Psychiatry Research, 39: 115- I27 Elsevier

115

Peripheral Opioid Secretory Pattern in Anorexia Nervosa

Francesca Brambilla, Ettore Ferrari, Felice Petraglia, Fabio Facchinetti, Marco Catalano, and Andrea R. Genazzani

Received October 4, 1990; revised version received March 4, 1991; accepted April 7, 1991

Abstract. The peripheral secretion of endogenous opioids was studied in 10 women with restrictive anorexia nervosa and 10 age- and sex-matched healthy controls. The circadian rhythm of /3-endorphin (/3-EP) and P-lipotropin (P-LPH), and their responses to the administration of corticotropin releasing hormone (CRH, 1 pg/ kg body weight, i.v.), clonidine (150 pg, i.v.), domperidone (10 mg, i.v.), and 5- hydroxytryptophan (SHTP, 200 mg, p.o.) were examined in patients and controls. The results revealed increased nocturnal secretion of p-EP and diurnal-nocturnal secretion of P-LPH with loss of circadian rhythmicity of both peptides, normal response to CRH stimulation, blunted response to clonidine and domperidine, and normal /3-EP and blunted /3-LPH response to S-HTP stimulation. The data suggest a complex alteration of peripheral opioids and of central aminergic mechanisms that regulate proopiomelanocortinderived peptide secretion and eating behavior.

Key Words. Eating disorder, /3-endorphin, /?-lipotropin, clonidine, corticotropin releasing hormone, Shydroxytryptophan, circadian rhythm.

Preclinical data indicate that endogenous opioids are involved in the regulation of eating behavior, through a facilitatory-stimulatory mechanism exerted on hunger, feeding, and taste preference. This effect is mediated by a complex interplay between multiple neurotransmitter-neuropeptide-neurohormone secretions in specific brain areas (Morley et al., 1985, 1988; Cooper et al., 1988; Olson et al., 1988). These observations have suggested the hypothesis that impairments of opioid secretion may intervene in the development and course of eating disorders in humans. Existing data, however, are controversial.

In anorexia nervosa (AN), opioid-like activity and P-endorphin (P-EP) concentra- tions in cerebrospinal fluid (CSF) have been reported to be normal, increased, or decreased (Catlin et al., 198 1; Pickar and Bunney, 198 1; Gerner et al., 1982; Kaye et al., 1982, 1985a, 1985b, 1987; Ebert et al., 1984). In plasma, P-EP levels have generally been reported to be increased (Baranowska et al., 1984; Brambilla et al., 1985, 1987; Rose et al., 1985). Plasma levels of /3-EP reflect pituitary secretion of the peptide and do not necessarily mirror the activity of the brain centers that regulate hunger and

Francesca Brambilla, M.D., is Professor and Chief of the Psychoneuroendocrine Center, Ospedale Psichia- trico Pini, Milan. Ettore Ferrari, M.D., is Professor and Director of Dip. Medicina Interna-Gerontologia Universitk. Pavia. Felice Petraglia, M.D., is Assistant; Fabio Facchinetti. M.D.. is Assistant Professor: and Andrea R. Genazzani, M.D.,-is Professor and Director of Clinica Ostetrica-Ginecologica Universith, Modena. Marco Catalano, M.D., is Assistant Professor, Dip. Scienze Neuropsiche, 1st Scientific0 S. Raffaele, Milan. (Reprint requests to Prof. F. Brambilla, Psychoneuroendocrine Center, Ospedale Psichia- trico Pini, Via lppocrate 45, Milan0 20161, Italy.)

0165-1781/91/$03.50 @ 1991 Elsevier Scientific Publishers Ireland Ltd

116

feeding. However, even though opioid control on eating behavior is primarily central, we cannot totally exclude that a peripheral action of the peptides on gastrointestinal or pancreatic opiate receptors would exert, through feedback mechanisms, an effect on central regulatory mechanisms (Margules et al., 1978). Moreover, the study of periph- eral P-EP secretory patterns may help to clarify the biochemical basis of eating

disorders. In fact, tonic and dynamic P-EP plasma secretion reflects the functional state of brain neurotransmitters-neuromodulators that influence both eating behavior and the secretory pattern of the parental molecule of P-EP, proopiomelanocortin

(POMC). In a preliminary study, we investigated the tonic and dynamic secretion of plasma

P-EP in a group of AN patients in an active phase of illness with three main objectives: (1) We wanted to see if the secretory tonus of peripheral /3-EP correlated with specific patterns of symptomatology-in particular, reduced food intake, weight loss, and

amenorrhea. (2) We wanted to investigate the function of the feeding system in its

central-peripheral complex through the pharmacological manipulation of the hypo-

thalamic monoaminergic systems that regulate both POMC secretion and feeding behavior. (3) We attempted to clarify the state of hypothalamic control of the opioid system by examining the circadian secretory pattern of /3-EP.

Methods

Ten women who met the criteria of Feighner et al. (1972) for AN entered the study. Major depression, as defined by DSM-III-R (American Psychiatric Association, 1987), was ruled out in all patients. The patients ranged in age from 14 to 35 years, and had been ill from 3 to 105 months. Ten age- and sex-matched healthy subjects, of normal weight and eating habits, served as controls. Table I presents demographic characteristics of the two groups. Both patients and controls were hospitalized in our Institute throughout the study; they were allowed free choice of diet and general activity. The patients had not received any type of treatment for at least 1 month before the study. During the study, they were still on their self-imposed restrictive diets and were neither gaining nor losing weight. We measured absolute weight and height, % reduction from ideal body weight (according to the Metropolitan Life Insurance Co., New York, NY), and body mass index (BMI) for both patients and controls.

We measured basal plasma levels of /3-EP and P-lipotropin (P-LPH), the circadian secretory pattern, and responses to stimulation with corticotropin releasing hormone (CRH, 1 pg/ kg body weight, i.v.), domperidone (10 mg/ i.v.), clonidine (150 I.cg, i.v.), and Shydroxytryptophan (5-HTP, 200 mg, p.o.), to control the dopaminergic, noradrenergic, and serotonergic regulation of the POMC system. Each test was performed at 48-hour intervals, after an overnight fast and 1 hour of bed rest. At 8:30 a.m., an indwelling catheter was inserted into a forearm vein and kept patent by saline infusion. At 9 a.m., each of the four stimulatory substances was either injected (CRH, domperidone, and clonidine) or administered orally (5-HTP). Heparinized-Trasylol treated blood for b-EP and /3-LPH was drawn at 9 a.m. before drug administration. After- wards, blood was drawn at 15-min intervals for 1 hour, and at 30-min intervals for another hour for the CRH, clonidine, and domperidone tests; for the 5-HTP test, blood was drawnat l-hour intervals for 4 hours. Blood was centrifuged immediately and plasma stored at -20 “C until assay.

The rhythmometric study was carried out 7 days after hospital admission to allow adaptation to the hospital routine. Heparinized blood samples were collected every 4 hours during the day (at 8 a.m., noon, 4 p.m., and 8 p.m.) and every 2 hours during the night (at 8 p.m., 10 p.m., midnight, 2 a.m., and 4 a.m.) through an indwelling catheter placed in an antecubital vein kept patent by saline infusion. Patients were allowed to maintain their usual sleep and feeding habits during the study. Concentrations of B-EP and P-LPH were measured radioimmunologically

117

Table 1. Characteristics of patients with anorexia nervosa and control subiects (mean f SD)

Duration Duration Age Weight Height % of anorexia amenorrhea (yr) (kg) (cm) BMI IBW (mo) (mo)

Anorexia netvosa patients (n = 10) Range 14-35 28.3-47.5 150-170 12-18 19-45 3-105 3-105 Mean 19.3 40.2 162 14.8 32 32 32 SD 5.8 6.2 5.5 1.8 8.6 42 42

Control subjects in = 10) Range 16-32 53-61 158-166 19-21 -5-+l3 Mean 20 58 161 20 2.3 SD 5 3 4 1 1

Note. BMI = body mass index. IBW = ideal body weight.

after extraction and separation of the two peptides in Sephadex G-75 columns by the method of Facchinetti and Genazzani (1979).

All the plasma samples from the various tests were frozen, and the radioimmunoassays (RIAs) were run together at the end of the study with the same batch of reagents. Sensitivity of the RIA for &LPH was 0.9 and that for /3-EP was 1.2 fmol. Interassay coefficients of variation (c.v.‘s) were 9.2% for /3-LPH and 9.8% for &EP; intra-assay c.v.‘s were 12.4% for /3-LPH and 11.4% for p-EP. Student’s t test, analysis of variance (ANOVA) for paired data, and multivar- iate analysis of variance (MANOVA) for repeated measures (SPSS PC+) were used in statistical analysis. The analysis of the circadian rhythm was carried out by conventional statistical methods and the population mean cosinor analysis (Halberg, 1969).

Results

Figs. l-5 present the main results. Baseline P-EP and P-LPH levels at 9 a.m. in patients did not differ significantly from those in controls. The values of the peptides assayed four times at 4%hour intervals during the four stimulation tests and during the circadian rhythm study showed no significant variability from day to day in either the patients or the controls (Figs. l-5).

After CRH administration, P-EP and /3-LPH increases were similar in both groups, with no statistically significant differences of values examined either at each point of the curve or as areas under the curve (AUCs) (&LPH [fmoliml] in patients: mean = 624, SD = 173; /3-LPH in controls: mean = 394, SD = 258; /3-EP [fmol/ ml] in patients: mean = 1209, SD = 375; P-EP in controls: mean= 1279, SD= 345). Mean peak times of P-EP were slightly delayed in patients as compared with controls (60 vs. 45 min) and those of P-LPH slightly advanced (15 vs. 30 min), but the difference was not statisti- cally significant (Fig. 1). MANOVA revealed a significant effect of time in both groups for both P-EP (F = 9.5, df = 6, p = 0.002) and P-LPH (F = 18, df = 6, p = 0.001) a significant interaction effect of diagnosis x time (P-EP: F= 6, df = 6,p = 0.009; P-LPH: F = 3.7, df = 6, p = 0.04), and a significant between-subject effect of time for P-LPH (time: F= 4.9, df = 6, p = 0.001; diagnosis x time: F= 3.2, df = 6, p q 0.007) but not for P-EP.

The /3-EP and P-LPH responses to clonidine administration (Fig. 2) were blunted in patients as compared with controls, and the difference between the mean AUC of the

118

Fig. 1. Opioid peptide responses to corticotropin releasing hormone in anorectic women and controls

01 , T Lo 0 w YI .w aI c mu 0’ ;w M’ .I‘ w w PO

c--e control - AN p.thnt*

8.Endorphin IO-EP) response is shown in the left panel; j34Ipotropin IP-LPHI response is shown in the right panel. Results for 10 patients with anorexia nervosa IANI are Indicated by a solid line; those for 10 controls are Indicated by a dashed line.

two groups was statistically significant for both peptides (P-EP [fmol/ml] in patients: mean q -9.3, SD = 435; b-EP in controls: mean = 787, SD = 602; F= 4.47, df = 13,~ = 0.001; P-LPH [fmol/ml] in patients: mean=-3.5, SD= 153;/I-LPH incontrols: mean= 945, SD = 667; F= 3.26, df = 13,~ = 0.006). MANOVA revealed a significant effect of time for both P-EP and P-LPH (/I-EP: F= 5.1, df= 6,p= 0.02; P-LPH: F= 5.1, df= 6,p = 0.0 I), a significant interaction effect of diagnosis x time@-EP: F=6,df=6,p=0.01; P-LPH: F= 5.6, df = 6,p q 0.0 1). The test involving time within subjects was significant for P-EP (for time: F= 2.6, df q 6, p = 0.03; for diagnosis x time: F= 2.8, df q 6,p = 0.02)

but not for P-LPH. After domperidone administration (Fig. 3), p-EP rises were significantly lower in

patients than in controls, both when examined at each point of the curve and as AUC (40 f 227 vs. 685 ? 498 fmol/ ml; F= 2.92, df= 16,~ = 0.01). Mean peak time occurred at 60 min in both groups of subjects. Mean P-LPH levels of controls were not modified by domperidone administration, as expected. In some of the patients, an increase of P-LPH levels occurred, with peak time at 30 min. The AUCs, however, were not significantly different in patients and controls (298 * 629 vs. -32 f 141 fmoljml). MANOVA revealed no significant interaction effect of diagnosis x time for &EP or P-LPH, a significant effect of time for p-EP (F= 4.4, df = 4,p = O.Ol), and a significant test involving time within subjects for both peptides (time for P-EP: F= 3.2, df= 6,p = 0.02; time for P-LPH: F= 4.2, df = 4,p 10.05; diagnosis x time for P-LPH: F= 3.1, df = 4, p = 0.02; diagnosis x time for P-EP: NS).

119

Fig. 2. Opioid peptide responses to clonidine in anorectic women and controls

PEP BLW fmol/ml fmol/ml

26-

24_

20..

16..

12_

6_

:

t

26

24

OJ ) I I I- , ( Lo 0’ 1s 3uo‘w 90’ 120

, , (

0 lI)‘3u4c6oO’ su 120

c - - ..* Controls _ AN patients

P-Endorphin ID-EPI response IS shown in the left panel; fl-lipotropin ID-LPHJ response is shown in the right panel. Results for 10 patients with anorexia nervosa IAN I are indicated by a solid line; those for 10 controls are indicated by a dashed line.

Administration of 5-HTP induced P-EP rises that did not differ in patients and controls when examined at each point of the curve or as AUC (757 f 1246 vs. 958 f 698 fmol/ ml), with a slightly advanced peak time in patients (120 vs. 180 min). The /3LPH rises were lower in patients than in controls, with the difference being statistically significant at 60 min and as AUC (503 f 657 vs. 1638 + 857 fmol/ ml; F= 2.61, a??= 10,~ = 0.02). Peak time was slightly delayed in patients (180 vs. 120 min) (Fig. 4). A MANOVA revealed a significant effect of time for both peptides (/?-EP: F= 6, df= 6,p = 0.002; P-LPH: F= 10, df= 4,p q 0.002) a significant interaction effect of diagnosis x

time for P-LPH (F = 4.4, df = 4, p = 0.003) but not for /3-EP, and a significant test involving time within subjects for /3-LPH (for time: F = 5.8, c!f= 4, p = 0.001; for diagnosis x time: F = 2.9, df= 4, p q 0.03) but not for &EP.

Both /3-EP and /3-LPH levels recorded throughout the 24-hour cycle were higher in AN patients than in controls, with significant differences being limited to the evening- early nighttime for fi-EP and extended into daytime for P-LPH (Fig. 5). ANOVA revealed that the circadian fluctuations of both peptides were highly significant in controls (P-EP: F: 5.8, df= 5,p= 0.001; P-LPH: F= 6.1, df= 5,p =O.OOl), but not in AN patients. The 24-hour mean levels of both peptides were significantly higher in AN patients than in controls (P-EP: 7.7 + 1 vs. 5.2 + 0.3; P-LPH: 9.0 + 2.3 vs. 5.2 * 0.4; F= 6,

120

Fig. 3. Opioid peptide responses to domperidone in anorectic women and controls

Q w 00' so law 0 30 w w 120

t- . Controls - AN patients . C.O.O1 .*c.am p-Endorphin (P-EP) response IS shown ln the left panel; P-lipotropin ID-LPH) response is shown In the right panel. Results for 10 patients with anorexia nervosa (AN I are indtcated by a solid line; those for 10 controls are indicated by a dashed line.

Fig. 4. Opioid peptide responses to Shydroxytryptophan in anorectic women and controls

Mti PEP

o- , . , 0 W (1Q’ --&--- 240 Q W lw (*r 14c ,--&anlnk -AN pmtimtr s ..C.Osl

I 11 wflml

PI

I8

0

0

p-Endorphin (D-EP) response is shown In the left panel; P-lipotropin ID-LPHI response IS shown in the right panel, Results for 10 patients with anorexia nervosa IAN) are indicated by a solid Ilne; those for 1Ocontrolsare indicated by a dashed line.

121

d’= 15, p = 0.02). The population mean cosinor analysis failed to validate a statistically significant circadian rhythm of P-EP and P-LPH in AN patients. The increase of the &EP circadian mesor correlated neither with the severity of the underweight condi- tion nor with the duration of anorexia or amenorrhea. The increase of the P-LPH mesor was significantly correlated with the severity of the underweight condition (r = 0.82, p = O.OOl), but not with the duration of anorexia or amenorrhea (Fig. 5).

Fig. 5. Circadian pattern of opioid secretion in AN & control subjects

f P J

& z

0 J, , p , 1 I I

0 8 16 24 h

Sludent’s 1 test c---a Conlrols In:71

r ii%SE l PC.06

- Anoroxin Nervosn (rr:lOb . . PC.08

. . . PC.001

p-Endorphin (/3-EP), above; fl-lipotropin (j3-LPH), below; anorexia nervosa (AN, n = lo), solid line; controls [n= 10). dashed line.

Discussion

Our data reveal that the peripheral secretion of /3-EP and fi-LPH is increased signifi- cantly in AN patients, thus confirming previous reports in the literature (Baranowska et al., 1984; Brambilla et al., 1985; Rose et al., 1985). In fact, even though morning concentrations of the peptides were normal in the AN patients, the 24-hour levels showed a significantly increased secretory pattern, especially at night. This result is not

122

unexpected, since investigators have repeatedly reported that cortisol concentrations, which reflect the secretory pattern of adrenocorticotropic hormone (ACTH), are similarly increased over 24 hours (for a review, see Garfinkel, 1984; Ferrari et al., 1986). On the whole, our data and those in the literature indicate a secretory activation of the POMC system.

The circadian pattern of opioid secretion reveals two interesting phenomena. First, the secretion of P-EP is partly dissociated from that of P-LPH, with the former increasing only in the late evening-early night hours and the latter during most of the day and night. Since P-EP and P-LPH derive from the same parental molecule (POMC) and are concomitantly secreted both at rest and under stimulation (Guil- lemin et al., 1977) the two peptides were expected to have a similar secretory profile. There are two possible explanations for this secretory dissociation. Because /I-EP is a metabolic product of P-LPH, it may be that in AN patients there is a defect in the processing of POMC-derived peptides, resulting in the &EP/P-LPH dissociation. If such were the case, however, we should have found different patterns of secretion of the two peptides not only during the day but also at night. Alternatively, it could be suggested that the two molecules derive from different sources and are regulated differently. In experimental animals, the anterior pituitary secretes equimolar

amounts of P-EP and /?-LPH, while the intermediate pituitary preferentially secretes /3-EP (Young et al., 1986). In human studies, no intermediate pituitary has been demonstrated, but some P-EP staining cells have been found in an intermediate layer between the anterior and the posterior lobes-cells that increase in specific physiologi- cal and pathological situations (Wilkes et al., 1980; Lamberts et al., 1982). Thus, we can tentatively suggest that in AN patients the dissociation of P-EP//3-LPH secretion is due to the different (anterior/ intermediate-like) points of origin of the two peptides in the pituitary.

The second observation stemming from our study is that the circadian rhythm of P-EP and P-LPH is lost in AN patients. The literature does not report any data in this regard. It has been repeatedly observed, however, that the circadian rhythm of cortisol, and therefore probably of ACTH, is preserved (for review, see Garfinkel, 1984; Ferrari et al., 1986). This again suggests a dissociation in the secretory pattern of the various POMC-derived peptides, together with a profound imbalance of the hypothalamic or suprahypothalamic mechanisms that regulate circadian rhythmicity. The results of the four dynamic tests reveal that the aminergic regulatory control of the opioid system is impaired.

The blunted response of both P-EP and P-LPH to stimulation with clonidine, a specific a,-adrenergic receptor agonist, suggests the existence of a subsensitivity of the postsynaptic noradrenergic receptor. The same phenomenon was also observed for the growth hormone response to clonidine stimulation, which revealed a reduced sensitivity of postsynaptic cY,-adrenergic receptors (Brambilla et al., 1989). The results of these two tests could suggest that the noradrenergic secretion is increased in AN patients, with ensuing downward regulation of postsynaptic o,-adrenergic receptors. The literature on the secretory tonus of norepinephrine (NE) in AN generally reports reduced concentrations of the amine and its main metabolite, 3-methoxy-4-hydroxy- phenylglycol, in biological fluids (Halmi et al., 1978; Gross et al., 1979; Biederman et

al., 1984; Ebert et al., 1984; Johnstone et al., 1984; Kaye et al, 1985a, 19856). These

123

data would exclude the possibility that the subsensitivity of a,-adrenergic receptors is due to increased NE secretion. Along these lines, we should not forget that the increased secretion of P-LPH in our patients may suggest that the secretion of NE is defective, since CRH and the dependent POMC are inhibited by NE (Holsboer, 1988). However, recent data suggest that NE has a stimulatory effect on the pituitary-adrenal system, acting at the pituitary level (Holsboer, 1988). Further exploration of the NE system is therefore needed to clarify this matter.

The results of the domperidone test reveal an impairment of the dopaminergic regulatory mechanism. Dopamine (DA) is a physiologic inhibitor of P-EP secretion in the intermediate pituitary (Hollt and Bergman, 1982). Domperidone, a peripheral DA inhibitor that does not cross the blood-brain barrier (Sharp et al., 1982) induced an increase in P-EP in controls, thus confirming that a DA-inhibited opioid secretion from an intermediate-like pituitary exists in normal human subjects (Genazzani et al., 1988). The blunted response in AN patients would suggest that DA receptors are subsensitive or, alternatively, that DA secretion is not sufficiently inhibited by the dose of domperidone used. An increased DA tone in AN patients has been hypothe- sized to justify the hyperactivity usually observed in these patients (Barry and Kla- wans, 1976). However, DA has been observed to be decreased in the CSF of AN patients, while urinary concentrations of homovanillic acid have generally been reported to be normal (Riederer et al., 1981; Gillberg, 1983; Gerner et al., 1984; Johnstone et al., 1984; Kaye et al., 1984b). Here again, the matter needs further clarification.

The 5-HTP test, used to explore the serotonergic system, revealed a reduced response of /3-LPH to the stimulus in AN patients. Since serotonin is a physiological stimulator of the POMC system in the anterior pituitary (Petraglia et al., 1984) the decreased response of fi-LPH to 5-HTP may suggest the existence of a deficient serotonergic system. Reduced CSF concentrations of 5hydroxyindoleacetic acid, the main metabolite of serotonin, have been observed in AN patients (Riederer et al., 1981; Ebert et al., 1984; Kaye et al., 1984~; Jimerson et al., 1988). Here again, the difference between the normal /3-EP and the reduced P-LPH response to 5-HTP suggests the existence of either a dysregulation in the peripheral POMC metabolism or a different source of origin of the two POMC fragments.

The normal responses of P-EP and fi-LPH to CRH stimulation that were observed in the AN patients suggest that the corticotroph function in the anterior pituitary is well preserved. Our data contrast with those of Cavagnini et al. (1986) who found a reduced /3-EP response to CRH. This difference, however, could have been due to patient selection. The normal opioid response to CRH stimulation indicates that the impairments observed in the clonidine and 5-HTP tests do not take place at the pituitary level, but reflect hypothalamic or even suprahypothalamic alterations. This hypothesis is reinforced by the observation that the circadian rhythm of secretion of /3-EP and P-LPH is lost.

The unexpected P-LPH secretory response to domperidone observed in some AN patients is difficult to explain. As mentioned above, DA inhibits P-EP secretion from the intermediate pituitary but does not modulate P-LPH production in the anterior pituitary. In this line, the data from the control subjects confirm those found in the

124

literature (Genazzani et al., 1988). Centrally acting DA antagonists such as halo- peridol, sulpiride, and metoclopramide do increase the secretion of /3-LPH and P-EP in parallel, thus revealing that a tonic central DA inhibitory mechanism of POMC- derived peptides exists (Abu-Samra et al., 1984; Murburg et al., 1986; Genazzani et al., 1988). Domperidone, however, does not cross the blood-brain barrier and acts periph- erally on DA receptors located in the intermediate pituitary. These normally secrete very little P-LPH, while no DA receptors have been demonstrated on corticotrophs from which most P-LPH derives. At the moment we have no explanation for this unusual /3-LPH response. Analysis of the relationship between the opioid secretory pattern and behavioral parameters has revealed that /3-EP hypersecretion is not linked to the degree of weight loss, while that of P-LPH is. Since increased ACTH-cortisol secretion also correlates with weight deficiency (for review, see Garfinkel, 1984) it is possible that anterior pituitary POMC hypersecretion may depend on, or be linked to, starvation, while that of P-EP, possibly deriving from an intermediate-like pituitary, would not.

The significance of the complex alterations observed in our patients in regard to the development and course of the eating disorder is difficult to pinpoint. It is worthwhile, however, to underscore that an analogous increase of P-EP, with disruption of circadian rhythmicity and dissociation from the other POMC-derived peptide secre- tory pattern, has been observed in hyperphagic obesity (Givens et al., 1980; Kyriakides et al., 1980; Genazzani et al., 1986). This analogy suggests that apparently different eating disorders, like AN and obesity, share similar impairments of peripheral opioid secretion that could interfere with sensations of hunger and satiety acting at the level of peripheral opioid receptors or through feedback mechanisms. This hypothesis needs confirmation, but it seems worthwhile to propose and test.

On the whole, our findings reveal a complex dysregulation of peripheral opioid secretion in AN, which may partly derive from central alterations of circadian rhyth- micity and neutotransmitter regulatory systems, and partly from production of pep- tides from unusual sources (intermediate-like pituitary). Further investigation with a larger group of patients and longitudinal studies over the course of the disorder are needed to clarify the significance of the impairments observed relative to the patho- genesis of the disorder, its prognosis, and its therapeutic course.

References

Abou-Samra, A.B.; Pugeat, M.; Dechaud, H.; Nachury, L.; and Toirniaire, J. Acute dopami- nergic blockade by sulpiride stimulates P-endorphin secretion in pregnant women. Clinical Endocrinology, 21:583-588, 1984.

American Psychiatric Association. DSM-III-R: Diagnostic and Statistical Manual of Men- tal Disorders. 3rd ed., revised. Washington, DC: Author, 1987.

Baranowska, B.; Rozbicka, G.; Jeske, W.; and Abdel-Fattah, M.J. The role of endogenous opiates in the mechanism of inhibited luteinizing hormone (LH) secretion in women with anorexia nervosa: The effect of naloxone on LH, follicle stimulating hormone, prolactin, and /3-endorphin secretion. Journal of Clinical Endocrinology and Metabolism, 59:412-417, 1984.

Barry, V.C., and Klawans, H.L. On the role of dopamine in the pathophysiologyofanorexia nervosa. Journal of Neural Transmission, 38: 107-122, 1976.

125

Biederman, J.; Herzog, D.B.; Rivinus, T.M.; Ferber, R.A.; Harper, G.P.; Orsulak, P. J.; Harmatz, J.S.; and Schildkraut, J.J. Urinary MHPG in anorexia nervosa patients with and without a concomitant major depressive disorder. Journalof Psychiatric Research, 18: 149-160, 1984.

Brambilla, F.; Cavagnini, F.; Invitti, C.; Poterzio, F.; Lampertico, M.; Sali, L.; Maggioni, M.; Candolfi, C.; Panerai, A.E.; and Mtlller, E.E. Neuroendocrine and psychopathological measures in anorexia nervosa: Resemblances to primary affective disorders. Psychiatry Research, 16:165-176, 1985.

Brambilla, F.; Ferrari, E.; Cavagnini, F.; Invitti, C..; Zanoboni, A.; Massironi, R.; Catalano, M.; Cocchi, D.; and Mtlller, E.E. cY,-Adrenoceptor sensitivity in anorexia nervosa: GH response to clonidine or GHRH stimulation. Biological Psychiatry, 25:256-264, 1989.

Brambilla, F.; Lampertico, M.; Sali, L.; Cavagnini, F.; lnvitti, C.; Maggioni, M.; Candolfi, C.; Panerai, A.E.; and Mtiller, E.E. Clonidine stimulation in anorexia nervosa: Growth hormone, cortisol, and beta-endorphin responses. Psychiatry Research, 20: 19-31, 1987.

Catlin, D.H.; Gerner, R.H.; and Gorelick, D.A. Beta-endorphin: Behavioral effect of single and multiple infusions-Measurement of CSFlevels. Abstracts, III World Congress of Biologi- cal Psychiatry, Stockholm, Sweden, I98 1.

Cavagnini, F.; Invitti, C.; Passamonti, M.; and Polli, E.E. Impaired ACTH and cortisol response to CRH in patients with anorexia nervosa. In: Ferrari, E., and Brambilla, F., eds. Disorders of Eating Behaviour: A Psychoneuroendocrine Approach. Oxford: Pergamon Press, 1986. pp. 229-234.

Cooper, S.J.; Jackson, A.; Kirkham, T.C.; and Turkish, A. Endorphins, opiates and food intake. In: Rodgers, R.J., and Cooper, S.J., eds. Endorphins, Opiates and Behavioral Pro- cesses. Chichester-New York: John Wiley 8c Sons, 1988. pp. 143-186.

Ebert, M.H.; Kaye, W.K.; and Gold, P.W. Neurotransmitter metabolism in anorexia ner- vosa. In: Pirke, K.M., and Ploog, D., eds. me Psychobiology> of Anorexia Nervosa. Berlin: Springer-Verlag, 1984. pp. 58-72.

Facchinetti, F., and Genazzani, A.R. Simultaneous radioimmunoassay for beta-lipotrophin and beta-endorphin in human plasma. In: Albertini, A.; DaPrada, M.; and Peskar, B.A., eds. Radioimmunoassay of Drugs and Hormones in Cardiovascular Medicine. Amsterdam: ElsevieriNorth-Holland, 1979. pp. 347-354.

Feighner, J.; Robins, E.; Guze, S.B.; Woodruff, R.A.; Winokur, G.;and Munoz, R. Diagnos- tic criteria for the use in psychiatric research. Archives of General Psychiatry, 26:57-63, 1972.

Ferrari, E.; Bossolo, P.A.; Marelli, G.; Foppa, S.; Magni, P.; and Zanoletti, M.G. Chrono- endocrinological aspects of anorexia nervosa. In: Ferrari, E., and Brambilla, F., eds. Disorders of Eating Behaviour: A Psychoneuroendocrine Approach. Oxford: Pergamon Press, 1986. pp. 207-218.

Garfinkel, P.E. Anorexia nervosa: An overview of hypothalamic-pituitary function. In: Brown, G.M.; Koslow, S.H.; and Reichlin, S., eds. Neuroendocrinology and Psychiatric Disorder. New York: Raven Press, 1984. pp. 301-314.

Genazzani, A.H.; Petraglia, F.; Facchinetti, F.; Golinelli, S.; Oltramari, P.; Santoro, V.; and Volpe, A. Evidence for a dopamine-regulated peripheral source of circulating /_I-endorphin. Journal of Clinical Endocrinology and h4etabolism, 66~279-282, 1988.

Genazzani, A.R.; Petraglia, F.; Pintor, C.; Santoro, V.; Giovannini, C.; Severi, F.; and Facchinetti, F. Dysregulation of plasma beta-endorphin in obese patients at different ages. In: Ferrari, E., and Brambilla, F., eds. Disorders of Eating Behaviour: A Psychoneuroendocrine Approach. Oxford: Pergamon Press, 1986. pp. 113-120.

Gerner, R.H.; Cohen, D.J.; Fairbanks, L.; Anderson, G.M.; Young, J.G.; Scheinin, M.; Linnoila, M.; Shaywitz, B.A.; and Hare, T.A. CSF neurochemistry of women with anorexia nervosa and normal women. American Journal of Psychiatry, 141:1441-1444, 1984.

Gerner, R.H., and Sharp, B. CSF /I-endorphin immunoreactivity in normal, schizophrenic, depressed, manic and anorexic subjects. Brain Research, 237~244-247, 1982.

Gillberg, C. Low dopamine and serotonin in anorexia nervosa. American Journalof Psychia- try, 140:948-949, 1983.

126

Givens, J.R.; Wiedeman, E.; Andersen, R.N.; and Kitabachi, A.E. Beta-endorphin and beta-lipotropin plasma levels in hirsute women: Correlation with body weight. Journal of Clinical Endocrinology and Metabolism, 50~975-977, 1980.

Gross, H.A.; Lake, C. R.; Ebert, M.H.; Ziegler, M.G.; and Kopin, I.J. Catecholamine metabolism in primary anorexia nervosa. Journal of Clinical Endocrinology and Metabolism, 49:805-809, 1979.

Guillemin, R.; Vargo, T.; Rossier, J.; Minick, S.; Ling, N.; Rivier, C.; Vale, V.; and Bloom, F. Beta-endorphin and adrenocorticotropin are secreted concomitantly by the pituitary gland. Science, 197:1367-1369, 1977.

Halberg, F. Chronobiology. Annual Review of Physiology, 31:675-725, 1969. Halmi, K.A.; Dekirmenjian, H.; Davis, J.M.; Casper, R.; and Goldberg, S. Catecholamine

metabolism in anorexia nervosa. Archives of General Psychiatry, 35:458-460, 1978. Hollt, V., and Bergman, M. Effects of acute and chronic haloperidol treatment on the

concentrations of immunoreactive /?-endorphin in plasma, pituitary and brain of rats. Neuro- pharmacology, 21: 147-154, 1982.

Holsboer, F. Implications of altered limbic-hypothalamic-pituitary-adrenocortical (LHPA) function for neurobiology of depression. Acta Psychiatrica Scandinavica, 77 (Suppl. 341):72- 111, 1988.

Jimerson, D.C.; Brandt. H.A.;and Brewerton, T.D. Evidence of altered serotoninfunction in bulimia and anorexia nervosa: Behavioral implications. In: Pirke, K.M.; Vandereychen, W.; and Ploog, D., eds. 7&e Psychobiology of Bulimia. Berlin: Springer Verlag, 1988. pp. 83-89.

Johnstone, J.L.; Leiter, L.A.; Burrow, G.N.; Garfinkel, P.E.; and Anderson, G.H. Excretion of urinary catecholamine metabolites in anorexia nervosa: Effect of body composition and energy intake. American Journal of Clinical Nutrition, 40:1001-1006, 1984.

Kaye, W.H.; Berrettini, W.H.; Gwirtsman, H.E.; Chretien, M.; Gold, P.W.; George, D.T.; Jimerson, D.C.; and Ebert, M.H. Reduced cerebrospinal fluid levels of immunoreactive pro- opiomelanocortin related peptides (including P-endorphin) in anorexia nervosa. Life Sciences, 41:2147-2155, 1987.

Kaye, W.H.; Ebert, M.H.; Gwirtsman, H.E.; and Weiss, S.R. Differences in brain sero- tonergic metabolism between nonbulimic and bulimic patients with anorexia nervosa. Ameri- can Journal of Psychiatry, 141: 1598-1601, 1984a.

Kaye, W.H.; Ebert, M.H.; Raleigh, M.; and Lake, CR. Abnormalities in CNS monoamine metabolism in anorexia nervosa. Archives of General Psychiatry, 41:350-354, 19846.

Kaye, W.H.; Gwirtsman, H.E.; George, D.T.; Jimerson, D.C.; and Ebert, M.H. CSF 5-HIAA concentrations in anorexia nervosa: Reduced values in underweight subjects normal- ize after weight gain. Biological Psychiatry, 23: 102-105, 1988.

Kaye, W.H.; Gwirtsman, H.E.; Lake, C.R.; Siever, L.J.; Jimerson, D.C.; Ebert, M.H.;and Murphy, D. L. Disturbances of norepinephrine metabolism and a,-adrenergic receptor activity in anorexia nervosa: Relationship to nutritional state. Psychopharmacology Bulletin, 2 I:4 19- 423, 1985a.

Kaye, W.H.; Jimerson, D.C.; Lake, C.R.; and Ebert, M.H. Altered norepinephrine metabo- lism following long-term weight recovery in patients with anorexia nervosa. Psychiatry Research, 14:333-342, 19856.

Kaye, W.H.; Pickar, D.; Naber, D.; and Ebert, M.H. Cerebrospinal fluid opioid activity in anorexia nervosa. American Journal of Psychiatry, 139:643-645, 1982.

Kyriakides, M.; Silverstone, T.; Jeffcoate, W.; and Laurence, B. Effect of naloxone on hyperphagia in Prader-Willi syndrome. Lancet, 1:876-877, 1980.

Lamberts, S.W.; De Lange, S.A.; and Stefanko, S.Z. Adrenocorticotropin secreting pituitary adenomas originate from the anterior or the intermediate lobe in Cushing’s disease: Differences in the regulation of hormone secretion. Journal of Clinical Endocrinology and Metabolism, 54:286-29 1, 1982.

Margules, D.L.; Moisset, B.; Lewis, M.J.; Shibuya, H.; and Pert, C.B. Beta-endorphin is associated with overeating in genetically obese mice (ob/ ob) and rats (fa/ fa). Science, 202:988- 991, 1978.

127

Morley, J.E., and Blundell, J.E. The neurobiological basis of eating disorders: Some formu- lations. Biological Psychiatry, 2353-78, 1988.

Morley, J.E.; Levine, A.S.; Gosnell, B.A.; Mitchell, J.E.; Krahn, D.D.; and Nizielski, S. Peptides and feeding. Peptides, 6 (Suppl. 2):181-192, 1985.

Murburg, M.; Paly, D.; Wilkinson, C.W.; Veith, R.C.; Malas, K.L.; and Dorsa, D.M. Haloperidol increases plasma P-endorphin-like immunoreactivity and cortisol in normal human males. Life Sciences, 39:373-381, 1986.

Olson, G.A.; Olson, R.D.; and Kastin, A.J. Endogenous opiates: 1988. Peptides, 10:1253- 1280, 1989.

Petraglia, F.; Facchinetti, F.; Martignoni, E.; Nappi, G.; Volpe, A.; and Genazzani, A.R. Serotoninergic agonists increase plasma levels of P-endorphin and P-lipotropin in humans. Journal of Clinical Endocrinology and Metabolism, 59:1138-l 142, 1984.

Pickar, D., and Bunney, W.E., Jr. Endogenous opioids and psychiatric illness. Neuroendo- crinology Letters. 3:80, 198 1.

Riederer, P.; Toifl, K.; and Boltzman, L. Excretion of biogenic amine metabolites in anorexia nervosa and other diseases accompanied by severe weight loss. Abstracts, III World Congress of Biological Psychiatry, Stockholm, Sweden, 198 I.

Rose, S.R.; Dahms, W.T.; and Goh, H.H. Increased androgen and beta-endorphin in anorexia nervosa. Abstracts, Endocrine Society Meeting, Baltimore, MD, 1985.

Sharp, B.; Karson, B.; Marshak, D.; Ross, R.; George, R.; Sowers, J.; and Yamada, T. Domperidone elevates rat plasma /3-endorphin immunoreactivity when administered peripher- ally but not intracerebroventricularly. Life Sciences, 3 1:98 I-985, 1982.

Wilkes, M.M.; Watkins, W.B.; Stewart, R.D.;and Yen, S.S.C. Localizationand quantitation of /?-endorphin in human brain and pituitary. Neuroendocrinology, 30: 113- 12 I, 1980.

Young, E.A.; Lewis, J.; and Akil, H. The preferential release of beta-endorphin from the anterior pituitary lobe by corticotropin releasing factor (CRF). Peptides, 7:603-607, 1986.