Consequences of lipid droplet coat protein downregulation in liver cells: abnormal lipid droplet metabolism and induction of insulin resistance

Abstract

Objective: Accumulation of intracellular lipid droplets (LDs) in non-adipose tissues is recognized as a strong prognostic factor for the development of insulin resistance in obesity. LDs are coated with perilipin, adipose differentiation-related protein, tail interacting protein of 47 kd (PAT) proteins that are thought to regulate LD turnover by modulating lipolysis. Our hypothesis is that PAT proteins modulate LD metabolism and therefore insulin resistance.

Research design and methods: We used a cell culture model (murine AML12 loaded with oleic acid) and small interfering RNA to directly assess the impact of PAT proteins on LD accumulation, lipid metabolism, and insulin action. PAT proteins associated with excess fat deposited in livers of diet-induced obese (DIO) mice were also measured.

Results: Cells lacking PAT proteins exhibited a dramatic increase in LD size and a decrease in LD number. Further, the lipolytic rate increased by approximately 2- to 2.5-fold in association with increased adipose triglyceride lipase (ATGL) at the LD surface. Downregulation of PAT proteins also produced insulin resistance, as indicated by decreased insulin stimulation of Akt phosphorylation (P < 0.001). Phosphoinositide-dependent kinase-1 and phosphoinositide 3-kinase decreased, and insulin receptor substrate-1 307 phosphorylation increased. Increased lipids in DIO mice livers were accompanied by changes in PAT composition but also increased ATGL, suggesting a relative PAT deficiency.

Conclusions: These data establish an important role for PAT proteins as surfactant at the LD surface, packaging lipids in smaller units and restricting access of lipases and thus preventing insulin resistance. We suggest that a deficiency of PAT proteins relative to the quantity of ectopic fat could contribute to cellular dysfunction in obesity and type 2 diabetes.

FIG. 1.

Identification and downregulation of all PAT proteins (ADFP and Tip47) in AML12 cells. A: Cells were incubated with 400 μmol/l oleic acid for 12 h before staining. Cells were costained with a polyclonal goat anti-ADFP antibody and with a rabbit polyclonal anti-Tip47 antibody and, respectively, alexa fluor 488–or 594–conjugated secondary antibodies. Fluorescent and phase images were generated by an LSM 510 confocal laser microscope. Bar represents 50 μm. B: Coimmunostaining with ADFP and Tip47 of AML12 cells treated for 4 days with siRNA ADFP, siRNA Tip47, both combined, or control. Cells were treated as above. C: Immunoblots of total cellular protein extract from AML12 cells treated with control siRNA (Qiagen) (lane 1); siRNA ADFP (lane 2); siRNA Tip47 (lane 3); or both combined (lane 4). Rabbit polyclonal anti-Tip47, anti-ADFP, and anti–β-actin antibodies were used as loading control. (Please see http://dx.doi.org/10.2337/db07-1383 for a high-quality digital representation of this figure.)

FIG. 2.

Morphological differences in AML12 cells following downregulation of PAT proteins. Cells were stained with bodipy fl568, which preferentially stained neutral lipid. Fluorescent and phase images were generated by a LSM 510 confocal laser microscopy. Bar represents 50 μm. (Please see http://dx.doi.org/10.2337/db07-1383 for a high-quality digital representation of this figure.)

FIG. 3.

Composition of the PAT proteins affects the LD size distribution. Number and diameter of LDs were measured following imaging cells treated with siRNA control (Ctl) (Qiagen) (□), both siRNAs Tip47 and ADFP (Both) (▪), siRNA ADFP (▒), and siRNA Tip47 (). Data are means ± SE from 11 separate experiments for control and ADFP, from three separate experiments for Tip47, and from 27 separate experiments for both. ***P < 0.0001; *P < 0.05 vs. control (t test).

FIG. 4.

Absence of exogenous NEFA reveals differences in accumulating lipids in cells lacking ADFP and Tip47. A: Triglycerides were measured in cells grown in culture media treated with siRNA control (Qiagen) (□), both siRNAs Tip47 and ADFP (▪), siRNA ADFP (▒), and siRNA Tip47 (). B: Triglycerides were measured in cells grown in culture media supplemented overnight with 400 μmol/l of oleic acid in cells treated as described above. Data are means ± SE from three separate experiments and P < 0.001 (one-way ANOVA) for cells lacking both ADFP and Tip47 grown without exogenous NEFA addition.

FIG. 5.

Increased lipolysis in AML12 cells lacking ADFP and Tip47. A: Cells were loaded overnight with two 106 dpm [³H]oleic acid and 400 μmol/l cold oleic acid; the efflux of [³H]oleic acid was tracked over 180 min in the presence of 10 μmol/l triacsin C. Values represent the means ± SE of triplicate determinations of nanomoles released per microgram of protein at time 0, 15, 30, 60, 120, and 180 min. Data are means ± SE from 16 separate experiments for siRNA control and ADFP and Tip47 combined. Data are means ± SE from eight separate experiments for single siRNA transfection with Tip 47 or ADFP. P < 0.0001 (one-way ANOVA) for time points 1, 2, and 3 h efflux and P < 0.002 (one-way ANOVA) for time point 15 min. B: Increase of ATGL and CGI-58 at the LD surface in cells lacking ADFP and Tip47. Immunostaining with ATGL of AML12 cells treated for 4 days with siRNA control or with both combined siRNA ADFP and Tip47. Cells were stained with a polyclonal rabbit ATGL anti-antibody. Alexa fluor 488–conjugated secondary antibodies were used. C: Fat cakes resulting from crude fractionation of AML12 cellular extract treated with siRNA: control (Ctl) (lane 1) or combined ADFP and Tip47 (lane 2). Rabbit polyclonal anti-ATGL and anti–CGI-58 antibodies were used. (Please see http://dx.doi.org/10.2337/db07-1383 for a high-quality digital representation of this figure.)

FIG. 6.

Effect of lack of PAT proteins on Akt. A: Decreased insulin sensitivity and responsiveness in AML12 cells lacking both ADFP and Tip47, measured by Akt phosphorylation (left) and inhibition of lipolytic activity by ATGL siRNA (right), returns the Akt phosphorylation to control levels. AML12 cells were treated with siRNA control 1 (Ctl1) or combined siRNAs ADFP and Tip47 (Both) for 4 days or combined siRNAs ADFP, Tip47, and ATGL (Triple). Akt phosphorylation was determined in the whole-cell lysate by immunoblotting with rabbit polyclonal antibody against Akt or against an Akt phospho–specific antibody. A: left and right: Representative Western blots. B: An insulin dose-response curve was generated by quantifying Western blot–enhanced chemiluminescence signals, with the help of an image software analysis UN-scan-it (Silk Scientific). Each data point is expressed in arbitrary units and represents the average of calculated ratio of the amount of phosphorylated Akt versus the total amount of Akt protein. All experiments were repeated at least three times. Each bar represents means ± SE. □, Ctl1; ▪, Both. *P < 0.02; **P < 0.01.

FIG. 7.

Insulin signaling steps upstream and downstream from Akt are also perturbed in cells lacking both Tip47 and ADFP. AML12 cells were cultured as above. On the day of the experiment, cells were treated with no or 10⁻⁸ mol/l insulin for 15 min prior to being scraped. A: PDK-1 phosphorylation is decreased in AML12 cells lacking ADFP and Tip47. PDK-1 phosphorylation was determined in the whole-cell lysate by immunoblotting with rabbit polyclonal antibody against PDK-1 or against a PDK-1 phospho-specific antibody. B: Cellular extracts were immunoprecipitated with an anti–IRS-1 antibody, and phosphatidylinositol 3-kinase activity was measured. Results are means ± SE arbitrary units. *P < 0.05 vs. basal of each group, PIP3, and phosphatidylinositol 3,4,5,-triphosphate. C: IRS-1 serine 307 phosphorylation was determined in the whole-cell lysate by immunoblotting with rabbit polyclonal antibody against IRS-1 or against an IRS-1 phospho–specific antibody (Upstate Cell Signaling Solutions). D: Activation of PKCθ in AML12 cells lacking both ADFP and Tip47. Activation of PKCθ was determined with phospho-specific antibodies to PKCθ (Thr538). Pellets were obtained by centrifuging the whole-cell lysate at 10,000g. The experiment was repeated three times with consistent results. E: Activation of Foxo-1 in AML12 cells lacking both ADFP and Tip47. F: Activation of GS3α kinase in AML12 cells lacking both ADFP and Tip47. G: Activation of ERK1/2 in AML12 cells lacking both ADFP and Tip47.

FIG. 8.

Content of PAT proteins in DIO mice liver fat cake appears insufficient to protect the LD against the presence of increased ATGL. Fat cakes extracted from three pooled liver tissues from eight mice fed ad libitum or nine DIO mice were loaded in lanes 1–3 or lanes 4–6, respectively. Fat cakes were isolated by ultracentifugation, 10-μl aliquots used for protein determination following cold acetone precipitation overnight, and the protein pellets washed. Equal amounts of total protein were loaded but for ADFP, and 1/20 dilution was performed before loading the samples to avoid overexposed blots.