Perfluorinated compounds exposure and atherogenic risk characteristics in a high-fat diet condition: In vitro/in vivo models and population panel study

Figures

• 

  Fig. 1.. Atherogenic risk in HFD-fed mice promoted by PFOS exposure. A) Representative images of ORO staining in the aorta (arrows indicate atherosclerotic plaques in the aorta) and B) the quantitative result of atherosclerotic plaque areas. Plaque areas were calculated as percentages of the total surface areas of the whole aorta. Scale bar, 3 mm. C) Representative images of VG-stained aortic and D) quantification for the aortic collagen fiber area. Scale bar, 50 μm. E) H&E staining for aortic roots and F) corresponding quantification of aortic wall thickness. The arrows indicate inflammatory infiltration of the aortic wall. Scale bar, 50 μm. G) Immunostaining for CD68 in mouse aortic roots and H) quantification of CD68-positive levels. Scale bar, 200 μm. I) Movat's pentachrome staining for cross-sections of aortic roots and (J) quantification of foam cells in plaque. The arrows indicate foam cells in the plaque. Scale bar, 50 μm. Data are presented as mean ± SD. *P < 0.05 and **P < 0.01 versus NCD-fed mice group, respectively. #P < 0.05 and ##P < 0.01 versus the HFD-fed mice group, respectively.

• 

  Fig. 2.. Phenotypic and functional alteration of macrophages promoted by PFOS exposure under the hyperlipidemic condition. Cell percentages (A and B) and number (C) of CAMs and AAMs. D) Relative mRNA expressions of pro- or anti-inflammatory genes in the aorta. Data are presented as mean ± SD, *P < 0.05 and **P < 0.01 versus the NCD-fed mice group, respectively. #P < 0.05 and ##P < 0.01 versus the HFD-fed mice group, respectively. E) Flowchart of primary murine BMDM treated by PFOS (0.1 or 1 μM) and oxLDL (25 μg/mL). F) Relative mRNA expressions of pro- or anti-inflammatory genes in the BMDM. NC was the negative control. G) Flowchart of co-culture of PFOS-treated BMDM and vascular endothelial cells. H) Representative images and I) quantitative analysis of adherent macrophages stained by the Dio fluorescent probe to vascular endothelial cells. Scale bar, 500 μm. J) Representative images and K) quantitative analysis of BMDM using ORO staining. Scale bar, 50 μm. Data are presented as mean ± SD, *P < 0.05 and **P < 0.01 versus the NC group, respectively. #P < 0.05 and ##P < 0.01 versus oxLDL treatment alone group, respectively.

• 

  Fig. 3.. Role of glycolysis in the pro-inflammatory phenotype of macrophages promoted by PFOS under the hyperlipidemic condition. A) Top 15 KEGG pathways of enrichment analysis for differentially expressed genes in BMDM co-treated with PFOS and oxLDL. B) Content of cellular ATP in BMDM. NC was the negative control. C) Different glycolytic metabolites in the oxLDL (25 μg/mL) treatment alone group, and co-treated group with PFOS and oxLDL, each dot represents a single sample. D) Contents of pyruvate and lactate in BMDM. E) Glucose uptake of BMDM. F) KEGG pathway mapping containing metabolites and rate-limiting enzymes in BMDM affected by PFOS. Data are presented as mean ± SD, *P < 0.05 and **P < 0.01 versus the NC group, respectively. #P < 0.05 and ##P < 0.01 versus oxLDL treatment alone group, respectively. G) Relative mRNA expressions of genes related to glycolysis, pro-inflammatory, and anti-inflammatory upon PFOS exposure with or without 2-DG (10 μM). aP < 0.05 versus oxLDL treatment alone group, respectively. bP < 0.05 versus the co-treated group with PFOS and oxLDL. H) Schematic representation for a critical role of glycolysis in the regulation of phenotype transformation of…

• 

  Fig. 4.. Regulation of iron-storage function of FTH in glycolytic and pro-inflammatory phenotype macrophages promoted by PFOS under the hyperlipidemic condition. A) Molecular docking. PFOS binds to the iron-storage site and catalytic site of FTH or FTL, two major subunits of ferritin. B) Calculated values of RMSD for PFOS and FTH from MD simulations. C) Representative fluorescence image of the intracellular Fe2+ in BMDM and D) corresponding quantification. Scale bar, 10 μm. NC was the negative control. **P < 0.01 versus NC group, ##P < 0.01 versus oxLDL treatment alone group. BMDM were cotreated with PFOS and or without CQ (25 μM), Fe2+ chelator, under the conditions of oxLDL treatment. E) Content of cellular ATP in BMDM. Level of (F) glucose uptake and (G) lactate in BMDM. H) Relative mRNA expressions of pro-inflammatory and anti-inflammatory genes. aP < 0.05 versus the oxLDL treatment alone group, bP < 0.05 versus the co-treated group with PFOS and oxLDL, respectively. I) Predicted change in percentage of secondary structure of FTH after biological coupling of PFOS to FTH. J) The interaction energy of PFOS binds to the key amino acid residues of FTH. K) Conformational diagram for the…

• 

  Fig. 5.. Higher atherogenic risk is caused by PFCs with acidic substituent type and short carbon chain under hyperlipidemic conditions. A) Flowchart for predicting the potential perturbation of PFCs on the iron-storage function of FTH using an XGBoost model. B) Key factors for binding affinity between 24 PFCs and iron-storage site of FTH. C) Comparison of binding affinity of PFCs with different substituent types under the same carbon chain length. D) Linear correlation analysis between carbon chain length and binding affinity of PFCs and FTH under the same substituent type. (E) Intracellular Fe2+ content of BMDM after typical PFCs with different substituent types and carbon chain lengths exposure. Data are presented as mean ± SD, *P < 0.05 and **P < 0.01 versus the control group (oxLDL treatment alone group). Heat map cluster analysis for typical PFCs with (F) different substituent types and (G) carbon chain lengths and the key indicators, which include binding affinity with FTH iron-storage sites, anti-inflammatory genes, Fe2+, glycolysis genes, and anti-inflammatory genes. Mixed-effects analysis for typical PFCs with (H) different substituent types and (I) carbon chain lengths and key…

• 

  Fig. 6.. Association between PFC exposure and atherogenic indices in populations with higher TG levels. A) The panel study design and workflow summary. B) Associations of individual PFCs with AI and CRI for T3 group by LME model. *P-FDR < 0.05. C) Joint effects of PFCs mixture on AI and CRI based on positive WQS regression model. D) Exposure–response relationships of PFOS with AI and CRI fitted by GAMs. E) Effects of 4 PFCs with different substituent types and carbon chain lengths on AI and CRI using multiple PFC LME models. Models were adjusted for age, gender, BMI, parental education level, passive smoking status, extra-school activity, season, and dietary variables, including the intake frequency of meat, seafood, packaged food, vegetable, and fruit.