MANY FACTORS INVOLVED in the regulation of body

Transcription

1 /03/$15.00/0 The Journal of Clinical Endocrinology & Metabolism 88(3): Printed in U.S.A. Copyright 2003 by The Endocrine Society doi: /jc Modulation of Hunger by Plasma Glucose and Metformin BERND SCHULTES, KERSTIN M. OLTMANNS, WERNER KERN, HORST L. FEHM, JAN BORN, AND ACHIM PETERS Departments of Internal Medicine I and Neuroendocrinology, University of Luebeck, D Luebeck, Germany The plasma glucose concentration is a major short-term regulator of hunger and food intake. In patients with diabetes, therapies lowering plasma glucose are frequently associated with body weight gain, suggesting that lowered plasma glucose leads to increased feelings of hunger and food intake. However, as many physiological and symptomatic responses to low plasma glucose are attenuated after repeated episodes of hypoglycemia, this may also pertain to feelings of hunger. Here we tested whether the stimulatory effect of low plasma glucose on feelings of hunger is likewise reduced by repeated episodes of hypoglycemia. As metformin has been shown to reduce plasma glucose levels without increasing body weight and also to decrease food intake, we tested for possible interacting effects of this substance with hypoglycemia-induced hunger. Feelings of hunger were assessed by rating scales during 3 consecutive hypoglycemic clamps performed on 2 consecutive d in 15 normal weight men. Subjects were tested once while being treated with 850 mg metformin twice daily Abbreviations: VO 2,O 2 uptake; REE, resting energy expenditure; VCO 2, exhaled CO 2. and once while receiving placebo. Treatment was started 14 d before the clamp experiments and was performed in a random order and double-blind fashion. Hypoglycemia markedly enhanced feelings of hunger (P < 0.001). However, rated feelings of hunger on the first and last hypoglycemic clamps were comparable (P 0.304). Compared with placebo, metformin decreased feelings of hunger during hypoglycemia (P 0.015). This reduction was not associated with a decrease in posthypoglycemic food intake as measured by the number of cookies consumed after the last clamp (P 0.676). Data indicate that the stimulatory effect of low plasma glucose on hunger is not attenuated after repeated episodes of hypoglycemia, which implies that, in contrast to other symptoms, hunger is not subject to adaptive attenuation upon repeated hypoglycemia. Metformin attenuates hypoglycemia-induced hunger, but does not appear to influence posthypoglycemic food intake. (J Clin Endocrinol Metab 88: , 2003) MANY FACTORS INVOLVED in the regulation of body weight and the control of food intake have been identified during the last years. There is growing evidence that this regulation involves complex and redundant central nervous pathways controlling feelings of hunger and satiety (1, 2). In this system plasma glucose appears to play a prominent role in the short-term regulation of food intake, as small spontaneous transient declines in the plasma glucose concentration within the physiological range have been found to induce meal initiation in rats (3 5) and humans (6, 7). Clinically, the regulatory effects of plasma glucose on food intake may be relevant for the frequently observed gain in body weight during therapies that lower plasma glucose, e.g. sulfonylurea, thiazolidinediones, and insulin, in patients with type 2 diabetes (8, 9). Although weight gain during improvement of glycemic control in type 2 diabetic patients has been attributed to decreased basal metabolic rate and glucosuria (10), increased food intake induced by lowered plasma glucose levels may also contribute to the weight gain. Evidence for this view derives also from findings that plasma glucose thresholds for the initiation of hormonal counterregulation and hypoglycemic symptoms in type 2 diabetic patients are shifted into the normal range, i.e. euglycemia (11). Thus, hunger may occur at higher plasma glucose levels in those patients than in healthy subjects and may be increased during plasma glucose-lowering therapies. On the other hand, repeated episodes of low plasma glucose have been shown to decrease plasma glucose thresholds for hormonal counterregulation and symptoms (12 22), suggesting a rapid adaptation of the brain to such conditions. Whether a similar adaptation also exists in regard to feelings of hunger is not known. If it does, an increase in feelings of hunger and food intake during plasma glucose-lowering therapies would be highly unlikely. Here we tested whether the stimulatory effect of low plasma glucose on hunger is attenuated after repeated episodes of hypoglycemia. Hunger was assessed by rating scales during three consecutive episodes of mild insulin-induced hypoglycemia in normal weight, healthy men. Also, short-term memory recall of food-related words in relation to recall of words not related to food, which can be considered an implicit measure of hunger motivation (23), was tested before and during the first and third hypoglycemic clamps. Metformin, in contrast to other plasma glucose-lowering agents, improves glycemic control without increasing body weight (24, 25). Moreover, the substance has been shown to decrease body weight in obese subjects with (26 29) and without (30, 31) type 2 diabetes and to reduce feelings of hunger (26) and food intake (10, 26, 31 33). The fact that metformin lowers plasma glucose levels without increasing body weight suggests that the substance may interact with the modulation of hunger by the plasma glucose concentration. To test this additional hypothesis, experiments were performed twice in all subjects, once while taking metformin and another time while taking placebo. To assess whether the supposed influence of metformin on feelings of hunger was also associated with concurrent changes in food intake, the number of standard cookies consumed by the subjects within 20 min after the 1133

2 1134 J Clin Endocrinol Metab, March 2003, 88(3): Schultes et al. Hunger Modulation by Plasma Glucose and Metformin last hypoglycemic episode was counted. Also, serum leptin and insulin concentrations were measured to control for a potential influence of these hormones on feelings of hunger. Because changes in resting energy expenditure (REE), in addition to eating behavior, represent another mechanism essentially contributing to the regulation of body weight (34), we additionally performed indirect calorimetry measurements during the experiments. Subjects Subjects and Methods Fifteen normal weight, healthy men participated in the experiments (mean sem age, yr; body mass index, kg/m 2 ). Exclusion criteria were chronic or acute illness, current medication of any kind, smoking, alcohol or drug abuse, obesity (body mass index, 27 kg/m 2 ), or diabetes and hypertension in first degree relatives. Each subject gave written informed consent, and the study was approved by the local ethics committee. Study design and measurements The study was performed in a placebo-controlled, double-blind, cross-over design with the order of conditions balanced across subjects. In one condition, subjects received 850 mg metformin twice daily for a 16-d period; in the other, placebo was administered with a 4-wk interval between the two treatment periods. Compliance with drug treatment was assessed by repeated telephone calls and in an interview at the end of the treatment period. On the last 2 d of thetreatment period (d 15 and 16), during which hypoglycemic clamps were performed, the subject took the study medication in the presence of the investigator. The subjects underwent a total of three hypoglycemic clamps, two on d 15 and the last on d 16. Symptomatic responses to hypoglycemia were assessed every 15 min by a standard questionnaire throughout all hypoglycemic clamps. The questionnaire required ratings from 0 (none) to 9 (severe) of the following 14 symptoms: hunger, tingling, tremor, palpitations, sweating, itching, blurred vision, physical indisposition, weakness, difficulty in thinking, anxiety, faintness, dizziness, and irritability. During the baseline period as well as after 140 min of the first and third hypoglycemic clamps, a short-term memory test was performed, comprising the oral presentation of a list of words. The words belonged to three semantic categories: neutral words such as tree and field, foodrelated words such as ham and eggs, and emotional words such as mother and friend. The list included five words from each semantic category in random order, and parallel lists were used for repeated testing. One minute after presentation of the wordlist the subject was required to recall verbally all words he remembered within 1 min. The percentage of correctly recalled food-related words in relation to the total number of recalled words was calculated and considered as an implicit measure of hunger motivation (23). Serum insulin and leptin concentrations were measured every 15 min during the baseline periods and the initial 120-min interval of the first and third clamps. Determination of these hormones was performed as previously described (35). The responses of epinephrine, norepinephrine, ACTH, cortisol, GH, and glucagon to hypoglycemia were also measured. These data together with results for pooled symptomatic responses have been previously reported (36). Briefly, hormonal and symptomatic responses were attenuated during the third hypoglycemia compared with those during the first hypoglycemia. Metformin did not influence hormonal and autonomic symptom responses to repeated hypoglycemia, except for slightly enhancing the response of GH and neuroglycopenic symptoms. Indirect calorimetry was performed during the baseline periods and twice (at 60 and 120 min) during the first and third hypoglycemic clamps. Respiratory gas exchanges [O 2 uptake (VO 2 ) and exhaled CO 2 (VCO 2 )] was measured by placing a ventilated hood (MBM-200 Deltatrac II, Datex, Helsinki, Finland) over the head of the subject for 10 min. After the end of the third hypoglycemic clamp, standard cookies (Leibnitz Butterkeks, Bahlsen, Hanover, Germany), each weighing 5 g (3.90 g carbohydrates, 0.55 g fat, and 0.45 g protein) and containing 22.3 kcal, were offered. The number of cookies the subject consumed within the next 20 min was counted. During the baseline periods (d 15 and 16), subjects were weighed while wearing only underwear, and body weight was recorded to the nearest 0.1 kg. Hypoglycemic clamps On d 15 of the treatment period subjects reported to the medical research unit at 0800 h and then underwent two hypoglycemic clamps, one in the morning and another in the afternoon. Subjects were instructed not to eat after 2200 h on the preceding day and not to have breakfast on d 15. Also, until the end of the clamps subjects were not allowed to eat. The setting and procedure of the hypoglycemic clamps have previously been described in detail (36). Briefly, after a 30-min baseline period starting at 0900 h, plasma glucose was lowered to a plateau of 2.8 mmol/liter by infusion of insulin (H-insulin, Hoechst Marion Roussel, Inc., Frankfurt, Germany) at a rate of 1.5 mu/min kg. Ninety minutes after the start of insulin infusion (1100 h), plasma glucose was further decreased to a plateau of 2.5 mmol/liter for another 60 min by increasing the rate of insulin infusion to 2.0 mu/min kg. Plasma glucose levels were measured (by glucose analyzer, Beckman, Munich, Germany) every 5 min, and a variable infusion of 20% dextrose solution was adjusted to maintain plasma glucose plateaus. At 1200 h the insulin infusion stopped, and plasma glucose levels were returned to normal by dextrose infusion. At 1300 h the insulin infusion was started again at a rate of 2.0 mu/min kg to induce the second hypoglycemia. During this hypoglycemic clamp a plasma glucose plateau of 2.5 mmol/liter was achieved for 90 min (until 1500 h). After the clamp, the subject went home. Eating was allowed until 2200 h on this day. On d 16 of the treatment period subjects again reported to the research unit at 0800 h. The third clamp was performed in the same way as the first hypoglycemic clamp (starting at 0900 h). Throughout the experiments the subject was not informed about his plasma glucose concentration. Statistical analysis Data are reported as the mean sem. Baseline data for d 15 and 16 were analyzed by ANOVA for repeated measures, including the factors day for comparisons between d 15 and 16 and treatment for the effects of metformin vs. placebo. For pairwise comparisons a t test was used. Data assessed during the clamps were also analyzed by ANOVA including the factors hypo for the multiple measurements during the clamps, treatment for the effects of metformin vs. placebo, and day for differences between the first vs. third hypoglycemia. A sample size of 15 subjects provided a statistical power of 0.90 to detect a difference of 20% in the increase in hunger ratings and that in other symptoms between the first and third hypoglycemia (i.e. the day by hypo interaction term) as well as between the metformin and placebo condition (i.e. the treatment by hypo interaction term). P 0.05 was considered significant. Results Baseline values on d 15 and 16 Table 1 summarizes the baseline values on d 15 and 16. Baseline values on d 16 indicated that after the hypoglycemic clamps on d 15, subjects had gained, on the average, 0.5 kg compared with values before these clamps (P 0.001) regardless of whether they took placebo or metformin (P 0.857). Compared with d 15 values, baseline serum leptin concentrations were found to be increased by 63% in the metformin condition and by 35% higher in the placebo condition on d 16 (P 0.001). The difference in leptin increase between the two treatment conditions was not significant (P 0.367). Baseline plasma glucose (P 0.001) and serum insulin (P 0.001) concentrations were also higher on d 16 than on d 15. Again, these changes were not dependent on the treatment (P 0.25 for both variables). Baseline hunger ratings did not differ between d 15 and 16 (P 0.886) or between the two treatment conditions (P 0.355). The per-

3 Schultes et al. Hunger Modulation by Plasma Glucose and Metformin J Clin Endocrinol Metab, March 2003, 88(3): TABLE 1. Mean SEM body weight, self-rated feelings of hunger, percentage of recalled food-related words, plasma glucose concentration, serum concentrations of insulin and leptin, exhaled CO 2 (VCO 2 ), O 2 uptake (VO 2 ), respiratory quotient (RQ), and resting energy expenditure (REE) in the morning of d 15 and d 16 during the 16-d treatment with metformin (1700 mg/d) and placebo in normalweight men Placebo Metformin P value for Day 15 Day 16 Day 15 Day 16 day P value for treatment P value for day treatment Weight (kg) Hunger score Recalled food-related words (%) Glucose (mmol/liter) Insulin (pmol/liter) Leptin (ng/dl) VCO 2 (ml/min) VO 2 (ml/min) RQ REE (kcal/d) Values of the hunger rating and glucose, leptin, and insulin concentrations were averaged across the 30-min baseline periods before the hypoglycemic clamps were started. P values refer to results of ANOVA including the factors day (d 15 vs. 16) and treatment (placebo vs. metformin). centage of recalled food-related words in relation to the total number of recalled words was slightly reduced on d 16 compared with that on d 15, but the difference did not reach significance (P 0.095). The VCO 2 was higher on d 16 than ond15(p 0.023), and REE tended to be increased on d 16 (P 0.071). These changes did not depend on the treatment condition (P 0.25 for both variables). Overall, there was no difference in body weight between the two treatment conditions (P 0.497), but leptin levels were significantly lower during metformin treatment (P 0.047). Plasma glucose (P 0.045) and serum insulin (P 0.049) concentrations were also slightly lower in the metformin than in the placebo condition, but only during the baseline period on d 15 and not during the baseline period ond16(p and P 0.247, respectively). Plasma glucose and self-rated feelings of hunger during the clamps The time courses of plasma glucose concentration during the first and third clamps were closely comparable and did not differ between the metformin and placebo conditions (Fig. 1A). The total amount of glucose infused per kilogram body weight on d 15 ( vs g/kg; P 0.398) and on d 16 ( vs g/kg; P 0.177) did not differ between the metformin and placebo conditions. The decrease in plasma glucose during the first hypoglycemic clamp induced a marked rise in feelings of hunger for both treatment conditions (P 0.001; Fig. 1B). After 150 min of the first hypoglycemic clamp, the insulin infusion was stopped, and plasma glucose levels began to rise and reached euglycemic values ( 4.2 mmol/liter) within the next 15 min. Euglycemia was then maintained for another 45 min. This transitory normalization of plasma glucose levels was accompanied by a rapid decrease in rated hunger from (during the last 30 min of hypoglycemia) to (P 0.003) in the placebo condition and from to (P 0.033) in the metformin condition. With the second hypoglycemia, feelings of hunger increased again for both treatment conditions (P 0.001). During the third hypoglycemic clamp, feelings of hunger also markedly increased (P 0.001), and this increase was remarkably comparable with that during the first hypoglycemia (P 0.304). This was true for both the placebo and metformin conditions (P for day treatment interaction). It is notable that overall the hypoglycemia-induced increased in hunger ratings was significantly lower in the metformin than in the placebo condition (P for treatment hypo interaction). Self-rated symptom scores during the clamps Analysis of the other symptomatic responses to hypoglycemia revealed that self-rated tremor (P 0.001), palpitations (P 0.001), sweating (P 0.001), blurred vision (P 0.003), physical indisposition (P 0.013), weakness (P 0.012), difficulty in thinking (P 0.014), faintness (P 0.001), dizziness (P 0.007), and irritability (P 0.002), but not anxiety (P 0.309), were lower during the third than during the first hypoglycemia (Fig. 2). Metformin treatment had no influence on these changes (P 0.23 for all variables). Selfrated scores for tingling and itching did not change during the hypoglycemic clamps (P 0.17 for both), and there was not a difference between the first and third hypoglycemia (P 0.39 for both) or between the placebo and metformin conditions for these symptoms (P 0.25 for both; data not shown). Recall of food-related words and consumption of cookies The percentage of recalled food-related words in relation to the total number of recalled words significantly increased during the first (placebo, from to ; metformin, from to ) and third (placebo, from to ; metformin, from to ) hypoglycemia (P 0.001), but this increase did not differ between the first and third hypoglycemia (P 0.634) or between the placebo and metformin conditions (P 0.256; Fig. 3). The number of cookies the subjects consumed within 20 min after the third hypoglycemic clamp did not differ between the metformin and placebo conditions ( vs ; i.e vs kcal; P 0.676).

4 1136 J Clin Endocrinol Metab, March 2003, 88(3): Schultes et al. Hunger Modulation by Plasma Glucose and Metformin FIG. 1. Mean SEM plasma glucose concentrations (A) and hunger scores (B) during the 30-min baseline periods and subsequent hypoglycemic clamp periods on d 15 and 16 of the 16-d treatment with placebo (E) and metformin (F). Serum insulin and leptin concentrations during the clamps During the first 120 min of the first and third hypoglycemic clamps, serum levels of insulin rapidly increased, as expected (P 0.001; data not shown). However, concentrations did not differ between the first and third hypoglycemia (P 0.597) or between the two treatment conditions (P 0.488). Overall, serum leptin concentrations slightly decreased during the hypoglycemic clamps (P 0.021; Fig. 4A). Compared with the first hypoglycemic clamp, serum leptin concentrations were distinctly higher during the third hypoglycemic clamp (P 0.001). According to values during the baseline periods, serum leptin concentrations were also lower in the metformin than in the placebo condition during the hypoglycemic clamps (P 0.048). As illustrated in Fig. 4B, the marked differences in serum leptin concentration between the baseline periods of d 15 and 16 as well as between the placebo and metformin conditions were not associated with the concurrent differences in hunger ratings. Also, despite distinctly higher serum leptin concentrations during the third than the first hypoglycemic clamp (e.g. at 120 min), hunger ratings during these clamps did not differ (Fig. 4B). Indirect calorimetry during the clamps During the hypoglycemic clamps, VCO 2, VO 2, respiratory quotient, and REE significantly increased on the average by 23 4%, 6 4%, 7 5%, and 10 5%, respectively (P for all variables; Fig. 5). However, the increase in these variables did not differ between the first and third hypoglycemic clamps (P 0.18 for all variables) and was not influenced by the treatment condition (P 0.40 for all variables). Discussion The present data confirm that hypoglycemia is a strong stimulus of hunger. Unlike other responses to hypoglycemia, the hunger-stimulating effect of a low plasma glucose concentration was found not to be attenuated after repeated episodes of hypoglycemia. Thus, there appears to be no adaptation to low plasma glucose levels with regard to feelings of hunger. However, as stepped hypoglycemic clamps were not performed here, we cannot exclude that the glycemic threshold for feelings of hunger is affected by antecedent hypoglycemia. Furthermore, given the multiple statistical comparisons performed, it is conceivable that the absence of a statistically significant difference for the symptom of hunger was a random statistical event. However, the finding that hunger motivation (assessed by the short-term memory task) also did not differ between the first and third hypoglycemia strengthens the view that hunger, in fact, does not adapt to repeated episodes of hypoglycemia. Although metformin had no effect on body weight and baseline hunger ratings, it significantly reduced feelings of hunger during hypoglycemia. This finding suggests that metformin may interact with the hunger signal induced by low plasma glucose. To our knowledge the lack of an adaptation of feelings of hunger to repeated episodes of hypoglycemia represents a novel finding with important physiological and clinical im-

5 Schultes et al. Hunger Modulation by Plasma Glucose and Metformin J Clin Endocrinol Metab, March 2003, 88(3): FIG. 2. Mean SEM self-rated symptom scores during the 30-min baseline periods and subsequent first and third hypoglycemic clamps. While self-rated tremor (P 0.001), palpitations (P 0.001), sweating (P 0.001), blurred vision (P 0.003), physical indisposition (P 0.013), weakness (P 0.012), difficulty in thinking (P 0.014), faintness (P 0.001), dizziness (P 0.007), and irritability (P 0.002) were significant lower during the third than during the first hypoglycemia, there was no difference in self-rated anxiety (P 0.309) and hunger (P 0.304). E, Placebo condition, first hypoglycemia;, placebo condition, third hypoglycemia; F, metformin condition, first hypoglycemia; f, metformin condition, third hypoglycemia. plications. Previous studies (12 18) investigating symptomatic responses to recurrent hypoglycemia did not focus on feelings of hunger. Typically, data on hunger ratings were not analyzed separately but were pooled with rating scores of other hypoglycemic symptoms. Accordingly, the symptomatic response to hypoglycemia was attenuated after antecedent hypoglycemia in these studies. In agreement with this finding, in the present study changes in rating scores for the majority of symptoms were distinctly smaller during the third than during the first hypoglycemia. This attenuation of symptomatic responses is thought to represent an adaptation of the brain to hypoglycemia. However, the present data

6 1138 J Clin Endocrinol Metab, March 2003, 88(3): Schultes et al. Hunger Modulation by Plasma Glucose and Metformin FIG. 3. Mean SEM percentage of recalled food-related words in relation to the total number of recalled words in a short-term memory test during the baseline periods and the first and third hypoglycemic clamps at a target plasma glucose concentration of 2.5 mmol/liter. E, Placebo condition; F, metformin condition. indicate that a comparable adaptation does not occur with regard to hunger feelings. One explanation for this finding could derive from the fact that various areas within the brain are involved in mediating different symptomatic responses to hypoglycemia. While autonomic nervous system activation, mediating many symptoms (37), is triggered by neuroglycopenia within the ventromedial hypothalamus (38, 39), there is evidence that the hunger-promoting effect of hypoglycemia results at least in part from an increased expression of prepro-orexin within the lateral hypothalamic area (40, 41). Thus, it may be possible that those hypothalamic areas differ in their adaptive responses to recurrent hypoglycemia. The present finding sheds new light on the widespread assumption that hunger in response to hypoglycemia is mediated by an activation of the autonomic nervous system. This assumption is based on a previous study by Towler et al. (42) showing that panautonomic blockade by phentolamine, propanolol, and atropine sulfate, but not - and adrenergic blockade alone, significantly reduces feelings of hunger during hypoglycemia. It is notable in this study that panautomomic blockade also enhanced hypoglycemiainduced cognitive dysfunction. This finding in conjunction with others (43) indicates that atropine sulfate, which is known to cross the blood-brain barrier, exerts a direct effect on central nervous system function. Thus, the reduction of hunger produced by atropine during hypoglycemia may result from the blockade of central nervous muscarinic acetylcholine receptors and does not necessarily indicate an autonomic mediation of this symptom. Further evidence for an important role of central nervous muscarinic acetylcholine receptors in the regulation of hunger derives from the recent finding that mice lacking the M3 subtype of this receptor are extremely hypophagic (44). This background may provide a plausible explanation for the hypoglycemiainduced hunger being preserved despite a markedly attenuated autonomic activation after antecedent hypoglycemia. Interestingly, the recovery of normal plasma glucose levels after the first hypoglycemia was associated with a rapid reduction in hunger ratings occurring within 5 10 min after FIG. 4. A, Mean SEM serum leptin concentrations during the 30- min baseline periods and the initial 120 min of the first and third hypoglycemic clamps. E, Placebo condition, first hypoglycemia;, placebo condition, third hypoglycemia; F, metformin condition, first hypoglycemia; f, metformin condition, third hypoglycemia. B, Mean SEM serum leptin concentrations plotted against mean SEM hunger ratings during the baseline periods and at 120 min of the hypoglycemic clamps. Baseline:, placebo condition before the first hypoglycemia;, placebo condition before the third hypoglycemia; Œ, metformin condition before the first hypoglycemia;, metformin condition before the third hypoglycemia. At 120 min: E, placebo condition, first hypoglycemia;, placebo condition, third hypoglycemia; F, metformin condition, first hypoglycemia; f, metformin condition, third hypoglycemia. euglycemic levels were reached. Obviously, feelings of hunger respond rather quickly to changes in the plasma glucose concentration. However, although the close association between plasma glucose levels and hunger observed here is remarkable, it should also be considered that the induced changes in plasma glucose concentration were much larger than the spontaneous small declines in glucose that have previously been found to initiate meal intake (6, 7). Metformin had no influence on baseline hunger ratings, but significantly reduced feelings of hunger during hypoglycemia. This finding points to a modulating influence of metformin on the hypoglycemia-induced hunger signal. However, it should also be noted that the reduction of hunger by metformin was only moderate, and subjects still experienced strong feelings of hunger during hypoglycemia. Possibly, the stimulatory effect of hypoglycemia on hunger, being much stronger than the counteracting influence of metformin, explains the failure of metformin to concurrently decrease posthypoglycemic weight gain (after the first two

7 Schultes et al. Hunger Modulation by Plasma Glucose and Metformin J Clin Endocrinol Metab, March 2003, 88(3): FIG. 5. Mean SEM exhaled CO 2 (VCO 2 ), O 2 uptake (VO 2 ), respiratory quotient (RQ), and REE during the baseline periods and the first and third hypoglycemic clamps at a target plasma glucose concentrations of 2.8 mmol/liter (60 min) and 2.5 mmol/liter (120 min). E, Placebo condition; F, metformin condition. clamps) and food intake (after the third clamp). Taken together, it remains questionable whether the modest reduction in hypoglycemia-stimulated hunger after metformin administration contributes to the beneficial effects of the substance on body weight. On the other hand, it should be pointed out that the observed effect of metformin on hunger during hypoglycemia could be harmful in patients with diabetes, because it may decrease the drive to eat, thereby delaying the recovery of normoglycemia. Although metformin has been shown to reduce food intake and hunger in animals (32, 33) as well as in humans (10, 26, 31), the mechanisms of this action remain obscure. While an effect via inhibiting hypothalamic neuropeptide Y expression can be excluded (32), the influence of metformin on other neuropeptides involved in the regulation of food intake has not been assessed to date. The present finding hints at a possible influence of metformin on hunger-regulating pathways activated by hypoglycemia. As mentioned above, hypoglycemia induced hyperphagia results at least in part from an increased expression of prepro-orexin within the lateral hypothalamic areas (40, 41), whereas the hypothalamic neuropeptide Y does not appear to be involved (45). However, it is not clear whether metformin after systemic administration is able to cross the blood-brain barrier to reach the hypothalamus. In this context, it is highly interesting that hypoglycemia-induced expression of preproorexin is acutely inhibited by ingestion of food, suggesting that visceral signals, presumably mediated via the nucleus of the solitary tract, also modulate hypothalamic prepro-orexin expression (46, 47). As metformin is known to accumulate in the gastrointestinal tract (48), it appears possible that the substance likewise exerts its anorectic effect via visceral signals to the brain. Leptin levels on d 16 were distinctly increased compared with those on d 15. Furthermore, confirming previous results

8 1140 J Clin Endocrinol Metab, March 2003, 88(3): Schultes et al. Hunger Modulation by Plasma Glucose and Metformin (31, 49, 50), leptin levels were markedly lower during metformin than during placebo treatment. Thus, the four different hypoglycemic clamp situations of the present study took place in the presence of four different levels of leptinemia. This is of relevance on the background that leptin has been claimed to be essentially involved the regulation of hunger. Data from the present study may argue against this view, as the marked differences in serum leptin concentration were not paralleled by concurrent differences in feelings of hunger under either hypoglycemic conditions or euglycemic baseline conditions. In fact, previous data were not entirely consistent with the idea that leptin directly influences feelings of hunger. Although changes in leptin levels upon food restriction in some studies (51 54) were found to be correlated to feelings of hunger and satiety, in other studies (55 57) the leptin response to a meal as well as fasting morning leptin levels were not. In conjunction with those data, the present results suggest that leptin is involved in long-term, rather than short-term, regulation of food intake. In conclusion, these data show that feelings of hunger do not adapt to repeated occurrence of mild hypoglycemic states, which is in contrast to many other subjective and physiological responses to hypoglycemia. These results may be relevant with regard to the body weight gain observed during improvement of glycemic control in type 2 diabetic patients, because lowered plasma glucose levels probably enhance feelings of hunger, thereby increasing food intake. However, to confirm this hypothesis it remains to be shown that the glycemic threshold for hunger feelings is elevated in type 2 diabetic patients and that the reduced plasma glucose concentration in these patients is, in fact, associated with increased food intake. Furthermore, the present study shows that metformin modestly attenuates the hunger signal induced by a low plasma glucose concentration. However, this effect probably cannot account for the weight reduction seen after prolonged treatment with metformin. Acknowledgments We thank Christiane Zinke, Barbara Toschek, and Stefan Sopke for their expert and invaluable laboratory assistance, and Anja Otterbein for her organizational work. We gratefully thank Dr. Thomas Kohlmann for his statistical advice. Received September 24, Accepted December 11, Address all correspondence and requests for reprints to: Bernd Schultes, M.D., Department of Internal Medicine I, Medical University Luebeck, Ratzeburger Allee 160, D Luebeck, Germany. address: schultes@kfg.mu-luebeck.de. References 1. Schwartz MW, Woods SC, Porte Jr D, Seeley RJ, Baskin DG 2000 Central nervous system control of food intake. Nature 404: Woods SC, Seeley RJ, Porte Jr D, Schwartz MW 1998 Signals that regulate food intake and energy homeostasis. Science 280: Louis-Sylvestre J, Le Magnen J 1980 Fall in blood glucose level precedes meal onset in free-feeding rats. Neurosci Biobehav Rev 4(Suppl 1): Campfield LA, Brandon P, Smith FJ 1985 On-line continuous measurement of blood glucose and meal pattern in free-feeding rats: the role of glucose in meal initiation. Brain Res Bull 14: Campfield LA, Smith FJ 1986 Functional coupling between transient declines in blood glucose and feeding behavior: temporal relationships. Brain Res Bull 17: Melanson KJ, Westerterp-Plantenga MS, Saris WH, Smith FJ, Campfield LA 1999 Blood glucose patterns and appetite in time-blinded humans: carbohydrate versus fat. Am J Physiol 277:R337 R Melanson KJ, Westerterp-Plantenga MS, Campfield LA, Saris WH 1999 Appetite and blood glucose profiles in humans after glycogen-depleting exercise. J Appl Physiol 87: UK Prospective Diabetes Study (UKPDS) Group 1998 Intensive bloodglucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). Lancet 352: Schwartz S, Raskin P, Fonseca V, Graveline JF 1998 Effect of troglitazone in insulin-treated patients with type II diabetes mellitus. Troglitazone and Exogenous Insulin Study Group. N Engl J Med 338: Makimattila S, Nikkila K, Yki-Jarvinen H 1999 Causes of weight gain during insulin therapy with and without metformin in patients with type II diabetes mellitus. Diabetologia 42: Spyer G, Hattersley AT, Macdonald IA, Amiel S, MacLeod KM 2000 Hypoglycaemic counter-regulation at normal blood glucose concentrations in patients with well controlled type-2 diabetes. Lancet 356: Lingenfelser T, Renn W, Sommerwerck U, Jung MF, Buettner UW, Zaiser Kaschel H, Kaschel R, Eggstein M, Jakober B 1993 Compromised hormonal counterregulation, symptom awareness, and neurophysiological function after recurrent short-term episodes of insulin-induced hypoglycemia in IDDM patients. Diabetes 42: Dagogo-Jack SE, Craft S, Cryer PE 1993 Hypoglycemia-associated autonomic failure in insulin-dependent diabetes mellitus. Recent antecedent hypoglycemia reduces autonomic responses to, symptoms of, and defense against subsequent hypoglycemia. J Clin Invest 91: Heller SR, Cryer PE 1991 Reduced neuroendocrine and symptomatic responses to subsequent hypoglycemia after 1 episode of hypoglycemia in nondiabetic humans. Diabetes 40: Davis MR, Shamoon H 1991 Counterregulatory adaption to recurrent hypoglycemia in normal humans. J Clin Endocrinol Metab 73: Widom B, Simonson DC 1992 Intermittent hypoglycemia impairs glucose counterregulation. Diabetes 41: Veneman T, Mitrakou A, Mokan M, Cryer PE, Gerich J 1993 Induction of hypoglycemia unawareness by asymptomatic nocturnal hypoglycemia. Diabetes 42: Fruehwald-Schultes B, Kern W, Deininger E, Wellhoener P, Kerner W, Born J, Fehm HL, Peters A 1999 Protective effect of insulin against hypoglycemiaassociated counterregulatory failure. J Clin Endocrinol Metab 84: Fruehwald-Schultes B, Born J, Kern W, Peters A, Fehm HL 2000 Adaptation of cognitive function to hypoglycemia in healthy men. Diabetes Care 23: Ovalle F, Fanelli CG, Paramore DS, Hershey T, Craft S, Cryer PE 1998 Brief twice-weekly episodes of hypoglycemia reduce detection of clinical hypoglycemia in type 1 diabetes mellitus. Diabetes 47: Fanelli CG, Paramore DS, Hershey T, Terkamp C, Ovalle F, Craft S, Cryer PE 1998 Impact of nocturnal hypoglycemia on hypoglycemic cognitive dysfunction in type 1 diabetes. Diabetes 47: Mellman MJ, Davis MR, Brisman M, Shamoon H 1994 Effect of antecedent hypoglycemia on cognitive function and on glycemic thresholds for counterregulatory hormone secretion in healthy humans. Diabetes Care 17: Pietrowsky R, Specht G, Fehm HL, Born J 1994 Comparison of satiating effects of ceruletide and food intake using behavioral and electrophysiological indicators of memory. Int J Psychophysiol 17: UK Prospective Diabetes Study (UKPDS) Group 1998 Effect of intensive blood-glucose control with metformin on complications in overweight patients with type 2 diabetes (UKPDS 34). Lancet 352: Hundal RS, Krssak M, Dufour S, Laurent D, Lebon V, Chandramouli V, Inzucchi SE, Schumann WC, Petersen KF, Landau BR, Shulman GI 2000 Mechanism by which metformin reduces glucose production in type 2 diabetes. Diabetes 49: Lee A, Morley JE 1998 Metformin decreases food consumption and induces weight loss in subjects with obesity with type II non-insulin-dependent diabetes. Obes Res 6: Hermann LS, Schersten B, Bitzen PO, Kjellstrom T, Lindgarde F, Melander A 1994 Therapeutic comparison of metformin and sulfonylurea, alone and in various combinations. A double-blind controlled study. Diabetes Care 17: Stumvoll M, Nurjhan N, Perriello G, Dailey G, Gerich JE 1995 Metabolic effects of metformin in non-insulin-dependent diabetes mellitus. N Engl J Med 333: DeFronzo RA, Goodman AM 1995 Efficacy of metformin in patients with non-insulin-dependent diabetes mellitus. The Multicenter Metformin Study Group. N Engl J Med 333: Fontbonne A, Charles MA, Juhan-Vague I, Bard JM, Andre P, Isnard F, Cohen JM, Grandmottet P, Vague P, Safar ME, Eschwege E 1996 The effect of metformin on the metabolic abnormalities associated with upper-body fat distribution. BIGPRO Study Group. Diabetes Care 19: Paolisso G, Amato L, Eccellente R, Gambardella A, Tagliamonte MR, Varricchio G, Carella C, Giugliano D, D Onofrio F 1998 Effect of metformin on food intake in obese subjects. Eur J Clin Invest 28: Rouru J, Pesonen U, Koulu M, Huupponen R, Santti E, Virtanen K, Rouvari T, Jhanwar-Uniyal M 1995 Anorectic effect of metformin in obese Zucker rats:

9 Schultes et al. Hunger Modulation by Plasma Glucose and Metformin J Clin Endocrinol Metab, March 2003, 88(3): lack of evidence for the involvement of neuropeptide Y. Eur J Pharmacol 273: Rouru J, Huupponen R, Pesonen U, Koulu M 1992 Subchronic treatment with metformin produces anorectic effect and reduces hyperinsulinemia in genetically obese Zucker rats. Life Sci 50: Leibel RL, Rosenbaum M, Hirsch J 1995 Changes in energy expenditure resulting from altered body weight. N Engl J Med 332: Wellhoener P, Fruehwald-Schultes B, Kern W, Dantz D, Kerner W, Born J, Fehm HL, Peters A 2000 Glucose metabolism rather than insulin is a main determinant of leptin secretion in humans. J Clin Endocrinol Metab 85: Fruehwald-Schultes B, Kern W, Oltmanns KM, Sopke S, Toschek B, Born J, Fehm HL, Peters A 2001 Metformin does not adversely affect hormonal and symptomatic responses to recurrent hypoglycemia. J Clin Endocrinol Metab 86: Towler DA, Havlin CE, Craft S, Cryer P 1993 Mechanism of awareness of hypoglycemia. Diabetes 42: Borg WP, Sherwin RS, During MJ, Borg MA, Shulman GI 1995 Local ventromedial hypthalamus glucopenia triggers counterregulatory hormone release. Diabetes 44: Borg MA, Sherwin RS, Borg WP, Tamborlane WV, Shulman GI 1997 Local ventromedial hypothalamus glucose perfusion blocks counterregulation during systemic hypoglycemia in awake rats. J Clin Invest 99: Griffond B, Risold PY, Jacquemard C, Colard C, Fellmann D 1999 Insulininduced hypoglycemia increases preprohypocretin (orexin) mrna in the rat lateral hypothalamic area. Neurosci Lett 262: Moriguchi T, Sakurai T, Nambu T, Yanagisawa M, Goto K 1999 Neurons containing orexin in the lateral hypothalamic area of the adult rat brain are activated by insulin-induced acute hypoglycemia. Neurosci Lett 264: Towler DA, Havlin CE, Craft S, Cryer P 1993 Mechanism of awareness of hypoglycemia. Perception of neurogenic (predominantly cholinergic) rather than neuroglycopenic symptoms. Diabetes 42: Dykes RW, Dancause N, Miasnikov AA, Agueev V 2001 Rapid differential conditioning of the somatosensory evoked potential by changed patterns of brief innocuous tactile stimuli in waking rats is altered by atropine sulfate. Brain Res 910: Yamada M, Miyakawa T, Duttaroy A, Yamanaka A, Moriguchi T, Makita R, Ogawa M, Chou CJ, Xia B, Crawley JN, Felder CC, Deng CX, Wess J 2001 Mice lacking the M3 muscarinic acetylcholine receptor are hypophagic and lean. Nature 410: Dryden S, Pickavance L, Henderson L, Williams G 1998 Hyperphagia induced by hypoglycemia in rats is independent of leptin and hypothalamic neuropeptide Y (NPY). Peptides 19: Cai XJ, Evans ML, Lister CA, Leslie RA, Arch JR, Wilson S, Williams G 2001 Hypoglycemia activates orexin neurons and selectively increases hypothalamic orexin-b levels: responses inhibited by feeding and possibly mediated by the nucleus of the solitary tract. Diabetes 50: Cai XJ, Widdowson PS, Harrold J, Wilson S, Buckingham RE, Arch JR, Tadayyon M, Clapham JC, Wilding J, Williams G 1999 Hypothalamic orexin expression: modulation by blood glucose and feeding. Diabetes 48: Bailey CJ, Turner RC 1996 Metformin. N Engl J Med 334: Pasquali R, Gambineri A, Biscotti D, Vicennati V, Gagliardi L, Colitta D, Fiorini S, Cognigni GE, Filicori M, Morselli-Labate AM 2000 Effect of longterm treatment with metformin added to hypocaloric diet on body composition, fat distribution, and androgen and insulin levels in abdominally obese women with and without the polycystic ovary syndrome. J Clin Endocrinol Metab 85: Morin Papunen LC, Koivunen RM, Tomas C, Ruokonen A, Martikainen HK 1998 Decreased serum leptin concentrations during metformin therapy in obese women with polycystic ovary syndrome. J Clin Endocrinol Metab 83: Keim NL, Stern JS, Havel PJ 1998 Relation between circulating leptin concentrations and appetite during a prolonged, moderate energy deficit in women. Am J Clin Nutr 68: Chin-Chance C, Polonsky KS, Schoeller DA 2000 Twenty-four-hour leptin levels respond to cumulative short-term energy imbalance and predict subsequent intake. J Clin Endocrinol Metab 85: Heini AF, Lara-Castro C, Kirk KA, Considine RV, Caro JF, Weinsier RL 1998 Association of leptin and hunger-satiety ratings in obese women. Int J Obes Relat Metab Disord 22: Doucet E, Imbeault P, St-Pierre S, Almeras N, Mauriege P, Richard D, Tremblay A 2000 Appetite after weight loss by energy restriction and a low-fat diet-exercise follow-up. Int J Obes Relat Metab Disord 24: Joannic JL, Oppert JM, Lahlou N, Basdevant A, Auboiron S, Raison J, Bornet F, Guy Grand B 1998 Plasma leptin and hunger ratings in healthy humans. Appetite 30: Karhunen L, Haffner S, Lappalainen R, Turpeinen A, Miettinen H, Uusitupa M 1997 Serum leptin and short-term regulation of eating in obese women. Clin Sci Colch 92: Romon M, Lebel P, Velly C, Marecaux N, Fruchart JC, Dallongeville J 1999 Leptin response to carbohydrate or fat meal and association with subsequent satiety and energy intake. Am J Physiol 277:E855 E861