The lipid-lowering effect of niacin has been attributed to the inhibition of cAMP production in adipocytes, thereby inhibiting intracellular lipolysis and release of nonesterified fatty acids (NEFA) to the circulation. However, long-term niacin treatment leads to a normalization of plasma NEFA levels and induces insulin resistance, for which the underlying mechanisms are poorly understood. The current study addressed the effects of long-term niacin treatment on insulin-mediated inhibition of adipocyte lipolysis and focused on the regulation of cAMP levels. APOE*3-Leiden.CETP transgenic mice treated with niacin for 15 wk were subjected to an insulin tolerance test and showed whole body insulin resistance. Similarly, adipocytes isolated from niacin-treated mice were insulin resistant and, interestingly, exhibited an increased response to cAMP stimulation by 8Br-cAMP, β1- and β2-adrenergic stimulation. Gene expression analysis of the insulin and β-adrenergic pathways in adipose tissue indicated that all genes were downregulated, including the gene encoding the cAMP-degrading enzyme phosphodiesterase 3B (PDE3B). In line with this, we showed that insulin induced a lower PDE3B response in adipocytes isolated from niacin-treated mice. Inhibiting PDE3B with cilostazol increased lipolytic responsiveness to cAMP stimulation in adipocytes. These data show that long-term niacin treatment leads to a downregulation of PDE3B in adipocytes, which could explain part of the observed insulin resistance and the increased responsiveness to cAMP stimulation.
- adipose tissue
- adenosine 3′,5′-cyclic monophosphate
niacin, also known as vitamin b3, is required for the synthesis of the cofactor nicotinamide adenine dinucleotide and is therefore essential for oxidative phosphorylation in energy metabolism (8). It has been used for more than 50 years for the treatment of dyslipidemias, since it decreases plasma triglycerides, low-density lipoprotein-cholesterol, and hepatic very low density lipoprotein (VLDL) triglyceride production (25), in addition to increasing high-density lipoprotein-cholesterol. Supplementation with niacin was shown to decrease risk of cardiovascular disease and atherosclerosis in dyslipidemic humans (1) and in dyslipidemic mouse models (23), using the APOE*3-Leiden.CETP cholesteryl ester transfer protein (CETP) transgenic female mouse.
The molecular mechanism by which niacin conveys its lipid-lowering effects is mostly unknown. The receptor for niacin, HCA2 (formerly known as GPR109A), has been shown to play an important role in acute antilipolytic effects (21) (10), but is not required for the long-term lipid-lowering effects (10). This receptor is expressed mostly in spleen, immune cells, and adipose tissue (20) where binding of niacin leads to the release of a Gα-inhibitory subunit that inhibits production of the secondary messenger cAMP by adenyl cyclase. Niacin is known to inhibit lipolysis in adipocytes via this reduction of cAMP (2), which leads to a reduced protein kinase A (PKA) activation and thus less phosphorylation and activation of the lipolytic enzyme hormone-sensitive lipase (HSL) (7). By inhibiting lipolysis, less nonesterified fatty acids (NEFAs) are released by adipocytes thus making less substrate available for VLDL-triglyceride production in the liver. Niacin may therefore lower lipid levels via its antilipolytic effect. However, during niacin administration the initial drop in NEFA and glycerol is followed within hours by a rebound and normalization of NEFA and glycerol levels and adipose tissue lipolysis rates (3, 10, 14, 24, 25).
In addition, long-term niacin administration leads to insulin resistance in liver, adipose, and muscle tissue (5, 17) for which the underlying mechanisms are poorly understood. In adipocytes, binding of insulin to its receptor leads to phosphorylation and activation of protein kinase B (PKB/Akt) and phosphodiesterase 3B (PDE3B), breaking down cAMP, which in turn reduces NEFA release from the cells.
We hypothesized that long-term niacin treatment modulates insulin signaling by acting at the level of PDE3B. Oh et al. (14) have previously shown that 24 h treatment with niacin induces insulin resistance and reduces expression of PDE3B in adipose tissue, hinting toward a possible role for PDE3B. In this study we functionally tested in APOE*3-Leiden.CETP transgenic female mice whether long-term niacin-induced downregulation of the enzyme PDE3B would lead to less PDE3B activity when stimulated by insulin, which would explain (part of) the observed insulin resistance in adipose tissue.
MATERIALS AND METHODS
In contrast to wild-type mice, female APOE*3-Leiden.CETP mice (27) have a human-like lipoprotein profile that makes them exquisitely suited to test hypolipidemic drugs, such as niacin (23). This mouse model was bred at the Leiden University Medical Center. At age 15 ± 1 wk, mice were fed a western type diet (Diet T with 0.1% cholesterol, which consisted of 17 kcal/100 kcal protein, 43 kcal/100 kcal carbohydrate, and 41 kcal/100 kcal fat; AB Diets, Woerden, the Netherlands) with or without niacin (0.3% wt/wt; Sigma Aldrich, St. Louis, MO). Body weight was registered weekly. Animals were housed in a controlled environment (21°C, 40–50% humidity) with a daily 12:12-h photoperiod (0700–1900). Food and tap water were available ad libitum during the whole experiment. All experiments were performed after a 15-wk dietary intervention period. All animal experiments were performed in accordance with the regulations of Dutch law on animal welfare. The institutional ethics committee for animal procedures from the Leiden University Medical Center, Leiden, The Netherlands, approved the protocols.
Intraperitoneal insulin tolerance test.
Food was withdrawn from all animals at 0800 for a period of 6.5 h. Subsequently, all animals received a single intraperitoneal injection with insulin (0.2 U/kg; Novo-Nordisk, Bagsværd, Denmark), and blood samples were taken every 30 min for a period of 2 h. Blood samples were taken from the tail tip in chilled paraoxon-coated capillaries and placed on ice to prevent ex vivo lipolysis.
All animals were euthanized, and organs were collected in the fed state between 0800 and 0930 unless otherwise indicated. Blood was collected by cardiac punction, and plasma was collected after centrifugation. Fresh subcutaneous (sWAT), gonadal (gWAT), and visceral (vWAT) white adipose tissue were harvested and used for determination of morphometry and lipolysis experiments as described below. In addition, a portion of the gWAT was frozen for qPCR and cAMP measurement. Liver and adrenal glands were harvested quantitatively and weighed before freezing. All tissues were stored at −80°C before further analysis.
Plasma and liver parameters.
Commercially available kits were used to determine plasma levels of triglycerides (1488872; Roche Molecular Biochemicals, Indianapolis, IL), total cholesterol (236691; Roche Molecular Biochemicals), phospholipids (Instruchemie, Delftzijl, The Netherlands), NEFAs (Wako Chemicals, Neuss, Germany), glucose (Accucheck, Roche, The Netherlands), and insulin (Crystal Chem, Downers Grove, IL), according to the manufacturer's instructions.
Adipose tissue from the gWAT, sWAT, and vWAT were minced and digested in 0.5 g/l collagenase in HEPES buffer (pH 7.4) with 20 g/l of dialyzed bovine serum albumin (BSA, fraction V; Sigma Aldrich) for 1 h at 37°C. The disaggregated WAT was filtered through a nylon mesh with a pore size of 236 μm. For the isolation of mature adipocytes, cells were obtained from the surface of the filtrate and washed several times. Cell size and volume of mature adipocytes were determined from micrographs (±1,000 cells/WAT sample) using image analysis software that was developed in-house in MATLAB (MathWorks, Natick, MA).
Lipolysis experiments in isolated adipocytes.
Adipocytes were incubated in DMEM-F-12 with 2% (wt/wt) BSA in 96-well plates with ∼10,000 adipocytes in 200 μl/well in the presence of 1) no additional reagents, 2) 10−3 M of the PDE3B-sensitive cAMP analog 8Bromo-cAMP, 3) 10−3 M 8Br-cAMP + 10−10 M insulin (Novo-Nordisk), 4) 10−6 M of the selective β2-adrenergic agonist terbutaline, or 5) 10−6 M of the selective β1-adrenergic agonist dobutamine. The wells of the adipocytes from niacin-treated mice contained 10−6 M niacin, whereas control adipocyte medium did not contain niacin. Basal lipolysis rates were determined in incubations without these reagents. PDE3B responsiveness in adipocytes from control and niacin-treated mice was assayed by lipolysis assay using the selective PDE3 inhibitor cilostazol (all obtained from Sigma Aldrich) under the following additional conditions: 6-8) 10−6, 10−5, or 10−4 M cilostazol or 9-11) 10−3 M 8Br-cAMP + 10−6, 10−5, or 10−4 M cilostazol. The adipocytes were incubated for 2 h at 37°C, after which 100 μl of lipolysis assay medium were frozen at −20°C. At the day of analysis, 10 μl of lipolysis assay medium were mixed with 100 μl of Free Glycerol Reagent (F6428; Sigma Aldrich) and 0.5 μl Amplex UltraRed Reagent (A36006; Invitrogen, Carlsbad, CA). After 10 min of incubation in the dark, the fluorescence was excited at 530 nm and measured at 590 nm on a Fluorometer (SpectraMAX Gemini; Molecular Devices).
Adipose tissue cAMP measurement.
cAMP concentrations were measured in 50 mg of gWAT, using the cAMP direct immunoassay kit (Biovision, Milpitas, CA).
RNA was isolated from gWAT using the Nucleospin RNA/Protein kit (MACHEREY-NAGEL, Düren, Germany). Subsequently, 1 μg of RNA was used for cDNA synthesis by iScript (Bio-Rad, Hercules, CA), which was purified by the Nucleospin Gel and PCR clean-up kit (Machery Nagel). Real-Time PCR was carried out on the IQ5 PCR machine (Bio-Rad) using the Sensimix SYBR Green RT-PCR mix (Quantace, London, UK) and QuantiTect SYBR Green RT-PCR mix (Qiagen, Venlo, the Netherlands). Target mRNA levels were normalized to glyceraldehyde-3-phosphate dehydrogenase mRNA levels. Primer sequences are listed in Table 1.
All data in Figs. 1–3 are represented as means ± SE. The mean of all data was tested between groups for differences by unpaired t-test for normally distributed data. A repeated-measures ANOVA was used when indicated. Threshold for significance was set at 5%. Tests were performed using GraphPad Prism version 5.00 for Windows (GraphPad Software, San Diego, CA).
Niacin does not alter body composition.
We have applied a humanized mouse model that has previously been shown (23) to react to niacin on plasma lipid parameters, atherosclerosis, and body weight in a similar manner as dyslipidemic humans. After 15 wk of niacin treatment body weight (Table 2) and weight gain (data not shown) did not differ between the niacin-treated and control mice. Adipose tissue depots were similar in weight, whereas the adrenal glands were heavier after niacin treatment (Table 2). Liver cholesterol content was significantly lower in niacin-treated mice (Table 2).
Niacin lowers plasma lipids, but increases fasting glucose and insulin.
Long-term niacin treatment resulted in the expected decrease of nonfasting plasma total cholesterol, triglycerides, and phospholipids (Table 2). In contrast to short-term niacin treatment, long-term niacin treatment did not lower NEFA levels (25) but slightly lowered glycerol levels compared with control mice (Table 2). Glucose and insulin concentrations were significantly higher after niacin treatment (Table 2), suggesting insulin resistance.
Niacin induces insulin resistance in vivo.
To verify the development of in vivo insulin resistance upon long-term niacin treatment we performed an insulin tolerance test. Insulin stimulates glucose uptake in muscle and adipose tissue and suppresses WAT glycerol release. After an intraperitoneal injection of insulin, blood was drawn and at indicated times, and glucose (Fig. 1A), glycerol (Fig. 1B), and insulin (Fig. 1C) were measured. Glucose levels dropped and insulin levels rose in both groups. Because glucose levels were already higher due to niacin treatment at time (t) = 0, the percentage change in glucose, glycerol, and insulin levels from the t = 0 values was calculated (data not shown). A repeated-measures ANOVA of the percentage change in insulin values indicated no significant effect of niacin, whereas a similar analysis of glucose values showed significant treatment differences. At 60 min, the percentage change in glucose was smaller for niacin-treated animals, indicating insulin resistance on a whole body level. Insulin induced a decrease in glycerol levels in the control group but failed to do so in the niacin-treated group, indicating adipose tissue insulin resistance.
Niacin reduces the antilipolytic effect of insulin in isolated adipocytes.
To study whether long-term niacin treatment also alters the response to insulin at the level of adipocyte lipolysis, we isolated gWAT adipocytes from control and niacin-treated mice and performed an ex vivo lipolysis assay. Basal lipolysis (i.e., unstimulated lipolysis) did not differ between the niacin-treated and control mice (Fig. 2A). 8Br-cAMP-stimulated glycerol release was higher in adipocytes from niacin-treated mice compared with control treatment. The antilipolytic effect of insulin, reflected by the percentage lipolysis suppression by insulin of 8Br-cAMP-stimulated lipolysis, was significantly smaller in the niacin-treated adipocytes (Fig. 2B). This implies a reduced insulin response of adipocytes from niacin-treated mice.
Niacin increases the lipolytic effect of β-adrenergic stimulation in isolated adipocytes.
Similar to testing the response of adipocyte lipolysis to inhibition by insulin, β1- and β2-adrenergic stimulation of gonadal adipocyte lipolysis was tested. Gonadal adipocytes from niacin-treated mice exhibited a greater response to both the β1-agonist dobutamine and the β2-agonist terbutaline, as well as the secondary lipolytic messenger 8Br-cAMP (Fig. 2A). These data show that lipolysis in niacin-treated adipocytes is more responsive to β-adrenergic and cAMP stimulation.
Niacin downregulates genes involved in both the insulin and the β-adrenergic signaling pathways that regulate lipolysis.
Expression of genes involved in the insulin and/or β-adrenergic pathways was analyzed by qPCR in gWAT. The niacin-treated adipose tissues showed a downregulation of the insulin signaling cascade: mRNA levels of the insulin receptor (Insr), insulin receptor substrate-1 (Irs1), and Pde3b were downregulated by 45, 97, and 53%, respectively, compared with control (Fig. 2C). Also mRNA levels of the lipolytic enzyme Hsl were downregulated (62%) in the niacin-treated mice. Furthermore, all β-adrenergic receptor genes were significantly downregulated. The β1-adrenoceptor (Adrb1) was downregulated by 68%, Adrb2 by 74%, and Adrb3 by 88%. The intracellular adrenergic adaptor β-arrestin1 (Arrb1) was also downregulated by 42%. These data indicate a decreased gene expression of the insulin and β-adrenergic receptors and their postreceptor signaling pathways to lipolysis.
Niacin-treated adipocytes display less PDE3B capacity.
Because the downregulation of the β-adrenoceptors was not consistent with the observed increased β-adrenergic response, we investigated whether the increased β-adrenergic response was the result of postreceptor signaling changes. Therefore we first tested the basal levels of the lipolytic postreceptor mediator cAMP. The cAMP concentration of nonfasted gWAT showed no difference after long-term niacin treatment (control: 19.9 ± 6.5, niacin: 21.6 ± 8.2 pmol/g adipose tissue, P = 0.63).
We then focused on the postreceptor enzyme PDE3B, catalyzing cAMP hydrolysis, whose gene expression was downregulated after niacin treatment. PDE3B is the most predominant PDE3 isoform in adipose tissue (12) that is activated by insulin. The role of PDE3B in lipolysis was characterized in gWAT adipocytes using the PDE3 selective inhibitor cilostazol (Fig. 3). Adding cilostazol to unstimulated adipocytes in basal medium did not increase lipolysis, indicating that under basal conditions PDE3B was not active. However, when 8Br-cAMP-stimulated adipocytes were subjected to increasing concentrations of cilostazol, there was a sharp rise in lipolysis. Therefore, inhibiting PDE3B capacity can increase the responsiveness to cAMP stimulation, e.g., by catecholamines. Furthermore, lipolysis of adipocytes from niacin-treated mice was stimulated to a lesser degree by 10−4 M cilostazol than control treated mice. A repeated-measures ANOVA indicated a significant interaction (P < 0.0001) between the concentration of cilostazol and the 8Br-cAMP-stimulated control/niacin adipocytes. This implied a lower responsiveness (slope) of PDE3B to its inhibitor in adipocytes from niacin-treated mice, independent of insulin signaling.
The present paper shows that long-term niacin induces insulin resistance in APOE*3-Leiden.CETP female transgenic mice fed a Western type diet. The insulin resistance by niacin has been shown before in dyslipidemic patients and is confirmed in our humanized mouse model, both in vivo on glucose and glycerol metabolism and ex vivo in isolated adipocytes on lipolysis. The niacin-induced insulin resistance was associated with a reduction in the expression of key insulin-signaling genes (i.e., Insr and Irs1) in WAT. This indicates that niacin downregulates the insulin signaling pathway at the transcriptional level, which led to the reduced antilipolytic effect of insulin. In contrast, gene expression of all β-adrenergic receptors, as well as Hsl and Arrb1, was downregulated in WAT, whereas an increased adrenergic lipolytic effect was observed. This might be because of desensitization after niacin-induced prolonged stress (9, 13, 14, 18, 26). Prolonged stress leads to increased adrenal gland size (22), also evident in our study. Decreased β-adrenoceptor gene expression combined with an increased β-adrenergic responsiveness suggest that the effects of niacin on β-adrenergic stimulation are the result of postreceptor signaling mechanisms that enhance cAMP stimulation and thus lipolysis.
Interestingly, 8Br-cAMP, which acts at the postreceptor level, induced lipolysis to a higher extent in niacin-treated mice, which strengthens the hypothesis that a postreceptor signaling mechanism is affected by niacin. The adipocyte lipolysis assays were not confounded by a difference in weight of the gWAT depots, nor in adipocyte mean cell volume (Table 1), which in itself might change responsiveness to reagents. It was shown by Oh et al. (14) that 24 h niacin infusion lowered the expression level of Pde3b in gWAT to a similar extent (−57%) compared with our study (−53%). PDE3B regulates intracellular cAMP levels and is the key enzyme involved in the antilipolytic action of insulin. Moreover, by modulating cAMP levels, PDE3B may interact with β-adrenergic signaling. Oh and coworkers suggested that the reduction in Pde3b may be an indirect effect of niacin that serves as a counterregulatory mechanism to increase the cellular cAMP-to-AMP ratio to maintain basal lipolytic rate. In the current paper we show that the reduction of expression of Pde3b in adipose tissue by niacin is accompanied by a reduced antilipolytic effect of insulin. We further show by using the selective PDE3 inhibitor cilostazol in an adipocyte lipolysis assay that the PDE3B responsiveness in adipocytes isolated from long-term niacin-treated mice was diminished. We also showed that cilostazol increased 8Br-cAMP-stimulated lipolysis in a concentration-dependent manner in adipocytes from control mice, whereas it did not have an effect on unstimulated cells.
These data suggest that in adipocytes PDE3B will only degrade cAMP when the lipolytic cascade is stimulated. These observations could explain the increased responsiveness to β-adrenergic or 8Br-cAMP stimulation in niacin-treated adipocytes. In the unstimulated basal condition PDE3B was not activated, which corresponds to the unchanged adipocyte lipolysis rate and basal adipose tissue cAMP levels between control and niacin-treated mice. In line with this, plasma NEFA levels did not differ between the groups at the start of the intraperitoneal insulin tolerance test, indicating that in vivo adipose lipolysis was also similar between the groups, despite decreased Pde3b expression.
We propose a molecular mechanism for long-term niacin treatment occurring in adipocytes (Fig. 4): lipolysis is regulated by HSL, which is controlled by cAMP levels. Production of cAMP can be stimulated via β-adrenergic pathways or inhibited by PDE3B, which is activated by the insulin-signaling cascade. PKA is activated by cAMP, which can activate PDE3B, creating a negative feedback loop (15, 16, 19) (Fig. 4). Long-term niacin treatment leads to downregulation of PDE3B (Fig. 4). This downregulation of PDE3B is also accompanied by a downregulation of the rest of the insulin signaling pathway, leading to reduced insulin responsiveness in the adipocyte. In addition, because the PDE3B responsiveness is smaller in adipocytes from niacin-treated mice, stimulated cAMP levels in the niacin-treated mice, after β-adrenergic stimulation, will be degraded to a lesser extent, leading to a net higher lipolysis.
In this study we have focused on adipocyte insulin response after long-term niacin treatment. However, the sequence of events leading up to adipocyte insulin resistance and decreased Pde3b expression remain unclear, as does the possible role of adipocyte insulin resistance in the development of whole body insulin resistance. Moreover, niacin has been shown to increase plasma adiponectin, an important insulin-sensitizing hormone. The increase in adiponectin plasma levels after niacin treatment was shown to correlate to the decrease in insulin resistance (6), indicating a possible compensatory insulin-sensitizing effect. We, however, did not detect an increase of adiponectin gene expression in gWAT after niacin treatment (data not shown). Whether niacin acutely interferes with the insulin-signaling cascade remains to be investigated, although PDE3B capacity was not affected in control adipocytes exposed to niacin for 2 h (data not shown). Choi et al. have shown a reduced phosphorylation of PKB/Akt in adipose tissue after 1.5 h infusion of niacin (4), which leads to a lower insulin response. Niacin thus seems to induce decreased adipocyte insulin responsiveness via more than one mechanism.
In conclusion, long-term niacin treatment resulted in insulin resistance and increased β-adrenergic responsiveness that may in part be explained by a niacin-induced downregulation of PDE3B.
This work was supported by grants from the Center of Medical Systems Biology, the Netherlands Consortium for Systems Biology established by The Netherlands Genomics Initiative/Netherlands Organization for Scientific Research. M. R. Boon is supported by the Board of Directors of the Leiden University Medical Center. P. C. N. Rensen is Established Investigator of the Netherlands Heart Foundation (NHS2009T038).
The authors declare no conflicts of interest, financial or otherwise.
Author contributions: M.M.H., S.A.v.d.B., and V.v.H. conception and design of research; M.M.H., S.A.v.d.B., A.C.P., M.R.B., and V.v.H. performed experiments; M.M.H., S.A.v.d.B., A.C.P., J.V.K., M.R.B., and V.v.H. analyzed data; M.M.H., S.A.v.d.B., and V.v.H. interpreted results of experiments; M.M.H. and S.A.v.d.B. prepared figures; M.M.H., S.A.v.d.B., and V.v.H. drafted manuscript; M.M.H., S.A.v.d.B., J.V.K., M.R.B., L.M.H., P.C.N.R., K.W.v.D., and V.v.H. edited and revised manuscript; M.M.H., S.A.v.d.B., A.C.P., J.V.K., M.R.B., L.M.H., P.C.N.R., K.W.v.D., and V.v.H. approved final version of manuscript.
We thank S. J. M. van der Tuin for discussions on gene expression analysis.
Present address of S. A. A. van den Berg: Department of Laboratory for Clinical Chemistry & Hematology, Amphia Hospital, Breda, The Netherlands.
- Copyright © 2014 the American Physiological Society