HgCl2 is a known toxic agent that can cause cell death at high concentrations (10, 11). To determine the cell toxicity of HgCl2 on immune cells, we treated mouse splenocytes with various concentrations (10μM to 1M) of HgCl2. Incorporation of propidium iodine as a marker of cell death was then determined by flow cytometry. Low concentrations of HgCl2 up to 100μM do not cause cell death ( Figures 1A, B ). Non-toxic doses under 20μM were used in this study for in vitro experiments.

To determine if mercury could regulate the secretion of IL-1β and IL-18, we treated LCWE-primed BMDMs with HgCl2. HgCl2 increases the secretion of IL-1β cytokine in LCWE-primed BMDMs ( Figure 2A ) compared to LCWE or HgCl2 treatment alone. Similar to previous observations (13), stimulation with LCWE alone does not lead to the secretion of IL-18. However, when LCWE-primed BMDMs were treated with HgCl2, there was the secretion of IL-18 cytokine. IL-18 secretion increases in a dose-dependent fashion with increasing concentrations of HgCl2 ( Figure 2B ). Furthermore, treatment with LCWE but not HgCl2 increases the protein expression of NLRP3 ( Figure 2C ). When LCWE-primed cells were treated with HgCl2, there is increased protein expression of NLRP3 from BMDMs ( Figure 2C ). When HgCl2 was chelated using Dimercaptosuccinic Acid (DMSA) ( Figure 2D ) or Dimercaptopropionic Acid (DMPS) (data not shown), secretion of inflammasome mediated cytokines was significantly reduced. These results suggest that HgCl2 amplifies activation in LCWE-primed immune cells leading to the secretion of inflammasome mediated cytokines IL-1β and IL-18.

Previous studies show that LCWE contains a TLR2 ligand (14). To determine if HgCl2 treatment could activate inflammasomes with a known TLR2 ligand, we used Pam3Cys to prime BMDMs. Similar to LCWE, treatment of HgCl2 with Pam3Cys primed cells leads to the secretion of both IL-1β and IL-18 cytokines ( Figures 3A, B ). Taken together, the results suggest that HgCl2 at low concentrations could act as a novel danger signal in providing optimal activation for the secretion of IL-1β and IL-18.

HgCl2 is a known agent of Ca²⁺ mobilization (15). Studies have shown that Ca²⁺ mobilization is critical for NLRP3 inflammasome activation (5, 16–19). Therefore, to assess whether HgCl2 mediated inflammasome activation is Ca²⁺ dependent, we used extracellular and intracellular Ca²⁺ inhibitors and measured the secretion of IL-1β and IL-18. When live BMDMs from C57BL/6 were assessed for intracellular Ca²⁺ levels using confocal spinning disc microscopy, HgCl2 treated cells acquired more intracellular Ca²⁺ when compared to untreated cells ( Figure 4A ). We next examined whether Ca²⁺ mobilization plays a role in HgCl2 mediated activation of the inflammasome.

Extracellular Ca²⁺ channels were blocked using Ca²⁺ channel inhibitor 2-APB and PLC pathway inhibitor U73122. Both IL-1β and IL-18 production in cells treated with 2-APB and U73122 was significantly reduced ( Figures 4B, C ), suggesting that extracellular Ca²⁺ mobilization is vital for HgCl2 mediated IL-1β and IL-18 secretion in LCWE-primed BMDM.

Intracellular Ca²⁺ channels and receptors were blocked using Ca²⁺ channel inhibitor Tg and IP3R-specific inhibitor XeC, respectively. Tg and XeC treatment inhibited the secretion of IL-1β and IL-18 cytokines ( Figure 4C ). The secretion of both IL-1β and IL-18 in cells treated with both extracellular and intracellular Ca²⁺ inhibitors was drastically downregulated ( Figures 4B, C ). These results emphasize the critical role of both extracellular and intracellular Ca²⁺ mobilization in HgCl2 mediated activation of inflammasomes, possibly NLRP3 inflammasome.

Inflammasome mediated secretion of IL-1β and IL-18 cytokines are caspase-1 mediated (20). Therefore, to determine if HgCl2 induced secretion of IL-1β is caspase-1 dependent, Z-YVAD- FMK, a known caspase-1 specific inhibitor, was used. LCWE-primed and HgCl2-activated BMDMs treated with the inhibitor produced significantly less IL-1β cytokine than untreated cells ( Figure 5 ). As HgCl2 mediated activation is Ca²⁺ dependent, we treated cells with XeC, a specific inhibitor for IP3R and Z-YVAD-FMK. There was further inhibition of secreted IL-1β cytokine ( Figure 5 ) compared to Z-YVAD-FMK or XeC treatment alone. These results show that HgCl2 induced secretion of IL-1β is caspase-1 mediated and is IP3R- dependent. Furthermore, when NLRP3 inflammasome activation is inhibited with a specific NLPR3 inhibitor MCC950 in LCWE with or without HgCl2-treated BMDMs, there is a complete abrogation of IL-1β and IL-18 cytokines secretion ( Figures 6A, B ).

To determine the effect of HgCl2 on the LCWE induced coronary arteritis disease model of KD, we administered HgCl2 to LCWE-injected mice. A single dose of HgCl2 was administered subcutaneously concurrent with LCWE. We assayed circulating IL-1β and IL-18 levels serially and determined the incidence and severity of coronary arteritis as per histologic grading protocol. In mice given LCWE and HgCl2, there was increased production of both IL-1β and IL-18 in the peripheral blood starting at 48 hrs post injection ( Figures 7A, B ). Treatment with mercury also resulted in more severe coronary arteritis. In comparison to mice injected with LCWE alone, those injected with LCWE and HgCl2 exhibited more severe and concentrated cellular infiltrates in the cardiac tissues. On the other hand, in animals injected with just HgCl2, only minimal cellular infiltrates were observed ( Figure 8 ). Furthermore, both disease incidence (n=10/10) and severity (Mean ± SD = 1.95 ± 0.15, p<0.0001) were significantly increased in mice injected with LCWE and HgCl2 when compared to LCWE (n=7/10) (Mean ± SD = 0.65 ± 0.47), or HgCl2 (n=0/10) injected mice alone (n=10 in each group) ( Table 1 ).

Four weeks old C57BL/6 mice were purchased from The Jackson Laboratory (Bar Harbor, ME). The mice were housed under specific pathogen-free conditions at the University Health Network (Toronto, Canada). All animal studies were performed according to guidelines and procedures approved by Animal Care Committee at both the Hospital for Sick Children Research Institute and the University Health Network (Toronto, Canada).

Splenocytes (1x10⁶ cells) were treated with various concentrations of HgCl2 (10μM-1M) for 45 minutes. Cells were loaded with propidium iodide (PI) (ThermoFisher) to determine live/dead cells. The cells were acquired using BD^TM LSR II Analyzer (BD Biosciences, Mississauga, ON, Canada). Data were analyzed using FlowJo v9 (Tree star, OR, USA) software.

BMDMs were obtained by differentiating bone marrow progenitors from the tibia and femur of 8-12 weeks old wild type C57BL/6 mice in RPMI-1640 (VWR, Mississauga, ON, Canada) supplemented with 10% FBS, 1mM Sodium pyruvate, 0.1mM non-essential amino acid, 50µM 2-ME, 2mM L-glutamine and 10mM HEPES, 100 U/ml penicillin and 100µg/ml streptomycin (ThermoFisher, Mississauga, ON, Canada) (complete medium), and 20ng/ml of M-CSF (Sigma Aldrich, MO, USA) for 7 days. The culture medium was replaced with fresh complete medium and M-CSF every 3 days. BMDMs were re-plated in 24 well plates (VWR International) in complete medium at 2.5x10⁵ cells/ml 1 day before experimental assays.

BMDMs from C57BL/6 were cultured at 2.5x10⁵ cells/ml in 24 well plates in complete RPMI alone, LCWE (10mg/ml), or Pam3Cys (2 or 4μg/ml) for 18 hours. HgCl2 (10μM-20μM) (Sigma Aldrich) was added during the last 30-45 min of incubation for activation of inflammasomes. Supernatants and cell lysates were collected and assayed for IL1-β and IL-18 by ELISA (eBiosciences, San Diego, CA) and for NLRP3 protein by immunoblotting, respectively. To chelate HgCl2, cells were treated with 50μM meso-2,3-Dimercaptosuccinic acid (DMSA) (Sigma Aldrich).

Intracellular and extracellular Ca²⁺ channels were inhibited with 5μM Xestospongin C (XeC), 100nM Thapsigargin (Tg), 10μM U73122, 100μM 2-Aminoethoxydiphenylborane (2-APB). All Ca²⁺ inhibitors were purchased from Tocris Biosciences, Bristol, UK. For inhibition of caspase-1, irreversible inhibitors Z-YVAD-FMK (40μM) (Abcam, Toronto, ON) and MCC950 (10nM) (Cayman Chemical, Ann Arbor, MI) were used to inhibit NLRP3 inflammasome. Where appropriate, inhibitor agents were added 15-20 minutes prior to the addition of HgCl2.

Four-week-old C57BL/6 mice were injected with either PBS alone or LCWE (1 mg/mouse) intraperitoneally as per protocol (27). A single injection of 0.1ml HgCl2 (1.25mg/Kg body weight) was administered subcutaneously (28) to animals treated with or without LCWE. Both LCWE and HgCl2 were administered concurrently. Animals were sacrificed on days 2, 4, 21, and 28. Blood was collected, and serum was extracted, aliquoted, and stored at -80oC until use. Heart sections were obtained on day 28, stained with Hematoxylin and Eosin (H&E), and assessed histologically by a blinded reviewer for coronary arteritis. A semi-quantitative score from 0 - 4 was assigned as follows: 0 = no infiltrate; 1 = minimal cellular infiltrate; 2 = moderate cellular infiltrate; 3 = extensive transmural cellular infiltrate; 4 = extensive cellular infiltrate extending well beyond the coronary artery wall. Inflammation was determined based on the presence of monocytes, lymphocytes, and neutrophils, as morphologically identified. LCWE was prepared, and concentration was determined as previously described (27, 29).

Proteins from lysates of BMDMs were extracted in Radio-Immunoprecipitation (RIPA) Buffer (Sigma Aldrich) containing Halt^TM protease and phosphatase inhibitor cocktail (Fisher Scientific, Napean, ON, Canada). Immunoblots were prepared with Bolt^® Bis-Tris Plus Gel (ThermoFisher), and western blot analysis was carried out according to standard protocols, with antibodies specific for rabbit polyclonal raised against human NLRP3 (1:6500, sc-66846, Santa Cruz Biotechnology Inc., Santa Cruz, CA). GAPDH was used as a loading control (AM4300, ThermoFisher). Relative protein levels were normalized to GAPDH as determined by densitometry using Image J software (1.8v, NIH).

All ELISAs were carried out as per the manufacturer’s protocol. The following ELISAs were used in this study: Mouse IL-1β and IL-18 in cell culture supernatants or serum were measured using mouse-IL-1β Ready-SET-Go!^® and mouse-IL-18 Platinum ELISA kits (ThermoFiasher).

BMDMs were plated on 4-chambered glass dishes (Fisher Scientific) at 0.1x10⁶ cells per chamber. Cells were loaded with Fluo-4/AM (ThermoFisher in Ca²⁺ free DMEM media (ThermoFisher) supplemented with 10% FBS, 1mM Sodium pyruvate, 0.1mM non-essential amino acid, 50μM 2-ME, 2mM L-glutamine and 10mM HEPES (complete-DMEM) and maintained in 37oC. Live images of untreated cells were taken at t=0. Cells were then treated with 20μM HgCl2 in Ca²⁺ free DMEM at t=0. Cells were imaged for 40 min with acquisition at 15-sec intervals. Ionomycin (1μM) (Sigma Aldrich) was added to the medium at the end of 30 min, and cells were imaged for the remaining time. Images were acquired using Olympus 1X81^® motorized inverted fluorescence microscope with a Hamamatsu C9100-13 black-thinned EM-CCD camera and Yokogawa CSU X1- spinning disc confocal imaging system using the 488-nm laser and emission in the range of 500-600 (Carl Zeiss, Toronto, ON, Canada). Images were analyzed using Volocity software (Perkin Elmer, Woodbridge, ON, Canada). Absolute intensity for cells in 5 independent fields at different time points was obtained, and the average is displayed as the mean fluorescence intensity (MFI) of all cells in the fields.

Paired t- test, unpaired t- test, or two-way ANOVA were used as appropriate for analysis of both in vivo and in vitro experiments using GraphPad Prism v5.0 and v9.0.