Regulation of Mitochondria-Associated Membranes (MAMs) by NO/sGC/PKG Participates in the Control of Hepatic Insulin Response

Abstract

Under physiological conditions, nitric oxide (NO) produced by the endothelial NO synthase (eNOS) upregulates hepatic insulin sensitivity. Recently, contact sites between the endoplasmic reticulum and mitochondria named mitochondria-associated membranes (MAMs) emerged as a crucial hub for insulin signaling in the liver. As mitochondria are targets of NO, we explored whether NO regulates hepatic insulin sensitivity by targeting MAMs. In Huh7 cells, primary rat hepatocytes and mouse livers, enhancing NO concentration increased MAMs, whereas inhibiting eNOS decreased them. In vitro, those effects were prevented by inhibiting protein kinase G (PKG) and mimicked by activating soluble guanylate cyclase (sGC) and PKG. In agreement with the regulation of MAMs, increasing NO concentration improved insulin signaling, both in vitro and in vivo, while eNOS inhibition disrupted this response. Finally, inhibition of insulin signaling by wortmannin did not affect the impact of NO on MAMs, while experimental MAM disruption, using either targeted silencing of cyclophilin D or the overexpression of the organelle spacer fetal and adult testis-expressed 1 (FATE-1), significantly blunted the effects of NO on both MAMs and insulin response. Therefore, under physiological conditions, NO participates to the regulation of MAM integrity through the sGC/PKG pathway and concomitantly improves hepatic insulin sensitivity. Altogether, our data suggest that the induction of MAMs participate in the impact of NO on hepatocyte insulin response.

Keywords: cyclic guanosine monophosphate (cGMP)-dependent protein kinase; hepatic glucose metabolism; metabolic flexibility; mitochondria-associated endoplasmic reticulum membranes; nitric oxide.

Figure 1

Nitric oxide (NO) regulates mitochondria-associated membranes’ (MAMs) integrity in vivo. (A) Representative scheme of strategies used to modulate NO concentration in C57Bl6/JOlash male mice. BH4 (tetrahydrobiopterin; a cofactor of endothelial NO synthase (eNOS), 12.5 mg/kg), l-Name (ω-nitro-l-arginine methyl ester hydrochloride; an inhibitor of eNOS, 25 mg/kg), and a combination of the two were given by intraperitoneal (ip) injection twice the day before and 2 h before euthanasia (n = 5 to 7 mice per condition). Impact of modulating in vivo NO production on (B) liver NO concentration assessed using Daf-FM (15 µM) on fresh homogenates; (C,D) amount of proteins in the MAM fraction isolated from fresh liver using differential ultracentrifugation expressed relative to (C) total and (D) mitochondrial proteins; (E) amount of proteins in the mitochondrial fraction isolated from fresh liver using differential ultracentrifugation expressed relative to total proteins; (F) endoplasmic reticulum (ER)–mitochondria interactions at different distances (from 0 to 50 nm) analyzed from transmission electronic microscopy (TEM) images. (F) Representative TEM images; (G) quantitative analysis of the interactions according to spacing (0–10 nm, 10–20 nm, 20–30 nm, 30–40 nm, 40–50 nm; (H) interactions in a 30–50 nm range; (I) and the number of mitochondria per field (I). A minimum of 10 images (scale bar 0.5 µm) was taken for each mouse (n = 160–260 ER–mitochondria interaction analyzed/mice group condition). £ and #, p < 0.05 vs. control.

Figure 2

NO regulates the hepatic insulin sensitivity in vivo. (A–C) Impact of modulating in vivo the NO production on hepatic insulin signaling. C57Bl6/JOlash male mice (n = 5 mice per condition) received ip injection of BH4 (12.5 mg/kg), l-Name (25 mg/kg), and a combination of the two, twice the day before and 2 h before the insulin test (0.75 U/kg). (A) Representative Western blots and quantitative analysis of insulin-stimulated (B) protein kinase B (Akt) and (C) glycogen synthase kinase 3 beta (GSK3β). (D) Impact of modulating in vivo the NO production on hepatic glucose production. C57Bl6/JOlash male mice (n = 8 mice per condition) received ip injection of BH4 (12.5 mg/kg), l-Name (25 mg/kg), and a combination of the two, twice the day before and 2 h before the pyruvate test (2 g/kg). £ and #, p < 0.05 vs. control. *p < 0.05, **p < 0.01, and ***p < 0.001 vs. appropriate control glycaemia, respectively.

Figure 3

NO regulates the ER–mitochondria contact points and the mitochondrial network architecture, density, and respiratory activity in hepatocytes in vitro. (A) Representative scheme of strategies used to modulate NO concentration in vitro. Arginine is a substrate of eNOS, l-Name in an inhibitor of eNOS, and NONOate (diethylamine NONOate sodium salt hydrate) is a NO donor. (B) Impact of l-Name (1 mM), arginine (1 mM), and a combination of the two on NO concentration in fresh Huh7 cells assessed using Daf-FM (15 µM). (C–G) Impact of modulating NO concentration for 24 h using l-Name (1 mM), arginine (1 mM), a combination of the two, and NONOate (1 mM), on (C,D) ER–mitochondria interactions assessed through VDAC1/IP3R1 (voltage dependent anion channel 1/inositol 1,4,5-trisphosphate receptor) interactions using in situ proximity ligation assay (PLA) in (C) Huh7 cells and (D) primary rat hepatocytes (n = 10 images minimum per experiment, three independent series per treatment, representative image at top and enlarged in Figure S7A,B, and quantitative analysis below, scale bar 15 µm, ×100); (E) mitochondrial network architecture, assessed in Huh7 cells using MitoTracker^® Green (500 nM) and calculation of aspect ratio (AR) and form factor (FF) (n = 10 images minimum per experiment, three independent series per treatment, representative image at top and enlarged in Figure S7C, scale bar 15 µm, ×100); (F) mitochondrial (COX-1, cyclooxygenase-1) DNA content relative to nuclear (PPIA, peptidylprolyl isomerase A) DNA assessed in Huh7 cells using RT-qPCR (n = three independent experiments); (G) whole cell oxygen consumption assessed using oxygraphy in Huh7 cells (n = minimum six independent experiments). £ and ^#, p < 0.05 vs. control; ***, p < 0.05 vs. respective uncoupled state.

Figure 4

NO regulates the insulin response in hepatocytes in vitro. (A,B) Impact of modulating NO concentration for 24 h using l-Name (1 mM), arginine (1 mM), a combination of the two, and NONOate (1 mM) on insulin signaling in (A) Huh7 cells and (B) primary rat hepatocytes. (C,D) Impact of inhibiting phosphoinositide 3-kinase (PI3K) using wortmannin (1 µM) on the regulation by NO of (C) ER–mitochondria interactions assessed using in situ PLA and (D) insulin signaling. For Western blot, representative image (at top) and quantitative analysis (below) of insulin-stimulated Akt (n = three independent experiments). For in situ PLA, representative image (at top and enlarged in Figure S7D) and quantitative analysis of VDAC1/IP3R1 interactions (below) (n = 10 images minimum per experiment, three independent series per treatment, scale bar 15 µm, ×100). £ and ^#, p < 0.05 vs. control. £ and #, p < 0.05 vs. control.

Figure 5

NO-activated subunits of guanylate cyclase/protein kinase G (sGC/PKG) pathway regulates the ER–mitochondria contact points and participates in the modulation of the mitochondrial network architecture, density, and respiratory activity in hepatocytes in vitro. (A) Representative scheme of strategies used to modulate the sGC/PKG pathway in vitro. Arginine is the substrate of eNOS, NONOate is a NO donor, BAY41-2272 (3-(4-amino-5-cyclopropylpyrimidin-2-yl)-1-(2-fluorobenzyl)-1H-pyrazolo [3–b]pyridine) is an activator of the sGC, 8-pCPT-cGMP (8-(4-chlorophenylthio)-guanosine 3′,5′-cyclic monophosphate sodium salt) activates PKG whereas KT5823 inhibits the kinase. (B) Impact of inhibiting the sGC/PKG pathway for 24 h in Huh7 cells using KT5823 (1 µM) in the presence or absense of arginine (1 mM) or NONOate (1 mM), on ER–mitochondria interactions assessed using in situ PLA (n = 10 images minimum per experiment, three independent series per treatment, representative images at top and enlarged in Figure S7E, and quantitative analysis below, scale bar 15 µm, ×100). (C) Impact of activating the sGC/PKG pathway for 24 h in Huh7 cells using 8-pCPT-cGMP (100 µM) and BAY41-2272 (2 µM) on ER–mitochondria interactions assessed using in situ PLA (n = 10 images minimum per experiment, three independent series per treatment, representative images at top and enlarged in Figure S7F, and quantitative analysis below, scale bar 15 µm, ×100). (D–F) Impact of modulating the sGC/PKG pathway for 24 h in Huh7 cells using KT5823 (1 µM), NONOate (1 mM), a combination of the two, 8-pCPT-cGMP (100 µM), and BAY41-2272 (2 µM) on (D) mitochondrial network architecture assessed using MitoTracker^® Green (500 nM) and calculation of aspect ratio (AR) and form factor (FF) (n = 10 images minimum per experiment, three independent series per treatment, representative image at top and enlarged in Figure S7G, scale bar 15 µm ×100); (E) mitochondrial (COX-1) DNA content relative to nuclear (PPIA) DNA assessed using RT-qPCR (n = three independent experiments); (F) whole cell oxygen consumption assessed using oxygraphy (n = minimum six independent experiments). £ and #, p < 0.05 vs. control; ***, p < 0.05 vs. respective uncoupled state.

Figure 6

NO-activated sGC/PKG pathway regulates the insulin response in hepatocytes in vitro. Impact on insulin signaling in Huh7 cells (A) inhibiting the sGC/PKG pathway using KT5823 (1 µM) in the presence or absence of arginine (1 mM) or NONOate (1 mM) for 24 h; and (B) activating the sGC/PKG pathway using 8-pCPT-cGMP (100 µM) and BAY41-2272 (2 µM) for 24 h. ^£ and ^#, p < 0.05 vs. control.

Figure 7

MAM integrity is required for mediating NO impact on hepatic insulin sensitivity in vitro. Impact of modulating NO concentration using l-Name (1 mM), arginine (1 mM), and NONOate (1 mM), in control and CypD-knockout (KO) Huh7 cells on (A) ER–mitochondria interactions assessed through VDAC1/IP3R1 interactions using in situ PLA and (C) insulin signaling. Impact of modulating NO concentration using arginine (1 mM), NONOate (1 mM), and 8-pCPT-cGMP (100 µM) in Huh7 cells transfected with a control adenovirus (green fluorescent protein; ad GFP) and FATE-1 (fetal and adult testis-expressed 1; ad FATE-1) on (B) ER–mitochondria interactions assessed through VDAC1/IP3R1 interactions using in situ PLA and (D) insulin signaling. For in situ PLA, representative image (at top and enlarged in Figure S7F,H,I) and quantitative analysis of VDAC1/IP3R1 interactions (below) (n = 10 images minimum per experiment, three independent series per treatment, scale bar 15 µm, ×63 or ×100). For Western blot, representative images (at top) and quantitative analysis (below) of insulin-stimulated Akt (n = three independent experiments). ^£ and ^#, p < 0.05 vs. control. *, ** and ***, p < 0.05, p < 0.01, and p < 0.001 vs. appropriate Huh7 control, respectively.