The ω6-fatty acid, arachidonic acid, regulates the conversion of white to brite adipocyte through a prostaglandin/calcium mediated pathway

Abstract

Objective: Brite adipocytes are inducible energy-dissipating cells expressing UCP1 which appear within white adipose tissue of healthy adult individuals. Recruitment of these cells represents a potential strategy to fight obesity and associated diseases.

Methods/results: Using human Multipotent Adipose-Derived Stem cells, able to convert into brite adipocytes, we show that arachidonic acid strongly inhibits brite adipocyte formation via a cyclooxygenase pathway leading to secretion of PGE2 and PGF2α. Both prostaglandins induce an oscillatory Ca(++) signaling coupled to ERK pathway and trigger a decrease in UCP1 expression and in oxygen consumption without altering mitochondriogenesis. In mice fed a standard diet supplemented with ω6 arachidonic acid, PGF2α and PGE2 amounts are increased in subcutaneous white adipose tissue and associated with a decrease in the recruitment of brite adipocytes.

Conclusion: Our results suggest that dietary excess of ω6 polyunsaturated fatty acids present in Western diets, may also favor obesity by preventing the "browning" process to take place.

Keywords: Calcium oscillation; PGE2; PGF2α; PGI2; Polyunsaturated fatty acids; UCP1.

Figure 1

Differential effect of ARA on white to brite adipocyte conversion. hMADS cells were differentiated into white (R3–9) or brite (R3–9/14–17) adipocytes, and treated or not between days 14 and 17 with 10 μM ARA. (A) UCP1 mRNA expression determined by RT-qPCR and (B) UCP1 protein level analyzed by Western blot (whole cell lysates, 80 μg/lane). β-tubulin was used as loading control. Histograms represent mean ± SEM of 3 independent experiments. a: p < 0.01 vs R3–9/14–17.

Figure 2

hMADS cells were maintained in the presence of 100 nM rosiglitazone from day 3 to day 9 (R3–9) or further exposed to rosiglitazone from day 14 to day 17 (R3–9/14–17) in the absence or presence of 10 μM ARA. Oil red O staining (A) and GPDH activity measurements (B) were carried out. (C) Immunodetection of perilipin (in red) was performed at day 17 and nuclei were counterstained with DAPI (in blue). (D) mRNA expression of adipogenic markers was determined by RT-qPCR. Histograms represent mean ± SEM of 3 independent experiments. a: p < 0.01 vs R3–9/14–17. Scale: 20 μm.

Figure 3

Changes in the brite phenotype of ARA-treated cells. hMADS cells were maintained in the presence of 100 nM rosiglitazone and treated or not from day 14 to day 17 with 10 μM ARA. Rosiglitazone-untreated cells were used as control. (A) Brown (CPT-1M and CIDEA) and mitochondrial markers (COX10, CYTC, MFN2, TOMM20 and TIMM23) mRNA expression were analyzed by RT-qPCR. (B) Mitochondria content analyzed by immunodetection of TIMM23 (red). Nuclei were counterstained with DAPI (blue). (C) Basal (untreated cells) and maximal (FCCP-treated cells) oxygen consumption levels measured in resuspended cells with an oxygraphic probe. (D) Cytochrome c oxidase activity was measured. Histograms represent mean ± SEM of 3 (A and D) or 4 (C) independent experiments. a: p < 0.01 vs R3–9/14–17. Scale: 20 μm.

Figure 4

Cyclooxygenases drive ARA effect in hMADS cells. (A) Cyclooxygenases (COX-1 and COX-2) mRNA expression assessed by RT-qPCR in cells exposed to 100 nM rosiglitazone and treated or not with 10 μM ARA between days 14 and 17. Untreated cells were used as control. (B) and (C) hMADS adipocytes were treated with 100 nM rosiglitazone and 10 μM ARA supplemented or not with 1 μM indomethacin (indo) or 100 nM celecoxib. mRNA expression of various adipogenic markers was analyzed by RT-qPCR. Histograms represent mean ± SEM of 3 independent experiments. a: p < 0.01 vs R3–9/14–17; b: p < 0.01 vs R3–9/14–17 + ARA.

Figure 5

Prostaglandins drive ARA effects on brite adipogenesis. hMADS cells were treated or not with 100 nM rosiglitazone from day 14 to day 17. (A) Rosiglitazone-treated cells were exposed to 10 μM ARA in the presence or not of 100 nM celecoxib for the 3 days. After a last change PGF2α, PGE2 and 6-keto-PGF1α were quantified from culture medium after 10 min incubation by EIA. (B) UCP1 mRNA expression analyzed by RT-qPCR or (C) UCP1 protein level assessed by Western blotting (whole cell lysates, 80 μg/lane, β-tubulin was used as loading control) in rosiglitazone-treated cells exposed to 10 μM ARA, 1 μM 16,16-dm-PGE2 (dm-PGE2), 10 nM fluprostenol (agonist of PGF2α receptor) or 1 μM cPGI2 (stable analog of PGI2). (D) Effects of ARA, dm-PGE2 and fluprostenol on others mRNA expressions analyzed by RT-qPCR. Untreated cells were used as control. Histograms represent mean ± SEM of 3 independent experiments. a: p < 0.01 vs R3–9/14–17; b: p < 0.01 vs R3–9/14–17 + ARA.

Figure 6

Intra-cellular Ca⁺⁺ oscillations are induced by ARA and its metabolites. hMADS cells were differentiated into brite adipocytes in the presence rosiglitazone and incubated 15 min with a fluorescent sensitive Ca⁺⁺ probe (Quest Fluo-8). Cells were analyzed by live fluoromicroscopy. (A) 100 μM ARA, 10 μM dm-PGE2 and 100 nM fluprostenol induce a transitory increase of i[Ca⁺⁺], followed by i[Ca⁺⁺] oscillations with a sustained frequency and intensity. (B) 100 μM LA and 10 μM cPGI2 did not trigger i[Ca⁺⁺] rise. Each track represents integrated imaging of an individual cell. These data are representative of 5 independent experiments (10–50 cells recorded in each experiment).

Figure 7

A Ca⁺⁺/ERK pathway is involved in the inhibition of white to brite adipocyte conversion. (A–C) hMADS cells were treated or not with 100 nM rosiglitazone from day 14 to day 17. Rosiglitazone-treated cells were exposed or not to 10 nM fluprostenol. Unstimulated brite (R3–9/14–17) cells were used as control. (A) UCP1 and Perilipin mRNA expression analyzed by RT-qPCR in the absence or presence of 300 μM EGTA. (B) Upper panel: hMADS adipocytes deprived of Ca⁺⁺ for 48 h in DMEM/BSA 0.5% and then exposed to 2 mM CaCl₂ or 10 nM fluprostenol. Lower panel: deprived hMADS adipocytes pretreated for 10 min with 300 μM EGTA or 10 μM U0126, and then stimulated with 10 nM fluprostenol. Cells were lysed at the indicated times and analyzed by western blotting (whole cell lysate, 25 μg protein/line). Values correspond to ERK1/2 phosphorylated vs ERK1/2 total protein intensity ratio evaluated with fluorescent signal quantification. (C) UCP1 and Perilipin mRNA expression analyzed by RT-qPCR in the absence or presence of 10 μM U0126. Histograms represent mean ± SEM of 3 independent experiments. a: p < 0.01 vs R3–9/14–17; b: p < 0.01 vs R3–9/14–17 + fluprostenol.

Figure 8

Impact of ARA metabolites production in scWAT on inducible brite adipocytes formation. 10-week-old C57BL/6RccHsd female mice fed standard diet supplemented with ARA or OA for 4 weeks. During the fourth week, mice subdivided in two groups, a first group was treated with CL316,243 and the second group with vehicle (NaCl). (A) Body weight gain between week 1 and 4 and plasma leptin levels were determined. (B) UCP1 protein level was assessed in scWAT (40 μg/lane). β-tubulin was used as loading control and iBAT protein extract (4 μg/lane) as UCP1 positive control. (C) UCP1, COX-2 and representative white and brite adipocyte markers mRNA were determined in scWAT by RT-qPCR. (D) Representative histological sections (4 μm, paraffin-embedded, HES staining) of scWAT from different treatment are shown. (E) Prostanoid amounts were measured by LC-MS/MS in scWAT of each group of mice. Histograms represent means ± SEM of 8–12 mice per group. a: p < 0.01 vs NaCl group and b: p < 0.01 vs OA group.