Metformin Targets Mitochondrial Electron Transport to Reduce Air-Pollution-Induced Thrombosis

Abstract

Urban particulate matter air pollution induces the release of pro-inflammatory cytokines including interleukin-6 (IL-6) from alveolar macrophages, resulting in an increase in thrombosis. Here, we report that metformin provides protection in this murine model. Treatment of mice with metformin or exposure of murine or human alveolar macrophages to metformin prevented the particulate matter-induced generation of complex III mitochondrial reactive oxygen species, which were necessary for the opening of calcium release-activated channels (CRAC) and release of IL-6. Targeted genetic deletion of electron transport or CRAC channels in alveolar macrophages in mice prevented particulate matter-induced acceleration of arterial thrombosis. These findings suggest metformin as a potential therapy to prevent some of the premature deaths attributable to air pollution exposure worldwide.

Keywords: air pollution; calcium channels; electron transport chain; mitochondria; proteostasis; reactive oxygen species; signaling; thrombosis.

Figure 1.. Metformin prevents the release of IL-6 and enhanced tendency to thrombosis induced by exposure to particulate matter air pollution.

(A,B) Mice were administered metformin in the drinking water (150 mg/kg/day) then exposed to concentrated ambient particulate matter air pollution < 2.5 μm in diameter (CAPS) via inhalation in a versatile aerosol concentrator for eight hours daily on three consecutive weekdays. At the end of the third day, a standardized ferric chloride injury was induced in the carotid artery and the time to thrombosis was assessed using an ultrasonic probe placed on the artery distal to the injury (n=6, 5, 10 and 5 mice per condition, correspondingly, p<0.05 for comparison with filtered air controls). (C) Mice treated as in (A) were harvested for measurement of IL6 mRNA in alveolar macrophages (n=4 animals per condition, * p <0.05). (D) Mice were treated with a standardized PM from the US National Institute of Standards and Technology (NIST) 10 μg/animal, intratracheally, and 24 hours later the levels of IL-6 in the BAL fluid (ELISA) were measured (n=3 mice per condition, * p < 0.05 for indicated comparison). (E) Mice were treated with metformin in the drinking water for 24 hours and the levels of oxidized and reduced nicotinamide adenine dinucleotide (NAD⁺/NADH) were measured in BAL fluid macrophages by mass spectroscopy (n=5 animals per condition, * p<0.05). (F,G) Alveolar macrophages from BAL fluid were allowed to adhere overnight to glass coverslips, loaded with the mitochondrially localized oxidant sensitive dye MitoSOX (5 μM) and then exposed to PM containing perfusate (10 μg/ml) in the presence or absence of metformin (1 mM) on the stage of an epifluorescent microscope and oxidation of the dye was recorded from the same cellular region over time (n=3 mice per condition, * p<0.05). See also Figure S1.

Figure 2.. Metformin inhibits mitochondrial electron transport complex I to limit PM-induced ROS generation from complex III.

(A-C) A murine alveolar macrophage cell line (MHS) was stably transfected with a lentivirus encoding GFP and NDI1, a yeast protein capable of transferring electrons from NADH to complex II/III but incapable of ROS generation, or GFP alone. These cells were exposed to PM (10 μg/m³) in the presence or absence of metformin and oxygen consumption (Seahorse XF Analyzer) and the oxidation of MitoSOX were measured 4 hours later, and the release of IL-6 into the media was measured 24 hours later (minimum of 8 replicates per measurement, * p <0.05). (D,E) MHS cells were treated with a suppressor of superoxide production from complex III (S3QEL) or complex I (S1QEL) (both at 5 μM) and mitochondrial ROS generation was measured immediately after PM exposure as in (F) and IL-6 release was measured 4 hours later (n=3, * p <0.05). (F,G) MHS cells were treated with a selective activator of AMPK A-79662 (2 μM) or vehicle and phosphorylation of AMPK and PM-induced IL-6 release were measured. (H,I) MHS cells were treated with a dose of Antimycin A (10 μM) or mitochondrially targeted paraquat (5 μM) selected to restore mitochondrial ROS to levels similar to PM alone, or in the presence of low dose PM, for measurement of MitoSOX oxidation (as in (F)) and IL-6 release after 4 hours (n=3–6 per condition, * p <0.01). See also Figure S2.

Figure 3.. Store operated calcium channels contribute to PM-induced IL-6 release.

(A,B). Primary alveolar macrophages isolated from wild-type mice were allowed to adhere to glass coverslips before loading with the calcium sensitive dye Fura-2 (2 μM) and treated with PM in calcium free followed by calcium replete (2 mM) media after 1 hour pretreatment with metformin (1 mM), Synta-66 (10 μM), 2-aminoethoxydiphenyl borate (2-ABP, 10 μM) or 1,2-bis(o-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA, 10 μM) and the change in intracellular calcium concentration in response to extracellular calcium was recorded (n=3–4 mice per condition, * p < 0.05 for comparison with PM treated cells). (C) Primary alveolar macrophages from wild-type mice were pretreated for 1 hour with the CRAC channel inhibitor Synya-66 (10 μM) and then treated with PM (10 μg/cm²) and the level of IL-6 in the media was measured 24 hours later (n=3, * p<0.05). (D-F) MHS cells were stably transfected with lentiviruses encoding shRNA against Stim1, Orai1 or a scrambled shRNA and treated with vehicle or PM (10 μg/cm²) and (D,E) the change in intracellular calcium was measured as in (A), and (F) the level of IL-6 in the media was measured 24 hours after PM treatment. See also Figure S3.

Figure 4.. PM-induced CRAC channel activation occurs downstream of mitochondrial ROS generation.

(A) MHS cells were stably transfected with lentiviruses encoding shRNA against Stim1 and Orai1 or a scrambled shRNA and treated with vehicle or PM (10 μg/cm²) and the oxidation of MitoSOX dye was measured (B,C) MHS cells transfected with GFP or GFP-NDI1 were treated with PM with or without metformin (1 mM) and changes in intracellular calcium upon the addition of calcium replete media were measured. (D-F) MHS cells were treated with S3QEL (5 μM) 1 hour before treatment with PM (10 μg/cm³) and immediate changes in intracellular calcium and IL-6 release into the media after 24 hours were measured. (K-N) MHS cells were pretreated for 1 hour with the PLC inhibitor U73122 or the inactive control compound U73343 (both 200 nM) and PM induced changes in cytosolic calcium, IL-6 release (24 hours), and MitoSOX oxidation were measured. For all calcium measures, data from a representative experiment, where each line represents an individual cell region, and summary data from 3 independent replicates are shown. Thapsigargin (TG) (25 nM) was added to the cells at the end of each experiment (* p<0.05). See also Figure S4 and S5.

Figure 5.. Mitochondrial electron transport is necessary for the PM-induced acceleration of carotid thrombosis.

(A-F) Mice deficient in Tfam, a nuclear encoded transcription factor necessary for mitochondrial DNA transcription, in alveolar macrophages (Cre^CD11c/Tfam^flox/flox) were compared with control mice. (A) Levels of Tfam mRNA in flow sorted AMs and neutrophils. (B) Mice were treated with PBS or PM (10 μg, intratracheally) and 24 hours later, the time to cessation of carotid artery blood flow after a standardized ferric chloride injury was measured. (C) Wild-type C57Bl/6 mice were treated with the mitochondrially targeted antioxidant MitoTempo (0.7 mg/kg/day, intraperitoneally) 24 hours before and simultaneous with the administration of PM (10 μg/mouse) or PBS and the time to cessation of carotid artery blood flow after standardized ferric chloride injury was measured. (D) Primary alveolar macrophages from the indicated strains of mice were treated with PM and IL-6 levels in the media were measured 24 hours later (n=3, * P<0.05). (E,F) Primary alveolar macrophages from the indicated strains were loaded with MitoSOX and mitochondrial ROS generation was measured continuously after the administration of PM on the stage of an epifluorescent microscope (n= 4, * P <0.05). (G,H) Primary alveolar macrophages were loaded with Fura-2 for measurement of intracellular calcium levels after PM exposure in calcium free followed by calcium replete media. Alveolar macrophages isolated from 3 mice per group (* p<0.05). See also Figure S5.

Figure 6.. CRAC channel activation is necessary for PM-induced acceleration of carotid thrombosis.

(A,B) Mice deficient in Orai1, a necessary component of CRAC channels, in alveolar macrophages (Cre^CD11c/Orai1^flox/flox) were compared with control mice. (A) Levels of Orai1 mRNA in flow sorted AMs and neutrophils. (B) Mice were treated with PBS or PM (10 μg, it) and 24 hours later, the time to cessation of carotid artery blood flow after a standardized ferric chloride injury was measured in Orai1^flox/flox mice and Cre^CD11c/Orai1^flox/flox mice. (C) Primary alveolar macrophages were treated with PM and the levels of IL-6 were measured 24 hours later (n=3, * P<0.05). (D,E) Primary alveolar macrophages were loaded with Fura-2 for measurement of intracellular calcium levels after PM exposure in calcium free followed by calcium replete media (n=3, * P<0.05). (F) Wild-type mice were treated with Synta-66 (10 μM) intratracheally or vehicle simultaneous with the intratracheal instillation of PM (10 μg/mouse) and the levels of IL-6 in the BAL were measured 6 hours later (n=3, * P<0.05).

Figure 7.. Metformin prevents PM-induced IL-6 release from primary human alveolar macrophages and modifies transcriptional response to PM in murine alveolar macrophages.

(A) Flow sorted alveolar macrophages were isolated from biopsies from donor lungs obtained at the time of lung transplantation and allowed to adhere to glass coverslips for 4-8 hours. The cells were loaded with MitoSOX (5 μM) and then pretreated with saline, metformin (1 mM) or Synta-66 (10 μM) prior to treatment with PM (10 μg/cm²) for measurement of mitochondrial ROS generation. (B,C) Alveolar macrophages were allowed to adhere to glass coverslips prior to loading with Fura-2 (2 μM) and then treated with metformin (1 mM) or saline 1 hour before treatment with PM and intracellular calcium levels were measured in calcium free followed by calcium replete (1 mM) media. Thapsigargin (TG, 25 nM) was added at the end of the experiment. (B) Representative time-series plots and (C) composite data from four individuals is shown. (D) Primary human alveolar macrophages were treated with metformin (1 mM) 1 hour before treatment with PM (10 μg/cm²) and IL-6 levels in the media were measured 4 hours later. (E-G) Mice were treated with metformin in the drinking water for 24 hours before and after instilling PM (10 μg /mouse) intratracheally and alveolar macrophages flow-sorted from whole lung homogenates 24 hours later were subjected to transcriptional profiling via RNA-seq. (E) Schematic of the experimental design. Numbers indicate differentially expressed genes (FDR p < 0.05). (F) Volcano plots demonstrating up- and down-regulated genes in mice treated with metformin before and after exposure to PM, numbers indicate up- and down-regulated genes (FDR p < 0.05). (G) k-means clustering identifies metformin-specific clusters of genes. Differentially expressed genes (1285 genes, identified using ANOVA-like test implemented in edgeR package, FDR p < 0.001) in alveolar macrophages exposed to PM in the presence and absence of metformin pretreatment in vivo were subjected to k-means clustering (number of genes per cluster is shown on the left). Representative gene names and GO processes for each cluster are shown on the right side. (H,I) Chaperones are enriched in cluster 3. Gene Set Enrichment Analysis shows enrichment for chaperone genes after treatment with metformin that is more significant in alveolar macrophages from PM exposed mice. (J) MHS cells stably transfected with a lentivirus encoding NDI1 or a control lentivirus (GFP) were treated with PM and the levels of mRNA encoding the indicated genes were measured 24 hours later using RT-qPCR (n=3, * P<0.05). See also Figure S6.