Effects of Nicotinic Acid on Gene Expression: Potential Mechanisms and Implications for Wanted and Unwanted Effects of the Lipid-Lowering Drug

Abstract

Context:

Nicotinic acid (NA), or niacin, lowers circulating levels of lipids, including triglycerides, very low-density lipoprotein-cholesterol, and low-density lipoprotein-cholesterol. The lipid-lowering effects have been attributed to its effect to inhibit lipolysis in adipocytes and thus lower plasma free fatty acid (FFA) level. However, evidence accumulates that the FFA-lowering effect may account for only a fraction of NA effects on plasma lipids, and other mechanisms may be involved. Recent studies have reported NA effects on gene expression in various tissues in vivo and in cultured cells in vitro.

Evidence Acquisition:

We reviewed articles reporting NA effects on gene expression, identified by searching PubMed, focusing on potential underlying mechanisms and implications for unexplained NA effects.

Conclusion:

The effects of NA on gene expression may be mediated directly via the NA receptor in the affected cells, indirectly via changes in circulating FFA or hormone levels induced by NA, or by activating the transcription factor FOXO1 in insulin-sensitive tissues. NA effects on gene expression provide new insights into previously unexplained NA effects, such as FFA-independent lipid-lowering effects, FFA rebound, and insulin resistance observed in clinics during NA treatment.

Nicotinic acid (NA), or niacin, is a B-group vitamin. In addition to its function as a vitamin, NA in high doses has been used as a lipid drug for five decades (1, 2). NA lowers circulating triglycerides (TG), very low-density lipoprotein (VLDL)-cholesterol, and low-density lipoprotein (LDL)-cholesterol levels while increasing high-density lipoprotein (HDL)-cholesterol (HDL-C) levels (3, 4). The lipid-lowering effects of NA have been attributed to its antilipolytic effect in adipocytes because this reduces plasma levels of free fatty acids (FFA), substrates for hepatic TG synthesis, and VLDL formation. However, circulating FFA levels rebound after long-term NA treatment, whereas its lipid-lowering effects persist (5). When this occurs, the hypolipidemic effects of NA may not be explained by reduced FFA delivery to the liver. In addition, the effect of NA to increase circulating HDL-C levels cannot be explained by the FFA-lowering effect of NA. Thus, despite a long history of clinical use, the precise mechanism by which NA modulates circulating lipids remains unclear. Also, insulin resistance often develops during NA treatment (6–8), but the underlying mechanism is unclear. We recently reported that NA has widespread effects on gene expression in major tissues of lipid metabolism, such as the liver, skeletal muscle, and adipose tissue (9), raising the possibility that the effects of NA on gene expression may explain some of the NA effects (wanted and unwanted) observed in clinical studies. In this review, we highlight literature data regarding NA effects on gene expression and discuss potential underlying mechanisms and implications for unexplained NA effects.

NA Action and Its Receptor

NA as a lipid drug

Dyslipidemias are major risk factors for cardiovascular disease, the number 1 cause of death in the United States. NA has been used as a drug for the treatment of dyslipidemias for five decades (1, 3). NA treatment results in a very desirable modification of atherogenic risk factors: decreases in VLDL-cholesterol, LDL-cholesterol, lipoprotein (a), and TG and increases in HDL-C (5, 10). Major clinical trials have demonstrated that NA treatment reduces the progression of atherosclerotic cardiovascular disease (3, 11). Despite these prominent beneficial effects and the fact that NA is both inexpensive and the most effective drug currently in use for increasing HDL-C (12), the use of NA has not been widespread because of its unwanted effects, such as the flush response (1, 3) and the development of insulin resistance. Clear understanding of the mechanisms underlying wanted and unwanted NA effects is critical to the development of a new strategy that maximizes the potentials of this class of drugs while eliminating unwanted side effects. Interest in NA has been rejuvenated by a recent discovery of its receptor (see NA receptor), which has enhanced our understanding of NA effects and has led to new strategies for treatment with NA or its derivatives. For example, partial agonists for the receptor are being developed as a drug that exerts the lipid-modifying effects of NA without causing the flush response (13, 14).

NA receptor

NA specifically binds to the plasma membrane of adipocytes via a specific G protein-coupled receptor (15, 16). In 2003, three independent groups discovered that the mouse orphan G protein-coupled receptor PUMA-G (protein up-regulated in macrophages by interferon γ) and its human ortholog HM74 are receptors for NA (17–19). These receptors (also known as GPR109) are expressed abundantly in adipose tissue but very little in other major tissues of lipid metabolism, such as the liver and skeletal muscle (18, 19). The NA receptor is coupled with the inhibitory G protein, and binding of NA to the receptor inhibits adenylate cyclase and thereby lowers cAMP levels (1). Decreased cAMP levels in turn reduce protein kinase A (PKA) activity, resulting in decreased phosphorylation and activity of hormone-sensitive lipase or other proteins involved in lipolysis, that is, hydrolysis of TG, in adipocytes.

Effects of NA on Gene Expression

In vivo studies

Recent studies have reported the effect of NA to alter gene expression in vivo in major tissues of lipid metabolism, such as the liver, adipose tissue, and skeletal muscle. NA treatment of obese dogs for 1 month was shown to decrease hepatic diacylglycerol O-acyltransferase (DGAT) 2 expression by approximately 60% (20). DGAT2 is a rate-limiting enzyme for hepatic TG synthesis (21), and decreased DGAT2 expression may explain NA-induced decreases in hepatic VLDL secretion and plasma lipoprotein levels in the absence of changes in plasma FFA levels. In addition, acute and chronic NA treatments in mice reduced hepatic expression of proliferator-activated receptor γ (PPARγ) coactivator (PGC)-1β and apolipoprotein C3 (APOC3) (22). In this study, PGC-1β was demonstrated to regulate plasma TG metabolism by increasing hepatic APOC3 expression and circulating APOC3 levels, suggesting that NA may lower plasma lipids by decreasing hepatic PGC-1β expression. NA also affects gene expression in vivo in adipocytes where the NA receptor is abundantly expressed. A 6-month treatment of patients with impaired glucose tolerance with extended-release (ER) NA led to a significant improvement of their atherogenic lipid profile, accompanied by changes in gene expression in adipose tissue, including increased adiponectin, C/EBPα, C/EBPδ, peroxisome PPARγ and decreased CPT2, hormone-sensitive lipase, NA receptor, and fatty-acid synthase mRNA expression, suggesting that some of the beneficial effects on lipid profiles may be due in part to altered gene expression (23). NA was shown to decrease circulating leptin concentrations in hypercholesterolemic rabbits, which was associated with a significant decrease in leptin mRNA expression in adipose tissue (24). The effect of NA to lower leptin mRNA expression was also demonstrated in cultured primary adipocytes where NA inhibited leptin expression and secretion and increased PPARγ expression in a dose-dependent manner (24). However, clinical studies also reported effects of acipimox, a long-acting NA analog, to increase circulating leptin levels (25, 26). Adiponectin, another adipokine, is known to exert atheroprotective effects and improve insulin action. Treatment of subjects with the metabolic syndrome with ER NA increased circulating adiponectin concentrations (26), primarily due to an increase in the active high-molecular weight complex (27, 28). Some of these effects may be due to increased adiponectin expression in adipose tissue, which increased by 35% after 6 months of ER NA treatment (23). NA was shown to acutely increase circulating IL-6 levels during exercise in humans, which was associated with increased IL-6 mRNA expression in adipose tissue (29). In skeletal muscle, NA infusion in humans acutely increased PPARα, PPARδ, and PGC-1α mRNA expression, and the effect on PGC-1α expression was suggested to be due to elevated epinephrine levels during NA infusion (30). Thus, numerous studies have demonstrated that NA treatment alters gene expression in vivo in various tissues. However, these in vivo studies are limited in testing whether the observed effects on gene expression represent direct effects of NA in the affected tissues or indirect effects mediated by changes in lipid or hormone levels induced by NA treatment.

In vitro studies

Direct effects of NA on gene expression have been studied in isolated cells. NA was shown to exert direct effects on adipokine expression in 3T3L1 adipocytes; NA inhibited TNF-α induction of proinflammatory chemokines, such as fractalkine, monocyte chemoattractant protein-1 and increased the expression of the atheroprotective adipokine, adiponectin, suggesting that some of the beneficial effects of NA on atherosclerosis and cardiovascular events may be mediated by these effects to regulate cytokine expression in adipocytes (31). NA was also shown to regulate monocyte adhesion to endothelial cells by altering the expression of cell-adhesion molecules in endothelial cells, such as ICAM-1 and PECAM-1, suggesting another potential mechanism underlying the beneficial effects of NA on atherogenesis (32, 33). Also, as discussed above, NA exerts direct effects to decrease leptin expression in primary adipocytes (24). ATP-binding cassette transporter A1 (ABCA1) plays a major role in the formation of HDL by transporting cellular cholesterol to lipid-poor apolipoprotein A-I (ApoA-I), a major apolipoprotein in HDL. NA was shown to promote ApoA-I-induced cholesterol efflux in adipocytes by increasing PPARγ, LXRα, and ABCA1 expression, revealing a potential mechanism by which NA increases circulating HDL-C levels (34). NA was also shown to increase PPARγ and ABCA1 expression in HepG2 cells and macrophages (35–37). In contrast, overexpression of the NA receptor GPR109A in mouse liver reduced ABCA1 expression, which was associated with decreased cholesterol efflux determined in primary hepatocytes (38).

Mechanisms of NA Effects on Gene Expression

NA receptor-mediated effects

Many of the in vitro studies on the effects of NA to alter gene expression have demonstrated that activation of NA receptors was required for the effects. As discussed above (see NA receptor), activation of the NA receptor in adipocytes gives rise to an inhibitory G protein signal that inhibits adenylate cyclase and lowers cAMP levels. Decreased cAMP levels then alter the activities of the two intracellular cAMP targets, PKA and exchange protein activated by cAMP (Epac) (39). PKA and Epac are known to be involved in the regulation of expression for numerous genes, and therefore PKA- or Epac-dependent pathways may be involved in the NA effects on gene expression observed in vitro. In our recent study with microarray analysis (9), we found 121 genes whose expression was altered more than 4-fold by NA in vivo in rats. Interestingly, all of these changes in gene expression occurred in fat tissue, and none in the liver, heart, or skeletal muscle. This striking confinement of the 121 independent gene-expression changes to adipose tissue suggests the existence of NA-activated pathways for gene regulation exclusively in adipocytes (Fig. 1), which may be explained by the NA receptor that is expressed abundantly in adipose tissue but not in the other tissues (18, 19).

Fig. 1.

Potential mechanisms by which NA may alter gene expression in adipocytes and other target cells in vivo. Numbers represent individual pathways of NA effects. See the text for details. CORT, Cortisol or corticosterone; EPI, epinephrine; GCR, glucocorticoid receptor; AR, adrenergic receptor.

FFA-mediated effects

Many of the effects of NA on gene expression observed in vivo may be indirectly mediated by humoral changes induced by NA. NA has the potent effect of lowering plasma FFA level, and some of the NA effects on gene expression may be mediated by altered plasma FFA (22). FFA are endogenous ligands for PPAR transcription factors, which play a major role in the regulation of genes involved in fuel metabolism and cellular signaling (40, 41). FFA are also known to be major regulators of PGC-1β, which regulates genes involved in hepatic lipid metabolism (42). Because insulin also lowers plasma FFA level (by inhibiting lipolysis in adipocytes), some effects on gene expression may be shared by NA and insulin. In fact, our study with microarray analysis (9) identified over 100 genes whose expression was altered similarly by NA and insulin. These genes may represent those regulated by plasma FFA. This idea is supported by the finding that this group included leptin and uncoupling protein 3, which are known to be regulated by plasma FFA (43, 44). Interestingly, we found that lipoprotein lipase (LPL) was up-regulated in skeletal muscle when plasma FFA levels were reduced by insulin or NA. Increased LPL expression in this major tissue increases the rate of VLDL removal and thereby lowers circulating lipid levels (10, 45).

Hormone-mediated effects

Our study with microarray analysis (9) showed that a 7-h NA infusion in rats had profound effects on mRNA expression for numerous genes, and the number of genes whose mRNA expression was altered by NA greatly exceeded the number of genes affected by insulin in skeletal muscle, fat, and liver tissues. Insulin is known for its profound effects to regulate gene expression in various tissues (46). Therefore, it was rather unexpected that the number of genes affected by NA exceeded that of genes affected by insulin in all of the insulin-sensitive tissues. However, many of the NA effects on gene expression may not represent direct effects of NA in the affected tissues. As discussed above (see NA receptor), the NA receptor is expressed in adipose tissue, but not in other insulin-sensitive tissues (18, 19). In addition, NA infusion or treatment increases plasma levels of GH (47), epinephrine (30, 48), corticosterone (49), and adiponectin (26, 27), which are known to regulate gene expression and cellular signaling in various tissues. Therefore, it is conceivable that many of the NA effects on gene expression may be mediated by changes in these hormones during NA infusion. In the same context, some unwanted NA effects may arise from these hormonal changes during NA treatment (see FFA rebound and Insulin resistance).

FOXO1-mediated effects

Our study (9) also showed that NA infusion in rats had profound effects to decrease FOXO1 phosphorylation (via Akt) and increase its transcriptional activity in insulin-sensitive tissues, including the liver, skeletal muscle, heart, and adipose tissue. These effects were opposite to the well-established effects of insulin to increase FOXO1 phosphorylation and inhibit its activity (50). In the same study, random, systematic analysis of microarray data revealed a few dozen genes whose expression was altered in opposing manners by insulin and NA, in correlation with FOXO1 or Akt phosphorylation. Interestingly, most of these genes encode proteins involved in energy metabolism and cellular signaling, consistent with the roles of Akt or FOXO1. Many of these genes may be regulated by FOXO1 and/or Akt. To support this idea, the gene list included PDK4, glucose 6-phosphatase, and IGF binding protein, known to be regulated by FOXO1 (51); and fatty acid synthase, ATP citrate lyase, 3-hydroxy-3-methylglutaryl-coenzyme A, and cAMP-responsive element modulator, reported to be regulated by Akt (52). The mechanism by which NA alters Akt and FOXO1 phosphorylation in insulin-sensitive tissues is unknown. These effects may not represent direct effects of NA in the affected tissues because NA had no significant effects on Akt and FOXO1 phosphorylation in vitro in adipose tissue explants or primary hepatocytes (9). These findings suggest that an in vivo (humoral) factor might have been activated during NA infusion and mediated the effects on Akt and FOXO1 phosphorylation.

Implications of NA Effects on Gene Expression for Unexplained NA Effects

FFA-independent lipid-lowering effects

Fasting plasma FFA levels were lowered during the first week of oral NA treatment in humans but rebounded to pretreatment levels after 2 wk of the treatment (6). After 1-month treatment with NA in female subjects, plasma TG and total cholesterol levels were significantly lowered despite FFA rebound or increased fasting FFA levels (53), suggesting FFA-independent lipid-lowering effects. DGAT catalyzes the final and rate-limiting step of hepatic TG synthesis. This is an important process because TG is used for VLDL formation and regulates cholesterol synthesis (54, 55). There are two isoforms of DGAT: DGAT1 and DGAT2. DGAT1 is involved in the formation of cytosolic lipid droplets and DGAT2 in TG synthesis for VLDL formation in the endoplasmic reticulum (55). Knockdown of DGAT2 in the liver using small interfering RNA led to hypolipidemia, indicating that DGAT2 plays a crucial role in hepatic synthesis and secretion of TG and VLDL (56, 57). In HepG2 cells, NA was shown to directly and noncompetitively inhibit DGAT2 but not DGAT1 enzyme activity, without altering DGAT2 expression (58), suggesting that this effect may account for FFA-independent effects of NA to decrease hepatic TG synthesis and VLDL secretion (4). In addition, as discussed above, NA treatment was shown to reduce hepatic DGAT2 expression by approximately 60% in dogs (20), providing another potential mechanism for FFA-independent lipid-lowering effects of NA. Our recent study showed that NA infusion (7–12 h) in rats significantly decreased hepatic DGAT2 mRNA and protein expression without altering DGAT1 expression (Oh, Y. and J. Youn, unpublished data). Taken together, these data suggest that NA may decrease plasma lipids not only by lowering substrates (i.e. FFA) but also by reducing the activity and/or the expression of the key enzyme of hepatic TG synthesis. NA was also shown to accelerate intracellular apolipoprotein B degradation by inhibiting TG synthesis in HepG2 cells (59). On the other hand, FFA-independent lipid-lowering effects of NA may be explained in part by the effect of NA to increase LPL expression and/or activity; increased LPL activity would increase the rate of VLDL removal and thereby lower circulating lipid levels (10, 45). However, there are very little data demonstrating NA effects on LPL expression or activity.

NA is the most effective drug currently available for raising HDL-C, but the underlying mechanism was unclear until recently when this effect could be demonstrated in animal models (60, 61). Cholesteryl ester transfer protein (CETP) is a protein, mainly produced by the liver, that transfers lipids (i.e. TG and cholesterol esters) between LDL and HDL particles. The importance of this protein in HDL-C metabolism was recognized by the discovery of genetic CETP deficiency in Japanese families with increased HDL-C and ApoA-I levels (62). Since then, CETP inhibition has been a major therapeutic strategy for raising HDL-C. Rodents naturally lack CETP, resulting in lipid profiles that are different from those of humans and are not altered by treatments with lipid drugs as in humans (60). When human CETP was introduced into mice, their plasma lipid profiles were changed to be closer to those of humans, and NA was able to increase HDL-C, demonstrating a critical role of this protein in mediating the effect of NA to raise HDL-C (60). In mice expressing human CETP under control of its natural flanking regions, NA was shown to increase HDL-C by decreasing hepatic expression and plasma levels of CETP (61). Taken together, these data suggest that the effect of NA to decrease hepatic expression and secretion of CETP may be a major, if not the major, mechanism by which NA raises HDL-C in humans. In addition, NA may increase plasma HDL-C by increasing ABCA1 expression, thereby increasing cholesterol efflux from various cells, including adipocytes and hepatocytes, as demonstrated in vitro (34–37). NA was also shown to inhibit surface expression of ATP synthase β chain, a receptor for HDL endocytosis, in HepG2 cells, associated with decreased HDL uptake (63), which may contribute to NA action to raise HDL. However, this effect occurred without changes in mRNA expression.

FFA rebound

Acute NA administration in humans (53) or rodents (49) results in a rapid decrease in the plasma FFA level, followed by a rebound and an overshoot above preinfusion levels. The rebound or overshoot of the plasma FFA level is probably due to the combination of waning antilipolytic effects of ingested or injected NA and stimulation of lipolysis by lipolytic hormones, such as epinephrine and cortisol, plasma levels of which increase after NA administration (30, 48, 49). Chronic treatment with NA is associated with increased basal (or fasting) FFA levels (8, 64). This may occur as a result of an overshoot of plasma FFA after NA administration the previous evening (65). However, when type 2 diabetic patients were treated extensively (i.e. with high and frequent dosages) for 3 d with acipimox, a long-acting NA analog, plasma FFA levels remained suppressed during the treatment period, but mean FFA levels gradually increased from d 1 to d 3 (65). These data suggest that slowly developing changes occur in adipocytes that increase lipolysis during repetitive or continuous exposure to acipimox (or NA). Consistent with this idea, fasting plasma FFA levels decreased in healthy subjects during the first week of NA treatment, but rebounded to control levels after 2 wk (6), and increased fasting FFA levels have been observed after 2 wk or more of NA treatment (8, 64).

The mechanisms underlying FFA rebound during chronic NA treatment have been unclear, in part due to a lack of appropriate animal models. Our recent studies showed that when NA was constantly infused in rats, the ability of NA to lower plasma FFA was maintained for at least 7 h (9), but when this infusion was extended to 24 h, plasma FFA levels rebounded to control levels (66). This was not due to a down-regulation of NA action because when the NA infusion was stopped, plasma FFA levels rapidly increased by more than 2-fold, indicating that basal lipolysis was increased. In the same study, microarray analysis revealed that the FFA rebound after 24-h NA infusion was associated with many changes in gene expression in adipocytes. Among these changes were 40–60% decreases in the expression of key enzymes of TG synthesis, including 1-acylglycerol-3-phosphate O-acyltransferase, DGAT1 and -2, GPD1, and PCK1. Decreased TG synthesis would help increase the net rate of lipolysis, that is, TG degradation into FFA and glycerol. In addition, perilipin expression decreased by approximately 40%. Perilipin coats lipid droplets in adipocytes and suppresses lipolysis by reducing lipase accessibility to the surface of lipid droplets (67). Therefore, decreased perilipin expression would contribute to the increase in lipolysis. Furthermore, phosphodiesterase (PDE) 3B gene expression decreased by approximately 60%. This change was specific because genes for other PDE isoforms were not affected. PDE3B degrades cAMP, a major regulator of lipolysis in adipocytes, and decreased PDE3B would help increase the cAMP level and thus lipolysis. Thus, these (and possibly other) changes in gene expression in adipocytes are likely to explain the increase in basal lipolysis in adipocytes (FFA rebound) after 24-h NA infusion in rats. Although a 24-h NA infusion in rats may be an excellent animal model in which to study FFA rebound, it may be crucial to test whether this acute model well represents similar phenomena observed during chronic oral NA treatments in humans. Because fat tissue sampling is feasible in humans, gene-expression analysis in fat tissue with or without NA treatment would provide answers to this critical question.

Insulin resistance

Insulin resistance often develops to deteriorate glucose tolerance during NA treatment, raising the concern of whether NA treatment is safe in glucose intolerant or diabetic patients despite its beneficial effects on lipid profile (thus reducing cardiovascular risk) (68, 69). The mechanism by which NA treatment results in insulin resistance remains unclear. As discussed above, changes in plasma adiponectin (26–28) levels cannot explain the development of insulin resistance during NA treatment. The possibility remains to be tested whether NA-induced increases in insulin antagonizing hormones, such as cortisol (49), epinephrine (30, 48), and GH (47), may cause insulin resistance. Previous studies have debated whether insulin resistance develops during NA treatment as a result of FFA rebound (7, 8). It is established that the plasma FFA level is a major determinant of insulin sensitivity (70); acute or chronic elevation of plasma FFA decreases insulin's action (71, 72). Lowering plasma FFA using NA or its analog acipimox for a short term (hours or days) has been shown to increase insulin sensitivity (73–75). In contrast, long-term treatments with NA have resulted in insulin resistance, often associated with increased basal FFA levels (8, 64), suggesting that FFA rebound and increased basal plasma FFA levels may be responsible for the development of insulin resistance. However, insulin resistance was observed during NA treatment even without increased basal FFA levels (6, 7), suggesting that other mechanisms may be involved. Our recent study showed that a continuous 24-h NA infusion in rats increased basal lipolysis in adipose tissue and basal plasma FFA levels by altering the expression of genes regulating lipolysis in adipocytes (66). If FFA rebound is responsible for insulin resistance during NA treatment, altered gene expression in adipocytes may drive the increases in adipocyte lipolysis and plasma FFA levels, resulting in insulin resistance. On the other hand, if insulin resistance develops independently of plasma FFA, there is the possibility that NA-induced changes in gene expression in other insulin-sensitive tissues, demonstrated in our recent study (9), may be responsible for the development of insulin resistance. Regarding this, the effect of NA to activate FOXO1 in insulin-sensitive tissues (9) may deserve attention from investigators because some FOXO1 target genes (e.g. PDK4, glucose 6-phosphatase, phosphoenolpyruvate carboxykinase, PGC-1, etc.) have been implicated in the regulation of blood-glucose control and/or insulin sensitivity (76–78).

Conclusion and Perspectives

The literature data reviewed here indicate that NA does more than merely decrease plasma FFA levels or suppress lipolysis in adipocytes. There is ample evidence that NA acutely or chronically alters gene expression and cellular signaling in various tissues and cells. The underlying mechanism may involve signaling through the NA receptor, but there is substantial evidence that many in vivo effects on gene expression may be mediated indirectly by changes in circulating lipid or hormone levels induced by NA treatment. Future studies are warranted to elucidate these mechanisms, which would contribute not only to the full understanding of NA effects (wanted and unwanted), but also to the discovery of novel pathways for gene regulation by lipids or hormones. Finally, NA has also been used as a tool for lowering plasma FFA (assuming that this is the only effect of NA) in studies on the role of plasma FFA. Data from such studies should be carefully interpreted because there are many ways NA could affect the target system under study.