Endotoxemia-mediated activation of acetyltransferase P300 impairs insulin signaling in obesity

Abstract

Diabetes and obesity are characterized by insulin resistance and chronic low-grade inflammation. An elevated plasma concentration of lipopolysaccharide (LPS) caused by increased intestinal permeability during diet-induced obesity promotes insulin resistance in mice. Here, we show that LPS induces endoplasmic reticulum (ER) stress and protein levels of P300, an acetyltransferase involved in glucose production. In high-fat diet fed and genetically obese ob/ob mice, P300 translocates from the nucleus into the cytoplasm of hepatocytes. We also demonstrate that LPS activates the transcription factor XBP1 via the ER stress sensor IRE1, resulting in the induction of P300 which, in turn, acetylates IRS1/2, inhibits its association with the insulin receptor, and disrupts insulin signaling. Pharmacological inhibition of P300 acetyltransferase activity by a specific inhibitor improves insulin sensitivity and decreases hyperglycemia in obese mice. We suggest that P300 acetyltransferase activity may be a promising therapeutic target for the treatment of obese patients.Elevated plasma LPS levels have been associated with insulin resistance. Here Cao et al. show that LPS induces ER stress and P300 activity via the XBP1/IRE1 pathway. P300 acetylates IRS1/2 and inhibits its binding with the insulin receptor. The consequent impairment of insulin signaling can be rescued by pharmacological inhibition of P300.

Fig. 1

HFD feeding results in P300 induction in the liver. a, b Insulin tolerance test (4 h fasting, 0.8 unit per kg, n  = 4) a and pyruvate tolerance test (5 h fasted, 2 g kg⁻¹, n  = 4) b in C57BL/6 mice fed a regular diet or HFD for 1–4 weeks. c WT mice were fed a regular diet or an HFD for 2 weeks. Basal glucose production in mice after 5 h fasting (left) and glucose disposal rates, glucose infusion rates, and hepatic glucose production during the clamp experiment (n  = 4–5). d Liver tissues were harvested in mice after feeding on an HFD for 1, 2 and 4 weeks, each lane represents a mouse sample. Data are presented as mean ± s.e.m. Statistical significance was calculated with a Student’s t-test

Fig. 2

Elevation of hepatic LPS levels by HFD feeding leads to P300 induction. a CD14 knockout mice were fed on a regular chow diet or an HFD for 2 weeks. b LPS levels in the liver of mice fed an HFD for up to 4 weeks (n  = 4). Each bar represents the mean ± s.e.m. *, P  < 0.05, Student’s t-test. c Hepa1-6 cells were treated with indicated amount of LPS for 24 h. d LPS (0.5 μg per g of body weight per day) was administrated to C57BL/6 mice through intraperitoneal injection for 2 weeks. e Hepa1-6 cells were treated with vehicle or LPS (500 ng ml⁻¹) for 24 h, followed by the treatment with (20 µM) MG132 for 4 h. Cell lysates were incubated with antibody against P300 (16 h, 4 °C). f C57BL/6 mice were fed on an HFD for 2 weeks, frozen liver tissues were sectioned and immunostained with P300 antibody. Scale bar, 20 µm. g Cytoplasmic and nuclear extracts were prepared from the liver of mice fed a regular diet or HFD for 2 weeks (n  = 3). h Densitometric analysis of P300 in the cytoplasm and nucleus in hepatocytes treated with LPS as in Fig. 2i (n = 8). i Forty-eight hours after the treatment with LPS (50 ng ml⁻¹), Hepa1-6 cells were subjected to immunofluorescence staining. Scale bar, 10 µm. a, d Each lane represents a mouse sample

Fig. 3

Activation of the IRE1-XBP1s pathway leads to P300 induction and cytoplasmic distribution. a Hepa1-6 cells were treated with 300 nM thapsigargin and harvested at indicated time points. b Hepa1-6 cells were pretreated with TUDCA (500 μg ml⁻¹) for 1 h, then, treated with indicated amounts of thapsigargin for 4 h. c Mice were fed an HFD along with the treatment of TUDCA (500 mg kg⁻¹ d⁻¹ i.p.) for 2 weeks. RD, regular diet. Veh, vehicle. d Immunoblots of lysates from Hepa1-6 cells treated with SCR or XBP1 adenoviral shRNAs for 24 h followed by treatment with LPS for 24 h. e Homozygous floxed XBP1 mice were injected with AAV8-TBG-Cre (1 × 10¹¹ GC/mouse) through the jugular vein, and fed on an HFD for 2 weeks. (lower) Densitometric analysis of P300 in the liver. f Hepa1-6 cells were transfected with equal amounts of expression plasmids of pcDNA and XBP1s for 48 h. g Activation of ER stress in the liver of age-matched male mice fed an HFD for 2–32 weeks. Right panel, densitometric analysis of XBP1s in the liver (n = 4). h The mRNA levels of ERdj4, p58^ipk, CHOP, and ATF6 in the liver of age-matched mice fed a regular diet or HFD (18 weeks). Each bar represents the mean ± s.e.m. *, P < 0.05, Student’s t-test. c, e, g Each lane represents a mouse sample

Fig. 4

Overexpression of XBP1s impairs insulin signaling and increases the cytoplasmic localization of P300. a Forty-eight hours after the addition of adenoviral-GFP and-FLAG-tagged XBP1s, Hepa1-6 cells were subjected to 4 h serum starvation and treated with insulin (10 nM) for 10 min. b WT mice were injected with AAV-TBG-GFP and AAV-TBG-XBP1s (3 × 10¹¹ GC/mouse). An insulin tolerance test (3 h fast, 0.5 unit per kg) was conducted 10 days after viral injection (n  = 4–5). Bottom, expression levels of XBP1s in the liver. Endo, endogenous. *, nonspecific. The indicated significance is the result of a paired sample t-test between mice injected with AAV-TBG-GFP and AAV-TBG-XBP1s. c Primary hepatocytes were isolated from mice injected with AAV-TBG-GFP and AAV-TBG-XBP1s and subjected to a glucose production assay (0.2 mM cAMP, 4 h) (n  = 3). d Hepa1-6 cells were treated with 400 nM thapsigargin for indicated time, cytosol and nuclear extracts were prepared. e Same amounts of bovine serum albumin and purified FLAG-tagged XBP1s protein were subjected to SDS–polyacrylamide gel electrophoresis (staining with Colloidal blue, lower) or transferred to membrane, then incubated with human P300 protein (1 µg) for 16 h, washed, followed by incubation with antibody against P300 (upper). f 48 h after the addition of adenoviral -GFP and -FLAG-tagged XBP1s, Hepa1-6 cells were subjected to immunofluorescence staining. Scale bar, 10 µm. Data are presented as mean ± s.e.m

Fig. 5

Depletion of P300 improves insulin sensitivity. a Hepa1-6 cells were treated with the indicated concentration of LPS for 48 h then insulin (10 nM, 20 min). (right) Densitometric analysis of pAKT and pGSK3 in cells treated with insulin (n = 3). b, c Hepa1-6 cells were treated with SCR, CBP and P300 adenoviral shRNA for 24 h, then LPS (500 ng ml⁻¹) for 24 h followed by treatment with 10 nM insulin (20 min) b and densitometric analysis of pAKT and pGSK3 in cells treated with insulin c (n = 3). d, e 48 h after the injection of adenoviral shRNAs, mice were subjected to 5 h fasting before the clamp (n = 5/group). Basal glucose production in mice after 5 h fasting d and Glucose disposal rates, glucose infusion rates and hepatic glucose production e during the clamp experiment. f PCA was performed on differentially expressed genes from the liver of mice with SCR or P300 AAV-shRNA injection. g Volcano plot is used to analyse differential expression of hepatic genes, and to show the fold change and statistical significance. SCR, scrambled shRNA. Each bar represents the mean ± s.e.m. *, P < 0.05, Student’s t-test

Fig. 6

Inhibition of P300 acetyltransferase activity improves insulin signaling. a After injection of AAV8-shRNAs for SCR or P300 via jugular vein, mice were fed on an HFD for 3 weeks, insulin tolerance test (6 h fasting, 0.5 unit per kg) was conducted (n = 5). b Densitometric analysis of the pAKT and pGSK3 in cells treated with curcumin as in Supplementary Fig. 5a. c, d Hepa1-6 cells were treated with 20 μM control C37 or inhibitor C646 for 1 h before the addition of 10 nM insulin (~4 h). Densitometric analysis of the pAKT and pGSK3 in hepatocytes treated with insulin for 4 h (n = 3) d. e 16 h after the planting of primary hepatocytes, vehicle (DMSO), C37 and C646 (20 µM) were added to medium during serum starvation. After washing with PBS, above agents and 0.2 mM Bt-cAMP were added in glucose production medium (n = 3). f, g 2 weeks after being fed on an HFD, mice were given either the vehicle or inhibitor C646 (15 nmol g⁻¹) via intraperitoneal injection for 2 weeks, insulin tolerance test (4 h fasting, 0.6 unit kg⁻¹, n = 4–5) and glucose tolerance test (4 h fasting, 1.5 g kg⁻¹, n = 4–5) were conducted. h Liver tissues from HFD-fed mice were collected after 16 days of treatment with vehicle or inhibitor C646. Each lane represents a mouse sample. i Blood glucose levels (6 h fasting) in ob/ob mice treated with vehicle or inhibitor C646 (30 nmol g⁻¹) for 8 days (n = 5). j Hepa1-6 cells were treated with 5 µM C37 or inhibitor C646 for 4 h. Pi3k P110α was immunoprecipitated (n = 3). k, l Hepa1-6 cells were treated with 20 µM C37 or inhibitor C646 for 4 h in DMEM without foetal bovine serum. m Liver lysates from mice fed an HFD (3 weeks) were incubated with P300-specific antibody and protein G beads overnight, washed, then incubated with peptide used to generate P300 antibody for 2 h. Each bar represents the mean ± s.e.m. *, P < 0.05, paired sample t-test between groups

Fig. 7

Identification and characterization of acetylation sites in IRS1 and 2. a–d To mimic the induction of hepatic P300, we overexpressed P300 along with FLAG-tagged-IRS1 or -IRS2 in Hepa1-6 cells, FLAG-tagged-IRS1 and -IRS2 proteins were purified and used to map the acetylation sites by mass spectrometry a, c. 2 μg of plasmids containing IRS1/2-WT or their mutants were transfected into Hepa1-6 cells, cells were harvested 48 h after transfection (n = 3) b, d. e, f Plasmids of IRS1 or 2 mutants with combined KR mutations and IRS1 or 2 WT plasmid were transfected into Hepa1-6 cells as above (n = 3). *, P < 0.05, paired sample t-test between groups transfected with WT and mutated IRS plasmids. g, h Purified Flagged-tagged IRS1/2-WT, and mutant proteins were incubated with 0.2 µg P300 protein for 1 h at 30 °C in the presence and absence of acetyl-CoA. i, j Forty-eight hours after the transfection of 2 µg IRS1-WT or –triKQ mutated plasmid, Hepa1-6 cells were treated with 10 nM insulin for 10 min i. Densitometric analysis of pAKT and pGSK3 in cells treated with insulin (n = 3) j. Data are presented as mean ± s.e.m

Fig. 8

Acetylation of IRS1/2 by P300 decreases their binding to IRβ. a, b After injection of adenoviral IRS1 and its mutants a or adenoviral IRS2 and its mutants b, mice were fed an HFD for 2 weeks, then insulin tolerance tests (6 h fasting, 0.5 unit per kg) were conducted (n = 5). *, P < 0.05, one-way analysis of variance test. c After injection of adenoviral IRS1/2-WT, adenoviral IRS1-triKR/IRS2-dKR, or adenoviral IRS1/2 panKR together with AAV8-TBG-Cre, mice were fed an HFD for 16 days, then insulin tolerance tests (6 h fasting, 0.5 unit per kg) were conducted (n = 5). d Liver tissues were collected from 4-month-old heterozygous (±) lean control and ob/ob mice. Immunoprecipitates were immunoblotted with IRS1/2, IRβ, and anti-acetylated lysine antibodies. e Hepa1-6 cells were treated with 20 µM C37 or inhibitor C646 for 4 h. f Liver tissues from HFD-fed mice were collected after 16 days of treatment with vehicle or inhibitor C646. g Forty-eight hours after the transfection of IRS1-WT, IRS1-panKR, IRS2-WT and IRS2-panKR expression plasmids, anti-FLAG magnetic meads were used to pull down these proteins. h Same amounts of IRS1-WT, IRS1-panKR or IRS2-WT, IRS2-panKR were employed in a SDS–polyacrylamide gel electrophoresis gel, and transferred onto a membrane, after renaturation, membranes were incubated with IRβ, followed by incubation with anti-IRβ antibody. i 36 h after the addition of adenoviral IRS1-WT to overexpress the IRS1 protein, Hepa1-6 cells were treated with inhibitor C646 for 16 h before harvest of cells to block IRS1 acetylation, and this protein was used in the Far-western blot. d, f Each lane represents a mouse sample. Data are presented as mean ± s.e.m