点击显示 收起
1 From the Minnesota Obesity Center, Minneapolis VA Medical Center, Minneapolis (ASL and CMK); the Geriatric, Research, Education, and Clinical Center, Minneapolis (CMK); the Department of Psychiatry, University of Minnesota, Minneapolis (ASL); the Department of Food Science & Nutrition, University of Minnesota, St Paul (ASL and CMK); the Neuropsychiatric Research Institute, Fargo, ND (BAG); and the Department of Neuroscience, University of North Dakota School of Medicine and Health Sciences, Fargo, ND (BAG).
2 Presented at the Sugars and Health Workshop, held in Washington, DC, September 18–20, 2002. Published proceedings edited by David R Lineback (University of Maryland, College Park) and Julie Miller Jones (College of St Catherine, St Paul). 3 Manuscript preparation supported by ILSI NA. Supported by the Department of Veterans Affairs and the National Institutes of Health. 4 Address reprint requests to AS Levine, Minnesota Obesity Center, 1 Veterans Drive (151), Minneapolis, MN 55417. E-mail: allenl{at}umn.edu.
ABSTRACT
The prevalence of obesity has increased dramatically in recent years in the United States, with similar patterns seen in several other countries. Although there are several potential explanations for this dramatic increase in obesity, dietary influences are a contributing factor. An inverse correlation between dietary sugar intake and body mass index has been reported, suggesting beneficial effects of carbohydrate intake on body mass index. In this review we discuss how sugars interact with regulatory neurochemicals in the brain to affect both energy intake and energy expenditure. These neurochemicals appear to be involved in dietary selection, and sugars and palatable substances affect neurochemical changes in the brain. For example, rats that drink sucrose solutions for 3 wk have major changes in neuronal activity in the limbic area of the brain, a region involved in pleasure and other emotions. We also investigate the relations between sucrose (and other sweet substances), drugs of abuse, and the mesolimbic dopaminergic system. The presence of sucrose in an animal’s cage can affect the animals desire to self-administer drugs of abuse. Also, an animal’s level of sucrose preference can predict its desire to self-administer cocaine. Such data suggest a relation between sweet taste and drug reward, although the relevance to humans is unclear. Finally, we address the influence of sugar on body weight control. For example, sucrose feeding for 2 wk decreases the efficiency of energy utilization and increases gene expression of uncoupling protein 3 in muscle, suggesting that sucrose may influence uncoupling protein 3 activity and contribute to changes in metabolic efficiency and thus regulation of body weight.
Key Words: WORDSSucrose • neuropeptides • dopamine • reward • energy expenditure • uncoupling proteins • substance abuse • carbohydrates
INTRODUCTION
Sugars provide energy and a pleasant taste. It should not be surprising, therefore, that the intake of sugars is influenced by 2 types of brain systems: those associated with the regulation of feeding and energy homeostasis and those associated with reward. During the past 3 decades, it has become clear that a host of neuromodulators are involved in the regulation of both energy and reward pathways. Many of these substances increase feeding (orexigenic agents) or decrease feeding (anorexic agents), and some also affect energy expenditure. In this review we focus on how sugars interact with regulatory neurochemicals in the brain to affect both energy intake and energy expenditure. We also examine the relations between sucrose (and other sweet substances), drugs of abuse, and the mesolimbic dopaminergic system. Finally, we discuss the influence of sugar on body weight regulation.
NEUROPEPTIDES AND REGULATION OF MACRONUTRIENT INTAKE
Soon after investigators reported that endogenous neuropeptides could alter food intake, there were reports that some peptides had a preferential effect on individual macronutrients. The classic methodology that Richter et al (1) used to study nutrient self-selection has been used to study the effects of biogenic amines and peptides on macronutrient intake. For example, opioid agonists increase fat intake to a greater extent than they do carbohydrate or protein intake (2–5). Another orexigenic agent, neuropeptide Y (NPY), increases carbohydrate intake more than it does fat or protein intake (6, 7). Galanin (8) and enterostatin (9) selectively increase or decrease fat intake. In fact, chronic consumption of fat is necessary for enterostatin to act as an anorexic agent (10). Thus, some data suggest that specific peptides may control the intake of single macronutrients. However, this idea is an oversimplification.
If one is interested in human or animal intake, one cannot speak about carbohydrate, fat, or protein as individual classes of compounds. This only works for the chemist. Most people do not select foods on the basis of whether they are carbohydrates, proteins, or fats. In fact, many laypeople and some scientists do not know the definition of a carbohydrate. They speak of carbohydrate craving and include foods such as candy and ice cream, which are as high in fat as in sugar content, and obesity is most associated with a craving for mixtures containing high amounts of fat and sugars (11). The macronutrient contents of hard candy and pasta are both primarily carbohydrates, yet their taste profiles are very different. Corn oil and lard are both 100% fat, yet their texture, taste, and preference profiles are not the same. Rats do not choose diets high in cornstarch, sucrose, or polycose equally when these diets are offered simultaneously although they may all contain the same percentage of energy derived from carbohydrates (12). To further complicate this issue, Tordoff (13) recently reported that rats given 5 bottles of sucrose solution and 1 bottle of water become more obese than do rats given 5 bottles of water and only 1 bottle of sucrose. This result, along with results with individual macronutrients, led Tordoff to conclude that the availability of food can negate the physiologic controls of ingestion.
In studies of peptides and macronutrients, it is important to consider baseline nutrient preferences. Evans and Vaccarino (14) deprived rats of either protein or carbohydrates and later offered them a choice between these 2 diets. They noted that morphine increased the consumption of the deprived diet. Gosnell et al (15) studied the relation between a rat’s diet preference and morphine’s effect on macronutrient selection. Rats were given both a high-carbohydrate and a high-fat diet and were then divided into carbohydrate, fat, or intermediate preferrers based on their daily intake patterns. The rats were then injected with morphine or saline and offered the 2 diets simultaneously. Four hours after the morphine injection, the rats that preferred fat ingested more fat, whereas those that preferred carbohydrates ate more carbohydrates. Morphine had no differential effect in the rats that did not have a clear preference for either diet.
Welch et al (16) also evaluated the interaction between baseline preference and drug-induced feeding. They studied the effect of food deprivation and various orexigenic agents on animals given individual macronutrients (fat, carbohydrates, and protein plus micronutrients) or complete diets high in fat or carbohydrates. Chronic food restriction (80% of ad libitum intake), 48-h food deprivation, and morphine each increased fat selection more than they increased carbohydrate or protein selection, whereas NPY and norepinephrine increased carbohydrate selection relative to fat and protein selection. Welch et al also found that baseline preference affected macronutrient selection, as indicated by covariate analyses. For example, fat intake was greater than carbohydrate and protein intakes in the chronically food-restricted rats; however, when means were adjusted for baseline macronutrient preference, the intakes of fat, carbohydrates, and protein were no longer different. Baseline preference significantly affected the effect of food deprivation and restriction, morphine, NPY, and norepinephrine on nutrient choices. Glass et al (17) found that baseline preference influenced food intake in food-deprived rats. They determined the baseline intakes of a high-carbohydrate and a high-fat diet, each of which was given to the rats for 3 d. They injected the rats with the opioid antagonist naloxone, which has anorexic activity, and determined the intake of either the preferred or nonpreferred diets. Extraordinarily low doses of naloxone (0.01 mg/kg, injected subcutaneously) significantly decreased preferred diet consumption, whereas doses as high as 3 mg/kg had no effect on the intake of the nonpreferred diet. Thus, baseline preference clearly affects feeding elicited by various means. Baseline preference is affected by various factors, including the animal’s diet history, genetics, strain, and energy state.
SUCROSE AND OPIOID-RELATED FEEDING
Sucrose appears to have an important relation with the opioidergic system. The opioid antagonist naloxone is more effective at reducing the intake of saccharin, sucrose, and saline than it is at reducing the intake of water or quinine solutions (18). Chronic infusion of the long-acting opioid antagonist naltrexone decreases intake more effectively in rats given chow plus a 32% sucrose solution than in those given chow alone (19). Yirmiya et al (20) found that opioid receptor–deficient mice (CXBK) had lower saccharin preferences than did control mice. Rockwood and Reid (21) noted that naloxone reduced the intake of a 10% sucrose solution in food-deprived and nondeprived sham-drinking rats (rats with a gastric fistula), indicating that the antidipsogenic actions of naloxone were not due to feedback from postabsorptional signals. Since this early study, others have found that sham intake of sucrose solutions is markedly decreased by naloxone. The intake pattern of sham-fed rats that were given a 10% sucrose solution and injected with naloxone was identical to that of rats that ingested a 5% sucrose solution without naloxone administration, suggesting a change in the perceived quality of the sucrose solution (22). Using a taste reactivity test, which measures the hedonic properties of tastants, Parker et al (23) found that naltrexone reduced the positive hedonic properties of sucrose solutions in rats. Lett (24) found that rats that consumed sucrose solutions had a stronger conditioned place preference associated with morphine. Lynch (25) reported that very low doses of naloxone (0.1 mg/kg) reduced the intake of saccharin solutions in nondeprived rats by 50%. Lynch also found that acquisition of preference for a saccharin solution was blocked by daily pretest naloxone administration. Philopena et al (26) found that the intake of sucrose (10% solution) depends on endogenous opioids as early as 10 d postnatally and that this opioid system operates when pups have no prior test experience. Shide and Blass (27) found that the development of a preference in 6-d-old rats for orange odor after it was paired with intraoral sucrose infusion was blocked by naloxone administration. Levine et al (28) studied the effect of peripheral naloxone administration on the intake of sweetened or unsweetened laboratory food pellets after various periods of food deprivation and restriction and in schedule-fed rats. They found that in all the groups of rats, naloxone decreased the intake of sweet food more effectively than it decreased the intake of unsweetened food. Compared with the intake of chronically deprived rats (rats having 85% of the body weight of rats that were given food ad libitum) that ingested diets containing cornstarch or polycose as the carbohydrate source, the intake of rats that ingested diets high in sucrose was also preferentially decreased by naloxone administration (12). Also, the intake of a diet high in fat and sucrose was found to increase gene expression of the opioid dynorphin in the arcuate nucleus and peptide content in the paraventricular nucleus of the hypothalamus (29). This diet did not alter NPY gene expression in the arcuate nucleus (30).
It is possible that blockade of the opioid receptor simply alters the taste of sweet substances. However, naloxone failed to affect sucrose discrimination in rats taught to discriminate 10% sucrose from water in a 2-lever operant chamber. That is, rats continued to choose the appropriate sucrose lever even after injection of naloxone. This of course does not indicate that the solutions are still palatable but simply that they are still discriminable. The same was found in human subjects given naltrexone (another opioid antagonist) via a taste discrimination procedure (31). On the other hand, naltrexone was found to decrease the pleasantness of a sucrose solution in humans (32, 33), and naloxone reduced the intake of sweet, high-fat food in binge eaters but not in control subjects (11).
Methadone addicts seem to prefer foods containing large amounts of sugar (34, 35). The relation between opioids and pleasant taste was also studied in rats by using the taste reactivity test, a test that indicates positive or negative responses associated with the intake of a given tastant (23). Naltrexone reduces positive ingestive reactions in response to sucrose but fails to affect aversive reactions after infusion of a quinine solution. Thus, although sucrose is recognized after blockade of opioid receptors, it may no longer be reinforcing.
NEUROCHEMICAL CHANGES INDUCED BY SUGARS
Several laboratories have published data suggesting that ingestion of sweet tastants results in neurochemical changes indicative of a change in opioid- or dopamine-mediated responses. Pomonis et al (36) gave rats access to a 10% sucrose solution or water for 3 wk and injected the rats with 10 mg naloxone or saline/kg. The rats’ brains were subsequently analyzed for c-Fos immunoreactivity in limbic and autonomic regions in the forebrain and hindbrain. c-Fos is an early gene transcription factor, and an increase in c-Fos concentrations is thought to reflect neural activation. Pomonis et al found that c-Fos concentrations in the central nucleus of the amygdala after naloxone injection were elevated more in the rats that drank 10% sucrose than in those that drank water. This suggests that this area is activated when sucrose is ingested. Thus, the central nucleus of the amygdala may participate in the integration of gustatory, hedonic, and autonomic signals as they relate to sucrose consumption.
Colantuoni et al (37) daily gave rats a 25% glucose solution with laboratory food pellets for 12 h, followed by 12 h of food deprivation. The rats doubled their glucose intake in 10 d and developed a pattern of excessive intake in the first hour of daily access. After 30 d, receptor binding in several brain areas of the glucose-exposed animals was compared with that in controls fed food pellets. Dopamine D-1 receptor binding increased significantly in the accumbens core and shell. In contrast, D-2 binding decreased in the dorsal striatum. Binding to dopamine transporter increased in the midbrain. Opioid mu-1 receptor binding increased significantly in the cingulate cortex, hippocampus, locus coeruleus, and accumbens shell. These data indicate neurochemical and region-specific responses to glucose intake and suggest a complex interplay between dopamine and opioids in the neural response to sugar. Colantuoni et al (38) also noted behavior changes associated with opiate withdrawal, such as teeth chattering in rats injected with naloxone after chronic glucose imbibition. Thus, chronic ingestion of sugars by laboratory animals may result in a state that resembles mild opioid dependence.
SUCROSE AND REWARD
Because sucrose (and most other sweeteners) are rewarding, it is worthwhile to examine whether the mechanisms underlying sweet reward are the same as those mediating the rewarding effects of drugs of abuse, such as cocaine, opiates, and alcohol. Several approaches have been applied to this issue. One approach has been to determine whether individual differences in the preference and consumption of sucrose are related to differences in self-administration of drugs. If sucrose reward is mediated by the same mechanisms that mediate drug reward, it is not unreasonable to expect some degree of correlation between the intake and preference for sucrose and the subsequent self-administration of the drug. A second approach has been to measure changes in drug intake when sucrose (or other sweetener) is presented as a concurrent alternative to the drug or to examine changes in sucrose intake when access to drugs is terminated. Finally, microdialysis techniques have been used to determine whether the ingestion of sweet foods or fluids causes changes in brain dopamine metabolism that resemble those produced by most drugs of abuse. The results of these approaches are discussed below.
Sweet preference as a predictor of substance abuse
Several laboratories reported positive relations between the intake of saccharin solutions and the intake of alcohol (eg, 39–42). Stewart et al (43) reported that this relation also appears to extend to sucrose solutions because rat strains selectively bred for high alcohol intake consumed more sucrose solution (and less sodium chloride solution) than did rats bred for low alcohol intake. The similarity of findings with sucrose and saccharin suggest that the relation to alcohol intake is not due to the postingestive consequences of sucrose. According to Scinska et al (44), however, the taste of ethanol is complex and has both a bitter and a sweet component; Scinska et al suggest that an overlap in taste qualities may account for the relation between sweet taste preference and alcohol intake. Kampov-Polevoy et al (45), however, have speculated that this relation is due to a similarity in the reward mechanisms activated by sweet solutions and alcohol.
The results of studies with intravenous self-administration support the suggestion that sucrose and some drugs of abuse activate similar mechanisms. A positive relation was reported between the intake of sucrose or saccharin solutions and the self-administration of morphine, cocaine, and amphetamine (46–48). For tests of the latter 2 drugs, rats were habituated to sucrose by presenting granulated sucrose in daily 1-h sessions conducted for 7 d. In a final test, rats were given saline injections (a mild stressor) before sucrose presentation. Based on sucrose intake on this day, groups of low- and high-sucrose-consuming rats were formed. The rate of acquisition of cocaine self-administration was positively related to sucrose intake, such that high-sucrose feeders required fewer sessions than did low-sucrose feeders to achieve a criterion level of cocaine self-administration (48). In further testing beyond the acquisition period, however, the groups did not differ. The acquisition of amphetamine self-administration was also more rapid in high-sucrose feeders (47). In contrast with the results with cocaine, however, high-sucrose feeders continued to self-administer more amphetamine in additional testing beyond the period of acquisition.
Human studies on the predictive potential of sucrose intake or preference on subsequent drug intake are relatively scarce and obviously cannot be conducted in a manner that precisely parallels the design of animal studies. Because of the established genetic influence on alcoholism (eg, 49), several studies examined taste preferences in subjects at risk for the development of alcoholism. Kampov-Polevoy et al (45) reported that subjects with a paternal history of alcohol dependence had a greater preference for strong sucrose solutions than did subjects with no paternal history of alcohol dependence. On the other hand, Scinska et al (50) and Kranzler et al (51) found no difference in the liking of sweets between sons of male alcoholics and control subjects. Scinska et al (50) did, however, note some differences in sensitivity to or preferences for salty and sour tastes. Therefore, the status of sweet taste preference as a marker for alcoholism is unclear, and we are not aware of any human studies in which sucrose preferences have been studied as potential markers for any other type of substance abuse.
Concurrent access to sweetened solutions and drugs of abuse
In rats, the intake of alcohol solutions with a concentration 6% is low (52), and sucrose is often added to alcohol solutions to facilitate intake (eg, 53). However, when sucrose is provided as an alternative to alcohol, alcohol intake is reduced (53, 54). This effect is not specific to sucrose, because reductions in alcohol intake have also been obtained by the provision of saccharin and fat (54, 55). Given the differences between fat, sucrose, and saccharin in terms of caloric value and postingestive consequences, one may speculate that palatability is the salient factor in causing a reduction in intake. The effect is also not limited to alcohol, because the availability of palatable substances has been shown to reduce the intakes of orally self-administered phencyclidine in monkeys (56) and amphetamine in rats (57). The intravenous self-administration of cocaine is reduced by the availability of a glucose or saccharin solution (58), and the acquisition of cocaine self-administration is slowed by concurrent access to this solution (59). In a large epidemiologic study by Colditz et al (60), a significant negative correlation between alcohol intake and sucrose intake was noted. In a study of alcohol and nutrient intake by subjects on an ocean-going oil tanker (3 crew members and 3 students), the percentage of total energy intake derived from alcohol and the percentage derived from protein and fat were significantly correlated (61). Overall, the percentages of energy derived from carbohydrates and alcohol were not correlated. However, when the intake of added sugars (defined as sugars added to beverages and sweets or candies) was considered separately, it was significantly negatively correlated with energy intake from alcohol. These types of studies suggest that, to some degree and for some people, sugars (and in some cases other palatable substances) may serve as a substitute for a drug.
The relation between sweet preference and opiate use by humans does not appear to follow the inverse relation that is observed with alcohol. Nonabstinent opiate users have been reported to have increased preferences for and intakes of sweets (35, 62). However, opiate availability may be more intermittent than is alcohol availability, such that concurrent access to opiates and sweets occurs less frequently. Weiss (63) found that opiate addicts generally reported greater craving and eating of sweets before drug use than after drug use.
Sweet preferences on termination of drug access
A somewhat different relation between the intake of sweets and drug intake has been observed when access to the drug is prevented. Carroll et al (58) reported that in rats that had concurrent access to intravenous cocaine and a glucose plus saccharin (G+S) solution, replacing the cocaine with saline caused a decreased intake of the G+S solution. However, in rats with only a short history of cocaine-reinforced responding and in rats that had not acquired a high rate of G+S-reinforced responding, termination of cocaine access caused an increase in G+S-reinforced responding. After repeated saline substitutions, G+S intake decreased on termination of cocaine availability. Similarly, the intake of and preference for saccharin solution decrease in rats for up to 6 d after the beginning of morphine withdrawal (64). It would be of interest to determine whether increases in sweet preference develop over longer periods of drug deprivation.
In humans, there is some indication that preference for sweets (or the intake of sweets) increases after termination of drug use. Compared with control subjects, detoxified alcoholics were recently reported to have an increased preference for concentrated sucrose solutions (45). As mentioned above, men with a family history of alcoholism also show this increased preference. Kampov-Polevoy et al (45) suggest that family history and alcoholic status operate as independent factors on preference for sweets. However, negative results were also reported for both factors in relation to preference for sweets (51, 65). In abstinent alcoholics, the duration of sobriety after entering an outpatient treatment program was related to the amount of sugar added to beverages (66). Patients who maintained sobriety for > 30 d used 2.5 times as much sugar per cup of beverage (coffee and tea) than did those who maintained sobriety for < 30 d. This increase in sugar intake was observed in the initial interview (on entering the program) as well as in the last interview with the patients (closer to the time of relapse). This finding suggests that increased sucrose intake was helpful in maintaining sobriety. The use of sweets to reduce the urge to drink is also recommended in literature designed to help those trying to reduce or quit drinking (67).
EFFECTS OF SWEET SUBSTANCES ON THE MESOLIMBIC DOPAMINERGIC SYSTEM
The evidence reviewed above suggests a relation between sweet taste and drugs of abuse. Under certain laboratory conditions, a high preference for or intake of sweet-tasting substances can predict subsequent drug use, and the intake of these substances may modify the amount of drug self-administration. These relations may be consistent with the possibility that the rewarding effects of sweets and drugs of abuse are mediated by similar or overlapping mechanisms. The mesolimbic dopaminergic system is thought to play an important role in mediating the rewarding and incentive or motivational properties of drugs (68, 69). One effect common to many drugs of abuse is an increase in extracellular dopamine in the nucleus accumbens (69). The ingestion of food can also cause an increase in dopamine release in the nucleus accumbens, although, as pointed out by Wise (70), the effects are not as large as those produced by cocaine, heroin, or amphetamine.
Martel and Fantino (71) reported that the ingestion of a palatable food caused a greater increase in dopamine release than did the ingestion of standard laboratory food. However, a subsequent study indicated that this difference may have been due in part to differences in the amount consumed (72). When ingested in solution (73) or in granulated form (74), sucrose was also shown to increase dopamine release in the nucleus accumbens. Interestingly, increased dopamine release in response to granulated sugar was only observed in rats classified as high-sugar feeders; rats classified as low-sugar feeders did not have increased dopamine release (74). This difference, like that observed in the comparison between palatable food and usual laboratory food (71), may have been partially due to differences in intake during the microdialysis sessions. Similarly, the response to the ingestion of sucrose solution was compared with the response to water intake, and the different amounts of water and sucrose ingested may have been a factor (73). Nevertheless, these studies indicate some degree of activation by the ingestion of palatable food of the same system thought to be important in mediating drug reward. The observations that dopaminergic antagonists cause a reduction in the intake of sucrose solutions are also consistent with this view (75, 76).
THE ROLE OF NEUROPEPTIDES ON SUGAR-INDUCED CHANGES IN ENERGY EXPENDITURE
Carbohydrates have differential effects on energy expenditure (77, 78), and as the above sections illustrate, several neuropeptides influence carbohydrate intake. Furthermore, sucrose feeding influences neuronal activation in brain areas involved in the regulation of energy expenditure (36). Although there is a paucity of data showing the relations between dietary carbohydrates, the specific activity of brain neuropeptide systems, and energy expenditure, differences in thermogenesis after carbohydrate feeding may be attributable to changes in neuropeptide activity. For example, compared with rats that ingested a high-fat diet, rats that ingested a high-carbohydrate diet (65% of total energy intake) had greater NPY gene expression in the arcuate nucleus (79). This change in NPY gene expression in rats was accompanied by a marked suppression of the expression of uncoupling protein 1 (UCP1) in brown adipose tissue (present in small amounts in humans), which indicates a decreased capacity for thermogenesis after high-carbohydrate feeding. Although this study did not directly show that NPY influenced the capacity for thermogenesis, NPY was previously shown to suppress UCP1 activity in brown adipose tissue (80). The mechanism underlying these changes in UCP activity is thought to involve the sympathetic nervous system, which is integral in mediating thermogenic responses. Bray (81, 82) showed that several neuropeptides that regulate feeding inversely affect the activity of the sympathetic nervous system. For example, NPY and ß-endorphin, which are neuropeptides that increase feeding, decrease sympathetic nervous system activation of brown adipose tissue (83, 84). Conversely, corticotrophin releasing hormone (CRH) and glucagon, which decrease feeding, result in activation of the sympathetic nervous system (85, 86). By influencing both energy intake and thermogenic potential, these neuropeptides can have powerful effects on energy balance.
Although the importance of UCP1 to human energy balance is unknown, several additional UCPs (UCP2, UCP3, UCP4, and UCP5, which is also referred to as brain mitochondrial carrier) were recently identified. These UCPs, like UCP1, have uncoupling capability and thus the potential to influence thermogenesis (87, 88). This uncoupling activity decreases the efficiency of energy use, and because an increased metabolic rate may result in a decreased propensity to gain weight (89), alterations in the activities of UCPs may have an important effect on energy balance. The tissue distribution of each UCP is distinct, which may reflect separate thermoregulatory roles for the UCPs. UCP4 and UCP5 are located primarily in the brain (90, 91), whereas the densest concentrations of UCP1, UCP2, and UCP3 are in brown fat, white fat, and muscle, respectively (87, 88, 92, 93). Of particular interest is the presence of UCP activity in muscle, which constitutes a large proportion of body mass and thus may contribute substantially to thermogenesis. Recent data from our laboratory indicate that sucrose feeding of rats for 2 wk decreases the efficiency of energy use (Figure 1) and increases UCP3 gene expression in muscle (Figure 2), suggesting that sucrose may influence UCP3 activity and contribute to changes in fuel efficiency. As noted in other studies, high-carbohydrate feeding appears to decrease the capacity for thermogenesis in brown adipose tissue by decreasing UCP1 activity (76). However, all the UCPs in several tissues are probably affected by sucrose (or other carbohydrates), and it is the sum total of activity of these UCPs that contributes to energy efficiency.
FIGURE 1. . Feed efficiency [change in body weight per energy intake (BW/EI)] in response to 2 wk of sucrose [10% (wt:vol) solution] or water feeding in rats (n = 10 per group). The rats had unlimited access to food. Body weight and intakes of food and fluid were measured every 2–3 d. *Significantly different from the group receiving water, P < 0.01.
FIGURE 2. . Effect of 2 wk of sucrose [10% (wt:vol) solution] or water consumption on uncoupling protein 3 (UCP3) gene expression in the biceps femoris muscle in rats (n = 10 per group). The rats had unlimited access to food. The rats were killed at the end of the 2-wk period, muscle tissue was dissected out, and UCP3 gene expression was determined by hybridization with a random primer–labeled complementary DNA probe as previously described (95–97). *Significantly different from the group receiving water, P < 0.05. mRNA, messenger RNA; OD, optical density.
Differences between the central regulation of the UCPs exist, and it is likely that differences also occur in the dietary regulation of the UCPs. Data from our laboratory indicate that NPY, cocaine-amphetamine–related transcript, leptin, and urocortin influence gene expression of the UCPs (80, 94–99) (Table 1). These changes in UCP gene expression may reflect changes in the thermogenic potential of the tissues and thus may affect overall energy expenditure. However, orexin A, a neuropeptide that stimulates feeding, had no effect on UCP gene expression (Table 1), which is surprising because orexin A stimulates the sympathetic nervous system (100, 101). With the exception of NPY (79), these neuropeptides have not been studied for their response to carbohydrate feeding in concert with concurrent effects on energy expenditure. However, what can be gleaned from the available data is that sucrose effects on energy expenditure may be mediated at least in part by effects on the sympathetic nervous system and UCPs.
View this table:
TABLE 1 . Effect of brain infusion of several neuropeptides on feeding and expression of uncoupling proteins (UCPs) in peripheral tissues in rats1
A recent study showed that sucrose feeding reverses the effects of adrenalectomy on CRH gene expression in the hypothalamic paraventricular nucleus and the central nucleus of the amygdala (102). Although the mechanism by which dietary sucrose induces these changes is unclear, these data suggest that dietary sucrose influences brain mechanisms regulating CRH expression, which could affect feeding behavior, stress, and anxiety. This same group of investigators has shown that voluntary sucrose overfeeding reverses the metabolic effects of adrenalectomy, suggesting that the metabolic effects of sucrose may substitute for those normally regulated by corticosteroids (103), which are under CRH control. These studies showing changes in neuropeptide activity after sucrose ingestion, together with data showing that these same neuropeptides may influence energy expenditure, indicate that carbohydrate feeding may influence energy expenditure through changes in central neuropeptide activity.
HOW SUGARS MIGHT AFFECT BODY WEIGHT CONTROL
The prevalence of obesity has increased dramatically in recent years in the United States (104), with similar patterns seen in several other countries (105, 106). Although several potential explanations exist for this dramatic increase in obesity, health behaviors, including diet and exercise habits, are of major importance (107–110). Acculturation into US society increases the prevalence of obesity, and a westernized diet is a major factor in the increased obesity observed (111). The relation between carbohydrate intake and obesity is unclear. An inverse correlation between the intake of sugars and body mass index was reported (112), suggesting a beneficial effect of carbohydrates on body mass index. However, in rats, excessive sucrose consumption resulted in hyperinsulinemia, hyperlipidemia, and excess abdominal fat (113, 114). In a 4-y prospective study, weight gain was not significantly influenced by dietary composition but rather by total energy intake (115).
Several human studies have investigated the differences between macronutrients in the thermic response to meals. The thermic response to high-carbohydrate meals (60% of total energy intake) is greater than that to high-fat meals (78, 116), and the thermogenic effect of sucrose is significantly greater than that of glucose (77). These data indicate enhanced thermogenesis after high-carbohydrate feeding, a finding also reported in animal studies (113). Differences in thermogenesis after ingestion of sucrose or glucose are blunted in obese persons (117), suggesting that this may be a factor contributing to their obesity.
More recent large-scale trials indicated that high-carbohydrate feeding is associated with lower body mass index and may have minor yet beneficial effects on body weight (89, 118–120). A 6-mo controlled trial of diets high in sugars and starch showed that, relative to control diets, diets high in polysaccharides resulted in moderate weight loss (2.4 kg) in overweight subjects (120). Subjects who consumed a diet high in simple carbohydrates also lost weight (1.9 kg), although the decrease was more moderate. In contrast with the results of previous studies, no significant changes in blood lipid profile occurred in either group. In a smaller study of patients with metabolic syndrome, which is characterized in part by abnormalities in blood lipids, a diet high in polysaccharides led to weight loss (4.7 kg) and a small improvement in serum cholesterol after 6 mo (119). As in the Saris study (120), weight loss in subjects who consumed a diet consisting mainly of sugars was lower in magnitude, suggesting that for weight loss, a diet containing mainly polysaccharides is more effective than is a diet containing carbohydrates in the form of sugars.
CONCLUSIONS AND FUTURE DIRECTIONS
In summary, ingestion of sugars induces neurochemical changes in brain areas involved in reward and energy. These effects on reward pathways may have implications for the prediction and treatment of substance abuse. However, studies related to substance abuse, intake of sugars, and reward pathways have not been conducted in humans. Brain imaging studies and more complete behavioral studies are needed to investigate the relation between the intake of sugars and reward and dependence in humans.
Most studies have focused on the relation between sucrose or glucose and reward. Studies are needed that compare and contrast the effect of sucrose, glucose, fructose, and saccharin on reward pathways. In addition, it would be useful to study preferred diets in further detail, ie, to include diets containing fat and sweet substances. Such data are needed to determine whether the proposed relation of sucrose and other sugars to reward and dependence is unique to sugars or can be generalized to all highly palatable diets.
Carbohydrate feeding may influence energy expenditure through changes in central neuropeptide activity. Data suggest that a diet rich in carbohydrates, especially polysaccharides, is beneficial for weight loss. Further studies are needed that directly examine the relations between neuroregulators, intake of sugars, thermogenesis, and body weight regulation.
ACKNOWLEDGMENTS
ASL is a member of the Scientific Advisory Board of the Dannon Foundation. There were no other potential sources of conflicts of financial or personal interest.
REFERENCES