<<

Molecular (2001) 6, 620–624  2001 Nature Publishing Group All rights reserved 1359-4184/01 $15.00 www.nature.com/mp HYPOTHESIS behavior in nervosa—an excess of both orexigenic and anorexigenic signalling? A Inui

Second Department of Internal Medicine, Kobe University School of Medicine, Kobe, Japan

Anorexia nervosa (AN) is a disorder characterized by abnormal eating behavior, weight regu- lation, and disturbances in attitudes and toward body weight and shape. Although progress has been made in the treatment of AN, a substantial portion of patients have a limited response to treatment. Multiple endocrine and metabolic changes occur after prolonged star- vation, conserving energy and . A number of the endocrine findings in patients with AN may be secondary to adaptive mechanisms. However, AN differs from simple in that excess of both feeding-stimulatory (orexigenic) and feeding-inhibitory (anorexigenic) signalling is characteristic, producing the ‘mixed’ signal about satiety and desire to feed. This leads to a failure of the adaptive feeding response that is initiated by a decrease in , an adiposity signal from tissue, and the resultant increase and decrease of orexigenic and anorexigenic signalling, respectively. The hypothesis of unbalanced shift of feeding-regulat- ory circuitry places anorexigenic corticotropin-releasing factor and orexigenic Y in the final common neurobiological substrate for AN. Therapeutic intervention using such receptor antagonists may lead to more successful and targeted psychopharmacological treat- ment. Molecular Psychiatry (2001) 6, 620–624. Keywords: ; orexigenic signal; anorexigenic signal; leptin; corticotropin-releasing factor; ; starvation

Anorexia nervosa (AN) and (BN) are Hypothalamic response to starvation disorders of unknown origin that most commonly occur in women (about 3% over their lifetimes) and The concept of the adipostat was originally proposed usually have their onset in . These ill- as a for the regulation of food intake, and it was nesses have become a major focus of attention in terms posited that the amount of energy stored in the form of both the research community and the general pub- of could be detected and held constant lic.1–5 Patients with eating disorders invariably have a by the . Leptin, a 16-kDa protein disturbed and an intense fear of weight secreted by adipocytes in proportion to body fat, was gain. Although endocrine disturbances such as those found to be a critical element of this system in pituitary –adrenal, –thyroid, and –gonadal axes are (Figure 1).15 Leptin acts via cell surface receptors in the common in AN and BN, it is more a consequence of , the effects of which are mediated by a starvation than a cause of their psycho- growing list of hypothalamic regulators on feeding, pathology. Little is known about the etiology and the including orexigenic signals such as neuropeptide Y intrinsic biological alterations. Although pathological (NPY), as well as anorexigenic signals such as cortico- eating behavior has tended to be viewed as derivative tropin-releasing factor (CRF).11,12,14,16 Leptin appears to and not the primary pathophysiology, recent develop- play an important role in triggering an adaptive ment in the understanding of feeding and body weight response to starvation since causes leptin regulation has shed new light on the role of feeding levels to fall in proportion to the loss of body fat mass. regulators in the systematic disturbance of the struc- Low leptin levels in the increase the activity of ture of feeding behavior in AN patients.6–14 the orexigenic signals that stimulate eating and sup- press energy expenditure, and decrease the activity of the anorexigenic signals that cause anorexia and weight loss (Figure 1a). NPY is a that is widely expressed throughout the brain. NPY path- ways from the arcuate to the paraventricular nuclei of Correspondence: Dr A Inui, Associate Professor, Second Depart- the hypothalamus are known to be activated in ment of Internal Medicine, Kobe University School of Medicine, response to , as well as in other conditions asso- Kusunoki-cho, Chuo-ku, Kobe 650–0017, Japan. E-mail: inuiȰmed.kobe-u.ac.jp ciated with negative energy balance such as caloric Received 16 February 2001; revised 12 April 2001; accepted 18 restriction, lactation, intense , and uncon- April 2001 trolled .11,12,16 NPY consists of an intercon- Orexigenic and anorexigenic signalling in AN A Inui 621

Figure 1 A simplified model of the hypothalamic feeding-regulatory circuitry in response to starvation (a) and in anorexia nervosa (b). Leptin acts as part of a feedback loop to maintain constant stores of fat. This is achieved by hypothalamic effector molecules downstream of leptin that regulate food intake and energy expenditure. Loss of body fat (starvation) leads to a decrease in leptin, which in turn leads to a state of positive energy balance wherein food intake exceeds energy expenditure. This compensatory response is mediated by increased production, release, and/or action of neuropeptide Y (NPY) and other orexigenic , as well as a decreased activity of anorexigenic signals such as corticotropin-releasing factor (CRF). In anorexia nervosa, simultaneous excess of orexigenic and anorexigenic signalling may produce a ‘mixed’ signal, leading to a failure of this adaptive feeding response. An increase in leptin or serotonergic function during or after the recovery may cause the patients difficulty in achieving and sustaining normal weight. Relative decrease in anorexigenic signalling such as cholecystokinin (CCK) and (PP) may underlie the occurrence and/or the maintenance of bulimic binging behavior. Anorexics and bulimics appear to have different disturbances in feeding-regulatory signals, such as increased and decreased CCK in the plasma22–24 and decreased and normal homovanillic acid, a major metabolite of in the cerebro- spinal fluid,33 respectively. These may suggest different neurobiological bases between the two subtypes. There are reports of other potential signals such as releasing hormone and thyrotropin releasing hormone that might have a role in the altered regulation in AN.34 Positive and negative regulators of food intake and body adiposity are expressed as red and blue, respectively, with the caliber of the arrows being proportionate to the impact. Abbreviations: +, stimulatory input; −, inhibitory input; AGRP, agouti-related protein; CART, - and -regulated transcript; GLP-1, glucagon-like ␣ peptide I7–36 amide; IL-6, interleukin-6; MCH, melanin-concentrating hormone; TNF- , tumor necrosis factor-alpha. nected orexigenic network that includes galanin, Abnormalities in feeding regulators in anorexia peptides, melanin-concentrating hormone nervosa (MCH), , and agouti-related protein (AGRP). The possible etiologic role of biological factors has Most of these peptides are up-regulated in ob/ob mice been difficult to study because good animal models do that lack functional leptin, and their expressions are not yet exist. Therefore, altered appetite or satiety sig- increased through fasting, suggesting an important role nals that may contribute to the development and/or the of these orexigenic signals in facilitating the recovery maintenance of AN have been investigated by measur- 11,12,16 of lost weight. The persistence of body weight ing a range of feeding regulators in the blood or the loss in AN, therefore, implies a failure of this adaptive cerebrospinal fluid (CSF) of the patients.1,2 Limited feeding response, but this has only been thought to be data suggest that AN patients have disturbances of cer- due to a relentless and successful pursuit of thinness tain orexigenic and anorexigenic signals in the brain by the patients. and the periphery (Figure 1b).

Molecular Psychiatry Orexigenic and anorexigenic signalling in AN A Inui 622 Anorexigenic signals Orexigenic signals

Malnourished and AN patients have con- Underweight anorexics had significantly elevated con- centrations of CSF NPY compared to normal con- sistently been found to have significantly reduced 18,26 plasma and CSF leptin concentrations compared to trols. NPY levels are not normalized until a long time after the weight has normalized and frequently at normal weight controls, suggesting a normal physio- the same time as the resumption of menstruation. CSF logical response to starvation.17,18 However, an elev- concentrations of PYY, a powerful feeding stimulator ated CSF to plasma leptin ratio was observed in AN structurally related to NPY, were significantly elevated patients, and CSF leptin concentrations reached nor- in bulimic women studied after a month of abstinence mal values before full weight restoration during feed- from binging and compared to controls and ing, possibly as a consequence of the relatively rapid patients with AN. It was suggested from animal experi- and disproportionate accumulation of fat during ments that abnormally elevated CNS PYY activity in refeeding. These findings suggest an implication in the the abstinent state may contribute to a persistent drive difficulty in achieving and sustaining a normal weight 17 in feeding behavior, particularly a desire for sweet in AN. Plasma leptin levels may be decreased in nor- foods and the resumption of binging behavior.1,2 BN mal weight bulimics, and the eating pattern may play patients had normal CSF NPY levels in contrast to AN. a role in the modulation of leptin secretion. lev- A radioreceptor assay, which can measure overall els may also be relatively increased in AN because of opioid activity, showed that underweight anorexics hypercortisolemia, which has been linked to the con- had an increase of CSF opioid activity.1,2 However, trol of food intake and to the regulation of body CSF concentrations of ␤-endorphin and dynorphin, 12,19 weight. CRF is a major anorexigenic signal in the some of the endogenous opioid peptides, as well as hypothalamus. In addition to its role as controller of those of galanin, were not increased in underweight the hypothalamic-pituitary-adrenal axis, CRF inhibits anorexics, suggesting an inhibitory influence from ano- food intake, increases energy expenditure, and pro- rexigenic signals or masking of the abnormality by the 9,11,12 duces sustained weight loss. Several studies have contributions of other brain areas to the total CSF reported elevated CRF levels in the CSF of immunoreactivity. underweight anorexics, indicating an implication for anorexia and weight loss, as well as for the accompanied neuroendocrine disturbance.20 The neur- Hypothesis otransmitter is involved in the physiological Few methods are available at present to assess CNS system relevant to AN.7,21 Increase in brain serotonin activity of orexigenic and anorexigenic signals in function leads to reduction in food intake in both ani- and to identify the specific pathways that may mals and humans. CSF levels of the major serotonin underlie the altered expression of the signals.1,2 How- metabolite, 5-hydroxyindoleacetic acid (5-HIAA), are ever, the data described here allow us to formulate a low in underweight AN but rise to above normal levels hypothesis for the role of feeding regulators in the in individuals who have made long-standing 21 development and/or the maintenance of the pathologi- recoveries. This suggests that a premorbid disturb- cal feeding behavior in AN. Even though most of the ance in serotonergic function may be a risk factor for disturbances may not be permanent features and thus the development of AN. Cholecystokinin (CCK), a hor- not trait-related disturbances, they are strongly mone released from the intestine after eating, is known entrenched and are not easily corrected by improved to be involved in the control of short-term satiety in nutrition or short-term weight normalization.1,2 many species, including man. Increased and decreased The arcuate NPY neurons are in a strategic position CCK responses to meals were reported in patients with to monitor the metabolic state of animals and are acti- 22–24 AN and those with BN, respectively. Pancreatic vated normally in a state of negative energy bal- polypeptide (PP), a hormone released from the pan- ance.9,11,12 NPY is a component of hypothalamic cir- creas after eating, is also thought to be a circulating cuitry integrating aspects of feeding behavior and 14 satiety factor in both humans and rodents. Basal as energy . The NPY orexigenic network is well as stimulated PP secretion is exaggerated in AN, composed of signals such as and galanin, as and the presence of bulimic symptoms is associated well as the newly identified MCH, orexin, and AGRP. with reduced PP and CCK responsiveness.24 Tumor In patients, the compensatory increase in necrosis factor-␣ (TNF-␣), also known as cachectin, is orexigenic signals is observed but is compromised by a cytokine that mediates weight loss in experimental the simultaneous excess of anorexigenic signalling animals and humans.13 Chronic TNF exposure causes such as CRF, CCK, PP, and TNF-␣, indicating dysregul- characterized by anorexia, weight loss and ation of hypothalamic feeding-regulatory circuitry. depletion of whole-body protein and lipid. Plasma con- Although information is needed on newly identified centrations of TNF-␣ and its soluble receptor were sig- orexigenic and anorexigenic signals, this may explain nificantly higher in anorectic patients than in controls, why many anorexics cannot easily reverse their illness and these remained altered even after weight resto- in contrast to simple starvation. It is possible that the ration.25 Most of these feeding inhibitory signals are ‘mixed’ signal about satiety and desire to feed could known to inhibit the NPY orexigenic network.11,12,14,16 contribute to the ambivalence about food, ie, the dis-

Molecular Psychiatry Orexigenic and anorexigenic signalling in AN A Inui 623 sociation that anorexics often display between reduced medial hypothalamic nuclei is associated with the caloric intake and obsessive thoughts about food. An stimulation of food intake in experimental animals.31 imbalance between orexigenic and anorexigenic signal- Therefore, in order to establish a causal relationship, ling may underlie the occurrence of bulimic binging the purported factors would need to be manipulated behavior in which a relative, but not absolute, decrease and shown to have an effect on the development or in anorexigenic signals such as CCK and PP may be course of the illness. A battery of psychological testing characteristic.24 It is another postulate that the alter- would be useful to evaluate a patient’s attitude toward ations in feeding-regulatory circuitry play a role in the food as well as other behavioral alterations.29 Although obsessional preoccupation with body weight and body selective serotonin reuptake inhibitors significantly shape, an aversively conditioned learning. NPY was reduced relapse and improved the outcome when reported to produce taste aversion at the same doses given after weight restoration,32 this may be due to that stimulate appetite and influence memory retention reduction in obsession, compulsion, and other associa- in rodents.27 Hypersecretion of vasopressin and ted psychological disturbances by the compound. reduced CSF oxytocin in underweight AN patients may Current treatments for AN are aimed at normalizing also maintain or exacerbate the persistent preoccu- body weight, correcting the irrational preoccupation pation with the adverse consequences of feeding.2 with weight loss, and preventing relapse.3–5 Many indi- viduals who develop AN during adolescence and young adulthood eventually make full recoveries. Testing the hypothesis However, for at least 50% of these patients, the long- Research over the past few years has provided an term outcome is not as good. Many individuals remain unprecedented expansion of our knowledge about the irrationally concerned about and never molecular mechanisms regulating body fat. Although achieve a normal body weight. The mortality, due to much of this knowledge has come from studies of obes- complications of starvation or from , is substan- ity, there appear to be implications for clinical con- tial, apporoximately 5% per decade of follow-up.4 It is ditions associated with weight loss.9,13 There is a large expected that correction of the abnormalities in feed- effort to develop new modulators of monoamine neuro- ing-regulatory circuitry improves not only pathological transmitters and neuropeptide receptor agonists or eating behaviors but also associated cognition and antagonists that have been associated with food intake mood alterations, leading to a better prognosis. and energy expenditure, some of which are presently Advances in molecular may permit the being evaluated in clinical trials for .28 It war- identification of genes that predispose individuals to rants evaluation in the treatment of AN or BN. Specifi- develop AN or BN. Although polymorphisms in genes cally, antagonists for the CRF2 receptor that mediate of the serotonin (S-HT) system such as the 5-HT2A feeding inhibition by CRF are expected to be effective receptor and susceptibility to AN need to be further for restricting AN if produced, and antagonists for Y1 clarified, greater knowledge of the biological factors in and Y5 receptors that mediate feeding stimulation by the development and maintenance of the pathology NPY and PYY, for a permanent drive in feeding will lead to innovative approaches to treatment and behavior, particularly the resumption of binging prevention. behavior. These neuropeptides appear to be the final common neurobiological substates for AN although there is increasing evidence that normally a dynamic Acknowledgements equilibrium exists between NPY and CRF neuronal I am indebted to Professor Masato Kasuga and Pro- activity (Figure 1). Application of CRF antagonists is fessor Shigeaki Baba (Kobe University) for many stimu- especially of theoretical interest since intracerebroven- lating discussions. The work was supported by grants tricular administration of CRF to experimental animals from the Ministry of Education, Science, Sports, and produces many of the physiologic and behavioral Culture of Japan. changes classically associated with AN, including decreased feeding behavior, hyperactivity, hypothal- amic , decreased sexual activity, and References decreased gastric emptying.1,2,6–9,12,13,29,30 Since the 1 Kaye WH, Klump KL, Frank GK, Strober M. Anorexia and bulimia final integration of food intake control is brought about nervosa. Annu Rev Med 2000; 51:299–313. by maintenance of a delicate balance between orexi- 2 Kaye WH, Gendall K, Kye C. The role of the central nervous system genic and anorexigenic signals, it is necessary to cor- in the psychoneuroendocrine disturbances of anorexia and bulimia rect for such unbalanced shift in feeding-regulatory cir- nervosa. Psychoneuroendocrinology 1998; 21: 381–396. cuitry as much as possible. However, we should keep 3 Walsh BT, Devlin MJ. Eating disorders: progress and problems. Science 1998; 280: 1387–1390. in mind that the hypothesis is mostly based on the 4 Sullivan PF. Mortality in anorexia nervosa. Am J Psychiatry 1995; measurement of CSF or plasma concentrations of the 152: 1073–1076. signals, which will give us rather limited information 5 Garner DM. Pathogenesis of anorexia nervosa. Lancet 1993; 341: on their activity in the feeding-regulatory center of the 1631–1635. 6 Morley JE. Neuropeptide regulation of appetite and weight. Endo- brain. It is known that enhanced release of both norepi- crinol Rev 1987; 8:256–287. nephrine and dopamine in the lateral hypothalamus is 7 Morley JE, Blundell JE. The neurobiological basis of eating dis- associated with the inhibition of food intake, but in the orders: some formulations. Biol Psychiatry 1988; 23:53–78.

Molecular Psychiatry Orexigenic and anorexigenic signalling in AN A Inui 624 8 Leibowitz SF. Hypothalamic neuropeptide Y, galanin, and amines. 22 Uhe AM, Szmukler GI, Collier GR, Hansky J, D’dea K, Young GP. Concepts of coexistence in relation to feeding behavior. Ann NY Potential regulators of feeding behavior in anorexia nervosa. Am J Acad Sci 1989; 575:221–235. Clin Nutr 1992; 55:28–32. 9 Schwartz MW, Seeley RJ. Neuroendocrine responses to starvation 23 Geracioti Jr TD, Liddle RA. Impaired cholecystokinin secretion in and weight loss. N Engl J Med 1997; 336: 1802–1811. bulimia nervosa. N Engl J Med 1988; 319: 683–688. 10 Flier JS. What’s in a name? In search of leptin’s physiological role. 24 Fujimoto S, Inui A, Kiyota N et al. Increased cholecystokinin and J Clin Endocrinol Metab 1998; 83: 1407–1413. pancreatic polypeptide responses to a fat rich meal in patients with 11 Kalra SP, Dube MG, Pu S, Xu B, Horvath TL, Kalra PS. Interacting restrictive but not bulimic anorexia nervosa. Biol Psychiatry 1997; appetite-regulating pathways in the hypothalamic regulation of 41: 1068–1070. body weight. Endocrinol Rev 1999; 20:68–100. 25 Nakai Y, Hamagaki S, Takagi R, Taniguchi A, Kurimoto F. Plasma 12 Inui A. Feeding and body weight regulation by hypothalamic neur- concentrations of tumor necrosis factor-␣ (TNF-␣) and soluble TNF opeptides—mediation of the actions of leptin. Trends Neurosci receptors in patients with anorexia nervosa. J Clin Endocrinol 1999; 22:62–67. Metab 1999; 84:1226–1228. 13 Inui A. anorexia-cachexia syndrome: are neuropeptides the 26 Kaye WH, Berrettini W, Gwirtsman H, George DT. Altered cerebro- key? Cancer Res 1999; 59:4493–4501. spinal fluid neuropeptide Y and peptide YY immunoreactivity in 14 Inui A. Transgenic approach to the study of body weight regulation. anorexia and bulimia nervosa. Arch Gen Psychiatry 1990; 47: Pharmacol Rev 2000; 52:35–61. 548–556. 15 Friedman JM, Halaas JL. Leptin and the regulation of body weight 27 Nakajima M, Inui A, Teranishi A, Miura M, Hirosue Y, Okita M et in mammals. Nature 1998; 395:763–770. al. Effects of PP family peptides of feeding and learning behavior 16 Schwartz MW, Woods SC, Porte Jr D, Seeley RJ, Baskin DG. Central in mice. J Pharmacol Exp Ther 1994; 268: 1010–1014. nervous system control of food intake. Nature 2000; 404:661–671. 28 Bray GA, Tartaglia LA. Medicinal strategies in the treatment of 17 Mantzoros C, Flier JS, Lesem MD et al. Cerebrospinal fluid leptin obesity. Nature 2000; 404:672–677. in anorexia nervosa: correlation with nutritional status and poten- 29 Inui A, Okano H, Miyamoto M, Aoyama N, Uemoto M, Baba S et tial role in resistance to weight gain. J Clin Endocrinol Metab 1997; al. Delayed gastric emptying in bulimic patients. Lancet 1995; 82: 1845–1851. 346: 1240. 18 Gendall KA, Kaye WH, Altemus M, McConaha CW, La Via MC. 30 Inui A, Namiki M, Morita S, Okada M, Kasuga M. Anorexia nervosa Leptin, neuropeptide Y, and peptide YY in long-term recovered eating disorder patients. Biol Psychiatry 1997; 46: 292–299. and intussusception. Lancet 1996; 347: 399. 19 Lucinio-Paixao J. Hyperinsulinemia: a mediator of decreased food 31 Meguid MM, Fetissov SO, Varma M, Sato T, Zhang L, Laviano A intake and weight loss in anorexia nervosa and major . et al. Hypothalamic dopamine and serotonin in the regulation of Med Hypothesis 1989; 28:125–130. food intake. Nutrition 2000; 16:843–857. 20 Hotta M, Shibasaki T, Masuda A et al. The responses of plasma 32 Kaye W, Gendall K, Strober M. Serotonin neuronal function and adrenocorticotropin and to corticotropin-releasing hor- selective serotonin reuptake inhibitor treatment in anorexia and mone (CRH) and cerebrospinal fluid immunoreactive CRH in anor- bulimia nervosa. Biol Psychiatry 1998; 44:825–838. exia nervosa patients. J Clin Endocrinol Metab 1986; 62:319–324. 33 Kaye WH, Frank GK, McConaha C. Altered dopamine activity after 21 Kaye WH, Gwirtsman HE, George DT, Ebert MH. Altered serotonin recovery from restricting type of anorexia nervosa. Neuropsycho- activity in anorexia nervosa after long-term weight restoration. pharmacology 1999; 21:503–506. Does elevated cerebrospinal fluid 5-hydroxyindoleacetic acid level 34 Stoving RK, Hangaard J, Hansen-Nord M, Hagen C. A review of correlate with rigid and obsessive behavior? Arch Gen Psychiatry endocrine changes in anorexia nervosa. J Psychiatr Res 1999; 33: 1991; 48:556–562. 139–152.

Molecular Psychiatry