Bisphenol AF promotes inflammation in human white adipocytes

Abstract

Endocrine-disrupting chemicals interact with transcription factors essential for adipocyte differentiation. Exposure to endocrine-disrupting chemicals corresponds with elevated risks of obesity, but the effects of these compounds on human cells remain largely undefined. Widespread use of bisphenol AF (BPAF) as a bisphenol A (BPA) alternative in the plastics industry presents unknown health risks. To this end, we discovered that BPAF interferes with the metabolic function of mature human adipocytes. Although 4-day exposures to BPAF accelerated adipocyte differentiation, we observed no effect on mature fat cell marker genes. Additional gene and protein expression analysis showed that BPAF treatment during human adipocyte differentiation failed to suppress the proinflammatory transcription factor STAT1. Microscopy and respirometry experiments demonstrated that BPAF impaired mitochondrial function and structure. To test the hypothesis that BPAF fosters vulnerabilities to STAT1 activation, we treated mature adipocytes previously exposed to BPAF with interferon-γ (IFNγ). BPAF increased IFNγ activation of STAT1 and exposed mitochondrial vulnerabilities that disrupt adipocyte lipid and carbohydrate metabolism. Collectively, our data establish that BPAF activates inflammatory signaling pathways that degrade metabolic activity in human adipocytes. These findings suggest how the BPA alternative BPAF contributes to metabolic changes that correspond with obesity.

Keywords: PPARγ; adipocyte; inflammation; metabolism; transcription.

Fig. 1.

Impacts of bisphenol AF (BPAF) on lipid accumulation in human adipocytes. A: human preadipocytes were induced to differentiate for 96 h in the presence or absence of BPAF concentrations up to 5 μM. Cells were then stained with DAPI (nucleus) and LipidTOX (lipid). Representative images are shown. Scale bars are 50 μm. B: lipid accumulation was quantified by image analysis. C: nuclei were counted to analyze toxicity (n ≥ 8 wells; means ± SE; *P < 0.05 relative to vehicle; #P < 0.05 relative to no differentiation). D: in parallel, relative mRNA levels were determined for canonical adipocyte marker genes (n = 6; means ± SE). E: to further validate effects on peroxisome proliferator-activated receptor-γ (PPARγ)-mediated transcription, HepG2 cells were treated with increasing doses of BPAF ± 1 µM rosiglitazone (rosi). Transcriptional activity was determined by measuring PPAR responsive element (PPRE)-luciferase normalized to β-galactosidase (β-gal; n = 4; means ± SE). ADIPOQ, adiponectin, C1Q and collagen domain containing; CIDEC, cell death-inducing DFFA-like effector c; diff, differentiation; FABP4, fatty acid-binding protein 4; RLU, relative light units.

Fig. 2.

Bisphenol AF (BPAF) does not alter occupancy of peroxisome proliferator-activated receptor-γ (PPARγ) near DNA sequences that regulate transcription of adipocyte marker genes. A and B: after 14-day treatment, PPARγ chromatin immunoprecipitation was performed in human adipocytes. A and B: quantitative PCR was used to analyze genomic occupancy using primers flanking PPARγ binding sites downstream of adiponectin, C1Q and collagen domain containing (ADIPOQ_dwnstrm; A) and within the uncoupling protein 1 enhancer (UCP1_enh; B). C: an intronic region of cyclin D1 (CCND1_intr4) served as a negative control (n = 4 independent experiments). Here, veh, vehicle.

Fig. 3.

Bisphenol AF (BPAF) reduces the metabolic performance of human adipocytes. A: mitochondria (MitoTracker), nuclei (DAPI), and lipids (LipidTOX) were labeled in human adipocytes differentiated ± 5 µM BPAF. Scale bars are 20 μm. B: respiration [as oxygen consumption rate (OCR)] was measured in human adipocytes differentiated 14 days in the presence or absence of BPAF concentrations up to 5 μM BPAF. OCR was measured over time with the addition of oligomycin (α), carbonyl cyanide p-trifluoromethoxyphenylhydrazone (FCCP; β), and antimycin A/rotenone (γ). C: maximal respiration capacity was calculated from OCR measurements performed after FCCP addition. Here, n ≥ 4 independent experiments per group. EtOH, ethanol. *P < 0.05 relative to treatment.

Fig. 4.

Inflammatory genes and proteins are not fully suppressed by bisphenol AF (BPAF) during human adipocyte differentiation. A–D: relative mRNA levels were determined for canonical white adipocyte marker genes (A), mitochondrial function genes (B), STAT1 target genes (C), and cytokine genes (D; n ≥ 6 experiments; means ± SE) in human preadipocytes treated ± differentiation (Diff) ± 5 µM BPAF for 14 days. Gene expression was normalized to TATA box-binding protein (TBP) and presented as relative mRNA. *P < 0.05 relative to vehicle; #P < 0.05 relative to no differentiation. E: immunoblot analysis of selected adipocyte metabolic markers and STAT1 in human preadipocytes treated ± differentiation ± 5 µM BPAF for 14 days. Quantification of STAT1 protein levels is expressed as relative change (Δ) vs. no-differentiation (−d) controls. ADIPOQ, adiponectin, C1Q and collagen domain containing; AGT, angiotensinogen; CIDEA, cell death-inducing DFFA-like effector a; Cyt C, cytochrome c; FABP4, fatty acid-binding protein 4; GLUT4, solute carrier family 2 member 4; Hsp60 and Hsp90, 60- and 90-kDa heat shock proteins, respectively; IRF1, interferon regulatory factor 1; ISG15, ISG15 ubiquitin-like modifier; LEP, leptin; PGC1A, peroxisome proliferator-activated receptor-γ coactivator 1α; PPARG, peroxisome proliferator-activated receptor-γ; PRDM16, PR/SET domain 16; rel, relative; UCP1, uncoupling protein 1.

Fig. 5.

The interferon-γ response in human adipocytes is primed by bisphenol AF (BPAF) treatment. A: immunoblots of total cell lysates from human preadipocytes differentiated 14 days ± 5 µM BPAF followed by treatment with vehicle (0.0001% BSA) or recombinant human IFNγ (100 ng/mL) for 10 min (short = 3–5-s film exposure; long = 60-s film exposure). B–D: quantitative PCR analysis of STAT1 (B), interferon regulatory factor 1 (IRF1; C), and uncoupling protein 1 (UCP1; D; n ≥ 3 experiments; means ± SE) expression in preadipocytes differentiated 14 days ± 5 μM BPAF treated with vehicle or 100 ng/mL IFNγ for 24 h. *P < 0.05 relative to vehicle; #P < 0.05 relative to IFNγ. E: HepG2 cells were transfected with a UCP1 −5.5-kb-luciferase reporter (UCP1-luc) and vectors expressing peroxisome proliferator-activated receptor-γ (PPARγ) and STAT1. Forty-eight hours after transfection, HepG2 cells were exposed to 1 μM rosiglitazone and BPAF concentrations up to 5 μM ± 100 ng/mL IFNγ. Transcriptional activity was determined by measuring UCP1-luc normalized to β-galactosidase (β-gal; n = 6; means ± SE, *P < 0.05 relative to vehicle; #P < 0.05 relative to IFNγ). F and G: oxygen consumption rate (OCR; F) and immunoblots of total cell lysates (G) in preadipocytes differentiated 14 days ± 5 µM BPAF treated with vehicle or 100 ng/mL IFNγ for 24 h. OCR was measured over time with the addition of oligomycin (α), carbonyl cyanide p-trifluoromethoxyphenylhydrazone (FCCP; β), and antimycin A/rotenone (γ). For Seahorse assays, statistical significance indicates changes in maximal respiration relative to vehicle controls (*P < 0.05, compared with the black curve in F). Cyt C, cytochrome c; Hsp60 and Hsp90, 60- and 90-kDa heat shock proteins, respectively; rel, relative; RLU, relative light units.