Sp1. Seven consensus Sp1-binding sequences embedded in a CpG island, located between −299 and 10 bp (Figure 2), have been found in the CIC gene promoter. Five of them, located at positions −299/290, −206/−197, −162/−153, −137/−128, and 109/−100 bp, are functionally active, as demonstrated by EMSA and transfection experiments [26]. Sp1-binding activity depends on the methylation status of the Sp1-binding sequence. Transfection experiments performed in HepG2 cells with a construct containing the methylated CpG region demonstrate a marked decrease in luciferase (LUC) activity as compared to the activity of cells transfected with unmethylated construct. Moreover, ChIP analysis revealed that Sp1 binds to the CIC proximal promoter only in HepG2, whereas no binding was observed in SK-N-SH cells [26], in agreement with the active role performed by CIC in liver. However, since Sp1 is a universally expressed transcription factor [27], it is likely that it acts in concert with other transcription factors to activate CIC gene expression.

FOXA1. An active FOXA site is present at position –1098/–1088 bp that acts as a strong activator of CIC gene expression (Figure 2) [28]. Transfection experiments with pGL3 basic-LUC vector containing the FOXA1 bp region of the CIC gene promoter revealed an increase in LUC activity. Modulation of FOXA1 levels by overexpression and silencing experiments determines a significant increase and decrease in CIC transcript and protein, respectively. FOXA1 belongs to the forkhead box transcription factors, which also include FOXA2 and FOXA3, and which also play a major role in carbohydrate and lipid metabolism and the differentiation of pancreas, liver, lung, and gut [29]. In particular, FOXA1 plays a specific role in hormonal regulation since it induces glucagon gene expression and insulin secretion in pancreatic cells. Some studies highlighted a relationship between insulin secretion and FOXA1 in INS-1 cells [30,31]. Joseph et al. [32] demonstrated an involvement of CIC in insulin secretion, and Iacobazzi et al. [28] demonstrated the key role performed by FOXA1 gene silencing in the inhibition of glucose-stimulated insulin secretion. In fact, CIC gene silencing, as well as FOXA1 silencing, inhibit glucose-stimulated insulin secretion in INS-1 cells without affecting the cytosolic ATP/ADP ratio.

Recently, it has been reported that the atypical antipsychotic drug clozapine increases CIC gene expression through up-regulation of FOXA1 in INS-1 and other cells [33]. In INS-1 cells, this effect is only exerted at a basal glucose concentration with a concomitant increase of insulin release. The mechanism of this finding is a matter of debate. Sasaki et al. [34] suggested an impairment of the TCA (tricarboxylic acids) cycle in mitochondria through inhibition of clozapine with a subsequent reduction of Krebs cycle metabolites, such as citrate, that act as signals/second messengers in insulin secretion. Interestingly, however, the abnormal insulin secretion under basal conditions is completely abolished by FOXA1 silencing in INS-1 cells treated with clozapine [33].

SREBP-1a. An active SREBP-1a (sterol responsive element binding protein) is present in the CIC gene promoter at position –1696/–1686 bp (Figure 2) [35]. SREBP-1a belongs to the basic helix–loop–helix leucine zipper transcription factors, along with SREBP-1c and SREBP-2. These factors regulate transcription of target genes, whose products play key roles in lipid metabolism [36]. In agreement with the reported role performed by SREBPs in the activation of genes involved in fatty acid biosynthesis [37], SREBP-1a activates CIC gene expression in HepG2 cells. It is known that lipogenic enzymes are coordinately regulated at the transcriptional level during starvation and the induction by nutrients [38,39]. Starvation decreases SREBP-1 gene expression and, consequently, the downstream regulated genes [40]. Polyunsaturated fatty acids (PUFAs) (docohexanoic and arachidonic acids) down-regulate CIC gene expression at the transcriptional and translational level. Since both glucose and insulin are required for the production of fatty acids via induction of hepatic lipogenic enzymes, some lines of evidence suggest that the role of glucose and insulin in this action is mediated by induction of SREBP-1c [41,42]. Up-regulation of the CIC gene by insulin is in line with this finding, although a different SREBP isoform, SREBP-1a, is involved. Insulin activates SREBP-1a gene transcription through phosphatidylinositol 3-kinase (PI 3-kinase) and the Akt-dependent pathway. In fact, the increase of CIC expression is abolished by the presence of LY-294002, an inhibitor of PI 3-kinase, which blocks the insulin-induced phosphorylation pathway [43].

NF-kB. Two specific binding elements for NF-kB (nuclear factor-kB) are present in the CIC gene promoter at −1314/−1305 bp and −414/−405 bp, respectively [44]. The NF-kB binding site activity was proven in the context of CIC involvement in the inflammation pathway. Infantino et al. [44] demonstrated that bacterial lipopolysaccharide (LPS)-induced macrophages greatly activate CIC gene expression through NF-kB, compared to undifferentiated monocytes. Using either siRNA to decrease the amount of CIC or the CIC inhibitor benzene-1,2,3-tricarboxylate (BTA), it was demonstrated that the LPS-induced increase of nitric oxide (NO), ROS and prostaglandins was reduced. These data indicate that CIC is critical for these LPS responses, and the most plausible explanation is that citrate must be transported by CIC into the cytosol for the production of NO, ROS and prostaglandins during inflammation. In fact, an increase of acetyl-CoA derived from CIC-exported citrate is needed to synthesize fatty acids and their derivatives, such as prostaglandins, and to supply NADPH + H⁺ , which is required for the production of inflammatory mediators, such as NO and ROS.

This somewhat unexpected role for CIC in LPS action adds to the growing literature on the role of metabolism in the regulation of signaling in inflammation [45]. However, the mechanism by which CIC transcription is regulated in inflammation needs further investigation, because apart from NF-kB, other transcriptional factors may be involved in CIC gene transcriptional activation.

PPARα/γ. CIC expression is regulated also by peroxysome proliferator-activated receptor (PPAR) alpha and gamma in rat hepatocytes (BRL-3A) and adipocytes (3T3-L1), respectively [46]. WY-14,643 and rosiglitazone, agonists of PPARα and PPARγ, respectively, greatly increase the expression of CIC. By utilizing LUC activity measurements, gel shift, and ChIP experiments, an active PPRE site was identified in the rat CIC gene promoter at position –625 bp [46]. Moreover, CIC gene expression increases during differentiation of adipocytes, most likely through an involvement of PPARγ.

ZNF224. By affinity chromatography and mass spectrometry, a zinc finger protein, ZNF224, that binds to the −595/−569 bp sequence in the CIC gene promoter, was identified in HepG2 cells (Figure 2) [47]. This promoter element displays a pronounced repression of CIC transcription, as it strongly decreases LUC transgene expression activity in transfected HepG2 cells. In contrast, ZNF224 silencing markedly activates LUC reporter activity in HEK293 cells transfected with the LUC vector harboring the −595/−569 bp region. Moreover, it exhibits highly specific protein-binding activity in the presence of HepG2 nuclear extracts, as shown by site-directed mutagenesis within the −595/−569 bp sequence. ZNF224 belongs to the Krüppel-like zinc finger family of transcriptional factors that are involved in cell differentiation, cell proliferation, morphogenesis and are expressed during embryonic development [48]. ZNF224 contains a Krüppel-associated box (KRAB) domain, which represses transcription depending on specific interaction with the KAP-1 co-repressor molecule. It was demonstrated that Kap1 binds to KRAB domains as an oligomer, functioning as a scaffold to recruit histone deacetylases [49].

The presence of negative regulatory elements in the CIC gene promoter is thought to be important to determine timing of transcription during development. ZNF224 and the CIC gene reveal a different pattern of expression: ZNF224 is present in small amounts in adult tissues and at very high levels in fetal tissues [49], whereas the CIC gene is expressed in small amounts in fetal liver and kidney as compared to the same adult tissues. This different regulation may be related to the fact that, in the fetus, fatty acid biosynthesis occurs at very low levels because the placenta is relatively permeable to free fatty acids, especially PUFAs. Maternal plasma triglycerides are taken up by the placenta where their intracellular hydrolysis facilitates the diffusion of fatty acids to the fetus [50]. By contrast, in advanced gestation, there is a gradual shift to de novo synthesis with an increase in CIC expression needed to transport the acetyl-CoA unit as a citrate outside mitochondria for fatty acid synthesis. Moreover, low expression of CIC in fetal tissue is also due to the higher concentration of PUFAs in fetal blood with respect to that of the mother. In fact, the human placenta membrane contains a fatty acid-binding protein, which selectively transports PUFAs [51]. In tissues such as brain, about half of total lipid content is composed of long-chain unsaturated fatty acids, of which arachidonic and docosahexaenoic acids are metabolically more important. PUFAs repress CIC gene expression via the SREBP-1 transcription factor [35]. It is not excluded that the CIC negative regulatory element ZNF224 is part of a more intricate DNA–protein interaction that plays a role in modulating CIC mRNA in fetal life.

As mentioned in the description of the role performed by Sp1, CpG methylation is a key factor in regulating CIC gene expression (Figure 2) [26]. Epigenetic regulation was demonstrated on different cell lines, HepG2, HEK293 and SK-N-SH, which express different amounts of CIC in relation to fatty acid biosynthesis activity. In hepatic cell lines, the CIC gene CpG island is hypomethylated and thus associated to Sp1, while histones are acetylated. This finding is based on the observation that CIC expression was affected neither by treatment with 5-Azacitidine (5-AzaC), an inhibitor of DNA methyltransferase, nor by treatment with tricostatin (TSA), an inhibitor of histone deacetylase. A CIC repressed status is present in neuroblastoma cells. ChIP assay revealed that CpG island methylation and H3 histone deacetylation inhibit the binding of Sp1. In these cells, demethylation of promoter due to exposure to 5-AzaC activates expression of CIC by 50%. However, simple demethylation is not sufficient to promote maximal expression, as long as chromatin condensation persists. Maximal induction of expression of the CIC gene in SK-N-SH cells requires the synergistic action of TSA and 5-AzaC, supporting the concept that alteration of the chromatin structure is necessary for the DNA demethylation process. This different regulation most likely reflects the metabolic role of CIC in different cells. In HepG2, CIC plays an important role in fatty acid metabolismand requires constant activation. By contrast, CIC is not so important to the metabolism of neuroblastoma cells and is therefore silenced by the epigenetic mechanism and other mechanisms and factors not yet identifed.

Two hormones are implicated in the control of CIC expression. One such hormone is insulin, an anabolic hormone that controls blood glucose levels and energy homeostasis and is involved in the pathology of metabolic disorders. Studies performed by Kaplan demonstrated that diabetic rats show reduced CIC activity, whereas the general mitochondrial function is not altered [52]. The reduced CIC activity observed in the diabetic state can be ascribed to either transcriptional or post-transcriptional CIC gene regulation [53]. Treatment of diabetic rats with insulin reverses the reduction of CIC activity [54]. In agreement with this finding, insulin increases human CIC gene expression in the HepG2 cell line through involvement of SREBP-1 [35], and CIC plays a regulatory role in glucose-stimulated insulin secretion [32].

Another hormone that influences CIC expression is 3,3’,5-triiodo-L-thyronine (T3) that mantains lipid homeostasis via its effect on gene expression in target organs, mainly in liver and adipose tissue. CIC activity, mRNA and protein are enhanced in hyperthyroidism and reduced significantly during starvation [55]. Change in CIC expression is due to the action of an unidentified cis-acting element, most likely located on exon 8, which inhibits the splicing process [56]. In the context of lipid metabolism, T3 increases the expression of several genes involved in hepatic lipogenesis, including fatty acid synthase, acyl-CoA synthase, fatty acid transporter protein, and malic enzyme [57]. T3 also induces the transcription of genes involved in fatty acid oxidation, such as lipoprotein lipase, fatty acid-binding protein [57], and CPT-1 [58].

PPARα. An active binding site for PPARα is present in the CAC gene promoter at position −99/−80 bp (Figure 3) [62]. The promoter activity of the CAC gene PPRE was proven by transfecting HepG2 cells with the pGL3 basic-LUC vector containing the –334/+3 bp region of the CAC gene. Mutations in the PPARα binding sites abolished CAC activation. In agreement with the role performed in the fatty acid catabolism pathway, PPARα is also a strong activator of CAC gene expression. It is known that PPARα., a member of the nuclear receptor super family of ligand-activated transcription factors, has a central role in the regulation of many genes involved in cellular fatty acid uptake and oxidation and provides a molecular link between the major pathways of energy metabolism [63]. The role of PPARα in fatty acid β-oxidation is highlighted by the observation that fasting results in an enhanced load of fatty acids in the liver to be used as an energy source [5].

CAC expression is further increased by hypolipidemic drugs such as statins, fibrates and 9-cis-retinoic acid [62]. HepG2 cells transfected with the wild-type PPRE-containing the pGL3 basic construct and incubated with GW7647 and/or fluvastatin exhibited an increase in LUC activity. Co-treatment of cells with both drugs triggered a synergistic rise in LUC activity by 304%. 9-Cis-retinoic acid increased CAC gene promoter activity only in the presence of wild-type PPRE as compared to control. Combination treatment with GW7647 and 9-cis-retinoic acid resulted in a still greater stimulation of gene reporter activity [62]. These drugs act according to different mechanisms. Fibrates exert their effect by binding to PPARα [64], statins by inhibiting the Rho-signaling pathway [65], and retinoic acid by activating PPRE when bound to the PPARα–RXRα heterodimer [66]. Finally, the PKA pathway also activates CAC gene expression. In HepG2 cells transfected with wild-type PPRE-containing the pGL3 basic construct and treated with forskolin, alone or in combination with GW7647, gene reporter activity was increased by 49% and 138%, respectively. These findings are particularly important because patients affected by CAC deficiency, who present a mild phenotype with some residual activity [67,68,69,70,71], might benefit from treatment with statins and fibrates acting via stimulation of CAC gene expression.

Sp1 and FOXA2. The CAC gene proximal promoter also contains two binding sites for Sp1 and FOXA2 at positions −232/−2248 bp and −300/−290 bp, respectively [72]. The role of the FOXA and Sp1 sites in the regulation of CAC gene expression was investigated in HepG2, HEK293 and SK-N-SH cell lines. Transfection of these cells with a construct containing the CAC promoter region from −334/+3 bp upstream of the LUC gene-coding sequence showed that HepG2 cells exhibit a much higher gene reporter activity than HEK293 and SK-N-SH cells. Consistently, HepG2 cells exhibited a higher level of CAC transcript and CAC protein than the other two cell types. ChIP experiments indicated that both FOXA2 and Sp1 in HepG2 cells, and only Sp1 in HEK293 and SK-N-SH cells, are bound to the CAC proximal promoter. This finding indicates that, in the context of fatty acid β-oxidation, Sp1 alone is not sufficient to activate the transcriptional machinery efficiently in liver, and that a liver-specific factor is needed for a high activation of the CAC gene expression.

FOXA2 is a liver- and pancreas-specific transcriptional factor essential for glucose and lipid homeostasis [73]. Other genes encoding hepatic enzymes involved in metabolism during fasting and energy deprivation are transcriptionally regulated by FOXA2. Among these enzymes are CPT-1, hydroxyacyl-CoA dehydrogenase and lipoprotein lipase (lipid catabolism), phosphoenolpyruvate carboxykinase and glucose-6-phosphatase (gluconeogenesis), and hydroxy-3-methylglutaryl-CoA synthase 1 (ketogenesis) [74]. The presence of FOXA2 in liver could explain, at least in part, the differences in CAC expression levels between liver and other tissues [61]. However, it may be that other transcriptional factors, not yet identified, contribute to an efficient activation of the transcriptional machinery.

SRC-3. Using a comprehensive metabolomics-based analysis, it was found that steroid receptor coactivator-3 (SRC-3) [75], but not SRC-1 and SRC-2, is necessary for the proper transport and metabolism of long-chain fatty acids through a specific regulation of CAC gene expression [76]. The metabolomics analysis was confirmed by transfection experiments. A LUC construct containing the mouse CAC proximal promoter (−995/+133 bp) induced a robust activation of LUC activity in the muscle cell line, C2C12. Furthermore, ChIP performed in vitro on C2C12 cells and in vivo using skeletal muscle isolated from SRC-3^+/+ and SRC-3^−/− mice demonstrated that SRC-3 is bound to the CAC promoter. Mice lacking SRC-3 show a marked deficiency in CAC expression, which is accompanied by phenotypes that mimic the clinical symptoms of CAC deficiency in humans [77]. Metabolomics analysis of plasma from SRC-3^+/+ and SRC-3^−/− mice demonstrated that SRC-3 knock-out causes an increase in the level of 15 amino acids and an elevation of the urea cycle intermediates ornithine, citrulline, and arginine, together with an elevation of urea cycle enzymes. SRC-3^−/− mice also display a variety of cardiac abnormalities. Interestingly, loss of SRC-3 conferred no alterations in 12 other genes essential for fatty acid β-oxidation and ketogenesis, thereby suggesting that CAC is a specific target for SRC-3, at least in mouse muscle. Moreover, SRC-3^−/− mice display increased insulin sensitivity, as determined by the hyperinsulinemic/euglycemic clamp, which has been reported in other cases of impaired fatty acid oxidation [78].

Because SRC-3 ablation mimics all clinical symptoms of CAC deficiency, it is possible to hypothesize a dysfunctional mutation in the SRC-3 gene in patients who present symptoms of CAC deficiency but harbor no pathogenic mutation in the CAC gene. To test this hypothesis, York et al. [76] analyzed 24 patients who exhibited CAC deficient-like symptoms, but failed to show any mutations in the CAC gene. In these patients, analysis of the SRC-3 coding region and flanking intron regions revealed no deleterious mutation, although some patients revealed heterozygous missense variations of unknown clinical significance. Although a number of studies have identified metabolic roles for SRC-3 in glycogenolysis, gluconeogenesis, dietary fat absorption and lipid storage [75], its new role as a key transcriptional activator of the CAC gene may contribute to understanding some unexplained disorders of lipid homeostasis.

ERR. An active consensus sequence for the estrogen-related receptor (ERR) is present in the mouse CAC proximal promoter 808 bp upstream of the transcription start site [79]. This sequence binds ERRα both in vitro and in vivo, as shown by gel shift and ChIP assays. However, ERRα alone is insufficient to activate CAC gene expression since nuclear receptor activity depends on the presence of co-regulatory proteins, such as the members of the PGC-1 co-activator family [80]. In fact, C2C12 myoblast infection with adenovirus expressing ERRα and PGC-1α or 1β showed that CAC gene activation is only evident in the presence of PGC-1α or 1β, co-activators of ERRα. An inverse agonist of ERRα, XTC790, specifically blocks CAC activation by PGC-1β, but not 1α, in C2C12 cells. All these data are consistent with previous results obtained by analyzing the effect of ERRα loss on global gene expression in heart, including genes involved in fatty acid metabolism [81]. In fact, loss of ERRα function causes CAC down-regulation in SAOS2 cells [81], and ERR silencing by siRNA reduces expression of genes up-regulated by PGC-1α, such as CAC [81].