- Email: [email protected]
Glucocorticoid receptors in anorexia nervosa and cushing’s disease
Glucocorticoid Receptors in Anorexia Nervosa and Cushing’s Disease Cecilia Invitti, Gabriella Redaelli, Gabriella Baldi, and Francesco Cavagnini Background: Patients with anorexia nervosa do not display cushingoid features in spite of elevated cortisol plasma levels. Whether a cortisol resistance or a reduced availability of the metabolic substrates necessary to develop the effect of glucocorticoids is responsible for this has not been established. Methods: Twenty-two patients with severe restrictive anorexia nervosa, 10 patients with active Cushing’s disease, and 24 healthy volunteers without psychiatric disorders or mood alterations were investigated. Glucocorticoid receptor characteristics were examined on mononuclear leukocytes by measuring [3H]dexamethasone binding and the effect of dexamethasone on [3H]thymidine incorporation, which represents an index of DNA synthesis. Results: The number of glucocorticoid receptors on mononuclear leukocytes (MNL) was comparable in patients with anorexia nervosa, patients with active Cushing’s disease, and normal subjects (binding capacity 3.3 6 0.23 vs. 3.7 6 0.30 and 3.5 6 0.20 fmol/106 cells). Conversely, glucocorticoid receptor affinity was significantly decreased in anorexia nervosa as well as in Cushing’s patients compared to control subjects (dissociation constant 4.0 6 0.31 and 4.1 6 0.34 vs. 2.9 6 0.29 nmol/L, p , .001) and inversely correlated with the levels of urinary free cortisol in both groups of patients. Basal [3H]thymidine incorporation in MNL was significantly reduced in anorexia nervosa as well as in Cushing’s patients compared to control subjects (p , .001) and was diminished by dexamethasone to an extent similar to control subjects in patients with anorexia nervosa, but significantly (p , .001) less in those with Cushing’s disease. In patients with anorexia nervosa, the incorporation of [3H]thymidine into the MNL was inversely correlated with urinary free cortisol levels. Conclusions: These data indicate that the lack of cushingoid features in patients with anorexia nervosa is not ascribable to a reduced sensitivity to glucocorticoids but is more likely due to the paucity of metabolic substrates. Biol Psychiatry 1999;45:1467–1471 © 1999 Society of Biological Psychiatry
2nd Chair of Endocrinology, University of Milan, Istituto Scientifico Ospedale San Luca, Milan, Italy. Address reprint requests to Prof. Francesco Cavagnini, Istituto Scientifico Ospedale San Luca, Via Spagnoletto, 3 20149 Milan, Italy. Received January 19, 1998; revised May 19, 1998; accepted May 26, 1998.
© 1999 Society of Biological Psychiatry
Key Words: Glucocorticoid receptor, anorexia nervosa, Cushing’s disease, glucocorticoids, hypercortisolism, cortisol
Introduction
P
atients with anorexia nervosa have a glucocorticoid hypersecretion pattern similar to patients with Cushing’s disease; they have elevated urinary free cortisol excretion and high serum diurnal and nocturnal cortisol levels, which are not suppressed by dexamethasone (Licinio et al 1996). Nevertheless, patients with anorexia nervosa do not develop the common clinical features of hypercortisolism. Two major explanations have been given for this phenomenon: 1) these patients have a reduced sensitivity to glucocorticoids due to a pre- or postreceptorial defect of glucocorticoid receptors; 2) their malnutrition led to a reduced availability of the metabolic substrates necessary to develop the effect of glucocorticoids. The studies hypothesizing alterations of glucocorticoid receptors are few. Kontula et al (1982) found reduced levels of glucocorticoid receptors not correlated with morning plasma cortisol concentrations in mononuclear leukocytes (MNL) of patients with anorexia nervosa. In agreement with this, a reduction of glucocorticoid receptor number was described in patients with depression who, like those with anorexia nervosa, have a biochemically but not clinically expressed hypercortisolism (Murphy 1991; Whalley et al 1986). Conversely, in patients with anorexia nervosa, Girardin et al (1991) reported virtually normal lymphocyte glucocorticoid receptor levels although, unlike in normal subjects, they were positively correlated with urinary free cortisol concentrations. None of these papers, however, have also explored the function of glucocorticoid receptors. We therefore decided to investigate glucocorticoid receptor binding capacity, affinity, and sensitivity to glucocorticoids on MNL in a group of patients with underweight anorexia nervosa and in a group of normal-weight control subjects. To evaluate glucocorticoid receptors in a typical condition of biochemical and clinical hypercortisolism, a group of patients with active Cushing’s disease was also investigated. 0006-3223/99/$20.00 PII S0006-3223(98)00189-9
1468
C. Invitti et al
BIOL PSYCHIATRY 1999;45:1467–1471
Methods and Materials Twenty-two patients with anorexia nervosa, 21 women and 1 man, aged 14 –33 years (mean 22.3 6 1.17 years), 10 patients with active Cushing’s disease, 7 women and 3 men, 11–59 years (mean 38.0 6 4.36 years), and 24 healthy volunteers, 22 women and 2 men, 10 –51 years (mean 27.3 6 1.61 years) without psychiatric disorders or mood alterations were investigated. None of the patients was receiving any medication when studied. All gave their informed consent to participate in this study, which was approved by our Institution’s Ethical Committee. Diagnosis of restrictive anorexia nervosa was established according to the criteria of DSM-IV (American Psychiatric Association 1994). Patients were not following a timed, prescribed diet and exhibited a severe food restriction without binging or vomiting behavior and were amenorrheal. Diagnosis of Cushing’s disease was posed on clinical and laboratory grounds, according to the conventional criteria [high serum cortisol concentration lacking circadian variations and suppressibility by low doses of dexamethasone, associated with measurable levels of adrenocorticotropic hormone (ACTH); evidence of an adenoma at pituitary imaging or center to periphery ACTH gradient greater than 3 at inferior petrosal sinuses sampling after administration of corticotropin-releasing hormone]. To assess the presence of depression anorectic and Cushing’s patients completed the self-reported Beck Depression Inventory (BDI) (Beck et al 1988). Rating scales on the BDI were below 20, between 20 and 30, and greater than 30 in 20%, 38%, and 42% respectively of patients with Cushing’s disease and in 20%, 26%, and 54% respectively of patients with anorexia nervosa. As a whole, depressive symptoms were present in 80% of both anorectic and Cushing’s patients. Three 24-hour urine collections were taken for free cortisol assay, and a blood sample was drawn for cortisol in the morning after an oral dose of 1 mg dexamethasone (Fdex) the evening before. Urinary free cortisol levels greater than 80 mg/24 hours as well as serum cortisol levels greater than 5 mg/dL indicate an overdrive of the hypothalamic–pituitary adrenal axis. All patients fasted overnight, and at 8 AM an indwelling venous catheter was inserted in an antecubital vein; a 30 mL blood sample was taken 20 min later in edetic acid (EDTA)containing siliconized glass tubes, and immediately processed to obtain MNL, which express glucocorticoid receptors (DelarcheHomo 1984) and provide an in vitro model to study the sensitivity to glucocorticoids. Ten milliliters of venous blood was taken for serum cortisol and free fatty acids (FFA) measurement.
Glucocorticoid Receptor Assay MNL were isolated by Ficoll-Hypaque sedimentation, washed with phosphate-buffered saline (PBS), and suspended in RPMI1640 medium (Gibco, Grand Island, NY), pH 7.4, containing 25 mmol HEPES, 2 nmol glutamine, 50 mg/mL gentamicin, and 10% fetal calf serum. Cell viability exceeded 95% as judged by trypan blue exclusion. MNL binding capacity (Bmax) was tested by adding increasing concentrations (0.25–16 nmol) of [3H]dexamethasone (83 Ci/ mmol, Amersham, Aylesbury, Buckinghamshire, U.K.) to cell
suspensions in the presence or absence of a 1000-fold molar excess of unlabeled dexamethasone. After incubation at 4°C for 18 hours, cells were harvested on glass fiber filters (Whatman GF/C, Clifton, NJ) and washed twice with cold PBS. Samples were counted in a Packard Tri-Carb liquid scintillation spectrometer (Downers Grove, IL) at an efficiency of 60%. Bmax was expressed as femtomoles of receptors per 106 cells, and the apparent dissociation constant (Kd) was obtained by Scatchard plot analysis. The reception intensity index Bmax/Kd was also used (Lyashko and Sukhikh 1987).
Thymidine Incorporation MNL sensitivity to glucocorticoids was determined in vitro by evaluating the inhibitory effect of dexamethasone on their ability to incorporate [3H]thymidine (25 Ci/mmol, Amersham, Aylesbury, Buckinghamshire, U.K.), which represents an index of DNA synthesis. MNL (1 3 106 cells/mL) were incubated at 37°C in a 5% CO2 incubator for 22 hours in the RPMI-1640 medium, as used before, containing phytohemagglutinin (PHA, 5 mg/mL) and increasing doses of dexamethasone (0.1–100 nmol). Incubation was continued for 6 hours after addition of [3H]thymidine (2 mCi/mL), and the cells were harvested on glass-fiber filters. The radioactivity was measured by liquid scintillation counting. Results were expressed as the percentage of isotope incorporation after dexamethasone treatment.
Biochemical Assays Serum cortisol was assayed by radioimmunoassay (Byk-Sangtec Diagnostica, Dietrenbach, Germany). Intra- and interassay coefficients of variation and sensitivity of the method were respectively 4.4%, 8.1%, and 0.6 mg/dL. To avoid differences due to interassay variability, samples taken from the three groups of subjects were run together in each of the assays described above. Urinary free cortisol was measured by radioimmunoassay (DPC, Los Angeles, CA) after extraction with dichloromethane. Intra- and interassay coefficients of variation and sensitivity of the method were respectively 3.5%, 6.2%, and 0.49 mg/dL. Normal values for our laboratory are 10 – 80 mg/24 hours. Since hypercortisolemia raises serum levels of FFA, and FFA may reduce glucocorticoid binding to their receptors (Haourigui et al 1994), we also measured serum FFA in all patients. Plasma FFA levels were determined using an acyl-CoA oxidase-based colorimetric kit (Wako, Osaka, Japan). The intraassay coefficient of variation was 1.6%, and normal values for our laboratory are 0.1– 0.6 mEq/L. To avoid differences due to interassay variability, samples taken from the three groups of subjects were run together in each of the assays described above.
Statistical Analysis Results were expressed as mean 6 SEM. The Mann–Whitney test was used for comparison among groups. Correlations were examined with linear regression analysis. A p value less than .05 was considered statistically significant.
Glucocorticoid Receptors in Anorexia Nervosa
BIOL PSYCHIATRY 1999;45:1467–1471
1469
Table 1. Body Mass Index (BMI), Defined as Body Weight Divided by Height Squared, Concentrations of 24-Hour Urinary Free Cortisol, Serum Cortisol Measured at 8 AM following 1 mg of Dexamethasone at Previous Midnight (Fdex), Serum Free Fatty Acids (FFA), Binding Capacity (Bmax), and Apparent Dissociation Constant (Kd) for Dexamethasone of Monuclear Leukocytes in Patients with Anorexia Nervosa and Cushing’s Disease and in Control Subjects
2
BMI (kg/m ) Urinary free cortisol (mg/24 hours) Fdex (mg/dL) FFA (mEq/L) Bmax (fmol/106 cells) Kd (nmol/L) Bmax/Kd (3 1023) [3H] thymidine incorporation (cpm 3 103)
Anorexia nervosa
Control subjects
Cushing’s disease
13.3 6 0.40 115.3 6 23.55b,c 9.3 6 1.78a 0.6 6 0.11 3.3 6 0.23 4.0 6 0.31a 0.9 6 0.05c 39.2 6 8.20a
22.1 6 2.01 38.1 6 3.67 1.0 6 0.11 0.4 6 0.06 3.5 6 0.20 2.9 6 0.29 1.3 6 0.07 82.2 6 11.18
30.9 6 2.88a 333.1 6 72.33a 13.3 6 2.36a 0.7 6 0.08d 3.7 6 0.30 4.1 6 0.34a 0.9 6 0.04a 38.0 6 8.39a
a,b
Results are expressed as mean 6 SEM. a p , .001 vs. control subjects. b p , .001 vs. Cushing’s disease. c p , .01 vs. control subjects. d p , .05 vs. control subjects.
Results Concentrations of urinary free cortisol, Fdex, FFA, body mass index (BMI), and results of glucocorticoid receptor binding are shown in Table 1. BMI and urinary free cortisol concentrations were significantly higher in patients with Cushing’s disease than in anorectic patients (p , .001). The number of glucocorticoid receptors, as expressed by Bmax, was comparable in MNL of patients with anorexia nervosa, patients with Cushing’s disease, and control subjects. No correlations were found between Bmax and urinary free cortisol, Fdex, and BMI in patients with anorexia nervosa and Cushing’s disease. On the contrary, glucocorticoid receptor affinity, as expressed by Kd, was significantly decreased in patients with anorexia nervosa as well as in those with Cushing’s disease compared to control subjects (p , .001). Kd was positively correlated with urinary free cortisol both in patients with anorexia nervosa and in patients with Cushing’s disease (r 5 .46, p , .05 in the former and r 5 .81, p , .01 in the latter patients). The integrative index of reception effectiveness (Bmax/Kd) was significantly lower in patients with anorexia nervosa (p , .01) and Cushing’s disease (p , .001) than in control subjects and positively correlated with BMI only in patients with anorexia nervosa (r 5 .60, p , .01). Basal [3H]thymidine incorporation was significantly reduced in MNL of patients with anorexia nervosa and Cushing’s disease compared to control subjects (p , .001). In patients with anorexia nervosa, but not in those with Cushing’s disease, [3H]thymidine incorporation was correlated positively with BMI (r 5 .52, p , .05) and negatively with urinary free cortisol (r 5 20.51, p , .05) and Fdex (r 5 2.53, p , .05). Dexamethasone dosedependently inhibited [3H]thymidine incorporation in all
groups (Figure 1). This effect was definitely lower in patients with Cushing’s disease than in control subjects and in patients with anorexia nervosa (radioisotope uptake after dexamethasone: 67.5 6 5.59% vs. 37.6 6 2.49%, p , .001 and 41.4 6 3.18%, p , .001, respectively). In both groups of patients the residual [3H]thymidine incorporation into MNL recorded after dexamethasone was positively correlated with Fdex (r 5 .54, p , .05 and r 5 .78, p , .05, respectively) and in patients with Cushing’s disease also with urinary free cortisol (r 5 .64, p , .05). Serum FFA levels were comparable in patients with anorexia nervosa and control subjects and significantly higher in patients with Cushing’s disease (p , .05). No correlations were found in each group between FFA levels
Figure 1. Dose-related inhibiting effect of dexamethasone on [3H]thymidine incorporation in MNL obtained from control subjects (●), patients with anorexia nervosa (Œ), and patients with Cushing’s disease (1).
1470
BIOL PSYCHIATRY 1999;45:1467–1471
and Bmax, Kd, and [3H]thymidine incorporation before and after dexamethasone. These latter parameters were not correlated with age both in the whole and in the separate groups of subjects.
Discussion The reason why patients with anorexia nervosa do not develop cushingoid features in spite of even markedly increased serum cortisol levels is unclear. Either a cortisol resistance secondary to alterations in number and/or function of glucocorticoid receptors or a lack of nutritional substrates necessary for the clinical expression of the cortisol effects might be in play. The present study has shown, in agreement with observations by others, that MNL of patients with either anorexia nervosa (Girardin et al 1991) or Cushing’s disease (Pardes et al 1991; Junker 1983; Kontula et al 1980) have a normal concentration of glucocorticoid receptors. This finding appears in contrast with the demonstration, given by some authors (Pardes et al 1991; Schlechte et al 1982) though not by others (Junker 1983), of a down-regulation of glucocorticoid receptor number in the lymphocytes of patients treated with corticosteroids; however, the model of patients on steroid treatment may differ from that of Cushing’s disease and anorexia nervosa, in whom complex neurotransmitter alterations might directly influence the glucocorticoid receptor function (Kabbaj et al 1995). On the other hand, care was exercised to run in the same experiment samples taken from each group of patients to overcome interassay variability. At variance with the authors mentioned above (Girardin et al 1991; Junker 1983; Pardes et al 1991; Kontula et al 1980, 1982) we found a reduced glucocorticoid receptor affinity in our patients with anorexia nervosa and Cushing’s disease. An impaired function of glucocorticoid receptors in patients with Cushing’s disease is also suggested by the observation that coincubation of their MNL with cortisol inhibited interleukin-2 production less than in normal subjects (Sauer et al 1994). A diminished glucocorticoid receptor affinity has been reported in other conditions characterized by activation of the hypothalamic–pituitary– adrenal axis such as the acquired immunodeficiency syndrome (Norbiato et al 1992) and late pregnancy (Junker 1983). The decreased glucocorticoid receptor affinity seen in our patients may easily be interpreted, especially in Cushing’s disease, as an adaptive phenomenon aimed at protecting the tissues from the high circulating levels of glucocorticoids. The possibility that FFA interfered with the estimation of glucocorticoid receptor affinity is unlikely, since their serum levels were normal in anorexia nervosa patients, and their elevated concentrations in patients with Cush-
C. Invitti et al
ing’s disease were not correlated with Kd values. Likewise, the age differences in the three groups of patients can hardly account for our findings, since aging seems to reduce in humans the number of glucocorticoid receptors without modifying their affinity (Grasso et al 1997). Further, in our patients there was no correlation between age on one side and glucocorticoid receptor characteristics and mononuclear leukocytes sensitivity to dexamethasone on the other side. Like the affinity of glucocorticoid receptors, the incorporation of [3H]thymidine into the MNL, an index of DNA synthesis, appeared to be reduced in patients with anorexia nervosa and Cushing’s disease. This observation is in line with a previous report that lymphocyte DNA synthesis is reduced in 53% of patients with anorexia nervosa (Polack et al 1993). Dexamethasone inhibited the [3H]thymidine incorporation into the MNL to a similar extent in patients with anorexia nervosa and controls, whereas this inhibition was definitely smaller in patients with Cushing’s disease. This finding seems to indicate that the sensitivity to glucocorticoids, at least as far as DNA synthesis is concerned, is preserved in anorexia nervosa. This is apparently in contrast with the reduced cortisol suppressibility by dexamethasone occurring in vivo in about 50% of these patients. It might be hypothesized that the effect of glucocorticoids on DNA synthesis, a crucial process for actively duplicating cells such as MNL, is preserved when hypercortisolism is of moderate degree. In contrast, DNA synthesis would not be normally modulated in patients with Cushing’s disease, due to the elevated levels of glucocorticoids. On the other hand, though a decrease in cortisol metabolism (Licinio et al 1996) and an alteration of mineralocorticoid receptors (Ratka et al 1989) might partly account for the impaired cortisol suppressibility by dexamethasone in anorexia nervosa, a primary activation of the hypothalamic–pituitary–adrenal axis, as shown by the increase of cortisol production rate and corticotropinreleasing hormone concentrations in cerebrospinal fluid (Licinio et al 1996), appears to play a major role. The diminished affinity of glucocorticoid receptors was directly related to the degree of hypercortisolism both in patients with anorexia nervosa and in patients with Cushing’s disease. In patients with anorexia nervosa [3H]thymidine incorporation into the MNL was also inversely related to urinary free cortisol and Fdex levels. It is highly likely that the high cortisol levels are responsible for the alterations of glucocorticoid receptors and of [3H]thymidine incorporation observed in this study, considering the antiproliferative effect of glucocorticoids (Cupps and Fauci 1982) and the low PHA-stimulated DNA synthesis found in the lymphocytes of patients with Cushing’s disease (Sauer et al 1994). In conclusion, we have demonstrated that patients with
Glucocorticoid Receptors in Anorexia Nervosa
anorexia nervosa, like those with Cushing’s disease, have a diminished affinity and a normal number of glucocorticoid receptors together with a reduced DNA synthesis in MNL. In patients with anorexia nervosa, unlike in patients with Cushing’s disease, [3H]thymidine incorporation can be further inhibited by dexamethasone. It appears therefore that the lack of cushingoid features in patients with anorexia nervosa is not ascribable to a reduced sensitivity to glucocorticoids. The possibility that it is due to the lack of metabolic substrates appears more likely, although a role of glucocorticoids in the development, in these patients, of some alterations also present in Cushing’s disease such as osteoporosis, muscle wasting, and cerebral atrophy cannot be excluded.
References American Psychiatric Association (1994): Diagnostic and Statistical Manual of Mental Disorders, 4th ed. Washington, DC: American Psychiatric Press. Beck AT, Epstein N, Brown G, Steer RA (1988): An inventory for measuring clinical anxiety: Psychometric properties. J Consult Clin Psychol 56:893– 897. Cupps TR, Fauci AS (1982): Corticosteroid-mediated immunoregulation in man. Immunol Rev 65:133–155. Delarche-Homo F (1984): Glucocorticoid receptors and steroid sensitivity in normal and neoplastic human lymphoid tissues: A review. Cancer Res 44:431– 437. Girardin E, Garoscio-Cholet M, Dechaud H, Lejeune H, Carrier E, Tourniaire J, et al (1991): Glucocorticoid receptors in lymphocytes in anorexia nervosa. Clin Endocrinol (Oxf) 35:79 – 84. Grasso G, Lodi L, Lupo C, Muscettola M (1997): Glucocorticoid receptors in human peripheral blood mononuclear cells in relation to age and to sport activity. Life Sci 61:301–308. Haourigui M, Vallette G, Martin ME, Sumida C, Benassayag C, Nunez EA (1994): In vivo effect of free fatty acids on the specific binding of glucocorticosteroids to corticosteroid binding globulin and liver receptors in immature rats. Steroids 59:46 –54. Junker K (1983): Glucocorticoid receptors of human mononuclear leukocytes in vitro. J Clin Endocrinol Metab 57:506 – 512. Kabbaj M, Piazza PV, Simon H, Le Moal M, Maccari S (1995):
BIOL PSYCHIATRY 1999;45:1467–1471
1471
Opposite effects on hippocampal corticosteroid receptors induced by stimulation of b and a1 noradrenergic receptors. Neuroscience 66:539 –549. Kontula K, Pelkonen R, Andersson L, Sivula A (1980): Glucocorticoid receptors in adrenocorticoid disorders. J Clin Endocrinol Metab 51:654 – 657. Kontula K, Andersson LC, Huttunen M, Pelkonen R (1982): Reduced level of cellular glucocorticoid receptors in patients with anorexia nervosa. Horm Metab Res 14:619 – 620. Licinio J, Wong ML, Gold PW (1996): The hypothalamicpituitary-adrenal axis in anorexia nervosa. Psychiatry Res 62:75– 83. Lyashko VN, Sukhikh GT (1987): Interstrain differences in cellular glucocorticoid reception in mice with suppressed natural killer activity during immobilisation stress. Biull Eksp Biol Med 103:273–276. Murphy BEP (1991): Treatment of major depression with steroid suppressive drugs. J Steroid Biochem Molec Biol 39:239 – 244. Norbiato G, Bevilacqua M, Vago T, Baldi G, Chebat E, Bertora P, et al (1992): Cortisol resistance in acquired immunodeficiency syndrome. J Clin Endocrinol Metab 74:608 – 613. Pardes E, De Yampey JE, Moses DF, De Nicola AF (1991): Regulation of glucocorticoid receptors in human mononuclear cells: Effects of glucocorticoid treatment, Cushing’s disease and ketoconazole. J Steroid Biochem Molec Biol 39:233–238. Polack E, Nahmod VE, Emeric-Sauval E, Bello M, Costas M, Finkielman S, et al (1993): Low lymphocyte interferongamma production and variable proliferative response in anorexia nervosa patients. J Clin Immunol 13:445– 451. Ratka A, Sutanto W, Bloemers M, de Kloet R (1989): On the role of brain mineralocorticoid (type I) and glucocorticoid (type II) receptors in neuroendocrine regulation. Neuroendocrinology 50:117–123. Sauer J, Stalla GK, Mu¨ller OA, Arzt E (1994): Inhibition of interleukin-2-mediated lymphocyte activation in patients with Cushing’s syndrome: A comparison with hypocortisolemic patients. Neuroendocrinology 59:144 –151. Schlechte AJ, Ginsberg BH, Sherman BM (1982): Regulation of the glucocorticoid receptor in human lymphocytes. J Steroid Biochem 16:69 –74. Whalley LJ, Borthwick N, Copolov D, Dick H, Christie JE, Fink G (1986): Glucocorticoid receptors and depression. Br Med J Clin Res Ed 292:859 – 861.
2nd Chair of Endocrinology, University of Milan, Istituto Scientifico Ospedale San Luca, Milan, Italy. Address reprint requests to Prof. Francesco Cavagnini, Istituto Scientifico Ospedale San Luca, Via Spagnoletto, 3 20149 Milan, Italy. Received January 19, 1998; revised May 19, 1998; accepted May 26, 1998.
© 1999 Society of Biological Psychiatry
Key Words: Glucocorticoid receptor, anorexia nervosa, Cushing’s disease, glucocorticoids, hypercortisolism, cortisol
Introduction
P
atients with anorexia nervosa have a glucocorticoid hypersecretion pattern similar to patients with Cushing’s disease; they have elevated urinary free cortisol excretion and high serum diurnal and nocturnal cortisol levels, which are not suppressed by dexamethasone (Licinio et al 1996). Nevertheless, patients with anorexia nervosa do not develop the common clinical features of hypercortisolism. Two major explanations have been given for this phenomenon: 1) these patients have a reduced sensitivity to glucocorticoids due to a pre- or postreceptorial defect of glucocorticoid receptors; 2) their malnutrition led to a reduced availability of the metabolic substrates necessary to develop the effect of glucocorticoids. The studies hypothesizing alterations of glucocorticoid receptors are few. Kontula et al (1982) found reduced levels of glucocorticoid receptors not correlated with morning plasma cortisol concentrations in mononuclear leukocytes (MNL) of patients with anorexia nervosa. In agreement with this, a reduction of glucocorticoid receptor number was described in patients with depression who, like those with anorexia nervosa, have a biochemically but not clinically expressed hypercortisolism (Murphy 1991; Whalley et al 1986). Conversely, in patients with anorexia nervosa, Girardin et al (1991) reported virtually normal lymphocyte glucocorticoid receptor levels although, unlike in normal subjects, they were positively correlated with urinary free cortisol concentrations. None of these papers, however, have also explored the function of glucocorticoid receptors. We therefore decided to investigate glucocorticoid receptor binding capacity, affinity, and sensitivity to glucocorticoids on MNL in a group of patients with underweight anorexia nervosa and in a group of normal-weight control subjects. To evaluate glucocorticoid receptors in a typical condition of biochemical and clinical hypercortisolism, a group of patients with active Cushing’s disease was also investigated. 0006-3223/99/$20.00 PII S0006-3223(98)00189-9
1468
C. Invitti et al
BIOL PSYCHIATRY 1999;45:1467–1471
Methods and Materials Twenty-two patients with anorexia nervosa, 21 women and 1 man, aged 14 –33 years (mean 22.3 6 1.17 years), 10 patients with active Cushing’s disease, 7 women and 3 men, 11–59 years (mean 38.0 6 4.36 years), and 24 healthy volunteers, 22 women and 2 men, 10 –51 years (mean 27.3 6 1.61 years) without psychiatric disorders or mood alterations were investigated. None of the patients was receiving any medication when studied. All gave their informed consent to participate in this study, which was approved by our Institution’s Ethical Committee. Diagnosis of restrictive anorexia nervosa was established according to the criteria of DSM-IV (American Psychiatric Association 1994). Patients were not following a timed, prescribed diet and exhibited a severe food restriction without binging or vomiting behavior and were amenorrheal. Diagnosis of Cushing’s disease was posed on clinical and laboratory grounds, according to the conventional criteria [high serum cortisol concentration lacking circadian variations and suppressibility by low doses of dexamethasone, associated with measurable levels of adrenocorticotropic hormone (ACTH); evidence of an adenoma at pituitary imaging or center to periphery ACTH gradient greater than 3 at inferior petrosal sinuses sampling after administration of corticotropin-releasing hormone]. To assess the presence of depression anorectic and Cushing’s patients completed the self-reported Beck Depression Inventory (BDI) (Beck et al 1988). Rating scales on the BDI were below 20, between 20 and 30, and greater than 30 in 20%, 38%, and 42% respectively of patients with Cushing’s disease and in 20%, 26%, and 54% respectively of patients with anorexia nervosa. As a whole, depressive symptoms were present in 80% of both anorectic and Cushing’s patients. Three 24-hour urine collections were taken for free cortisol assay, and a blood sample was drawn for cortisol in the morning after an oral dose of 1 mg dexamethasone (Fdex) the evening before. Urinary free cortisol levels greater than 80 mg/24 hours as well as serum cortisol levels greater than 5 mg/dL indicate an overdrive of the hypothalamic–pituitary adrenal axis. All patients fasted overnight, and at 8 AM an indwelling venous catheter was inserted in an antecubital vein; a 30 mL blood sample was taken 20 min later in edetic acid (EDTA)containing siliconized glass tubes, and immediately processed to obtain MNL, which express glucocorticoid receptors (DelarcheHomo 1984) and provide an in vitro model to study the sensitivity to glucocorticoids. Ten milliliters of venous blood was taken for serum cortisol and free fatty acids (FFA) measurement.
Glucocorticoid Receptor Assay MNL were isolated by Ficoll-Hypaque sedimentation, washed with phosphate-buffered saline (PBS), and suspended in RPMI1640 medium (Gibco, Grand Island, NY), pH 7.4, containing 25 mmol HEPES, 2 nmol glutamine, 50 mg/mL gentamicin, and 10% fetal calf serum. Cell viability exceeded 95% as judged by trypan blue exclusion. MNL binding capacity (Bmax) was tested by adding increasing concentrations (0.25–16 nmol) of [3H]dexamethasone (83 Ci/ mmol, Amersham, Aylesbury, Buckinghamshire, U.K.) to cell
suspensions in the presence or absence of a 1000-fold molar excess of unlabeled dexamethasone. After incubation at 4°C for 18 hours, cells were harvested on glass fiber filters (Whatman GF/C, Clifton, NJ) and washed twice with cold PBS. Samples were counted in a Packard Tri-Carb liquid scintillation spectrometer (Downers Grove, IL) at an efficiency of 60%. Bmax was expressed as femtomoles of receptors per 106 cells, and the apparent dissociation constant (Kd) was obtained by Scatchard plot analysis. The reception intensity index Bmax/Kd was also used (Lyashko and Sukhikh 1987).
Thymidine Incorporation MNL sensitivity to glucocorticoids was determined in vitro by evaluating the inhibitory effect of dexamethasone on their ability to incorporate [3H]thymidine (25 Ci/mmol, Amersham, Aylesbury, Buckinghamshire, U.K.), which represents an index of DNA synthesis. MNL (1 3 106 cells/mL) were incubated at 37°C in a 5% CO2 incubator for 22 hours in the RPMI-1640 medium, as used before, containing phytohemagglutinin (PHA, 5 mg/mL) and increasing doses of dexamethasone (0.1–100 nmol). Incubation was continued for 6 hours after addition of [3H]thymidine (2 mCi/mL), and the cells were harvested on glass-fiber filters. The radioactivity was measured by liquid scintillation counting. Results were expressed as the percentage of isotope incorporation after dexamethasone treatment.
Biochemical Assays Serum cortisol was assayed by radioimmunoassay (Byk-Sangtec Diagnostica, Dietrenbach, Germany). Intra- and interassay coefficients of variation and sensitivity of the method were respectively 4.4%, 8.1%, and 0.6 mg/dL. To avoid differences due to interassay variability, samples taken from the three groups of subjects were run together in each of the assays described above. Urinary free cortisol was measured by radioimmunoassay (DPC, Los Angeles, CA) after extraction with dichloromethane. Intra- and interassay coefficients of variation and sensitivity of the method were respectively 3.5%, 6.2%, and 0.49 mg/dL. Normal values for our laboratory are 10 – 80 mg/24 hours. Since hypercortisolemia raises serum levels of FFA, and FFA may reduce glucocorticoid binding to their receptors (Haourigui et al 1994), we also measured serum FFA in all patients. Plasma FFA levels were determined using an acyl-CoA oxidase-based colorimetric kit (Wako, Osaka, Japan). The intraassay coefficient of variation was 1.6%, and normal values for our laboratory are 0.1– 0.6 mEq/L. To avoid differences due to interassay variability, samples taken from the three groups of subjects were run together in each of the assays described above.
Statistical Analysis Results were expressed as mean 6 SEM. The Mann–Whitney test was used for comparison among groups. Correlations were examined with linear regression analysis. A p value less than .05 was considered statistically significant.
Glucocorticoid Receptors in Anorexia Nervosa
BIOL PSYCHIATRY 1999;45:1467–1471
1469
Table 1. Body Mass Index (BMI), Defined as Body Weight Divided by Height Squared, Concentrations of 24-Hour Urinary Free Cortisol, Serum Cortisol Measured at 8 AM following 1 mg of Dexamethasone at Previous Midnight (Fdex), Serum Free Fatty Acids (FFA), Binding Capacity (Bmax), and Apparent Dissociation Constant (Kd) for Dexamethasone of Monuclear Leukocytes in Patients with Anorexia Nervosa and Cushing’s Disease and in Control Subjects
2
BMI (kg/m ) Urinary free cortisol (mg/24 hours) Fdex (mg/dL) FFA (mEq/L) Bmax (fmol/106 cells) Kd (nmol/L) Bmax/Kd (3 1023) [3H] thymidine incorporation (cpm 3 103)
Anorexia nervosa
Control subjects
Cushing’s disease
13.3 6 0.40 115.3 6 23.55b,c 9.3 6 1.78a 0.6 6 0.11 3.3 6 0.23 4.0 6 0.31a 0.9 6 0.05c 39.2 6 8.20a
22.1 6 2.01 38.1 6 3.67 1.0 6 0.11 0.4 6 0.06 3.5 6 0.20 2.9 6 0.29 1.3 6 0.07 82.2 6 11.18
30.9 6 2.88a 333.1 6 72.33a 13.3 6 2.36a 0.7 6 0.08d 3.7 6 0.30 4.1 6 0.34a 0.9 6 0.04a 38.0 6 8.39a
a,b
Results are expressed as mean 6 SEM. a p , .001 vs. control subjects. b p , .001 vs. Cushing’s disease. c p , .01 vs. control subjects. d p , .05 vs. control subjects.
Results Concentrations of urinary free cortisol, Fdex, FFA, body mass index (BMI), and results of glucocorticoid receptor binding are shown in Table 1. BMI and urinary free cortisol concentrations were significantly higher in patients with Cushing’s disease than in anorectic patients (p , .001). The number of glucocorticoid receptors, as expressed by Bmax, was comparable in MNL of patients with anorexia nervosa, patients with Cushing’s disease, and control subjects. No correlations were found between Bmax and urinary free cortisol, Fdex, and BMI in patients with anorexia nervosa and Cushing’s disease. On the contrary, glucocorticoid receptor affinity, as expressed by Kd, was significantly decreased in patients with anorexia nervosa as well as in those with Cushing’s disease compared to control subjects (p , .001). Kd was positively correlated with urinary free cortisol both in patients with anorexia nervosa and in patients with Cushing’s disease (r 5 .46, p , .05 in the former and r 5 .81, p , .01 in the latter patients). The integrative index of reception effectiveness (Bmax/Kd) was significantly lower in patients with anorexia nervosa (p , .01) and Cushing’s disease (p , .001) than in control subjects and positively correlated with BMI only in patients with anorexia nervosa (r 5 .60, p , .01). Basal [3H]thymidine incorporation was significantly reduced in MNL of patients with anorexia nervosa and Cushing’s disease compared to control subjects (p , .001). In patients with anorexia nervosa, but not in those with Cushing’s disease, [3H]thymidine incorporation was correlated positively with BMI (r 5 .52, p , .05) and negatively with urinary free cortisol (r 5 20.51, p , .05) and Fdex (r 5 2.53, p , .05). Dexamethasone dosedependently inhibited [3H]thymidine incorporation in all
groups (Figure 1). This effect was definitely lower in patients with Cushing’s disease than in control subjects and in patients with anorexia nervosa (radioisotope uptake after dexamethasone: 67.5 6 5.59% vs. 37.6 6 2.49%, p , .001 and 41.4 6 3.18%, p , .001, respectively). In both groups of patients the residual [3H]thymidine incorporation into MNL recorded after dexamethasone was positively correlated with Fdex (r 5 .54, p , .05 and r 5 .78, p , .05, respectively) and in patients with Cushing’s disease also with urinary free cortisol (r 5 .64, p , .05). Serum FFA levels were comparable in patients with anorexia nervosa and control subjects and significantly higher in patients with Cushing’s disease (p , .05). No correlations were found in each group between FFA levels
Figure 1. Dose-related inhibiting effect of dexamethasone on [3H]thymidine incorporation in MNL obtained from control subjects (●), patients with anorexia nervosa (Œ), and patients with Cushing’s disease (1).
1470
BIOL PSYCHIATRY 1999;45:1467–1471
and Bmax, Kd, and [3H]thymidine incorporation before and after dexamethasone. These latter parameters were not correlated with age both in the whole and in the separate groups of subjects.
Discussion The reason why patients with anorexia nervosa do not develop cushingoid features in spite of even markedly increased serum cortisol levels is unclear. Either a cortisol resistance secondary to alterations in number and/or function of glucocorticoid receptors or a lack of nutritional substrates necessary for the clinical expression of the cortisol effects might be in play. The present study has shown, in agreement with observations by others, that MNL of patients with either anorexia nervosa (Girardin et al 1991) or Cushing’s disease (Pardes et al 1991; Junker 1983; Kontula et al 1980) have a normal concentration of glucocorticoid receptors. This finding appears in contrast with the demonstration, given by some authors (Pardes et al 1991; Schlechte et al 1982) though not by others (Junker 1983), of a down-regulation of glucocorticoid receptor number in the lymphocytes of patients treated with corticosteroids; however, the model of patients on steroid treatment may differ from that of Cushing’s disease and anorexia nervosa, in whom complex neurotransmitter alterations might directly influence the glucocorticoid receptor function (Kabbaj et al 1995). On the other hand, care was exercised to run in the same experiment samples taken from each group of patients to overcome interassay variability. At variance with the authors mentioned above (Girardin et al 1991; Junker 1983; Pardes et al 1991; Kontula et al 1980, 1982) we found a reduced glucocorticoid receptor affinity in our patients with anorexia nervosa and Cushing’s disease. An impaired function of glucocorticoid receptors in patients with Cushing’s disease is also suggested by the observation that coincubation of their MNL with cortisol inhibited interleukin-2 production less than in normal subjects (Sauer et al 1994). A diminished glucocorticoid receptor affinity has been reported in other conditions characterized by activation of the hypothalamic–pituitary– adrenal axis such as the acquired immunodeficiency syndrome (Norbiato et al 1992) and late pregnancy (Junker 1983). The decreased glucocorticoid receptor affinity seen in our patients may easily be interpreted, especially in Cushing’s disease, as an adaptive phenomenon aimed at protecting the tissues from the high circulating levels of glucocorticoids. The possibility that FFA interfered with the estimation of glucocorticoid receptor affinity is unlikely, since their serum levels were normal in anorexia nervosa patients, and their elevated concentrations in patients with Cush-
C. Invitti et al
ing’s disease were not correlated with Kd values. Likewise, the age differences in the three groups of patients can hardly account for our findings, since aging seems to reduce in humans the number of glucocorticoid receptors without modifying their affinity (Grasso et al 1997). Further, in our patients there was no correlation between age on one side and glucocorticoid receptor characteristics and mononuclear leukocytes sensitivity to dexamethasone on the other side. Like the affinity of glucocorticoid receptors, the incorporation of [3H]thymidine into the MNL, an index of DNA synthesis, appeared to be reduced in patients with anorexia nervosa and Cushing’s disease. This observation is in line with a previous report that lymphocyte DNA synthesis is reduced in 53% of patients with anorexia nervosa (Polack et al 1993). Dexamethasone inhibited the [3H]thymidine incorporation into the MNL to a similar extent in patients with anorexia nervosa and controls, whereas this inhibition was definitely smaller in patients with Cushing’s disease. This finding seems to indicate that the sensitivity to glucocorticoids, at least as far as DNA synthesis is concerned, is preserved in anorexia nervosa. This is apparently in contrast with the reduced cortisol suppressibility by dexamethasone occurring in vivo in about 50% of these patients. It might be hypothesized that the effect of glucocorticoids on DNA synthesis, a crucial process for actively duplicating cells such as MNL, is preserved when hypercortisolism is of moderate degree. In contrast, DNA synthesis would not be normally modulated in patients with Cushing’s disease, due to the elevated levels of glucocorticoids. On the other hand, though a decrease in cortisol metabolism (Licinio et al 1996) and an alteration of mineralocorticoid receptors (Ratka et al 1989) might partly account for the impaired cortisol suppressibility by dexamethasone in anorexia nervosa, a primary activation of the hypothalamic–pituitary–adrenal axis, as shown by the increase of cortisol production rate and corticotropinreleasing hormone concentrations in cerebrospinal fluid (Licinio et al 1996), appears to play a major role. The diminished affinity of glucocorticoid receptors was directly related to the degree of hypercortisolism both in patients with anorexia nervosa and in patients with Cushing’s disease. In patients with anorexia nervosa [3H]thymidine incorporation into the MNL was also inversely related to urinary free cortisol and Fdex levels. It is highly likely that the high cortisol levels are responsible for the alterations of glucocorticoid receptors and of [3H]thymidine incorporation observed in this study, considering the antiproliferative effect of glucocorticoids (Cupps and Fauci 1982) and the low PHA-stimulated DNA synthesis found in the lymphocytes of patients with Cushing’s disease (Sauer et al 1994). In conclusion, we have demonstrated that patients with
Glucocorticoid Receptors in Anorexia Nervosa
anorexia nervosa, like those with Cushing’s disease, have a diminished affinity and a normal number of glucocorticoid receptors together with a reduced DNA synthesis in MNL. In patients with anorexia nervosa, unlike in patients with Cushing’s disease, [3H]thymidine incorporation can be further inhibited by dexamethasone. It appears therefore that the lack of cushingoid features in patients with anorexia nervosa is not ascribable to a reduced sensitivity to glucocorticoids. The possibility that it is due to the lack of metabolic substrates appears more likely, although a role of glucocorticoids in the development, in these patients, of some alterations also present in Cushing’s disease such as osteoporosis, muscle wasting, and cerebral atrophy cannot be excluded.
References American Psychiatric Association (1994): Diagnostic and Statistical Manual of Mental Disorders, 4th ed. Washington, DC: American Psychiatric Press. Beck AT, Epstein N, Brown G, Steer RA (1988): An inventory for measuring clinical anxiety: Psychometric properties. J Consult Clin Psychol 56:893– 897. Cupps TR, Fauci AS (1982): Corticosteroid-mediated immunoregulation in man. Immunol Rev 65:133–155. Delarche-Homo F (1984): Glucocorticoid receptors and steroid sensitivity in normal and neoplastic human lymphoid tissues: A review. Cancer Res 44:431– 437. Girardin E, Garoscio-Cholet M, Dechaud H, Lejeune H, Carrier E, Tourniaire J, et al (1991): Glucocorticoid receptors in lymphocytes in anorexia nervosa. Clin Endocrinol (Oxf) 35:79 – 84. Grasso G, Lodi L, Lupo C, Muscettola M (1997): Glucocorticoid receptors in human peripheral blood mononuclear cells in relation to age and to sport activity. Life Sci 61:301–308. Haourigui M, Vallette G, Martin ME, Sumida C, Benassayag C, Nunez EA (1994): In vivo effect of free fatty acids on the specific binding of glucocorticosteroids to corticosteroid binding globulin and liver receptors in immature rats. Steroids 59:46 –54. Junker K (1983): Glucocorticoid receptors of human mononuclear leukocytes in vitro. J Clin Endocrinol Metab 57:506 – 512. Kabbaj M, Piazza PV, Simon H, Le Moal M, Maccari S (1995):
BIOL PSYCHIATRY 1999;45:1467–1471
1471
Opposite effects on hippocampal corticosteroid receptors induced by stimulation of b and a1 noradrenergic receptors. Neuroscience 66:539 –549. Kontula K, Pelkonen R, Andersson L, Sivula A (1980): Glucocorticoid receptors in adrenocorticoid disorders. J Clin Endocrinol Metab 51:654 – 657. Kontula K, Andersson LC, Huttunen M, Pelkonen R (1982): Reduced level of cellular glucocorticoid receptors in patients with anorexia nervosa. Horm Metab Res 14:619 – 620. Licinio J, Wong ML, Gold PW (1996): The hypothalamicpituitary-adrenal axis in anorexia nervosa. Psychiatry Res 62:75– 83. Lyashko VN, Sukhikh GT (1987): Interstrain differences in cellular glucocorticoid reception in mice with suppressed natural killer activity during immobilisation stress. Biull Eksp Biol Med 103:273–276. Murphy BEP (1991): Treatment of major depression with steroid suppressive drugs. J Steroid Biochem Molec Biol 39:239 – 244. Norbiato G, Bevilacqua M, Vago T, Baldi G, Chebat E, Bertora P, et al (1992): Cortisol resistance in acquired immunodeficiency syndrome. J Clin Endocrinol Metab 74:608 – 613. Pardes E, De Yampey JE, Moses DF, De Nicola AF (1991): Regulation of glucocorticoid receptors in human mononuclear cells: Effects of glucocorticoid treatment, Cushing’s disease and ketoconazole. J Steroid Biochem Molec Biol 39:233–238. Polack E, Nahmod VE, Emeric-Sauval E, Bello M, Costas M, Finkielman S, et al (1993): Low lymphocyte interferongamma production and variable proliferative response in anorexia nervosa patients. J Clin Immunol 13:445– 451. Ratka A, Sutanto W, Bloemers M, de Kloet R (1989): On the role of brain mineralocorticoid (type I) and glucocorticoid (type II) receptors in neuroendocrine regulation. Neuroendocrinology 50:117–123. Sauer J, Stalla GK, Mu¨ller OA, Arzt E (1994): Inhibition of interleukin-2-mediated lymphocyte activation in patients with Cushing’s syndrome: A comparison with hypocortisolemic patients. Neuroendocrinology 59:144 –151. Schlechte AJ, Ginsberg BH, Sherman BM (1982): Regulation of the glucocorticoid receptor in human lymphocytes. J Steroid Biochem 16:69 –74. Whalley LJ, Borthwick N, Copolov D, Dick H, Christie JE, Fink G (1986): Glucocorticoid receptors and depression. Br Med J Clin Res Ed 292:859 – 861.