Fatty Acid Metabolism, the Central Nervous System, and Feeding
Department of Neuroscience, 1006B Preclinical Teaching Building, Johns Hopkins University School of Medicine, 725 North Wolfe Street, Baltimore, MD 21205. E-mail: email@example.com
A potential role for fatty acid metabolism in the regulation of energy balance in the brain or in the periphery has been considered only recently. Fatty acid synthase (FAS) catalyzes the synthesis of long-chain fatty acids, whereas the breakdown of fatty acids by β-oxidation is regulated by carnitine palmitoyltransferase-1, the rate-limiting enzyme for the entry of fatty acids into the mitochondria for oxidation. While the question of the physiological role of fatty acid metabolism remains to be resolved, studies indicate that inhibition of FAS or stimulation of carnitine palmitoyltransferase-1 using cerulenin or synthetic FAS inhibitors reduces food intake and incurs profound and reversible weight loss. Several hypotheses regarding the mechanisms by which these small molecules mediate their effects have been entertained. Centrally, these compounds alter the expression of hypothalamic neuropeptides, generally reducing the expression of orexigenic peptides. Whether through central, peripheral, or combined central and peripheral mechanisms, these compounds also increase energy consumption to augment weight loss. In vitro and in vivo studies indicate that at least part of C75's effects is mediated by modulation of adenosine monophosphate-activated protein kinase, a member of an energy-sensing kinase family. These compounds, with chronic treatment, also alter gene expression peripherally to favor a state of enhanced energy consumption. Together, these effects raise the possibility that pharmacological alterations in fatty acid synthesis/degradation may serve as a target for obesity therapeutics.
Obesity has emerged as a worldwide health problem and is, therefore, a major focus for therapeutic intervention (1). The increase in the prevalence of obesity is attributed to multiple factors, including decreased physical activity, poor dietary habits, and an overall increase in caloric intake (2, 3, 4, 5, 6). These factors no doubt combine to dysregulate normal homeostatic mechanisms that control the balance between food intake and energy expenditure. Given the multifactorial causes for the epidemic of obesity, it seems important to view obesity as an overall disorder of energy balance that results from a disarray of both appetite regulation and energy metabolism, as has been proposed by investigators in the field (7, 8, 9). As the regulation of energy balance is central to survival of the organism and of the species, energy balance is unlikely to be relegated to one signaling system.
Indeed, we are beginning to appreciate the redundancy of signaling systems that exist to control energy balance, a topic considered in many recent reviews, and a concept that must be taken into account when considering weight loss therapies (1, 7, 10, 11, 12). Our growing understanding of the multiple organs and signaling pathways involved is a prerequisite to designing rational therapies for obesity but also requires us to acknowledge that targeting one signaling system may have limited use, as other pathways may rapidly restore a state of dysregulation once it has been established (13). Here, we consider studies that unfolded from in vivo observations that modulation of fatty acid metabolism can alter food intake and decrease weight (14, 15, 16, 17). These studies may provide us with clues as to pathways that impinge on the regulation of energy balance at a proximal point and, thus, may afford therapeutic opportunities for obesity.
Metabolic Pathways of Fatty Acid Synthesis and Degradation/Oxidation
The fatty acid synthase (FAS)1 pathway is a unique pathway responsible for the de novo biosynthesis of the fatty acid palmitate in the cytosol with the aid of a fatty acid synthase enzyme, FAS. Previously, this pathway was considered to be of little importance in man because of its low activity in the setting of high-fat diets; however, this is no longer the case (18). In mammals, FAS is the product of a single gene that produces a 250-kDa protein that contains a phosphopantetheine arm to anchor the nascent fatty acyl chain (19). FAS executes a complex, sequential, seven-step, NADPH-dependent synthesis of the 16-carbon free fatty acid palmitate from acetyl-CoA and malonyl-CoA (20). FAS is active in states of energy surplus, because it requires 7 ATP and 14 NADPH per molecule of palmitate produced. The synthesis of malonyl-CoA from acetyl-CoA by acetyl-CoA carboxylase (ACC) is ATP dependent and is actually the regulated step in the pathway. FAS is regulated transcriptionally, whereas ACC is regulated transcriptionally and is subject to rapid regulation by phosphorylation (21) and by the presence of long-chain acyl-CoAs, which competitively inhibit ACC (22). There are significant differences in the tissue distribution of ACC isoforms (23). ACCα is mainly expressed in liver and adipose tissue, whereas ACCβ is highly expressed in heart and muscle.
During times of energy depletion, fatty acids are degraded in the mitochondria by β-oxidation. Carnitine palmitoyltransferase-1 (CPT-1) is the rate-limiting enzyme that facilitates the entry of fatty acids into the mitochondria for oxidation (24). To avoid futile cycling of newly synthesized fatty acids into the mitochondrion, CPT-1 activity is competitively inhibited by malonyl-CoA. Malonyl-CoA is, therefore, a molecule of interest, because it acts both as a substrate for FAS and as a regulator of the rate of fatty acid oxidation. This becomes important when considering the potential role of the fatty acid pathway in energy modulation. CPT-1 has three isoforms: L-CPT-1 (liver form/CPT-1α) is expressed in most tissues, including liver, kidney, lung, and heart (25), M-CPT-1 (muscle form/CPT-1β) is expressed in skeletal muscle, heart, and adipose tissue (25), and CPT-1C (CPT-1γ) is a recently reported brain specific isoform of unclear activity (26). As with other metabolic pathways, enzymes within the fatty acid pathway must shift from anabolic to catabolic states in response to changes in energy availability. Hence, FAS and CPT-1 function at a metabolic “cross-roads,” balanced between energy storage (anabolism) and consumption of stored energy (catabolism). This suggests that activity through the fatty acid synthesis pathway reflects the energetic state of the cell.
Effects of Fatty Acid Metabolism on Appetite and Weight
The concept that fatty acid metabolism might serve as a target for the modulation of food intake originated in studies of pathways that were up-regulated in human cancers. FAS was found to be highly expressed in many common human tumors (27). Cerulenin ([2S, 3R]2,3-epoxy-4-oxo-7E, 10E-dodecadienamide), a natural product FAS inhibitor of Cephalosporium ceruleans (28), served as a model for inhibitor design. Early studies showed that cerulenin was selectively cytotoxic to human cancers in vitro, whereas it did not affect non-malignant cells (29). However, the epoxide structure of cerulenin made it less applicable to in vivo studies, prompting the design of synthetic small molecule FAS inhibitors. Because of its rather unique mechanism of action, investigators speculated that it might be possible to design synthetic FAS inhibitors with relatively specific actions. Among the seven moieties of FAS, the β-ketoacyl synthase moiety of FAS that joins acetate and malonate was chosen (27, 30). C75, an α-methylene-γ-butyrolactone, interferes with the binding of malonyl-CoA to the active site of FAS, although other studies indicate it may interact with other FAS components (31). C75 was the first such FAS inhibitor tested in vivo (32). Unexpectedly, its dose-limiting toxicity was reversible weight loss.
While one study has suggested that C75's actions reflect a non-specific mechanism of action (33), many investigators have explored the effects of C75 or cerulenin on food intake using rats, lean mice, diet-induced obese (DIO) mice, and genetic models of obesity, such as ob/ob mice (16, 34, 35, 36, 37, 38, 39, 40, 41). Their results have shown that these compounds induced profound weight loss and decreased food intake. The profound weight loss led to the identification of another site of C75 action, which was determined to be CPT-1 (38). Thus, not only did C75 block palmitate synthesis, but it potentially accelerated the rate of fatty acid oxidation. While intraperitoneal administration of C75 to rats may cause local irritation leading to an aversive response (41), this effect has not been observed in other species or with intracerebroventricular administration. Acute dosing studies have been followed up with studies that used a chronic treatment regimen of daily dosing to show that chronic treatment produces persistent weight loss over the treatment interval, without a rebound hyperphagia once treatment has been stopped (39, 42).
Mechanisms for Weight Loss and Appetite Reduction with Modulation of Fatty Acid Metabolism
The compelling biological effects of C75 treatment on weight loss and appetite control caused a rethinking of the role of fatty acid metabolism in energy regulation and perception. The first issue to consider was how an enzyme, best known for its role in lipogenic tissues such as liver and adipose, could effect changes in feeding behavior, which at some point must influence pathways in the brain, specifically the hypothalamus (43, 44, 45). The first problem was the central nervous system expression, or lack thereof, of FAS. A previous study localized FAS to non-neuronal cells in the brain (46). A subsequent study showed that FAS was actually expressed in neurons (14). FAS displayed a similar distribution in human hypothalamus, with particularly high levels of expression in the arcuate nucleus of the hypothalamus. A second issue was that in peripheral lipogenic tissues, FAS expression is regulated in response to diet, and after 24 hours of starvation, FAS mRNA is reduced (14); FAS down-regulation in the brain in a similar manner would raise doubt as to FAS being the target of C75 actions, especially in the chronic dosing paradigm. However, brain FAS is regulated quite differently than peripheral FAS. FAS message and protein are unaffected after 24 hours of starvation (14). Thus, FAS is present and serves as a target for C75. Moreover, these data suggest that FAS may play a different role in the brain than its canonical role of energy storage in the periphery.
The initial hypothesis was that interfering with fatty acid synthesis would reduce fat accumulation and, thus, would account for weight loss (36). While this was reasonable, the weight loss in C75-treated animals exceeded that of pair-fed controls, indicating that other mechanisms must be at work. Mechanistic studies have been complicated by the finding that C75 affects both FAS and CPT-1 activities. The relative contributions of FAS inhibition vs. CPT-1 stimulation to C75's effects have been debated. While progress has been made on the mechanisms of action of C75, it is hoped that newer compounds shown to be selective FAS inhibitors or CPT-1 stimulators will clarify the issues.
The neuronal expression of FAS in hypothalamic nuclei associated with the regulation of food intake allowed investigators to consider central nervous system mechanisms of action for C75. C75 reduced hypothalamus neuropeptide Y (NPY) message levels compared with fasted animals (14, 16). Radiolabeled C75 injected intraperitoneally was recovered in brain tissue with a half-life of ∼24 hours, showing that C75 crossed the “blood–brain barrier.” Administration of C75 intracerebroventricularly at a 100-fold lower dose than the dose used peripherally also led to weight loss, supporting a central nervous system mechanism of action. These data suggested that C75 acted, at least in part, centrally through FAS inhibition to prevent NPY production. Makimura et al. (34) also showed that cerulenin had effects similar to C75, and Lane and colleagues (35, 37, 47) have provided additional characterizations of the effects of C75 on food intake and neuropeptide expression in mice. Next, it became essential to examine the chronic effects of C75. Chronic C75 administration causes sustained weight loss and changes in expression of other neuropeptides (39, 48). Interestingly, this weight loss was accompanied by an increase in cocaine and amphetamine-related transcript expression (39, 42), and taken together, these data suggested that C75 could act centrally to alter feeding.
However, the molecular mechanism of action was unclear, although several possibilities were considered. Investigators initially proposed that hypothalamic malonyl-CoA was the mediator of C75's effects (16). This was logical, because malonyl-CoA is a key regulator of fatty acid metabolism in muscle (24, 49, 50, 51, 52). Therefore, FAS inhibition by C75 may cause malonyl-CoA levels to rise, thus decreasing CPT-1 activity and fatty acid oxidation. Fatty acyl-CoA might signal increased nutrient availability and decrease food intake. This hypothesis was supported by studies in which inhibition of CPT-1 or the addition of specific exogenous lipids (oleate) into the hypothalamus caused a decrease in food intake, with no dramatic weight loss (53).
Confounding the hypothesis that malonyl-CoA was the mediator of C75 actions was the recognition that CPT-1 was also a target of C75 actions. C75 stimulated the activity of CPT-1, thus increasing fatty acid oxidation (38, 42). Whole animal calorimetry revealed that C75-treated DIO mice displayed increased fatty acid oxidation, explaining the greater weight loss and energy production in C75-treated animals compared with pair-fed controls (38, 42). Etomoxir, an inhibitor of CPT-1, blocked the effects of C75 on fatty acid oxidation, energy expenditure, and weight loss in DIO mice. In vitro, C75 increased CPT-1 activity, fatty acid oxidation, and ATP levels in many cell types, including liver and skeletal muscle, even in the presence of high levels of malonyl-CoA (38, 54, 55). Thus, malonyl-CoA cannot mediate C75 actions, unless one hypothesizes that the regulation of CPT-1 in specific hypothalamic neurons is different than regulation in the periphery. The novel CPT-1γ recently reported (26) could have a different sensitivity to C75. A recent study has implicated malonyl-CoA as a mediator of anorectic effects in the central nervous system (56). The role of malonyl-CoA as a regulator of fatty acid disposition is most well studied in muscle (57, 58), where malonyl-CoA is instrumental in fuel switching, depending on fatty acid availability. In muscle, FAS expression is low, and, thus, malonyl-CoA levels can accumulate and affect CPT-1 activity; in tissues that contain higher levels of FAS expression, malonyl-CoA may be directly incorporated into fatty acids and, thus, are not available to regulate CPT-1. Thus, the role of malonyl-CoA may vary in different tissues, based on the presence of FAS.
While the modulation of FAS activity was a novel target for the regulation of food intake, modulation of fatty acid oxidation has been previously studied. Pharmacological inhibition of fatty acid oxidation at multiple enzyme targets showed that systemic inhibition of fatty acid oxidation did not inhibit, but stimulated, food intake (59, 60, 61, 62). The decreased food intake was transient, resolving with continued administration such that there were no significant changes in weight or carcass composition after 2 weeks of treatment intracerebroventricularly. However, these studies did not alter CPT-1 activity directly because the effects of CPT-1 stimulation have not been as thoroughly studied because of the lack of CPT-1 stimulators. Thus, the role of central fatty acid oxidation in the regulation of food intake remains unclear.
While the consequences of altering fatty acid oxidation in the central nervous system remain an open issue, the effects of C75 on peripheral fatty acid oxidation are more apparent. The dramatic decrease in hepatic fat accumulation in C75-treated DIO mice probably occurs through stimulation of CPT-1 (38). Most likely, this occurs through a direct stimulation of CPT-1 in the periphery to increase fatty acid oxidation in peripheral tissues, accounting for the increase in energy consumption and reduced fat mass. However, it is possible that central FAS inhibition may increase peripheral fatty acid oxidation, because central administration of cerulenin increased peripheral CPT-1 activity in muscle and liver. Cerulenin also increased core temperature (63), a phenomenon hypothesized to be caused by an alteration in sympathetic activity.
Molecular Mediators of the Effects of Fatty Acid Metabolism Alteration on Food Intake
After the initial observations showing that C75 affected CPT-1 and FAS activities, the question remained as to how a change in the flux of fatty acids in neurons in the hypothalamus could alter neuropeptide expression to signal satiety and possibly signal the periphery to increase peripheral fatty acid oxidation. If we consider the effect of simultaneous inhibition of FAS and stimulation of CPT-1, we realize that C75 “short-circuits” the normal protective regulatory mechanisms that prohibit futile fatty acid cycling. The overall effect is inhibition of fatty acid synthesis and promotion of fatty acid oxidation. The net effect is to increase energy availability in the form of ATP and increased availability of reducing equivalents in the form of NADPH. While fatty acid oxidation is not appreciated as a major pathway in neurons for generating energy, studies have shown that C75 does increase fatty acid oxidation in neurons, resulting in a dramatic increase in ATP levels in vitro (15). The hypothesis was put forward that C75 affects cellular energy balance by inhibiting FAS and/or stimulating CPT-1, which could be sensed in specific neurons within regions concerned with appetite or energy homeostasis to ultimately alter food intake, and perhaps, sympathetic outflow to the periphery.
Several candidates that might be affected by a change in energy availability were considered. One candidate is AMP-activated protein kinase (AMPK). AMPK is a known sensor of peripheral energy balance and a member of a metabolite-sensing kinase family (64). Increases in AMP/ATP ratio and changes in pH or redox status can lead to phosphorylation and activation of AMPK (65). Once activated, AMPK alters cellular pathways and gene expression to inhibit anabolic pathways and stimulate catabolic pathways to restore energy balance and ATP levels (66). AMPK contains an α-catalytic subunit and regulatory β and γ subunits (66). There are two α isoforms: α1 and α2 (67). In peripheral tissues, AMPK is regulated by exercise, starvation, and hypoxia (65). AMP activates AMPK kinase to phosphorylate AMPK on Thr172 on the α subunit (68).
Recent studies have examined the role of AMPK in neuronal energy metabolism. In primary neuronal cultures, C75 inhibits FAS and activates CPT-1 to alter ATP levels (15). Although ATP levels initially display a transient decrease, within 15 minutes, ATP levels significantly increase and remain increased for many hours. These fluctuations in ATP levels were accompanied by the anticipated changes in AMPK phosphorylation and activity: the initial drop in ATP was accompanied by an increase in AMPK phosphorylation and activity, whereas the prolonged increase in ATP correlated with a decrease in phosphorylation and AMPK activity. An in vivo role for hypothalamic AMPK as an effector of C75 actions was also studied (69). 5-aminoimidazole-4-carboxamide-1-β-d-ribofuranoside (AICAR) mimics the effects of AMP on AMPK activation (70), thus stimulating AMPK activity. While AICAR increased food intake, C75 and compound C, an inhibitor of AMPK, decreased food intake. C75 rapidly reduced levels of phosphorylated AMPKα in the hypothalamus, and AICAR was able to reverse both C75-induced anorexia and the decreased phosphorylated AMPKα levels in the hypothalamus. Of note, AMPK is highly localized to neurons in the hypothalamus, which provides a mechanism for C75's effects on weight and appetite.
Recent studies from several laboratories have revealed that hypothalamic AMPK indeed serves as a neuronal energy sensor for the regulation of food intake (69, 71, 72). Leptin was shown to inactivate hypothalamic AMPK, leading to anorexia (72), whereas a different effect was seen in skeletal muscle where leptin activated AMPK (73). Central administration of anorexigenic agents such as insulin, glucose, or MC3 and 4 agonists also inactivated AMPK (71).
Our studies suggest a mechanism by which C75 can affect feeding behavior, at least in part, by modulating AMPK activity (69). By inhibiting FAS and stimulating CPT-1, C75 increases ATP levels in hypothalamic neurons. This would signal a positive energy balance, inactivating AMPK, while contributing to a decrease in NPY expression. When energy stores are depleted or decreased by fasting or increased activity, AMPK is activated, which in turn activates several other downstream signals, including the CREB–NPY pathway to influence food intake. Under physiological conditions (normal feeding), it seems that there is relatively little change in the level of phosphorylated AMPK in the hypothalamus; a prolonged period of decreased food intake seems to be required before hypothalamic phosphorylated AMPK levels increase. Thus, AMPK may function as a “fuel sensor” in the central nervous system, as it does in peripheral tissues such as muscle (73, 74). Although several pathways have been proposed to function upstream of the AMPK pathway (75, 76), the contributions of these pathways remain to be elucidated.
Potential for Modulation of Fatty Acid Metabolism as a Target for Obesity Intervention
The sustained weight loss and decreased food intake seen with peripheral C75 treatment raises the possibility of FAS and CPT-1 as targets for obesity intervention. Recent studies indicate that chronic C75 treatment induces changes in gene expression to favor increased energy production and fatty acid oxidation use to create a more favorable phenotype (38, 77). Thus, the weight loss and increase in energy expenditure actually persist for weeks after C75 treatment is stopped. Although used as an index compound for many studies that investigated the roles of FAS and CPT-1 in the perception of neuronal energy balance, it is unlikely that C75 will be developed as a drug. More likely, it will be one of the hundreds of compounds with diverse molecular scaffolds that may go forward for drug development. A key issue is target specificity. A limitation of the current approach is that mice null for FAS are not available, because targeted deletion of the gene for FAS is lethal (78). Thus, the consequences of FAS inhibition cannot be studied using a genetic model. One may, however, question the relevance of chronic gene deletion to transient pharmacological enzyme inhibition, because many compensatory changes no doubt occur in genetic models, especially with regard to a life-sustaining function such as energy balance.
An appreciation of the complexity of organismal energy balance has caused us to rethink the strategies we need to pursue to address the daunting problem of obesity and obesity-associated diseases. Results reviewed here from many laboratories show that modulation of fatty acid metabolism, specifically FAS and CPT-1, can affect neuronal energy metabolism to influence energy perception, which act, at least partially, through AMPK. Understanding the consequences of altering fatty acid metabolism on neuronal energy balance and, thus, energy perception, could facilitate the understanding of mechanisms that regulate overall feeding behavior.
This work was supported by grants from the NIH NIDDK, NINDS, and NIDCD to G.V.R. and an NINDS F32 to L.E.L. Funding for work reviewed in this article was provided by FASgen. Under a licensing agreement between FASgen and the Johns Hopkins University, G.V.R. and L.E.L. are entitled to a share of royalty received by the University on sales of products related to work described in this article. G.V.R. serves without compensation on the Board of FASgen. G.V.R. has an interest in FASgen stock, which is subject to certain restrictions under University policy. The Johns Hopkins University, in accordance with its conflict of interest policies, is managing the terms of this arrangement.
Nonstandard abbreviations: FAS, fatty acid synthase; ACC, acetyl-CoA carboxylase; CPT-1, carnitine palmitoyltransferase-1; DIO, diet-induced obesity; NPY, neuropeptide Y; AMPK, AMP-activated protein kinase; AICAR, 5-aminoimidazole-4-carboxamide-1-β-d-ribofuranoside.