Protection From Obesity and Insulin Resistance in Mice Overexpressing Human Apolipoprotein C1

Homozygous APOC1 transgenic mice of line 11/1 were previously generated in our laboratory (3). Heterozygous ob/− mice obtained from the Jackson Laboratories were intercrossed with the APOC1 transgenic mice to obtain APOC1 homozygous overexpressors on a homozygous ob/ob background (ob/ob APOC1). Ob/ob and wild-type littermates were used as controls in the respective experiments. All mice were males and housed individually under standard conditions with free access to water and food. Food intake was monitored individually in metabolic cages. Core body temperature was measured using a rectal thermometer (YSI Tele-thermometer). All experiments were performed after 5 h of fasting at 1:00 p.m., with food withdrawn at 8:00 a.m.

Absorption of the major fatty acids (palmitic [C16:0], oleic [C18:1], and linoleic [C18:2] acids) was determined by subtracting the amount of the respective fatty acid excreted in feces of 72 h from the amount of dietary fatty acid ingested during this 72-h period and expressed as a percentage of the amount of fatty acid ingested during 72 h (8).

Glucose tolerance tests were performed on conscious mice that received an intraperitoneal injection of 10% glucose in sterile saline at a dose of 2 g/kg. All blood samples were centrifuged at 4°C, and the separated plasma was measured for glucose and insulin as described below.

Reproductive white adipose tissue was obtained mid-distally from the testis from mice (3 months of age) that had been maintained on a standard diet. After formalin fixation, adipose tissue was cut in sections (3 μm) and stained with hematoxylin phloxin saffron. For the quantification of adipocyte size, five sectional areas of adipose tissues per mouse were analyzed by counting 27–56 cells per section using Leica Qwin image analysis software.

BMIPP (provided by Dr. F. F. Knapp) was radioiodinated by the iodide exchange method as previously described (7). Purification was done over a Sep-Pak Light C18 cartridge (Waters), and the specific activity of ¹²⁵I-BMIPP was typically ∼0.5 mCi/mg. The radiochemical purity was determined by thin-layer chromatography using a silica gel G plate developed in 6% (vol/vol) methanol in chloroform by comparison with a BMIPP standard and was always found to be >95%. BMIPP was dissolved in 25 μl of warm absolute ethanol and added dropwise to a stirred solution of 0.9% NaCl containing 6% bovine serum albumin (BSA; essentially FFA free) at 40°C. The solution was sterile-filtered (0.22 μm, Millipore) before injection.

To investigate tissue distribution of BMIPP, we performed experiments as previously described by Coburn et al. (9). In short, fasted mice received an intravenous injection via the tail vein of 5 μCi of ¹²⁵I-BMIPP in a 200-μl 0.9% NaCl solution containing 6% BSA (essentially FFA free). Mice were bled from the retro-orbital plexus 2 h after BMIPP injection and killed by cervical dislocation. The respective tissues were removed, washed with ice-cold saline, and blotted dry. Subsequently, tissues were weighed and counted in a NaI auto-γ counter. A sample (10 μl) of the injected BMIPP solution was counted to determine the total injected dose. Total recovery of the BMIPP was >60% after estimation of the total weights of the organs collected. The rest of injected radiolabel probably resides in organs that were not collected (e.g., bone, skin, brain).

In all experiments, blood was collected from the tail vein in chilled paraoxonized capillary tubes as described previously (10). The tubes were placed in ice and centrifuged at 4°C, and the obtained plasma was immediately assayed for cholesterol, triglycerides (without free glycerol), and FFA using enzymatically available kits #236691 (Boehringer Mannheim, Mannheim, Germany), #337-B (GPO-trinder kit; Sigma, St. Louis, MO), and a NEFA-C kit (Wako Chemicals, Neuss, Germany), respectively. Levels of plasma glucose were determined using an enzymatically available kit (#315-500; Sigma) and plasma insulin levels using a radioimmunoassay kit (Linco Research, St. Charles, MO).

The Mann-Whitney nonparametric test for two independent samples (11) was used to define differences between the groups of mice. The criterion for significance was set at P < 0.05. All data are presented as the mean ± SD.

Because previous studies revealed a marked reduction in the insulin-mediated whole-body FFA uptake upon APOC1 overexpression (6), we first aimed to identify the peripheral tissues showing impaired FFA uptake upon APOC1 overexpression. Therefore, we compared the biodistribution of the slowly oxidized fatty acid analog ¹²⁵I-BMIPP between APOC1 transgenic mice and wild-type controls. The usefulness of BMIPP as a metabolic tracer for FFA uptake in vivo recently was clearly demonstrated in CD36 knockout mice (9). The biodistribution data for BMIPP are shown in Fig. 1. Of all tissues examined, significant impairment of BMIPP uptake was observed only in the different fat depots, i.e., renal, mesenteric, and subcutaneous adipose tissue of APOC1 transgenic mice as compared with wild-type mice. Reproductive adipose tissue did not show differences between the two mouse genotypes. Ob/ob APOC1 overexpressing mice showed decreased BMIPP uptake in different fat pads, i.e., reproductive, renal, mesenteric, and subcutaneous adipose tissue, and higher BMIPP remained in plasma, kidney, and lung compared with ob/ob mice (Fig. 1B).

To assess the potential effects of APOC1 overexpression on body weight regulation, we measured total body weight, adipose tissue weight, and adipocyte size in APOC1 transgenic mice both on a wild-type and ob/ob background at 3 months of age. As shown in Fig. 2A, APOC1 transgenic mice have a slightly but significantly lower body weight compared with their wild-type littermates; values are depicted in Table 1. Remarkably, APOC1 overexpression completely prevented the development of obesity on an ob/ob background (Fig. 2A). Total fat pad weights were significantly reduced by APOC1 overexpression (Fig. 2B). The reduction in fat pad weights in APOC1 transgenic mice were reflected by a decrease in reproductive, renal, mesenteric, and subcutaneous fat (Fig. 2C). Histological analysis of reproductive fat pads revealed that adipocytes from APOC1 transgenic mice were significantly smaller than those from control mice, both on the wild-type background (2,516 ± 254 vs. 3,696 ± 390 μm²; P < 0.05) and on the ob/ob background (7,689 ± 624 vs. 15,295 ± 1,289 μm²; P < 0.05).

To determine whether the reduction in body weight upon APOC1 overexpression was caused by severe growth retardation, decreased food intake, and/or lipid malabsorption, we measured body length, daily food intake, and lipid absorption of these animals. Table 1 shows that the nasal-anal length is significantly decreased upon APOC1 overexpression, on both the wild-type and ob/ob background, indicating that growth retardation may partly contribute to the lower body weights in APOC1 transgenic mice. Furthermore, no significant differences were found between APOC1 transgenic mice and controls with respect to food intake or total lipid absorption, on both the wild-type and ob/ob backgrounds (Table 1). In addition, on the wild-type background, core body temperatures were slightly decreased in APOC1 transgenic mice compared with controls, whereas similar values for body temperature were found in ob/ob APOC1 mice compared to ob/ob only mice (Table 1).

We next investigated whether decreased obesity with APOC1 overexpression would affect plasma lipid levels and insulin sensitivity. Plasma cholesterol, triglycerides, and FFA concentrations were significantly increased in APOC1 transgenic mice, on both the wild-type and ob/ob backgrounds (Table 2). On the wild-type background, basal plasma glucose levels were slightly reduced in APOC1 transgenic mice compared with controls, whereas plasma insulin levels were similar. In contrast, ob/ob APOC1 mice, which showed complete protection from obesity, exhibited extremely low plasma glucose levels and normal plasma insulin levels, comparable to wild-type ob/ob mice (Table 2). These results suggest that ob/ob APOC1 mice remain insulin-sensitive, whereas ob/ob mice are indeed insulin-resistant. This is sustained by an intraperitoneal glucose tolerance test (Fig. 3). Although little effect was observed on a normal wild-type background (Fig. 3A), ob/ob APOC1 mice were much more efficient in the clearance of an intraperitoneally injected bolus of glucose than ob/ob mice (Fig. 3B), demonstrating increased insulin sensitivity upon APOC1 overexpression.

Insulin resistance, as part of the “insulin resistance syndrome,” is often associated with hyperlipidemia (12). However, previous studies have shown that primary hyperlipidemia in human APOC1 transgenic mice is not necessarily associated with insulin resistance (5). Indeed, hyperinsulinemic-euglycemic clamp studies have shown that whole-body glucose turnover in APOC1 transgenic mice is enhanced compared with wild-type mice (6). Furthermore, whole-body FFA uptake was found to be decreased upon APOC1 overexpression (6). In the present study, the decreased tissue FFA uptake upon APOC1 expression, on either wild-type or ob/ob background, was confined to the various depots of white adipose tissue. This was shown by comparing the biodistribution of BMIPP in tissues from APOC1 transgenic mice and controls. BMIPP uptake was reduced in various white fat depots of APOC1 transgenic mice, whereas BMIPP uptake was normal in liver, heart, and skeletal muscle (Fig. 1).

To further explore a potential direct link between apoC1 and tissue FFA uptake, we next investigated whether APOC1 overexpression can modulate the development of obesity and insulin resistance in genetically obese mice. On a normal wild-type background, APOC1 transgenic mice displayed a slightly reduced body weight compared with controls. Strikingly, on the genetically obese ob/ob background, APOC1 overexpression fully protected mice from the development of obesity as reflected by a marked reduction in body weight, adipose tissue mass, and adipocyte size. APOC1 transgenic mice were also protected from obesity as induced by dietary intervention (high-fat diet for a period of 20 weeks (13); data not shown). Although body weights of APOC1 transgenic ob/ob mice were similar to APOC1 transgenic mice, fat pad weight added up to ∼15% in APOC1 ob/ob mice, whereas it amounts to only ∼3% in APOC1 transgenic mice. These data indicate a lower lean body mass in APOC1 ob/ob mice (∼15% less). Similarly, in ob/ob mice, lean body mass decreased by 20% as compared with control wild-type mice. It has been described that leptin, which is absent in ob/ob mice, exerts stimulatory effects on growth and development of different tissues, e.g., skeletal muscle (14,15).

In addition, APOC1 transgenic ob/ob mice were fully protected from insulin resistance. This was reflected by lower plasma glucose and insulin levels and improved glucose tolerance (Fig. 3B) in these animals compared with their ob/ob littermates. In contrast to the observations on the ob/ob background, no clear differences were found in the glucose tolerance test between APOC1 transgenic mice and controls on the wild-type background (Fig. 3A). Thus, it can be argued that the protection from insulin resistance with APOC1 overexpression is secondary to the protection from obesity. However, basal plasma glucose levels were significantly reduced in APOC1 transgenic mice compared with wild-type mice (Table 2). Furthermore, in previous studies, we observed that whole-body glucose uptake in hyperinsulinemic-euglycemic conditions is markedly increased in APOC1 transgenic mice on the wild-type background (6). The decreased fatty acid turnover found in these APOC1 transgenic mice (6) suggests that the increased whole-body glucose uptake compensates for the decreased whole-body free fatty acid uptake. Taken together, these data suggest that the improved glucose tolerance observed in APOC1 transgenic mice may be, at least partly, independent of the inhibitory effect of APOC1 overexpression on adipose tissue development and despite higher plasma FFA concentrations. We cannot exclude that altered insulin secretion in APOC1 transgenic mice also contributes to the improved glucose tolerance measured during the glucose tolerance test.

The observations that APOC1 transgenic mice exhibit 1) a marked decrease in the insulin-mediated whole-body FFA uptake (6), 2) elevated basal plasma FFA levels (Table 2), and 3) a marked reduction in adipose tissue BMIPP uptake (Fig. 1) suggest that impaired adipocyte FFA uptake may underlie the protection from obesity and insulin resistance induced by APOC1 overexpression. We explored alternative explanations pertaining to food intake, lipid absorption, and energy expenditure. No significant differences were found in food intake or in total lipid absorption between APOC1 transgenics and controls, on both the wild-type and ob/ob backgrounds (Table 1). Although not statistically significant, APOC1 ob/ob mice tended to eat 20% less food than ob/ob mice. This might, at least in part, contribute to the lower body weights found in these APOC1 ob/ob mice. Previously, it was shown that obesity in ob/ob mice results partly from decreased energy expenditure, reflected by a reduction in core body temperature (16). Although body temperatures in ob/ob APOC1 mice were similar to those in ob/ob mice (Table 1), lean ob/ob APOC1 mice exhibited increased cage activity compared with ob/ob only mice (gross observations, data not shown). These latter observations suggest that APOC1 transgenic mice may have increased energy expenditure, which might protect against development of obesity. This phenomenon was described by Smith et al. (17), who showed that DGAT (diacylglycerol acyl-transferase) knockout animals had increased activity and were also protected from obesity. After an overnight fast, lower core body temperatures were observed in APOC1 transgenic mice (data not shown), suggesting decreased energy expenditure as a result of low adipose tissue depots reserves. However, additional analyses are required to investigate whether increased energy expenditure contributes to the protection from obesity by APOC1 overexpression.

It has been proposed that FFA play a central role in the link among obesity, insulin resistance, and hyperlipidemia (18). Most insulin-resistant obese individuals exhibit elevated levels of FFA and triglycerides (19). Furthermore, elevated plasma levels of FFA may inhibit insulin-stimulated peripheral glucose uptake (20) and enhance VLDL-triglycerides production (21). However, in the present study, it was shown that primary hyperlipidemia and elevated levels of FFA in APOC1 transgenic mice are associated with neither insulin resistance for glucose metabolism nor hyperinsulinemia. In contrast, our study provides evidence that the entry of FFA into adipose tissues can be effectively inhibited by apoC1 and that this process may underlie the protection against genetically induced obesity and insulin resistance. Previously, Reaven et al. (22) investigated whether hypertriglyceridemia induced by overexpression of the human APOC3 gene is associated with insulin resistance for glucose metabolism in mice. These APOC3 transgenic mice showed a normal whole-body insulin-mediated glucose disposal. Because hyperlipidemia associated with obesity and type 2 diabetes is usually caused by hepatic overproduction of VLDL (18) and apoC3 primarily impairs VLDL lipolysis (23), it was concluded that insulin resistance will develop only in cases in which hypertriglyceridemia is primarily caused by enhanced VLDL secretion. In the current study, we used APOC1 ob/ob mice that are homozygous for APOC1. In these mice, the plasma triglycerides levels are too high to measure the hepatic triglycerides production after Triton WR1339 injection in a reliable way. We have measured VLDL production in heterozygous APOC1 ob/ob mice. These mice also have elevated plasma FFA levels and do indeed display increased hepatic VLDL-triglycerides production (5.1 ± 0.6 vs. 2.9 ± 0.5 μmol · min⁻¹ · kg⁻¹ in heterozygous APOC1 ob/ob and ob/ob mice, respectively). It is highly likely that the current homozygous APOC1 ob/ob mice also exhibit an enhanced VLDL production rate.

The molecular mechanisms by which APOC1 overexpression lead to impaired FFA uptake remain unknown at the present time. A schematic representation of the various metabolic steps in which apoC1 is involved is presented in Fig. 4. In APOC1-overexpressing mice, the uptake of VLDL by the liver is inhibited, leading to increased plasma VLDL-triglycerides levels. In addition, APOC1 overexpression leads to an impaired uptake of fatty acids by skeletal muscle and adipose tissue. This will lead to an enhanced utilization of glucose by these tissues and, eventually, hypoglycemia and improved insulin sensitivity, which is in accordance with the Randle cycle (24,25). Concomitantly, a decreased uptake of fatty acids in the periphery will lead to an enhanced flux of fatty acids to the liver. This will, in turn, lead to an enhanced production and secretion of VLDL by the liver. Thus, central to this scheme is that in APOC1-overexpressing mice, the plasma fatty acid levels are high because of an impaired uptake of (VLDL-derived) fatty acids by peripheral tissues, in particular adipose tissue. As we reported previously, apoC1 is only a weak inhibitor of the LDL receptor (3,4); thus, the effect of apoC1 on binding of VLDL to the LDL receptor is not a major contributor to the phenotype of APOC1 mice.

It has been suggested that the VLDL receptor, which is highly expressed in fatty acid–utilizing tissues (adipose and skeletal tissue), acts as a docking protein to facilitate peripheral lipolysis of VLDL-triglycerides. Because APOC1 is a very strong inhibitor of the binding of VLDL to the VLDL receptor (3,4), we hypothesized that this strong inhibition might be fundamental to the observed impaired peripheral uptake of fatty acids. If this is the case, then how are the BMIPP data as presented in Fig. 1 explained? BMIPP is not transported in the plasma by lipoproteins but by albumin, like native fatty acids. We hypothesized that the decreased uptake of BMIPP, as a representative for fatty acid uptake, fits with a role for the VLDL receptor in this respect, provided that the pool of fatty acids derived from VLDL-triglycerides lipolysis is freely exchangeable with the albumin-bound plasma pool of fatty acids. For this hypothesis, we recently obtained ample experimental evidence. In addition, such an explanation is sustained by the data reported previously by Hultin et al. (26) showing that VLDL-derived fatty acids generated by lipoprotein lipase are leaked into the plasma rather than directly taken up by the adipocyte tissues.

Next to the VLDL receptor, the action fatty acid transporters (e.g., CD36, fatty acid transporting protein) may also be affected by apoC1. Our data indicate that apoC1 is not likely to inhibit FFA tissue uptake through interference with the fatty acid transporter CD36. Recent studies with CD36 knockout mice showed reduced uptake of BMIPP in heart, skeletal muscle, and adipose tissue (9), whereas apoC1 seems to inhibit BMIPP uptake in white adipose tissue only.

In addition, it is possible that apoC1, either bound to VLDL or present in a free form in plasma, is able to bind fatty acids, thereby preventing rapid uptake by peripheral tissues. In vitro experiments using lipoprotein fractions, plasma, and freshly isolated adipocytes from the respective transgenic mice will shed more light on the molecular mechanisms that underlie the full protection from obesity in APOC1-overexpressing mice.

We thank J. Goudriaan and M. de Vries for excellent technical assistance. We also thank Dr. F. F. Knapp, Jr. (Oak Ridge National Laboratory, Oak Ridge, TN) for providing the substrate and procedures for preparing radioiodinated BMIPP, through support from the Laboratory Directed Research and Development Program of Oak Ridge National Laboratory, managed by UT Battelle, LLC, for the U.S. Department of Energy under contract no. DE-AC05-00OR22725. This work was supported by the Netherlands Hearth Foundation (#97067) and the Netherlands Organization for Scientific Research (#903-39-194).