Uncoupling protein-2 attenuates glucose-stimulated insulin secretion in INS-1E insulinoma cells by lowering mitochondrial reactive oxygen species

Abstract

Glucose-stimulated insulin secretion (GSIS) by pancreatic β cells is regulated by mitochondrial uncoupling protein-2 (UCP2), but opposing phenotypes, GSIS improvement and impairment, have been reported for different Ucp2-ablated mouse models. By measuring mitochondrial bioenergetics in attached INS-1E insulinoma cells with and without UCP2, we show that UCP2 contributes to proton leak and attenuates glucose-induced rises in both respiratory activity and the coupling efficiency of oxidative phosphorylation. Strikingly, the GSIS improvement seen upon UCP2 knockdown in INS-1E cells is annulled completely by the cell-permeative antioxidant MnTMPyP. Consistent with this observation, UCP2 lowers mitochondrial reactive oxygen species at high glucose levels. We conclude that UCP2 plays both regulatory and protective roles in β cells by acutely lowering GSIS and chronically preventing oxidative stress. Our findings thus provide a mechanistic explanation for the apparently discrepant findings in the field.

Fig. 1

UCP2 knockdown lowers mitochondrial respiratory rate and increases the coupling efficiency of attached INS-1E cells. (A and B) Time-resolved absolute respiratory rates (J_o), normalized to cell number as measured by nuclear staining, of cells transfected with Ucp2-targeted (closed symbols) or scrambled siRNA (open symbols) and treated as described under Experimental procedures. After four Seahorse assay cycles at 2 mM glucose, cells were subjected to 30 mM glucose (G30) and, in parallel experiments, either oligomycin (O) or a mixture of oligomycin, rotenone, and myxothiazol (ORM). Each time point represents the average ± SEM (n = 8–12) of 8–12 separate wells sampled from seven independent Seahorse XF24 plates. (C) Respiratory rates in the presence of the oligomycin/rotenone/myxothiazol mix were averaged and subtracted from all other rates to yield mitochondria-specific respiratory activities. Glucose-stimulated activity (G30) is the average of the last two rate measurements before oligomycin addition, and the oligomycin-insensitive activity (oligomycin) is the average of the two rates after this addition. (D) Coupling efficiencies were calculated as the fraction of glucose-stimulated mitochondrial oxygen uptake that is sensitive to oligomycin and thus reflects respiratory activity used to drive ATP synthesis. The data in (C) and (D) are averages ± SEM (n = 13–15) of 13 (scrambled siRNA, open bars) or 15 (Ucp2-targeted siRNA, shaded bars) separate wells sampled from seven independent XF24 plates. *p < 0.05.

Fig. 2

UCP2 knockdown improves GSIS in a MnTMPyP-sensitive manner, amplifies a glucose-induced rise in coupling efficiency, and improves the cells’ respiratory response to glucose. Cells transfected with Ucp2-targeted or scrambled siRNA (shaded and open bars, respectively) were first starved of glucose as described under Experimental procedures and then subjected to glucose in the absence (G2, G5, G30) or presence (G5-PyP, G30-PyP) of 20 μM MnTMPyP; G2, G5, and G30 reflect 2, 5, and 30 mM glucose, respectively. (A) Insulin secretion data are averages ± SEM of four independent experiments with each condition assayed six times. (B) Mitochondrial coupling efficiencies and (C) the respiratory responses of the cells to glucose are averages ± SEM of 4–11 separate XF24 runs. These bioenergetic parameters were calculated from Seahorse traces as described under Experimental procedures. **p < 0.01 and ***p < 0.001.

Fig. 3

MnTMPyP lowers the rate of MitoSOX oxidation by INS-1E cells. (A and B) Typical time-resolved fluorescence observed upon the oxidation of 5 μM MitoSOX by nontransfected INS-1E cells that were not subjected to periods of glucose starvation. Traces are shown to illustrate assays that were done without any effector (circles) or added probe (A, diamonds) and assays that were performed in the presence of 15 μM antimycin A (squares), 15 μM FCCP (A, triangles), 20 μM MnTMPyP (B, diamonds), or 15 μM antimycin A and 20 μM MnTMPyP (B, triangles). Data are averages ± SEM of 4–6 wells sampled from a single 96-well plate and, to aid clarity, symbols are shown only for every 10th measurement. (C) MitoSOX oxidation rates in the presence (shaded bars) and the absence (open bars) of 20 μM MnTMPyP were calculated from the slopes (after the first 1000 s) of progress curves exemplified in (A) and (B). Data are averages ± SEM of 34–38 wells sampled from seven plates and 18–22 wells sampled from four plates in the absence and the presence of MnTMPyP, respectively. (D) MitoSOX oxidation by nontransfected glucose-starved INS-1E cells (see Experimental procedures) subjected to 2, 5, and 30 mM glucose (G2, G5, and G30, respectively) in the presence (shaded bars) and the absence (open bars) of 20 μM MnTMPyP. Data are averages ± SEM of 22–24 wells sampled from three plates. RFU, relative fluorescence units. ***p < 0.001.

Fig. 4

UCP2 activity lowers mitochondrial ROS at high glucose levels. Cells transfected with Ucp2-targeted (shaded bars) or scrambled (open bars) siRNA were first starved of glucose as described under Experimental procedures and then subjected to glucose in the absence (G2, G5, G30) or presence (G2-PyP, G5-PyP, G30-PyP) of 20 μM MnTMPyP; G2, G5, and G30 reflect 2, 5, and 30 mM glucose, respectively. (A) MitoSOX (5 μM) oxidation rates were obtained as described for Fig. 3 and (B) DHE (100 μM) oxidation rates were calculated from the slopes of the first 500 s of the fluorescence traces. Data are averages ± SEM of 25–30 wells sampled from six 96-well plates. *p < 0.05, **p < 0.01 and ***p < 0.001.