 inflammasome activation in platelets.

NLRP3 inflammasome mediates platelet activation and thrombocytopenia in HS rats

To investigate whether HS-activated NLRP3 inflammasome mediates platelet activation and thrombocytopenia, we injected rats i.p. with MCC950, a specific inhibitor of the NLRP3 inflammasome, before the rats subjected to heat stress. The results showed that inhibited NLRP3 inflammasome significantly down-regulated the levels of cleaved caspase-1 (Fig. 3A), IL-1β secretion (Fig. 3B) and CD62P expression (Fig. 3C) and reduced the drop of platelet count (Fig. 3D) at 9 h after HS onset. Moreover, pharmacological inhibition of caspase-1 activity with ac-YVAD-cmk also markedly down-regulated the IL-1β secretion (Fig. 3E) and CD62P expression (Fig. 3F) as well as the drop of platelet count (Fig. 3G) at 9 h after HS onset.

HMGB1 activates platelet NLRP3 inflammasome in HS rats

Previous studies found that early increased extracellular HMGB1 levels in HS were negatively correlated with its prognosis (Tong et al., 2013a). HMGB1 activates the NLRP3 inflammasome in liver cells, which in turn promotes hepatocyte pyroptosis (Geng et al., 2015). Consistent with previous studies (Tong et al., 2011), we found that plasma HMGB1 concentrations in HS rats were gradually increased (Fig. 4A). We also unraveled the role of HMGB1 in inducing NLRP3 inflammasome activation, platelet activation, and thrombocytopenia in HS. We injected rats i.p. with EP to inhibit HMGB1 release and used anti-HMGB1 neutralizing antibody to block the effect of HMGB1 before submitting rats to HS.

EP inhibiting or neutralizing antibody against HMGB1 significantly reduced the expression of NLRP3 (Fig. 4B) and cleaved caspase-1 (Fig. 4C) at 9 h after HS onset. Notably, levels of secreted IL-1β (Fig. 4D), expression of CD62P (Fig. 4E), and thrombocytopenia (Fig. 4F) at 9 h after HS onset were significantly alleviated after pretreatment with EP and anti-HMGB1 antibody. These results indicate that HMGB1 plays an important role in activating NLRP3 inflammasome and possibly participates in platelet activation and thrombocytopenia in HS.

In the early stages of sterile inflammation, HMGB1 activates the NLRP3 inflammasome through TLR4 or RAGE receptors (Xu et al., 2013). As shown in Figs. 4G–4I, expressions of TLR2, TLR4, and RAGE in platelets were gradually increased after HS onset. To define the receptors that mediate HMGB1-induced NLRP3 inflammasome activation, we applied anti-TLR4 and anti-RAGE neutralizing antibodies before heat exposure. Either anti-TLR4 or anti-RAGE neutralizing antibody significantly inhibited HS-induced NLRP3 (Fig. 4J) and cleaved caspase-1 (Fig. 4K) expression in platelets at 9 h after HS onset. In addition, the combination of anti-TLR4 and anti-RAGE neutralizing antibody induced greater inhibition of NLRP3 and cleavage of caspase-1 expression. Synchronous changes were also observed in HMGB1-induced upregulated expression of IL-1β (Fig. 4L), and thrombocytopenia (Fig. 4N) at 9 h after HS. Anti-TLR4 or anti-RAGE neutralizing antibody inhibited HS-induced CD62P expression, while combination of anti-TLR4 and anti-RAGE neutralizing antibody made no significantly effect compared with single neutralizing antibody (Fig. 4M). These data indicate that HMGB1 activates NLRP3 inflammasome via both TLR4 and RAGE receptors, which is possibly associated with platelet activation and thrombocytopenia in HS.

HMGB1 activates platelet NLRP3 inflammasome by upregulating ROS in HS rats

ROS mainly includes the superoxide anion (O2–), hydrogen peroxide (H2O2), and the hydroxyl radical (HO–) (Leytin, 2012). Many studies confirmed HMGB1 could up-regulate ROS (Tsung et al., 2007; Pietraforte et al., 2014), which was the key to activate NLRP3 inflammasome (Dostert et al., 2008; Zhou et al., 2011). To study the changes of ROS in platelets and the effects of ROS on the HMGB1-induced NLRP3 inflammasome in HS, we used a fluorescence probe to detect the expression of mitochondrial O2•-, cytoplasmic O2•-, and H₂O₂ in platelets in HS rats. We found that mitochondrial O2•- (Fig. 5A), cytoplasmic O2•- (Fig. 5B), and H₂O₂ (Fig. 5C) increased at 3 h after HS onset. Inhibiting HMGB1 significantly reduced the level of ROS, which indicated that HMGB1 was involved in high-level of ROS in HS (Figs. 5D–5F).

To explore the relationship between ROS and the NLRP3 inflammasome activation, we pretreated rats with the antioxidant NAC. In platelets, NAC significantly inhibited the expression of NLRP3 (Fig. 5G) and cleaved caspase-1 (Fig. 5H) at 9 h after HS onset. Similar results were observed for secreted IL-1 β (Fig. 5I), CD62P expression (Fig. 5J), and thrombocytopenia (Fig. 5K). These results indicate that ROS is responsible for platelet NLRP3 inflammasome-induced activation and thrombocytopenia in HS.