Very low-density lipoprotein receptor mediates triglyceride-rich lipoprotein-induced oxidative stress and insulin resistance

Abstract

Obesity is associated with excess lipid deposition in nonadipose tissues, leading to increased oxidative stress and insulin resistance. Very low-density lipoprotein receptor (VLDLR), a member of the LDL receptor family, binds and increases the catabolism of triglyceride-rich lipoproteins. Although VLDLR is highly expressed in the heart, its role in obesity-associated oxidative stress and insulin resistance is unclear. Here, we used lean (wild type), genetically obese leptin-deficient (ob/ob), and leptin-VLDLR double-null (ob/ob-VLDLR^−/−) mice to determine the impact of VLDLR deficiency on obesity-induced oxidative stress and insulin resistance in the heart. Although insulin sensitivity and glucose uptake were reduced in the hearts of ob/ob mice, VLDLR expression was upregulated and was associated with increased VLDL uptake and excess lipid deposition. This was accompanied by an upregulation of cardiac NADPH oxidase (Nox) expression and increased production of Nox-dependent superoxides. Silencing the VLDLR in ob/ob mice had reduced VLDL uptake and prevented excess lipid deposition in the heart, in addition to a reduction of superoxide overproduction and the normalization of insulin sensitivity and glucose uptake. In isolated cardiomyocytes, VLDLR deficiency had prevented VLDL-mediated induction of Nox activity and superoxide overproduction while improving insulin sensitivity and glucose uptake. Our findings indicate that VLDLR deficiency prevents excess lipid accumulation and moderates oxidative stress and insulin resistance in the hearts of obese mice. This effect is linked to the active role of VLDLR in VLDL uptake, which triggers a cascade of events leading to increased Nox activity, superoxide overproduction, and insulin resistance.

NEW & NOTEWORTHY Obesity is associated with excess lipid deposition in muscles, which is considered as a leading cause of metabolic dysfunction and oxidative stress. Cellular uptake of lipids is regulated by several membrane receptors, among which is the very low-density lipoprotein receptor (VLDLR). This article provides information on the role of VLDLR in cardiac muscle and how its expression regulates insulin resistance and oxidative stress in the obese mouse model.

INTRODUCTION

Obesity is associated with excess lipid deposition in nonadipose tissues, which instigates metabolic disorders and cell dysfunction. Obesity is also accompanied by a state of oxidative stress, defined as excess production of reactive oxygen species (ROS) relative to antioxidant defense (1). Increased oxidative stress and insulin resistance are characteristics of lipotoxicity widely reported in obese humans and animal models (2, 3). The accumulation of intracellular lipids increases superoxide production, inflammation, and insulin resistance, all of which alter metabolic and cell function. Understanding the mechanisms leading to increased lipid uptake and oxidative stress is important in disease prevention and treatment.

Triglyceride (TG)-rich lipoproteins, including chylomicrons and very low-density lipoproteins (VLDL), are the primary carriers of TGs in the bloodstream, and as such, they represent a major source of fatty acid (FA)-derived energy for the body. The value of this source is mostly important for the heart, where energy needs are predominantly derived from FA oxidation (3, 4). Cardiac lipid metabolism is greatly affected by obesity and diabetes (3–6), conditions under which blood concentration of TG-rich lipoproteins is elevated, leading to increased lipid supply. Because the heart’s ability to store lipids is limited, the increase in lipid influx may lead to metabolic dysfunction, including insulin resistance and cardiac oxidative stress (7–10).

In addition to substrate availability, the heart is equipped with multiple regulatory mechanisms that contribute to maintaining an adequate supply of FAs. These mechanisms include a high lipolytic ability and the presence of multiple cell membrane receptors that facilitate lipid uptake (3, 4). Among these receptors is the very low-density lipoprotein receptor (VLDLR), a member of the low-density receptor family, which regulates the catabolism and uptake of TG-rich lipoprotein (11). Several investigations (12–15) have shown that VLDLR has an impact on the catabolism of TG-rich lipoprotein at multiple levels. First, VLDLR is required for optimal functioning of lipoprotein lipase (LPL) either by increasing LPL and TG-rich lipoprotein interactions at the capillary surface (16) or by serving as a helper for LPL transportation to the luminal surface of vascular endothelial cells (17). Second, VLDLR binds and internalizes TG-rich lipoprotein remnants through specific binding of apolipoprotein-E (18). Through this dual role, VLDLR controls the availability of lipids to the cell both as FA liberated from TG-rich lipoproteins and as remnant particles (13, 15, 19). Although VLDLR is highly expressed in the heart, it is not known if its expression is altered by obesity and whether it participates in obesity-induced lipotoxicity. Accordingly, we have undertaken this study to investigate the impact of VLDLR expression on obesity-induced lipid deposition, insulin resistance, and oxidative stress.

EXPERIMENTAL PROCEDURES

Animals

Wild-type (WT), VLDLR-null (VLDLR^−/−), and ob/+ (all with C57Bl/6 background) mice were obtained from Jackson Laboratory (Bar Harbor, ME). Leptin and VLDLR double-null mice were generated by breeding VLDLR^−/− with b/+ mice, and double heterozygotes were then mated to generate ob/ob-VLDLR^−/− mice. Breeding male and female heterozygous b/+ mice generated homozygotes ob/ob mice. All mice were fed a standard chow diet throughout the study, and experiments were conducted in 5–6-mo-old WT, ob/ob, and ob/ob-VLDLR^−/− mice. WT mice were used as reference lean controls, and ob/ob mice served as control obese for ob/ob-VLDLR^−/− mice. All procedures were approved by the Institutional Animal Care and Use Committee of Vanderbilt University.

Glucose and Insulin Tolerance Tests

These tests were performed as previously described (20), starting with the glucose tolerance test (GTT) and then the insulin tolerance test (ITT) 7 days later.

Echocardiography Measurement

Cardiac performance was assessed at Vanderbilt Core Center with noninvasive transthoracic echocardiography on unanesthetized conscious mice using an echocardiogram with a 35-MHz probe (Acuson, Mountain View, CA) according to the procedure described earlier (21, 22). Pulse Doppler images were collected with the apical four-chamber view to record the mitral Doppler flow spectra. All data and images were saved and analyzed by Advanced Cardiovascular Package Software to determine intraventricular septum (IVS), left ventricular internal diameter (LVID), left ventricular posterior wall (LVDW), percent ejection fraction (%EF), and percent fractional shortening (%FS) under long-axis M-mode. Mitral-valve Doppler was used to establish MV E/A ratio MV decel time.

Tissue Collection, Cardiac Histology, and Lipid Extraction

Overnight fasted mice were anesthetized with an intraperitoneal injection of 100 mg/kg ketamine and 10 mg/kg xylazine, and blood was collected by cardiac puncture. For removal of any residual blood, hearts were perfused through the aorta with saline, and tissues were excised, immediately frozen in liquid nitrogen, and stored at −80°C for later analysis. In addition, samples of myocardium were collected in 4% formaldehyde for histology. Lipids were extracted from tissues with chloroform, methanol, and 0.9% NaCl, as described earlier (20). Lipid extract was dried under a nitrogen stream, and the residue was dissolved in 2-propanol. Lipids in tissue extracts and plasma were measured enzymatically (13), and plasma insulin was assessed with the ELISA procedure (20). Cardiac glycogen and ATP contents were measured in freeze-clamped tissues. Glycogen was assayed enzymatically after hydrolysis into glucose residues, as reported earlier (20), and ATP was measured by enzymatic assay using a commercial kit (Sigma, St Louis, MO), as described by Irie et al. (23). To assess lipid deposition in ventricular myocardium, cardiac sections were fixed with 10% formalin in PBS (pH 7.4) for 1 h and then stained with 0.3% Oil Red O in 60% isopropanol for 30 min, as described elsewhere (13). After being rinsed twice with distilled water, stained sections were viewed and photographed using a microscope with phase-contrast optics and a digital camera (Carl Zeiss).

Plasma Lipoprotein Isolation

Plasma lipoprotein fractions were isolated from pooled plasma by discontinuous density gradient ultracentrifugation, as reported elsewhere (24). Lipoprotein classes were collected according to their density: VLDL (density < 1.006 g/L), low-density lipoprotein (LDL) (1.006 < d < 1.063 g/L), and high-density lipoprotein (1.063 < d < 1.21 g/L). Total cholesterol was assayed by enzymatic assay (Sigma).

Isolation and Treatment of Adult Cardiomyocytes

Cardiomyocytes were isolated using the Langendorff perfusion system, as previously reported (25, 26). Then, myocytes were either seeded onto laminin-coated coverslips for culture or left in suspension and used within 2 h of isolation to investigate metabolite uptake (25).

Assessment of Glucose Uptake

Glucose tissue uptake was measured in vivo using 2-deoxy-[³H]-glucose (PerkinElmer Life Sciences) using the procedure previously detailed (20, 25, 26). Noninsulin-dependent and insulin-dependent glucose uptake was measured with the same tracer (0.5 μCi/mL) in cardiomyocyte suspension preincubated in a medium with or without insulin (100 nM) for 15 min at 37°C (20, 27). Deoxyglucose uptake was found to be linear with time for at least 40 min. To assess the effects of VLDL and VLDLR on insulin sensitivity, another set of experiments was conducted in cardiomyocytes plated in laminin-coated plates. Cells were preincubated for 12 h in a culture medium with or without VLDL (final concentration: 250 µM of TG). Mouse VLDL was separated from pooled plasma of WT mice with ultracentrifugation, as previously described (13). After VLDL preincubation, cells were washed and incubated in a medium with or without insulin (100 nM) for 15 min. Then, glucose uptake was assayed, as described earlier.

Assessment of VLDL Uptake and Incorporation of VLDL-Derived Fatty Acid in Cellular Lipids

Blood clearance and tissue uptake of VLDL were assessed in vivo using double-labeled VLDL, as previously described (13). Mouse VLDL was double labeled by a two-step procedure: first labeling with [³H]-TG VLDL in vivo by intravenous injection of [³H]-palmitate and then in vitro incorporation of [¹⁴]-cholesteryl oleoyl ether ([¹⁴C]-COE), as previously described (13). Kinetics of double-labeled ([³H]-TG, [¹⁴C]-COE) VLDL, blood and tissue collection, and radioactivity measurement were conducted, as previously described (13). Radioactivity recovered in plasma was expressed as a percentage of initial (30-s) counts. Incorporation of VLDL-derived ³H radioactivity in cellular lipids was assessed after extraction and separation of lipid classes with thin-layer chromatography, as previously described (25, 26). Final results were normalized to total proteins determined by the Bradford method (13).

Assessment of Albumin-Bound Fatty Acid Uptake

The uptake of FA was examined using FA analog β-methyl-p-123I-iodophenyl-pentadecanoic acid ¹²⁵I-BMIPP, as previously published (25, 26). Briefly, each mouse was injected with 200 μL of the radioisotope solution of [¹²⁵I]-BMIPP (15 μCi) via the lateral tail vein. Collection of blood and tissues, measurements of radioactivity, and calculation of uptake were performed, as previously described (26).

Measurement of Oxidative Stress Markers in the Heart

Homogenates were prepared by mixing frozen hearts with ice-cold 0.1 M phosphate buffer (pH 7.5) containing 1 mM Na2EDTA and 500 mM butylated hydroxytoluene in acetonitrile to prevent autoxidation. The supernatant, collected from homogenates after centrifugation at 2,000 rpm for 5 min at 4°C, was used to evaluate lipid peroxidation products and hydrogen peroxides with enzymatic tests, as previously described (25).

NADPH-Dependent Superoxide Assay

NADPH-dependent superoxide production was measured in tissue homogenates with the lucigenin chemiluminescent assay using electron donor NADPH and lucigenin, as previously detailed (25). To assess potential sources of superoxide, homogenates were preincubated for 15 min with the following agents before addition of NADPH: nitric oxide synthase inhibitor N^G-nitro-l-arginine methyl ester (l-NAME, 100 μM), xanthine oxidase inhibitor oxypurinol (100 μM), complex I mitochondrial electron chain inhibitor rotenone (20 μM), and NADPH oxidase (Nox) inhibitor VAS2870 (10 nM). In addition, superoxide scavenger superoxide dismutase (SOD, 200 U/mL) was used as a positive control to confirm the specificity of superoxide detection. Superoxide production was expressed as arbitrary light units after subtraction of background reading set as reactions without NADPH.

Quantification of Reactive Oxygen Species in Cardiomyocytes

Estimation of reactive oxygen species (ROS) production was performed using the membrane-permeable fluorescent probes 5-(6)-chloromethyl-2′,7′-dichlorodihydrofluorescein diacetate (CM-H₂DCFDA) (25). After reaction with ROS, H₂DCFDA was converted to a highly fluorescent 2′,7′-dichlorofluorescein (DCF) and used to assess total cell ROS (tROS). To examine the effect of VLDL, cells were cultured for 12 h in low-glucose medium with or without VLDL (final concentration, 250 µM of TG) and then incubated with CM-H₂DCFDA (1 μM) for 30 min at 37°C. Positive control preparations consisted of cells cultured in medium plus hydrogen peroxide (H₂O₂, 40 nmol/mL). After incubation, cells were washed with PBS and fluorescence was quantified using a fluorescence microplate reader with excitation at 485 nm and emission at 530 nm (25). The DCF fluorescent intensity was expressed in arbitrary units (AU) after background subtraction.

Cell Treatment and Fractionation

Cardiomyocytes seeded onto laminin-coated dishes were cultured in a medium with or without VLDL for 12 h, as described earlier. Cells were then homogenized in an ice-cold buffer, and the supernatant collected after centrifugation at 600 g for 5 min was recentrifuged at 12,000 g for 15 min to sediment mitochondria. The resulting supernatant was transferred into another tube and centrifuged for 20 min at 30,000 g at 4°C to separate the membrane fraction in the pellet and the cytosol fraction in the supernatant. Both membrane and cytosol fractions were used for Western blot detection of p47^phox and p67^phox.

Biochemical Analysis

Triglycerides, cholesterol, free fatty acids, and glucose were assayed enzymatically with available commercial kits, as previously described (13, 19). Plasma level of insulin was measured by enzyme-linked immunosorbent assays (ELISA), according to the manufacturer’s recommendation as previously reported (19). Homeostatic model assessment for insulin resistance (HOMA-IR) was calculated by using the mathematical equation from Matthew: [fasting insulin (µIU/mL) × fasting glucose (mg/dL)]/405. Quantitative insulin sensitivity check index (QUICKI) was calculated using the following formula: 1/[log fasting serum insulin (μU/mL) + log fasting plasma glucose (mg/dL)].

Immunoblotting and Protein Determination

Tissue proteins were analyzed by Western blotting, as previously described (13, 19), and the following primary antibodies were applied: VLDLR, protein kinase (Akt), phosphor-(Ser473)-Akt, insulin receptor substrate 1 (IRS1), phospho-(Tyr608)-IRS-1, nicotinamide adenine dinucleotide phosphate oxidase 2 (Nox2), Nox4, p22^phox protein, and β-actin. The specificity and reproducibility of these antibodies were validated before this study (13, 19, 25). The list of antibodies is reported in Supplemental Table S1; data supplements can be accessed here at https://doi.org/10.6084/m9.figshare.25800496). Band intensity for each protein was analyzed by densitometry (ImageJ v.1.37), and corrections were made using β-actin intensity reading.

Gene Expression

Quantitative polymerase chain reaction (qPCR) of selected genes was performed using SYBR Green Supermix with iTaq DNA polymerase on the IQ5 thermocycler and specifically designed and optimized oligonucleotides; the sequences are reported in Supplemental Table S2. Data of qPCR were obtained as threshold cycle (CT) values, and the difference in the CT values (ΔCT) was derived from the specific gene tested and CT of the control gene (β-actin) according to the equation 2^{[CTactin − CT(target gene)]}, as previously described (13).

Statistical Analysis

Averaged values are presented as means ± SD. Statistical significance between groups was performed by a one-way ANOVA test followed by Tukey’s test using GraphPad Prism 4 software (GraphPad Software).

RESULTS

Weight and Metabolic Parameters

As expected, the body weight of ob/ob mice was significantly higher than that of lean WT mice but was significantly lower (P < 0.01) in ob/ob-VLDLR^−/− mice (Table 1). In addition, weights of liver, fat depots, and heart were significantly higher in obese ob/ob mice than lean WT mice. All these parameters, except the liver weight, were significantly reduced in ob/ob-VLDLR^−/− mice. Because of the difference in body weight, the ratio of heart mass-to-body weight was significantly lower in ob/ob mice than WT and ob/ob VLDLR^−/− mice. However, the ratio of heart weight-to-tibia length, a more reliable index, was ∼21% higher in ob/ob mice than WT mice, indicating mild cardiac hypertrophy. This ratio was relatively reduced in ob/ob-VLDLR^−/− mice (Table 1). These differences were detected in both male and female mice, and there was no apparent sex-specific impact of VLDLR deficiency on all tested parameters (Table 1). Within each group, blood parameters were comparable between males and females, and there was no sex-related significant difference (Table 2). When compared with WT mice, the concentrations of total cholesterol in plasma and lipoprotein classes were higher in ob/ob and ob/ob-VLDLR^−/− mice. Because of the small volume of blood in a single mouse, lipoprotein separation and analysis were performed in pooled plasma from male and female mice of the same group. Therefore, cholesterol concentrations in lipoprotein classes were performed in samples pooled form both males and females (Table 2). Knowing that total cholesterol concentration in plasma was similar in males and females of each group, we did not expect sex differences in lipoprotein cholesterol concentration. The concentrations of plasma TG were significantly higher in ob/ob-VLDLR^−/− mice than ob/ob and WT mice, but FFA concentrations were comparable in all groups (Table 2). When compared with WT mice, fasting serum insulin level was significantly higher in ob/ob mice (P < 0.001), but it was reduced by over 60% in ob/ob-VLDLR^−/− mice. To assess insulin sensitivity status, we calculate HOMA-IR and QUICKI, both are reliable indicators of insulin activity. Overall, there was no significant difference between males and females in each group. For both genders, HOMA-IR was over four times higher in ob/ob mice than WT mice (Table 2) but was significantly reduced (P < 0.01) in ob/ob-VLDLR^−/− mice (∼2.8 times lower in both males and females). By opposite to HOMA-IR, low values of the QUICKI index indicate insulin resistance. When compared with control WT mice, the QUICKI value was significantly lower (P < 0.01) in both male and female ob/ob mice but was normalized in ob/ob-VLDLR^−/− mice (Table 2). In agreement with HOMA-IR and QUICKI indices, there was a significant improvement of GTT (Fig. 1A) and ITT (Fig. 1B) in ob/ob-VLDLR^−/− mice compared with ob/ob mice.

Table 1.

Body and organ weights of lean WT, leptin null (ob/ob), and leptin VLDLR double null (ob/ob-VLDLR^−/−) mice

	WT			ob/ob			ob/ob-VLDLR−/−
	Female	Male	All	Female	Male	All	Female	Male	All
Body weight, g	27 ± 5	33 ± 4	29 ± 5	64 ± 10**	69 ± 9**	67 ± 10**	36 ± 7a	39 ± 7a	38 ± 8a
Liver weight, g	2.3 ± 0.6	2.6 ± 0.7	2.4 ± 0.6	3.7 ± 1.0**	3.8 ± 1.0**	3.8 ± 1.0**	3.6 ± 0.9	3.9 ± 0.5	3.7 ± 0.8
Gonadal fat, g	1.82 ± 0.56	2.39 ± 0.52	2.11 ± 0.61	3.37 ± 0.79**	3.74 ± 0.79**	3.55 ± 0.79**	2.24 ± 0.37a	2.48 ± 0.49a	2.36 ± 0.44a
Peri-renal fat, g	0.68 ± 0.21	1.17 ± 0.39	0.92 ± 0.39	1.69 ± 0.57**	1.79 ± 0.47**	1.74 ± 0.51**	0.97 ± 0.39a	1.20 ± 0.28a	1.09 ± 0.35a
Heart weight, mg	120.3 ± 11.5	124.5 ± 0.39	122.4 ± 10.0	143.6 ± 13.5*	147.1 ± 16.7*	145.3 ± 14.9*	126.6 ± 7.9b	130.4 ± 9.9b	128.5 ± 8.9b
Heart-to-tibia ratio, mg/mm	7.9 ± 0.5	7.7 ± 0.7	7.8 ± 0.6	9.6 ± 1.1*	9.5 ± 0.9*	9.6 ± 1.0*	8.1 ± 0.9b	8.1 ± 0.9b	8.1 ± 0.8b

Values are means ± SD; data generated from n = 20 female and male mice. Statistical significance was performed by a one-way ANOVA test followed by Tukey’s test. There was no significant difference between females and males in each group. *P < 0.05 and **P < 0.01, statistical difference between wild-type (WT) and leptin-null (ob/ob) mice. ^aP < 0.01 and ^bP < 0.05, statistical difference between ob/ob and ob/ob-VLDLR^−/− mice.

Table 2.

Plasma parameters and insulin resistance indices of lean WT, ob/ob, and ob/ob-VLDLR^−/− mice

	WT			ob/ob			ob/ob-VLDLR−/−
	Female	Male	All	Female	Male	All	Female	Male	All
Cholesterol, mg/dL	84 ± 15	85 ± 16	84 ± 14	143 ± 34**	139 ± 33**	140 ± 33**	141 ± 21	142 ± 27	142 ± 24
VLDL cholesterol, mg/dL	ND	ND	3.8 ± 0.6	ND	ND	7.2 ± 0.7**	ND	ND	6.2 ± 08
LDL cholesterol, mg/dL	ND	ND	6.2 ± 0.7	ND	ND	37.8 ± 2.8**	ND	ND	37.0 ± 5.9
HDL cholesterol, mg/dL	ND	ND	75 ± 19	ND	ND	96 ± 5*	ND	ND	101 ± 10
Triglycerides, mg/dL	98 ± 24	87 ± 19	92 ± 22	106 ± 25	111 ± 25	109 ± 24	189 ± 32a	186 ± 42a	188 ± 36a
Free fatty acids, mmol/L	1.01 ± 0.26	1.06 ± 0.12	1.03 ± 0.19	0.94 ± 0.18	0.92 ± 0.18	0.92 ± 0.17	1.05 ± 0.19	1.09 ± 0.25	1.07 ± 0.22
Glucose, mg/dL	99 ± 14	96 ± 13	97 ± 13	106 ± 7	106 ± 11	105 ± 9	102 ± 15	98 ± 13	100 ± 13
Insulin, ng/mL	0.28 ± 0.05	0.32 ± 0.08	0.30 ± 0.07	1.12 ± 0.33**	1.28 ± 0.43**	1.20 ± 0.38**	0.49 ± 0.09a	0.49 ± 0.09a	0.45 ± 0.09a
HOMA-IR	1.59 ± 0.32	1.76 ± 0.51	1.68 ± 0.42	6.79 ± 2.11**	7.58 ± 3.01**	7.18 ± 2.56**	2.39 ± 0.45a	2.82 ± 0.44a	2.61 ± 0.49a
QUICKI	0.36 ± 0.04	0.35 ± 0.02	0.36 ± 0.030	0.29 ± 0.01**	0.28 ± 0.01**	0.29 ± 0.01**	0.33 ± 0.01a	0.34 ± 0.02a	0.33 ± 0.02a

Values are means ± SD; data were generated from 10 female and 10 male mice. Plasma was pooled from male and female mice of the same group; lipoprotein classes were separated by gradient ultracentrifugation, and the concentration of cholesterol in lipoprotein classes was performed by enzymatic assay, as described in experimental procedures (n = 5 samples per group). LDL, low-density lipoprotein; HDL, high-density lipoprotein; HOMA-IR, homeostatic model assessment for insulin resistance; QUICKI, quantitative insulin sensitivity check index. Statistical significance was performed by a one-way ANOVA test followed by Tukey’s test. There was no significant difference between females and males in each group. *P < 0.05 and **P < 0.01, statistical difference between wild-type (WT) and leptin-null (ob/ob) mice for females, males, and all mice. ^aP < 0.01 and ^bP < 0.05, statistical difference between ob/ob and leptin very low-density lipoprotein receptor double-null (ob/ob-VLDLR^−/−) mice for females, males, and all mice. Cholesterol concentration in lipoproteins was not determined (ND) for females and males separately.

Figure 1.

Glucose metabolism parameters. A: glucose tolerance test (GTT) in overnight fasted control wild-type (WT), leptin null (ob/ob), and leptin and VLDLR double-null (ob/ob-VLDLR^−/−) mice. B: insulin tolerance test (ITT) in 4-h fasted mice. Results were generated from n = 6 per group and presented as means ± SD. Statistical analysis was performed by a one-way ANOVA test followed by Tukey’s test. ^aP < 0.01 and ^bP < 0.05, statistical difference between ob/ob and ob/ob-VLDLR^−/− mice. C: uptake of ³H-2-deoxyglucose in the heart of mice, n = 5–7 per group. Results are presented as individual data and means ± SD. *P < 0.05, statistical difference between ob/ob and WT mice. ^aP < 0.05, statistical difference between ob/ob and ob/ob-VLDLR^−/− mice. D: representative blots of phosphorylated (p) and total (t) Akt and IRS1. E: uptake of ³H-2-deoxyglucose in isolated cardiomyocytes cultured with or without insulin (n = 6 in duplicates), as described in experimental procedures. #P < 0.05 and ##P < 0.01, statistical difference between insulin-treated and untreated cells of the same genotype (E). *P < 0.01, statistical significance between insulin-treated ob/ob and WT cells. ^aP < 0.01, statistical significance between insulin-treated ob/ob and ob/ob-VLDLR^−/− cells.

VLDLR Deficiency Increased Glucose Uptake

Considering the results of GTT and ITT, we examined the impact of VLDLR deficiency on organ glucose uptake using glucose analog 2-deoxyglucose. Although cardiac glucose uptake was reduced in ob/ob mice compared with WT mice, it was normalized in ob/ob-VLDLR^−/− mice (Fig. 1C). A significant increase in glucose uptake was also noticed in skeletal muscle and adipose tissue but not in liver of ob/ob-VLDLR^−/− mice (Fig. 1C). Improvement of insulin resistance in the hearts of ob/ob-VLDLR^−/− mice was also demonstrated in isolated cardiomyocytes showing a marked increase in insulin-induced phosphorylation of protein kinase Akt and insulin receptor substrate-1 (IRS1) proteins (Fig. 1D) accompanied by a significant enhancement of 2-deoxyglucose uptake (Fig. 1E).

VLDLR Deficiency Prevented Obesity-Induced Excess Lipid Deposition in the Heart

In the second step, we examined the impact of VLDLR expression on cardiac lipid content (Fig. 2A). Cardiac phospholipid contents were similar in all groups. When compared with WT mice, FA and TG contents were considerably increased in ob/ob mice but were normalized in ob/ob-VLDLR^−/−mice (Fig. 2A). These results were also confirmed with lipid-specific Oil Red O staining of cardiac sections (Fig. 2B). Although qualitative, this procedure showed that the heart sections of ob/ob mice exhibited stronger staining than that of WT and ob/ob-VLDLR mice (Fig. 2B). To gain further information about the molecular mechanisms that could explain increased lipid accumulation in ob/ob mice, we assessed the expression of key proteins. Interestingly, the protein level of VLDLR was markedly higher in the hearts of ob/ob mice than WT mice (Fig. 2, C and D). Similarly, mRNA abundance of vldlr, as well as low-density lipoprotein receptor-related protein (lrp), cluster of differentiation 36 (cd36), and fatty acid transport protein 1 (fatp1), was significantly increased in the hearts of ob/ob mice; however, the expression of the three latter genes was modestly reduced in ob/ob-VLDLR^−/− mice (Fig. 2E).

Figure 2.

Lipid contents and protein expression in cardiac muscle. A–E: myocardial lipid contents (n = 8 per group) (A), representative histological sections stained with Oil Red O (black line in each picture represents 50 μm) (B), representative Western blot (C), relative ratio of optic density of very low-density lipoprotein receptor (VLDLR) protein (D), and relative expression of selected genes assayed by quantitative polymerase chain reaction (qPCR) (n = 12) (E). A–D: results are presented as individual data and means ± SD. *P < 0.01, statistical difference between ob/ob and WT mice. ^aP < 0.01, statistical difference between ob/ob and ob/ob-VLDLR^−/− mice.

VLDLR Deficiency Reduced VLDL Clearance and Tissue Uptake but Did Not Alter Albumin-Bound FA Uptake

We next examined the clearance of plasma VLDL and tissue uptake using [³H-TG,¹⁴C-COE]-double labeled VLDL (Fig. 3). Although ³H-label traced the catabolism of VLDL-TG, ¹⁴C-COE provided information on the uptake of VLDL remnants. Plasma clearance of ³H-TG (Fig. 3A) and ¹⁴C-COE (Fig. 3B) was clearly delayed in ob/ob-VLDLR^−/− mice compared with WT and ob/ob mice, resulting in lower uptake of ³H (Fig. 3C) and ¹⁴C (Fig. 3D) in the heart. In addition, uptake of ³H and ¹⁴C was lower in skeletal muscles and adipose tissue but not in liver of ob/ob-VLDLR^−/− mice (Fig. 3, C and D). The distribution of VLDL-derived ³H in myocardial lipids (Fig. 4A) indicated that the proportions of ³H recovered in FA, diglycerides (DG), and TG were higher in ob/ob than WT mice, indicating that part of VLDL-derived FA was channeled toward the esterification pathway, thus increasing cardiac lipid content. Silencing VLDLR in ob/ob mice significantly reduced the proportions of ³H recovered in cardiac FA, DG, and TG (Fig. 4A). In addition to VLDL-derived lipids, albumin-bound FAs circulating in blood represent another source of lipids readily taken by the heart. To assess whether VLDLR deficiency altered albumin-bound FA uptake, we examined the in vivo uptake of nondegradable fatty acid analog ¹²⁵I-BMIPP complexed with albumin. When compared with WT mice, the uptake of ¹²⁵I-BMIPP was significantly higher in the tissues of ob/ob mice (Fig. 4B). There was a modest reduction of ¹²⁵I-BMIPP FA uptake in tissues of ob/ob-VLDLR mice compared with those of ob/ob mice but did not reach statistical significance (P > 0.05).

Figure 3.

Plasma clearance and tissue uptake of double-labeled very low-density lipoprotein (VLDL). A–D: plasma decay of ³H-triglyceride (³H-TG; A) and [¹⁴]-cholesteryl oleoyl ether (¹⁴C-COE; B) and tissue uptake of ³H-TG (C) and ¹⁴C-COE (D) in mice that received intravenous injection of double-labeled VLDL. Experiments were conducted, as described in experimental procedures and results are presented as means ± SD (n = 6 per group). A and B: no significant difference between leptin-deficient (ob/ob) mice and wild-type (WT) mice at all points. ^aP < 0.05, statistical difference between ob/ob and ob/ob-VLDLR^−/− mice. C and D: results are presented as individual data and means ± SD with n = 6 per group. *P < 0.05, statistical difference between ob/ob and WT mice. ^aP < 0.05, statistical difference between ob/ob and ob/ob-VLDLR^−/− mice.

Figure 4.

Incorporation of very low-density lipoprotein-triglyceride (VLDL-TG) in cardiac lipids and organ uptake of albumin-bound [¹²⁵I]-BMIPP. A: distribution of VLDL-derived ³H in cardiac lipids was determined following extraction and separation with thin layer chromatography. Polar lipids, including phospholipids and monoglycerides; FA, fatty acids; DG, diglycerides; TG, triglycerides. Results are presented as individual data and means ± SD with n = 6 per group. *P < 0.05 and **P < 0.01, statistical difference between leptin-deficient (ob/ob) and wild type (WT). ^aP < 0.01 and ^bP < 0.05, statistical difference between ob/ob and ob/ob-VLDLR^−/− mice. B: [¹²⁵I]-BMIPP distribution in organs following injection of [¹²⁵I]-BMIPP complexed with albumin. Tissues were removed 2 h after injection. Uptake is expressed as the percentage of injected dose per gram of tissue. Results are presented as individual data and means ± SD with n = 6 per group. *P < 0.05, statistical difference between ob/ob and WT. ^aP < 0.05, statistical difference between ob/ob and ob/ob-VLDLR^−/− mice.

VLDLR Deficiency Improved Heart Performance

When compared with WT mice, cardiac performance parameters were negatively altered in obese ob/ob mice, as reflected by a significant reduction of fractional shortening and increased left ventricular end-diastolic dimension (Table 3). These parameters were notably improved in ob/ob-VLDLR^−/− mice. In association with this improvement, the contents of glycogen and ATP in the heart were significantly higher in ob/ob-VLDLR^−/− mice compared with ob/ob mice, although they remained ∼31 to 36% lower than WT mice (Table 4).

Table 3.

Cardiac function parameters of conscious mice

	WT			ob/ob			ob/ob-VLDLR−/−
	Female	Male	All	Female	Male	All	Female	Male	All
Heart rate, beats/min	610 ± 26	613 ± 22	612 ± 23	575 ± 37	577 ± 29	576 ± 30	595 ± 36	599 ± 21	597 ± 28
LV mass, mg	92 ± 7	95 ± 8	93 ± 6	130 ± 14*	134 ± 15*	132 ± 15*	101 ± 10a	107 ± 13a	104 ± 12a
FS, %	52 ± 7	53 ± 6	53 ± 6	38 ± 3*	39 ± 3*	39 ± 3*	47 ± 6a	50 ± 4a	49 ± 5a
IVSd, mm	0.77 ± 0.09	0.83 ± 0.13	0.79 ± 0.10	0.88 ± 0.12	0.90 ± 0.11	0.89 ± 0.11	0.91 ± 0.15	0.92 ± 0.12	0.92 ± 0.12
LVIDd, mm	2.86 ± 0.16	2.98 ± 0.18	2.91 ± 0.13	3.80 ± 0.17*	3.84 ± 0.14*	3.82 ± 0.15*	3.01 ± 0.17b	3.02 ± 0.15a	3.02 ± 0.14a
LVPWd, mm	0.82 ± 0.11	0.83 ± 0.15	0.82 ± 0.13	0.67 ± 0.13	0.68 ± 0.14	0.68 ± 0.13	0.73 ± 0.13	0.75 ± 0.17	0.74 ± 0.12
IVSs, mm	1.58 ± 0.10	1.62 ± 0.13	1.60 ± 0.12	1.39 ± 0.11	1.40 ± 0.15	1.39 ± 0.13	1.49 ± 0.15	1.51 ± 0.13	1.50 ± 0.14
LVIDs, mm	1.77 ± 0.37	1.78 ± 0.48	1.77 ± 0.49	2.08 ± 0.39	2.03 ± 0.38	2.05 ± 0.29	1.78 ± 0.41	1.79 ± 0.42	1.78 ± 0.32
LVPWs, mm	1.05 ± 0.14	1.07 ± 0.14	1.06 ± 0.14	0.96 ± 0.13	0.96 ± 0.14	0.96 ± 0.13	1.02 ± 0.10	1.03 ± 0.12	1.03 ± 0.11

Values are means ± SD with n = 12 per group (6 males and 6 females) for wild-type (WT) and n = 14 (7 females and 7 males) for leptin-null (ob/ob) and leptin very low-density lipoprotein receptor double-null (ob/ob-VLDLR^−/−) mice. LV, left ventricular; FS, fractional shortening; IVSd, interventricular septal thickness in diastole; LVIDd, LV end-diastolic dimension; LVPWd, LV posterior wall thickness in diastole; IVSs, interventricular septal thickness in systole; LVIDs, LV end-systolic dimension; LVPWs, LV posterior wall thickness in systole. There was no significant difference between females and males in each group. *P < 0.05, statistical difference between ob/ob and WT for males, females, and all mice. ^aP < 0.05, statistical difference between ob/ob and ob/ob VLDLR^−/− for males, females, and all mice.

Table 4.

Cardiac glycogen and ATP contents of conscious mice

	WT			ob/ob			ob/ob-VLDLR−/−
	Female	Male	All	Female	Male	All	Female	Male	All
ATP, mmol/g weight	6.2 ± 1.8	6.1 ± 1.1	6.1 ± 1.3	2.7 ± 0.7*	3.2 ± 1.1*	3.0 ± 1.0*	5.0 ± 1.0a	5.2 ± 1.2a	5.1 ± 1.1a
Glycogen, mg/mg protein	69 + 10	71 ± 11	70 ± 10	27 ± 9*	28 ± 10*	27 ± 9*	62 ± 13a	62 ± 15a	62 ± 14a

Data are means ± SD; n = 16 per group (8 males and 8 females). There was no significant difference between females and males in each group. *P < 0.05, statistical difference between wild-type (WT) and leptin-null (ob/ob) for females, males, and all mice. ^aP < 0.05, statistical difference between ob/ob and leptin very low-density lipoprotein receptor double-null (ob/ob-VLDLR^−/−) for females, males, and all mice.

VLDLR Deficiency Reduced Obesity-Associated Oxidative Stress

To assess oxidative stress, we examined tissue contents of hydrogen peroxides and lipid peroxides, both of which are used as time-integrated markers of oxidative stress. When compared with WT mice, cardiac contents of hydrogen peroxides (Fig. 5A) and lipid peroxides (Fig. 5B) were strongly increased in ob/ob mice. All these markers were reduced in ob/ob-VLDLR^−/− mice to levels comparable with those in WT mice.

Figure 5.

Oxidative stress markers and superoxide production: myocardial contents of hydrogen peroxides (A) and lipid peroxides (B) (n = 6 per group). Results are presented as individual data and means ± SD. *P < 0.01, statistical difference between leptin-deficient (ob/ob) and wild type (WT). *P < 0.01. ^aP < 0.05, statistical difference between ob/ob and ob/ob-VLDLR^−/− mice. C: NADPH-dependent and superoxide dismutase (SOD)-inhibitable superoxide was measured in cardiac homogenates by the lucigenin chemiluminescent method. Measurements were performed in preparations without inhibitors (basal) or in the presence of SOD used to determine the specificity of measurements, N^G-nitro-l-arginine methyl ester (l-NAME, nitric oxide synthase inhibitor), oxypurinol (oxidase inhibitor), rotenone (complex I mitochondrial electron chain inhibitor), and VAS280 inhibitor of NADPH oxidase (Nox). Measurements were conducted in triplicates from an n = 5 per group. Results are presented as individual data and means ± SD. *P < 0.05, statistical difference between ob/ob and WT. ^aP < 0.05, significance difference between ob/ob and ob/ob-VLDLR^−/− mice. NS, not significantly different.

VLDLR Deficiency Reduced Superoxide Production

Several pathways are capable of generating reactive oxygen species (ROS), some of which are associated with mitochondrial activity (28), and others are linked to enzyme activity (29, 30). To examine the contribution of these pathways, we measured superoxide production in cardiac tissue homogenates using the lucigenin chemiluminescent superoxide assay in the presence of specific inhibitors (Fig. 5C). Measurements without inhibitors (basal) provided total superoxide production that was markedly higher in the hearts of ob/ob mice than those of WT mice but was significantly lower in ob/ob-VLDLR^−/− hearts. The addition of SOD strongly inhibited lucigenin signal in all groups, thereby confirming that the signals measured in the assay were in fact superoxide-induced chemiluminescence. Inhibitors of nitric oxide synthase (l-NAME) and xanthine oxidase (oxypurinol) did not significantly alter superoxide production, while mitochondrial electron transport inhibitor rotenone reduced superoxides in WT (−28% from basal), ob/ob (−39%), and ob/ob-VLDLR^−/− (−19%) hearts. Interestingly, Nox inhibitor (VAS2870) strongly reduced superoxide production in all groups, but the inhibitory effect was more pronounced in WT (−73% from basal) and ob/ob (−79%) mice than ob/ob-VLDLR^−/− mice (−47%). These experiments also show that the effect of Nox inhibitor VAS2870 was more pronounced than that of mitochondrial inhibitor rotenone.

NADPH Oxidase Expression Was Upregulated in Obese Mice and Was Reduced by VLDLR Deficiency

To further examine the role of VLDLR in obesity-induced oxidative stress, we assessed the expression of Nox. Interestingly, protein levels of Nox2, Nox4, and p22^phox were noticeably higher in the hearts of ob/ob mice but were reduced in ob/ob-VLDLR^−/− mice (Fig. 6, A and B). These changes were also reflected at the level of mRNA abundance showing higher expression in ob/ob mice and a significant reduction in ob/ob-VLDLR^−/− mice (Fig. 6C).

Figure 6.

NADPH oxidase (Nox) expression. A–C: representative blots (A), ratio of optic density of proteins (n = 5 per group) (B), and relative expression of cardiac Nox 2, Nox 4, and p22^phox genes (n = 10–12 per group) (C). Results are presented as individual data and means ± SD. *P < 0.05, statistical difference between ob/ob and WT mice. ^aP < 0.05, statistical differences between ob/ob and ob/ob-VLDLR^−/− mice.

VLDLR Mediated VLDL-Induced ROS Production in Cardiomyocytes

Because VLDLs are the natural ligand of VLDLR, we questioned the involvement of VLDL in ROS production. To answer this question, cardiomyocytes were cultured with VLDL, and then, total intracellular ROS (tROS) production was examined with CM-H₂DCFDA probe (Fig. 7A). When compared with untreated cells, chronic exposure to VLDL induced a marked increase of ROS production in cardiomyocytes of ob/ob (+76%) and WT (+62%) mice but significantly less in ob/ob-VLDLR^−/−(−36%) cardiomyocytes.

Figure 7.

Superoxide production in cardiomyocytes culture with or without the presence of very low-density lipoprotein (VLDL). A: production of reactive oxygen species (ROS) measured with 5-(6)-chloromethyl-2′,7′-dichlorodihydrofluorescein diacetate (CM-H₂DCFDA) in cardiomyocytes cultured with or without the presence of VLDL, as described in experimental procedures. Results are presented as individual data and means ± SD. #P < 0.05 and ##P < 0.01, statistical differences between untreated and VLDL-treated cells of the same genotype are indicated with pound symbols. B and C: superoxide production is determined by the lucigenin chemiluminescent method in cardiomyocytes cultured with VLDL alone (B) or in the presence of Nox inhibitor VAS2870 (C). B and C: data were means ± SD of triplicates from two experiments. *P < 0.05, statistical difference between ob/ob and WT mice. ^aP < 0.05, statistical differences between ob/ob mice ob/ob-VLDLR^−/− mice.

VLDLR Deficiency Reduced the Production of Nox-Derived Superoxides

Having shown that VLDLR mediates VLDL-induced ROS production and knowing that VLDLR deficiency reduced Nox expression, we question the role of Nox in VLDL-induced superoxide production. Accordingly, VLDL-induced superoxide production was assessed in cardiomyocytes cultured with or without membrane permeable Nox inhibitor VA285 using the lucigenin chemiluminescence procedure and NADPH as a stimulator of superoxide production and indicator of Nox activity (Fig. 7, B and C). Although NADPH induced superoxide production, the addition of SOD strongly inhibited lucigenin signal in all groups, confirming that the signal measured was superoxide-induced chemiluminescence. Chronic treatment with VLDL induced a strong increase in superoxide production in ob/ob cardiomyocytes but relatively less in WT and ob/ob-VLDLR^−/− cardiomyocytes (Fig. 7B). Treatment with Nox inhibitor VAS2870 reduced VLDL-induced superoxide production in all groups, but the strongest inhibitory effect was in ob/ob cardiomyocytes (∼3-fold reduction compared with VLDL alone), followed by WT and ob/ob-VLDLR^−/− cardiomyocytes (Fig. 7C). These results indicate that higher production of superoxides in ob/ob cardiomyocytes is mostly related to Nox activity, in agreement with protein upregulation shown in Fig. 6.

VLDLR Mediated VLDL-Induced Translocation of p47^phox and p67^phox to the Cell Membrane

The translocation of p47^phox and p67^phox subunits from the cytosol to the cell membrane is required to form an active Nox2 enzymatic complex (31). Therefore, the ratio of the cell membrane to the cytosol is indicative of protein complex activation. In the absence of VLDL, the amounts of p47^phox and p67^phox in the cell membrane fraction were low in both ob/ob and ob/ob-VLDLR^−/− cardiomyocytes (Fig. 8, A and B). VLDL overload, however, noticeably increased the amounts of p47^phox and p67^phox in plasma membranes of ob/ob cardiomyocytes but was clearly less effective in plasma membranes of ob/ob-VLDLR^−/− cardiomyocytes (Fig. 8, A and B). Consequently, the ratio of membrane to cytosol of these proteins was markedly lower in ob/ob-VLDLR^−/− cardiomyocytes (Fig. 8, C and D).

Figure 8.

Effects of very low-density lipoprotein (VLDL) on the distribution of NADPH oxide (Nox) subunits in membrane and cytosol fractions. Representative blots (A and B) and means of membrane-to-cytosol ratios of optic density (C and D) of p47^phox (A) and p67^phox (B) proteins in cardiomyocyte culture with or without VLDL. Data were generated from two experiments performed in triplicates. C and D: results are presented as individual data and means ± SD. #P < 0.05 and ##P < 0.01, statistical difference between untreated and VLDL-treated cells of the same genotype. ^aP < 0.01, statistical difference between VLDL-treated cells of leptin-deficient (ob/ob) and ob/ob-VLDLR^−/− mice.

VLDLR Deficiency Prevents VLDL-Induced Insulin Resistance in Cardiomyocytes

To assess the effects of VLDL and VLDLR expression on insulin sensitivity, we examined insulin-induced glucose uptake in ob/ob and ob/ob-VLDLR^−/− cardiomyocytes exposed to VLDL (Fig. 9). Basal glucose uptake (without insulin) was low in both ob/ob and ob/ob VLDLR^−/− cardiomyocytes and was not significantly altered by VLDL exposure (Fig. 9A). Insulin-stimulated glucose uptake was markedly smaller in ob/ob cardiomyocytes (1.2-fold increase compared with ob/ob nonstimulated cells) than ob/ob-VLDLR^−/− cardiomyocytes (2.7-fold increase compared with ob/ob-VLDLR^−/− nonstimulated cells). Adding VLDL significantly reduced insulin-stimulated glucose uptake in ob/ob cells (−63%, P < 0.05) but had a minor effect in ob/ob-VLDL^−/− cells (−21%, nonsignificant) (Fig. 9A). In agreement with this, although pAkt was low in both ob/ob and ob/ob-VLDLR^−/− unstimulated cells (Fig. 9B), insulin-induced Akt phosphorylation was markedly higher in ob/ob-VLDLR^−/− cardiomyocytes (+4-fold compared with no insulin stimulation) than ob/ob cardiomyocytes (+1.4-fold compared with no insulin stimulation). Exposure to VLDL significantly reduced insulin-induced Akt phosphorylation in ob/ob cardiomyocytes (−60% compared with insulin alone) and had a minor effect in ob/ob-VLDLR cardiomyocytes (−12% compared with insulin alone) (Fig. 9, C and D).

Figure 9.

Effects of very low-density lipoprotein (VLDL) on insulin-induced glucose uptake and Akt phosphorylation. A–C: ³H-2-deoxyglucose uptake (A) and representative blots (B and C) of phosphorylated (p) and total (t) Akt in isolated cardiomyocytes cultured with or without insulin and VLDL, as described in experimental procedures. D: ratio of phosphorylated Akt to total Akt (pAkt to tAkt) in isolated cardiomyocytes. A and D: results are presented as individual data and means ± SD. #P < 0.05, statistical difference between insulin-treated and (VLDL + insulin)-treated cells of the same genotype. ^aP < 0.01, statistical significance between insulin-treated leptin-deficient (ob/ob) and ob/ob-VLDLR^−/− cells. ^AP < 0.05, statistical significance between (VLDL + insulin)-treated ob/ob cells and (VLDL + insulin)-treated ob/ob-VLDLR^−/− cells.

DISCUSSION

In this study, we asked the question of whether VLDLR expression had an impact on obesity-associated oxidative stress and insulin resistance. The results indicate that upregulation of VLDLR expression in the hearts of obese mice is associated with excess lipid deposition, insulin resistance, and oxidative stress, whereas silencing VLDLR expression markedly decreased cardiac lipid deposition, improved insulin sensitivity, and reduced superoxide overproduction. The role of VLDLR was further validated in isolated cardiomyocytes, which demonstrated that VLDLR regulated Nox activity and mediated VLDL-induced superoxide overproduction and insulin resistance.

The process of ectopic lipid deposition has been linked to increased availability of lipids in blood as well as upregulation of several genes required for cellular lipid uptake and processing (2–4). We now provide new evidence demonstrating that VLDLR is another potential mediator of obesity-induced lipid deposition in the heart. VLDLR regulation of lipid deposition is linked to VLDL lipids and not to albumin-bound FA, in that knockdown of VLDLR expression reduced VLDL uptake and prevented lipid deposition but did not significantly alter the uptake of albumin-bound FAs. Although the expression of LRP, another member of the LDLR family that can mediate TG-rich lipoprotein uptake (32), was not significantly altered by VLDLR deficiency, it did not fully compensate for the loss of VLDLR expression. All together, these results confer to VLDLR a prominent role in mediating VLDL-induced lipid deposition.

In line with previous reports (2, 13, 33), ob/ob mice displayed excess fat deposition and a marked increase in body weight compared with control WT mice. Silencing VLDLR in ob/ob mice markedly reduced body weight and fat mass but did not significantly decrease liver weight and its fat content (Table 1). In addition, silencing VLDLR significantly reduced VLDL uptake in peripheral tissues, including adipose tissue, heart, and skeletal muscles, but did not change the already high influx of lipids into the liver, including high uptake of VLDL lipids (Fig. 3) and FFA (Fig. 4). Therefore, it is possible that VLDLR deficiency while disrupting lipid uptake in peripheral tissues of ob/ob-VLDLR^−/− mice, which may be the underlying mechanism for reduced body weight and fat mass, neither reduced lipid deposition nor improved insulin resistance in the liver. These findings are consistent with the fact that VLDLR is expressed mostly in peripheral tissues, including heart, skeletal muscles, and adipose tissue. Moreover, several proteins known to regulate lipid fluxes into the liver, including LRP, CD36, and FATP (2, 13), may account for hepatic lipid uptake in the absence of VLDLR. Another important finding is that VLDLR deficiency did not improve hepatic insulin resistance, while significantly increasing insulin sensitivity in peripheral tissues, including heart, skeletal muscles, and adipose tissue. Because lipid deposition in tissues regulates insulin activity, it is reasonable to expect that high lipid content in the liver of VLDLR-ob/ob mice is linked to hepatic insulin resistance, as reflected with the low uptake of deoxy-glucose (Fig. 3). These findings indicate that silencing VLDLR is associated with a tissue-specific improvement of insulin sensitivity, which may be linked to lipid fluxes.

Excess lipid accumulation in the cell induces multiple metabolic dysfunctions, including insulin resistance, which is a major factor in the pathogenesis of type 2 diabetes associated with obesity. Previous investigations have shown that the use of TG-rich lipoproteins is increased in diabetes and insulin resistance state (5, 34). Our results are in line with these studies and further highlight the role of VLDLR in obesity-induced insulin resistance. In obese mice, cardiac upregulation of VLDLR expression is associated with increased uptake of VLDL, creating a mismatch between the supply and the use of FAs, leading eventually to excess cardiac lipid accumulation and insulin resistance. By contrast, VLDLR deficiency reduced TG-rich lipoprotein uptake and induced a marked improvement of insulin resistance and glucose uptake. In agreement with this, exposure of cardiomyocytes to VLDL induced significant reduction of insulin-stimulated glucose uptake in ob/ob cells but had clearly less effect on ob/ob-VLDLR^−/− cells (Fig. 9). Currently, there is no evidence to directly implicate VLDLR in insulin signaling pathway, but one possible mechanism of action could be related to the role of VLDLR in controlling the rate of TG-rich lipoprotein uptake and intracellular lipid deposition. The causative effect of lipids in the development of insulin resistance has been widely reported (35, 36) and was linked to increased concentration of intracellular DGs and FAs, among other lipids and lipid derivatives (37, 38). It is thus possible that downregulation or deficiency of VLDLR improves obesity-associated insulin resistance indirectly through the reduction of toxic lipids.

Although the function of VLDLR in peripheral tissues, to facilitate TG-rich lipoprotein catabolism and uptake, is well described (15–18), its role in obesity-induced oxidative stress remains poorly understood. The present study provides new findings that establish a link between VLDLR expression and Nox-dependent superoxide production. First, the upregulation of VLDLR expression in obese mice was associated with increased expression of Nox and oxidative stress. Second, VLDLR deficiency reduced VLDL uptake and alleviated Nox-dependent superoxide production. In addition and in agreement with previous findings in endothelial (39) and beta cells (40), exposure of cardiomyocytes to VLDL induced Nox-dependent superoxide production. Silencing VLDLR, however, reduced VLDL uptake and abrogated VLDL-induced superoxide production. These results demonstrate that VLDLR mediates VLDL-induced superoxide production and highlight the evidence of a regulatory mechanism between VLDLR expression and Nox-dependent superoxide production. This mechanism may contribute to the increase in oxidative stress associated with obesity.

Increasing evidence indicates that obesity is associated with both excess lipid deposition and cardiac dysfunction (41, 42), but whether this association is a direct cause-effect relationship is still debated. In line with previous reports (33, 43, 44), cardiac steatosis in ob/ob mice is associated with mild cardiac hypertrophy and dysfunction. It has been reported that obesity may negatively alter cardiac performance (28, 33, 44). Therefore, we do not exclude the possibility that weight reduction in ob/obVLDLR^−/− mice could contribute to the improvement of cardiac function parameters (Table 3). Although the impact of VLDLR expression on cardiac function is not evident in lean mice, it seems that VLDLR deficiency preserves cardiac function under physiological and metabolic stress associated with obesity, as revealed in this study, or under hypoxic ischemia, as reported in a previous study (14). The question of whether this protective effect is linked to the improvement of cardiac lipotoxicity is not clear and requires further investigation. Of note, reduction of cardiac performance in obese mice has been associated with a depletion of energetic reserves (43, 44). This is possibly related to the switch of substrate use associated with insulin resistance (44). It is thus conceivable that improvement of insulin sensitivity in ob/ob-VLDLR^−/− mice restores energy reserves and improves cardiac function. Cardiac hypertrophy and dysfunction is a complex disease involving multiple factors that may extend far beyond lipid infiltration (8). Interestingly, in this study, the reduction of cardiac performance in ob/ob mice is associated with the induction of Nox expression. Of note, target overexpression of Nox2 or Nox4 to cardiomyocytes induced cardiac hypertrophy and dysfunction (45–47). Altogether, these findings suggest the existence of possible regulatory mechanisms between Nox expression and cardiac function that may also be linked to lipotoxicity.

Nox is a family of membrane-bound enzyme complexes that transfer electrons from NADPH to oxygen, generating superoxides (31, 48). Multiple isoforms have been identified, among which are Nox2 and Nox4, the two most abundant isoforms in myocytes (48). Activation of Nox2 and Nox4 requires the assembly of a multiple-unit complex that includes membrane-bound p22^phox. In addition, and in contrast to Nox4, Nox2 activation necessitates the association of cytosolic activators p47^phox and p67^phox to the complex, a mechanism that requires relocation and phosphorylation of these subunits to the membrane (31, 48). Therefore, Nox activation is under the control of two mechanisms: chronic upregulation of expression and acute increase in enzymatic complex formation secondary to the translocation and assembly of regulatory subunits (31). In the present study, upregulation of Nox2, Nox4, and p22^phox in obese mice was associated with increased activity, a result that is in line with previous studies in mice with genetic or diet-induced obesity (49–51). Moreover, our study provides new findings, in that VLDLR deficiency normalized Nox2 and Nox4 expression, suggesting that pretranslational regulation is involved. Furthermore, VLDL increased membrane-associated p47^phox and p67^phox in ob/ob cardiomyocytes but did not significantly alter the membrane-cytosol distribution of these proteins in ob/ob-VLDLR^−/− cardiomyocytes. These results indicate that VLDL uptake may also enhance Nox2 activity acutely by increasing the translocation and assembly of required subunits and that VLDLR is required for this process. Of interest to these results, it has been reported that FAs induce Nox2 activity (52) and increase acute translocation of cytosolic subunits p47^phox and p67^phox to the cell membrane (25, 53). Because VLDLR deficiency suppresses VLDL uptake, and hence diminishes intracellular FAs, it may also delay the translocation of cytosolic subunits and thus reduce the acute formation of the Nox2 complex.

In conclusion, the current study provides evidence that VLDLR expression promotes cardiac lipotoxicity in obese mice. In addition, experiments in cardiomyocytes demonstrate that VLDLR-mediated VLDL uptake triggers a cascade of events leading to insulin resistance and induction of Nox-dependent superoxide production. Such a mechanism highlights the role of VLDLR in the regulation of Nox activity and obesity-induced oxidative stress.

DATA AVAILABILITY

Data will be made available upon reasonable request.

SUPPLEMENTAL MATERIAL

Supplemental Figs. S1–S7 and Supplemental Tables S1 and S2: https://doi.org/10.6084/m9.figshare.25800496.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

T.H. conceived and designed research; T.H. and M.G. performed experiments; T.H., M.G., T.F., and A.M. analyzed data; T.H., M.G., T.F., and A.M. interpreted results of experiments; T.H., M.G., and A.M. prepared figures; T.H. drafted manuscript; T.H., M.G., T.F., and A.M. edited and revised manuscript; T.H., M.G., T.F., and A.M. approved final version of manuscript.