Prevention of crystalline silica-induced inflammation by the anti-malarial hydroxychloroquine.

C57Bl/6 mice (Jackson Laboratories, Bar Harbor, ME) were housed in the University of Montana’s specific-pathogen-free Laboratory Animal Resources facility. The mouse room was maintained on a 12 hr light/dark cycle, and mice were provided with mouse feed and deionized water ad libitum . Male and female mice between 8-16 weeks of age were used for the studies. Euthanasia was administered through intraperitoneal injection of a lethal dose of pentobarbital sodium (Euthasol, Virbac, Fort Worth, TX). The University of Montana Institutional Animal Care and Use Committee (Missoula, MT) approved all procedures performed on the animals. Crystalline silica (Min-U-Sil-5, average particle size 1.5-2 μm in diameter) was obtained from Pennsylvania Sand Glass Corporation (Pittsburgh, PA), and acid washed in 1N HCl. The cSiO 2 was washed with sterile water four times and dried in an oven at 200°C. Before use, cSiO 2 particles were suspended in PBS or dispersion media (for in vivo experiments) and sonicated for at least 2 min (550 watts at 20 kHz) by a cup-horn sonicator in a circulating water bath (Misonix, Inc. Farmingdale, NY). Dispersion media consisted of PBS containing 0.6 mg/ml mouse serum albumin and 0.01 mg/ml 1,2-dipalmitoyl-sn-glycero-3-phosphocholine. Bone Marrow Derived Macrophages (BMDM) were generated as described previously ( Migliaccio et al., 2008 ; Pfau et al., 2004 ). C57Bl/6 mice were sacrificed and the hind legs were removed. Complete media (RPMI, 10% FBS, Penicillin/Streptomycin, β-Me) was washed through the femur and tibia to collect the bone marrow cells. Cells were incubated in T75 flasks at a density of 3.0×10 7 overnight at 37°C for stromal elimination by adherence. The next day, nonadherent cells were transferred to a new flask at a density of 1.5×10 7 cells per flask. Macrophage colony stimulating factor (M-CSF) (20 ng/mL R&D Systems) was added and cells were spiked with M-CSF (10 ng/ml R&D Systems) every 3-4 days. BMDM were used on day 10. BMDM were platted in a flat-bottom, tissue culture- treated 96-well plates at 1×105 cells/well in 100 μL of RPMI complete media and incubated with lipopolysaccharide (LPS) (20 ng/mL) for inflammasome priming. Cells were treated with 25 μM of hydroxychloroquine sulfate (Sigma-Aldrich cat. H0915-5MG), 25 μM imipramine hydrochloride (Sigma-Aldrich cat. 10899-5G), or 100 nM bafilomycin (EnzoLife Sciences cat. BML-CM110-0100) for 30 minutes prior to addition of cSiO 2 . Cells were exposed to various doses (0-50 μg/mL) of cSiO 2 and plates were incubated in a 37°C water-jacketed CO 2 incubator (ThermoForma, Houston, TX) for 24 hours. Toxicity induced by cSiO 2 was determined by two complementary assays, a lactate dehydrogenase assay (Promega, cat. G1780) and a common colorimetric tetrazolium viability (MTS) assay (Promega, cat. G3580), and read on a plate reader (Molecular Devices SpectraMax M4 colorimetric microplate reader). In order to avoid artifacts in the optical density values, the MTS reaction was transferred to a clean plate to separate it from the cell/particle mixture adhered to the plate bottom. Data were normalized to a percent relative to the no particle, no HCQ control cells. NLRP3 inflammasome activation was assayed by measuring the release of IL-1β by using a commercially available ELISA kit (R&D Systems, cat. DY201). C57Bl/6 mice were treated with either hydroxychloroquine (Hydroxychloroquine Sulfate, Calbiochem cat. 509272) reconstituted in PBS (10 mg HCQ/kg/per day) or vehicle (PBS) by oral gavage once a day for seven days in order to more closely approximate normal oral use of this pharmaceutical. After seven days, mice were instilled with 1 mg or 0.25 mg of cSiO 2 through oropharyngeal aspiration. Prior to gavage and instillation of cSiO 2 , mice were briefly anesthetized by isoflurane. For instillation, the sedated mouse is positioned vertically while 50 μL of dispersion media with or without cSiO 2 was dispensed with a pipette to the back of the throat. By holding the tongue to the side, the mouse could not swallow and, therefore, aspirated the volume into the lungs. Mice were sacrificed 24 hours after cSiO 2 exposure and a whole lung lavage was conducted. The lungs and heart were removed from the chest cavity, and 1 mL of cold PBS was washed in and out of the lungs four times in order to collect concentrated lung fluid. Samples were centrifuged at 400 × g for 5 minutes and the supernatant was saved for analysis. The lungs were lavaged with an additional 4 mL of PBS in order to collect the maximum number of cells. Lavage cells were counted using a Coulter Z2 particle counter (Beckman Coulter, Brea, CA) and resuspended in RPMI 1640 culture media supplemented with 10% fetal bovine serum, sodium pyruvate, and an antibiotic-antimycotic solution (Mediatech, Manassas, VA) for differential analysis or flow cytometry. Cells were stained for differential analysis with a Wright-Geimsa stain in a Hematek 2000 autostainer (Miles-Bayer-Siemens Diagnostics, Deerfield, IL). Toxicity was measured in the lung lavage fluid by assaying for LDH as above. Cytokines from lung lavage fluid were measured by a multiplex immunoassay (MSD U-PLEX, Meso Scale Diagnostics, U-PLEX, cat K15069L-2) because of the sensitivity and small sample size used. Seven cytokines IFN-γ, IL-1β, IL-6, IL-10, IL-13, IL-33, and TNF-α were included in the 7-plex plate used for this experiment. Protein in the lung lavage fluid was assayed by the Pierce BCA Protein Assay Kit (ThermoFisher Scientific cat. 23225) following manufacturer’s specifications. Extracellular cathepsin B release in the lavage fluid was analyzed as previously described by our laboratory ( Sager et al., 2016 ). In a 96 well plate, 50 μL of whole lung lavage was combined with 50 μL of a cathepsin B inhibitor (Calbiochem cat. 219385) or 50 μL PBS and incubated at room temperature for 15 min. Then the cathepsin substrate Z-LR-AMC (specific to cathepsin B, cathepsin L and cathepsin V; R&D systems cat. ES008) at 20 μL in PBS was added to 50 μL of whole lung lavage fluid in a total reaction volume of 150 μL. The assays were incubated at 37°C for 1 hour then fluorescence was measured using a plate reader at 380 nm excitation and 460 nm emission. The wells with the only the cathepsin substrate measured the total (cathepsin B, cathepsin L and cathepsin V) cathepsin activity in the lavage fluid while wells with the cathepsin B inhibitor measured the remaining cathepsin L and cathepsin V activity. By subtracting the inhibitor wells from the total cathepsin well, the level of cathepsin B was calculated. Cells collected during the whole lung lavage of mice treated with 0.25 mg of cSiO 2 in vivo were plated for ex vivo IL-1β culture. Lung lavage fluid was centrifuged and the supernatants saved for other analyses. Cells were resuspended in 250 μL of complete media, counted, and then 100 μL of cell suspension was plated in a 96 well plate. Cells were primed for inflammasome activity with LPS (20 ng/mL) and incubated for 24 hours at 37°C. After incubation, supernatants were collected and assayed with a commercial IL-1β kit (R&D Systems, cat. DY201). Data were normalized to IL-1β release per 1×10 5 AM using the total cell count and percent of macrophages in the whole lung lavage. Flow cytometry was used to assess phagocytosis of cSiO 2 particles and lysosomal uptake of LysoTracker dye both in BMDM and AM collected in the lung lavage of treated mice. Internalization of cSiO 2 was determined by using a side-scatter technique previously described ( Hamilton et al., 2006 ). BMDM or lung lavage cells were cultured or collected as described above. BMDM were treated with 25 μM HCQ for 30 minutes and then exposed to 50 μg/mL cSiO 2 for 4 hours using 1.5 mL microfuge tubes and end over end tumbling (Lab Quaker Shaker, Thermo Forma). Lung lavage cells had no additional treatment of HCQ or cSiO 2 after the in vivo exposures. Both lung lavage cells and BMDM were treated with LysoTracker Red DND-99 (ex/em 577⁄590 nm, ThermoFisher Scientific cat. L7528) at 50 nM for 30 min. The cells were centrifuged, resuspended in PAB, and transferred to filter-top flow cytometry tubes (BD Biosciences, San Jose, CA) for analysis. Phagocytosis and LysoTracker uptake data were expressed as mean fluorescent intensity. Cells were analyzed on a Life Technologies Attune NxT Acoustic Focusing Cytometer with the YL-1 585/16 nm laser. Lysosome membrane permeabilization (LMP) was assessed using methods modified from Aits et al. (2015) and as described previously by our laboratory ( Jessop et al., 2017 ). BMDM were plated in 24 well plates at a density of 2×10 5 cells per well. Cells were treated with or without HCQ (25 μM) and with or without cSiO 2 (50 μg/mL). Cells were washed twice with PBS and placed on ice. BMDM were then incubated with 200 μL of cytosol extraction buffer, which consisted of 250 mM sucrose, 20 mM Hepes, 10 mM KCl, 1.5 mM MgCl2, 1 mM EDTA, 1 mM EGTA, 0.5 mM pefabloc (Sigma-Aldrich cat. 76307-100mg), pH 7.5, and digitonin (15 μg/mL ) (Sigma-Aldrich cat. D141-100MG), for 15 min on ice with rocking. The concentration of digitonin for optimal extraction of the cytosolic fraction was determined by titration. β-N-acetylglucosaminidase (NAG) activity was measured by adding 30 μL cytosolic extract to 100 μL of NAG reaction buffer (0.2 M sodium citrate, pH 4.5 with 300 μg/mL 4-methylumbelliferyl-2-acetamido-2-deoxy-β-D-glucopyranoside (Sigma-Aldrich cat. 37067-30-4) and assessed on a plate reader (20 min; 45s intervals; 356 nm excitation; 444 nm emission). Extracted cytosolic LDH activity was measured as described above and used as a control to which the NAG activities were normalized. Depending on the data type, a parametric one or two-way analysis of variance (ANOVA), followed by post hoc mean comparison (Holms-Sidak etc.) was used throughout the research described. Statistical significance will be defined as a two-tailed probability of type I error at less than 5% ( p < 0.05) unless otherwise stated. The minimum number of experimental replications was 3.

The ability of HCQ to decrease toxicity and cell death in vitro was assessed using C57Bl/6 BMDM. BMDM were primed for inflammasome activation with LPS (20 ng/mL) and treated with either HCQ (25 μM), imipramine (IMP, 25 μM), or bafilomycin A1 (BAF, 100 nM) for 30 min prior to cSiO 2 exposure. IMP and BAF were positive controls for blocking cSiO 2 -induced toxicity. Bafilomycin A1 (BAF) specifically inhibits vATPases to prevent lysosomal acidification, which previous studies have indicated is necessary for particle induced toxicity ( Jessop et al., 2017 ). In biological systems small particles develop a protein shell called a protein corona. The protein corona is degraded in the acidic environment of the lysosomes which allows the particle to directly interact with the lipid membrane causing LMP. Preventing acidification is thought to prevent degradation of the protein corona and protect the lysosome from LMP ( Wang et al., 2013 ). Imipramine, a tricyclic anti-depressant, was previously shown by our laboratory to block cSiO 2 -induced toxicity; therefore, it was used as a comparison to HCQ ( Biswas et al., 2017 ). Cells were exposed to 0, 50, or 100 μg/mL of cSiO 2 and incubated for 24 hrs at 37°C. HCQ, IMP, and BAF were all found to block toxicity at both the 50 μg/mL and 100 μg/mL exposures of cSiO 2 as determined by the LDH assay and MTS assay ( Fig. 1A and B , respectively). No toxic effects from either HCQ or IMP alone were found in cells not exposed to cSiO 2 . Protective effects were more pronounced at the 50 μg/mL cSiO 2 exposure as there was significant cell death at the 100 μg/mL cSiO 2 dose which could not be completely overcome by the HCQ, IMP, or BAF. Cell supernatants were assayed for IL-1β as a representative measure for NLRP3 inflammasome activation. IL-1β production was markedly decreased in exposed cells receiving HCQ, IMP, and BAF ( Fig. 1C ). In order to confirm that HCQ was not blocking cSiO 2 toxicity by reducing phagocytosis of the cSiO 2 particles, the amount of cSiO 2 taken up by the BMDM was quantified by flow cytometry ( Fig. 2A ). BMDM which have taken up cSiO 2 have higher side-scatter (SSC) due to the deflection of laser light by the cSiO 2 particles. BMDM exposed to cSiO 2 (50 μg/mL) with or without HCQ (25 μM) had significantly higher side scatter (**** p < 0.0001) than BMDM not exposed to cSiO 2 . There was no difference between the SSC of control vs HCQ-treated cells nor cSiO 2 vs HCQ + cSiO 2 cells. This indicates that HCQ does not affect the ability of BMDM to take up cSiO 2 . Additionally, BMDM were also treated with LysoTracker Red DND-99 at 50 nM for 30 min to assess the relative acidity of the lysosomes ( Fig. 2B ). HCQ caused a decrease in the amount of LysoTracker taken up by lysosomes. Cells treated with HCQ or HCQ + cSiO 2 had significantly less (* p < 0.05) LysoTracker uptake than control cells or cells exposed cSiO 2 without HCQ treatment. This suggests that when HCQ accumulates in the lysosomes it causes the lysosomes to be less acidic which may help protect the lysosomes from being permeabilized by cSiO 2 . Together, the in vitro data collected in these experiments are consistent with HCQ being able to reduce LMP after cSiO 2 exposure. Direct quantification of the ability of HCQ to block LMP in vitro was assessed through a method modified from Aits et al . which measures the release of β- N -acetyl-glucosaminidase (NAG) into the cytosol after cell permeabilization by the detergent digitonin ( Aits et al., 2015 ). NAG is a lysosomal hydrolase which is not present in the cytosol unless the lysosomal membrane has been compromised. Using BMDM, the level of digitonin is titrated so that the cell membrane is permeabilized while the lysosomal membrane remains intact. The lysosomal membrane contains less cholesterol than the cell membrane making the digitonin required to permeabilize the lysosomal membrane higher than that required for the cell membrane ( Aits et al., 2015 ). BMDM were treated with or without HCQ (25 μM) and with or without cSiO 2 (50 μg/mL) and incubated for 4 or 24 hours. Cytosolic LDH activity was measured as described above and used as a control to which the NAG activities were normalized. HCQ was able to significantly block the release of NAG from lysosomes after both 4 (* p < 0.05) and 24 (*** p < 0.001) hours ( Fig 3A and B , respectively). This shows that HCQ blocks cSiO 2 -induced LMP in vitro. The above in vitro studies demonstrated that HCQ accumulated in macrophages and blocked the ability of cSiO 2 to cause LMP and downstream events including toxicity and NLRP3 inflammasome activation as determined by IL-1β release. However, in vivo studies have not been previously conducted in order to determine whether the in vitro studies predict in vivo outcomes. Therefore, C57Bl/6 mice were treated with HCQ at 10 mg/kg/day or PBS by oral gavage for 7 days. This dose of HCQ was chosen based on previous in vivo studies with HCQ ( Chen et al., 2017 ; Gómez-Guzmán et al., 2014 ). At the end of the 7-day pre-treatment period, mice were exposed to either 0.25 mg or 1 mg of cSiO 2 by oropharyngeal instillation (in 50 μL of dispersion media). Control mice received only dispersion media. Mice were sacrificed 24 hours post cSiO 2 exposure. The 1 mg of cSiO 2 exposure, has been commonly used in previous studies allowing comparison to previous results ( Biswas et al., 2017 ). Naïve mice generally have around 3-4×10 5 AM residing in their lungs; therefore, the 0.25 mg dosage more closely matches the in vitro dose used in this study of 50 μg/mL cSiO 2 per 1×10 5 BMDM. In order to confirm that HCQ was incorporated into the AM during the 7-day HCQ in vivo treatment, AM from whole lung lavage were treated with LysoTracker Red DND-99 and the cells analyzed by flow cytometry. As in the in vitro experiments, AM from mice treated with HCQ showed significantly less uptake of the LysoTracker dye. This suggests that receiving HCQ at 10 mg/kg by oral gavage for 7 days was sufficient for HCQ to become sequestered in the lysosomes of the AM ( Fig. 4A ). Additionally, in order to rule out that HCQ was reducing toxicity due to changes in cSiO 2 uptake by the AM in vivo , side scatter data of the whole lung lavage cells were collected by flow cytometry. Side scatter was increased as expected in the cSiO 2 -treated mice, and there were no differences between the control and HCQ-treated mice nor the cSiO 2 and HCQ + cSiO 2 -treated mice ( Fig 4B ). This matches the in vitro data showing that changes in cSiO 2 -induced toxicity in HCQ-treated mice was not due to a change in the ability of the AM to phagocytize the cSiO 2 ( Fig. 2A ). Instillation of cSiO 2 in mice is known to cause lung inflammation and infiltration of immune cells. Therefore, analysis of cells from the whole lung lavage was conducted ( Fig. 5 ). At the 1 mg cSiO 2 dose, there were no differences in the percentage of cells or the raw cell numbers between the HCQ-treated mice and control mice ± cSiO 2 exposure ( Fig 5A and B ). Mice exposed to cSiO 2 displayed the expected neutrophilic infiltration. At the 0.25 mg cSiO 2 dose, cSiO 2 -exposed mice which were treated with HCQ had a significant increase in the percentage of macrophages and a significant decrease in the percentage of neutrophils recruited to the lungs as compared to cSiO 2 -exposed mice with no HCQ treatment ( Fig 5C and D ). Additionally, cSiO 2 -exposed mice with HCQ treatment had a greater number of macrophages in their lungs 24 hours after cSiO 2 exposure than did cSiO 2 exposed mice without HCQ treatment ( Fig 5D ). This suggests that macrophages may be protected from cSiO 2 -induced cell death with HCQ treatment. In order to evaluate lung injury following instillation of cSiO 2 , lavage supernatants were assayed for LDH release (HCQ expressed as a % relative to without HCQ), total protein, relative total cathepsin, and relative cathepsin B levels ( Fig. 6A - D , respectively). Although not significant, LDH, total cathepsin, and cathepsin B release were all lower in HCQ-treated mice at both the 0.25 mg and 1 mg cSiO 2 exposures. Total protein in the lung lavage supernatants, as measured by a BCA assay, was used to assess lung injury. The BCA assay showed a slight decrease in the amount of protein in the lung lavage fluid at the 1 mg cSiO 2 dose in HCQ-treated animals versus control. As expected the lung protein level increased as the cSiO 2 dose increased in both control and HCQ-treated mice. Cytokines from lung lavage supernatants were measured by a multiplex immunoassay: IFN-γ, IL-1β, IL-6, IL-10, IL-13, IL-33, and TNF-α ( Figs. 7 and 8 ). IL-10 and IL-13 levels were below detection levels for both 0.25 mg and 1 mg cSiO 2 exposures (data not shown). For mice exposed to 1 mg of cSiO 2 , IFN-γ ( Fig 7A ) was only detectable in the cSiO 2 -exposed group suggesting that HCQ was able to reduce IFN-γ after cSiO 2 exposure (* p < 0.05). While not significant, there was a trend that TNF-α, IL-33, and IL-6 ( Fig 7 B - D , respectively) appeared to be lower with HCQ treatment in the 1 mg cSiO 2 exposed mice. IL-1β did not appear to be blocked with HCQ treatment ( Fig 7E ). For the 0.25 mg dose, IFN-γ again appeared to be decreased by HCQ treatment ( Fig 8A ), although in this case it was not statistically significant due to a large variance in the cSiO 2 -exposed group. As in the 1 mg dose group, there was a trend that HCQ + cSiO 2 mice seem to have lower levels of TNF-α as compared to cSiO 2 only mice ( Fig 8B ). In all graphs, no visible bar signifies that levels were below detection. Since HCQ was effective in blocking silica-induced IL-1β release from BMDM and the data showed that HCQ was taken up by lysosomes in AM, the AM from the mice were evaluated ex vivo . Cells collected during the whole lung lavage of mice treated with 0.25 mg of cSiO 2 in vivo were plated for ex vivo IL-1β culture and primed with LPS. After incubation, supernatants were assayed for IL-1β by ELISA. Data were normalized to IL-1β release per 1×10 5 AM using the total cell count and percent of macrophages ( Fig 9 ). Control and HCQ-treated cells had very little IL-1β production as expected. Consistent with the in vitro data, cells from HCQ + cSiO 2 -treated mice had significantly less IL-1β production when compared to cSiO 2- exposed mice not treated with HCQ. These data are consistent with HCQ being able to prevent LMP after cSiO 2 exposure in vivo .

Silicosis is a progressive, fibrotic pulmonary disease that occurs after exposure to inhaled cSiO 2 . The loss of pulmonary function resulting from silicosis causes significant morbidity and, in severe cases, mortality. Treatment options are limited to supporting respiratory function through supplemental oxygen and decreasing inflammation with corticosteroids ( Leung et al., 2012 ). A growing body of evidence suggests that a key step in cSiO 2 -induced lung inflammation is activation of the NLRP3 inflammasome resulting in the release of pleiotropic cytokines such as IL-1β and IL-18. At this time there is limited information on possible pharmaceutical approaches to regulate NLRP3 inflammasome activity. However, it may be possible to block LMP, which has been proposed to precede NLRP3 inflammasome activation ( Sayan & Mossman, 2016 ; Serrano-Puebla & Boya, 2016 ). When cSiO 2 is phagocytized by resident AM in an attempt to clear the particles from the lungs, the cSiO 2 compromises lysosomal membranes allowing lysosomal proteases to leak into the cytosol and induce NLRP3 inflammasome assembly ( Boya & Kroemer, 2008 ; Bunderson-Schelvan et al., 2016 ; Jessop et al., 2017 ). Therefore, blocking LMP would attenuate NLRP3 inflammasome assembly and could be a potential mechanism to target in order to treat cSiO 2 -induced inflammation. HCQ was chosen for this research because it is cationic amphiphilic drug (CAD) that is known to accumulate in lysosomes ( Ben-Zvi et al., 2012 ; Boya et al., 2003 ). CADs diffuse through membranes in an unprotonated state but become protonated in the acidic environment of lysosomes; losing their ability to diffuse back out through lipid membranes. HCQ is used extensively as an anti-malarial and also to treat autoimmune diseases such as systemic lupus erythematosus (SLE). However, the precise mechanism to explain its action in SLE remain unclear ( Ben-Zvi et al., 2012 ). The in vitro work from this study strongly supports the hypothesis that HCQ is able to block cSiO 2 -induced LMP. HCQ was able to rescue BMDM from cell death after exposure to 50 or 100 μg/mL of cSiO 2 ( Fig. 1A & B ). HCQ also caused significant reduction in the production of IL-1β in BMDM after cSiO 2 exposure ( Fig. 1C ). These data are consistent with a previous study that reported chloroquine (which differs from HCQ by one hydroxyl group) suppresses NLRP3 inflammasome activation in a murine model of endotoxic shock ( Chen et al., 2017 ). They showed that chloroquine was able to decrease IL-1β and IL-18 release from BMDM stimulated with LPS. They also found that chloroquine inhibits transcription of Nlrp3 genes. Their research, however, did not examine the effect of HCQ on cSiO 2 -induced toxicity. Our research also directly measured the ability of HCQ to prevent LMP by measuring the release of β- N -acetyl-glucosaminidase (NAG) into the cytosol. HCQ was able to block NAG release in BMDM exposed to cSiO 2 at both 4 and 24 hours ( Fig. 3 ). This data conflicts with data published by Boya et al. (2003) where they argue that HCQ induces LMP. When cells were stained by immunofluorescence for cathepsin B, they reported that HCQ caused cathepsin B staining to change from being contained in lysosomes to being diffuse throughout the cell. They also concluded that HCQ causes mitochondrial membrane permeabilization and apoptosis. However, their research used HeLa cells and some concentrations of HCQ which were higher than the concentrations used in this study (up to 60 μg/ml). Furthermore, they did not directly measure LMP as in our study, but instead used a microscopic approach for LMP. Therefore, it is hard to make direct comparisons between their research and the research presented in the current study. The above in vitro results were predicated on the assumption that HCQ was being taken up into lysosomes. In order to confirm that HCQ was becoming sequestered into the lysosomes, BMDM and AM were treated with LysoTracker. The amount of LysoTracker taken up by both the BMDM in vitro and AM in vivo was significantly decreased from control which is consistent with HCQ sequestration into the lysosomes of these cells ( Fig 2B and 4A , respectively). The most likely mechanism by which HCQ is preventing the uptake of LysoTracker is due to HCQ increasing the pH of the lysosomes. Additionally, flow cytometry side scatter data from in vitro and in vivo experiments also confirmed that HCQ was not affecting the ability of cells to phagocytize the cSiO 2 particles ( Fig 2A and 4B , respectively); therefore, the toxicity results were not due to changes in cell interactions with particles. Mice treated with HCQ in vivo showed modest reductions in cSiO 2 -induced toxicity. Overall, LDH, total cathepsins, cathepsin B, IFN-γ and TNF-α trended lower in HCQ-treated mice ( Fig. 6 - 8 ). It was surprising that cSiO 2 -induced increase in IL-1β in lung lavage fluid was not blocked in vivo by HCQ treatment; however, ex vivo culture of AM from these same studies did result in significantly less IL-1β release in mice treated with HCQ. Several factors could have been responsible for the fact that there was not as pronounced of an effect in vivo as there was in vitro . It is possible that the dosing strategy and exposure time were not sufficient. Additionally, twenty-four hours after cSiO 2 exposure may be too brief of a time to see differences in the immune response. While these results could indicate that the dose of HCQ was simply too low, the ex vivo results suggest that there was enough HCQ sequestered into the lysosomes to block inflammasome formation and IL-1β production. The results from the ex vivo experiment may be explained by two differences: 1) the AM were primed for inflammasome activation with LPS when they were plated and 2) they were incubated for 24 hours in culture after isolation following in vivo cSiO 2 exposure. LPS stimulates the NF- κ B pathway increasing formation of the pro-form of IL-1β and increases the expression of NLRP3 inflammasome proteins ( Sayan & Mossman, 2016 ). Therefore, it could be suggested that HCQ blocked IL-1β release by acting on the NF- κ B pathway as previously proposed ( Chen et al., 2017 ). In their studies, Chen et al. pre-treated BMDM with chloroquine before adding LPS and saw a reduction of phospho-NF- κ B p65 protein levels in chloroquine-treated cells. However, in the in vitro experiments in this study LPS was added prior to HCQ treatment thereby allowing activation of the NF- κ B pathway by LPS to proceed prior to any potential inhibition by HCQ. Since, there was still a reduction in IL-1β production, it supports the notion that the action of HCQ to block IL-1β is not at NF- κ B but rather by blocking LMP. Previous research from our laboratory has shown that the tricyclic antidepressant, imipramine, was able to decrease cSiO 2 -induced toxicity in vitro and in vivo ( Biswas et al., 2017 ). Because HCQ and IMP are both lysosomotropic drugs that sequester in lysosomes, we hypothesized that they would have a similar ability to protect lysosomes from cSiO 2 -induced LMP. The imipramine study showed similar results as HCQ in the ability to prevent IL-1β production in vitro and in an ex vivo culture. IMP was also able to decrease IL-1β levels in the lavage fluid in vivo after a 24-hour cSiO 2 exposure, which was not seen with HCQ. In long-term exposure studies, IMP-treated mice had less lung pathology and hydroxyproline levels. It is unknown whether other CAD may have similar protective effects on lysosomal membrane stability after particle exposure and further work should be done to better understand the interactions between CAD and cSiO 2 exposure. To our knowledge, no other work has been done linking a pharmaceutical intervention with the ability to block cSiO 2 -induced LMP. The results of this study support the hypothesis that HCQ can prevent lysosomal membrane permeability from exposure to cSiO 2 . The in vitro data strongly suggests that HCQ attenuates activation of the NLRP3 inflammasome and is doing this, at least in part, by blocking LMP. Cell death after cSiO 2 exposure was decreased with HCQ treatment in vitro . HCQ was able to block the production of IL-1β and the release of NAG into the cytosol in cSiO 2 -exposed BMDM indicating that LMP was being prevented. This was the first study to show that HCQ is able to block LMP. While the in vivo data was less definitive, it also suggests that HCQ may be protective against cSiO 2 toxicity in vivo . These studies were able to confirm that HCQ is accumulating in lysosomes of AM when mice are treated with HCQ by gavage for 7 days. Data also indicated that HCQ was not affecting AM’s ability to phagocytize cSiO 2 particles. LDH, total cathepsins, and cathepsin B, IFN-γ, and TNF-α levels trended lower in HCQ-treated mice. This study was limited by the short time-frame for the in vivo studies. Mice in this study were sacrificed 24 hours after cSiO 2 exposure in order to assess the acute inflammatory response; however, this time frame is too short to observe histological changes in the lungs. Future studies would benefit from longer-term exposures to discern if HCQ is able to block the development of fibrosis and cSiO 2 -induced pulmonary pathology. Additionally, the dosage and treatment schedules were determined based on previous literature; however, in order to see more profound effects from HCQ, it could be beneficial to increase the dose or lengthen the treatment period ( Chen et al., 2017 ; Gómez-Guzmán et al., 2014 ). Additionally, this study only assessed the ability of HCQ to block cSiO 2 toxicity by pre-treating cells or mice with HCQ before exposure to cSiO 2 . While it is beneficial to have options for prophylactic treatment for individuals who know they will be exposed to cSiO 2 , such as military personnel deployed to arid regions, most treatments for cSiO 2 inhalation occur after an exposure has occurred. Therefore, further studies could examine the ability of HCQ to block cSiO 2 -induced lung pathology after exposure.

Silicosis is a fibrotic lung disease that develops after inhalation of crystalline silica (cSiO 2 ) particles. Most silicosis cases occur due to occupational exposures which remain common in the United States and across the world. Silicosis is characterized by pulmonary edema, chronic interstitial inflammation, and the development of silicotic nodules ( Kawasaki 2015 ). Current treatments for silicosis are, at best, supportive in nature, and no treatments exist to prevent progression of disease or reverse pulmonary damage. When cSiO 2 is inhaled, particles less than 2.5 μm can reach the alveolar spaces where they are phagocytized by the resident alveolar macrophages (AM). The phagocytized cSiO 2 particles interact with and disrupt the AM’s phagolysosomal membrane allowing acidic hydrolases, including cathepsin B, normally contained by the lysosome to leak into the cytosol ( Boya and Kroemer 2008 ). This lysosomal disruption is referred to as lysosomal membrane permeability (LMP) and leads to caspase-1 dependent cell death ( Bunderson-Schelvan et al., 2016 ; Serrano-Puebla & Boya, 2016 ). Following activation of caspase-1, a product of NLRP3 inflammasome assembly, AM may undergo pyroptosis and propagate a chronic inflammatory response. Damage-associated molecular patterns (DAMPs) and inflammatory cytokines such as TNF-α, IL-1β, and IL-6 are released and recruit other inflammatory cells, primarily other macrophages and polymorphonuclear neutrophils, and stimulate the fibrogenic cascade ( Beamer & Holian, 2007 ; Kawasaki, 2015 ; Pollard, 2016 ; Rimal, Greenberg, & Rom, 2005 ). Fibroblasts surround the cSiO 2 particles with collagen creating fibrotic nodules which increase in size as disease progresses. Within the cell, LMP-induced inflammation is directed by the NLRP3 inflammasome which drives the production of the inflammatory cytokines IL-1β and IL-18 ( Jo et al., 2016 ; Sayan & Mossman, 2016 ). Two signals are required for the induction of the NLRP3 inflammasome. The first signal is a priming signal which upregulates expression of the NLRP3 components. It is dependent on activation of the NF-κB pathway through TLR signaling initiated by various pathogen-associated molecular patterns (PAMPs) or damage-associated molecular patterns (DAMPs) ( Biswas et al., 2014 ; Sayan & Mossman, 2016 ). A second signal, in this case LMP, induces the assembly of the NLRP3 inflammasome into the multi-protein complex containing the apoptosis-associated speck-like protein (ASC) which binds to procaspase-1. Procaspase-1 is cleaved into an active form and in turn cleaves pro-IL-1β and pro-IL-18 into their mature, secreted forms ( Hornung et al., 2008 ). Together, the first signal and the second signal induce an inflammatory cascade which ultimately results in caspase-1 dependent cell death ( Bunderson-Schelvan et al., 2016 ) Many existing pharmaceuticals are known to become sequestered in lysosomes through an ion trapping mechanism. These pharmaceuticals tend to be weak bases that easily diffuse through membranes in an unprotonated state yet lose their ability to diffuse through membranes after becoming protonated in the acidic lysosomal environment. This results in a high concentration of drug stored within the lysosome which may provide an opportunity to prevent particle-induced lysosomal damage ( Villamil Giraldo et al., 2014 ). Previously, our laboratory has shown that imipramine (IMP) can reduce pulmonary inflammation after exposure to cSiO 2 both in vivo and in vitro ( Biswas et al., 2017 ). IMP is a tricyclic antidepressant that accumulates in lysosomes and has been shown to have some protective effects against lung injury ( Yang et al., 2010 ). Interestingly, a drug with a similar ability to become sequestered in lysosomes, hydroxychloroquine (HCQ), is commonly used in the treatment of the autoimmune disease systemic lupus erythematosus. Exposure to cSiO 2 has been shown to contribute to the development of systemic lupus erythematosus; however there has been no research examining the effect of HCQ on cSiO 2 -induced lung injury. HCQ has been extensively used to prevent malaria and as treatment for autoimmune diseases such as rheumatoid arthritis and systemic lupus erythematosus ( Olsen et al., 2013 ). Recently HCQ has been investigated for use in diseases including endometriosis, cancer, cardiovascular disorders, diabetes, and infectious diseases ( Ben-Zvi et al., 2012 ; Chen et al., 2018 ; Ruiz et al., 2016 ). There are many proposed mechanisms to explain how HCQ could be acting in such a diverse number of diseases. Some evidence has shown that HCQ can reduce inflammatory cytokine production by macrophages and alter macrophage phenotype ( Chen et al., 2018 ). Other reports show changes in B and T cell signaling and inhibition of phospholipase A2 ( Ben-Zvi et al., 2012 ). HCQ’s effect on cancer and endometriosis have been attributed to modulation of autophagy. Furthermore, in a murine model of endotoxic shock, Chen et al. showed that chloroquine can decrease activation of the NLRP3 inflammasome and block production of IL-1β and IL-18 (D. Chen et al., 2018 ). The objective of this research was to determine if HCQ has the ability to attenuate pulmonary injury by decreasing cSiO 2 -induced LMP.