Glucocorticoid receptor-mediated effects on rat fibrosarcoma growth.

Glucocorticoid receptors are present in most normal and malignant mammalian cells. To examine the hypothesis that the growth of methylcholanthrene-induced malignant sarcoma is glucocorticoid dependent, we evaluated the behavior of malignant fibrosarcoma (MCA) in adrenalectomized rats treated with either normal saline or deoxycorticosterone acetate and in intact rats treated with placebo or with the glucocorticoid receptor antagonist RU 486. Survival, tumor weight, and loss of body weight (an index of cachexia) were measured. In MCA-bearing rats, neither survival nor loss of body weight was affected by bilateral adrenalectomy or by treatment with RU 486. Tumor weight and time-integrated tumor volume, however, were significantly less in bilaterally adrenalectomized rats without deoxycorticosterone acetate replacement than in animals treated with deoxycorticosterone acetate. Similarly, tumor weight and time-integrated tumor volume were less in intact animals treated with RU 486 than in intact animals treated with placebo. The glucocorticoid receptors in the tumor cells had similar binding capacity (Ro) and equilibrium dissociation constant (Kd) as in control rat fibroblasts. These results suggest that the growth of MCA sarcoma cells is partially dependent upon glucocorticoids. This effect of glucocorticoids, however, was not of sufficient magnitude to improve survival and prevent cachexia. We conclude that glucocorticoids appear to influence MCA sarcoma growth in the rat, and that glucocorticoid receptor blockade, perhaps in combination with other antitumor agents, merits future study in the treatment of malignant tumors.

Compound heterozygous mutations of the luteinizing hormone receptor gene in Leydig cell hypoplasia.

The human LH receptor (hLHR) is a member of the G protein-coupled receptors characterized by the presence of seven-transmembrane (TM) helices. Inactivating mutations of the hLHR lead to Leydig cell hypoplasia (LCH), a form of male pseudohermaphroditism resulting from the failure of fetal testicular Leydig cell differentiation. We have identified three mutations of the hLHR in a patient with LCH: deletion of exon 8 (delta Exon 8), A872G transition resulting in Asn291Ser substitution in the extracellular domain, and C1847A transversion resulting in Ser616Tyr substitution in the seventh TM helix. Nucleotide sequencing, gene dosage, and allele-specific amplification analyses revealed that exon 8 deletion and the two missense mutations are present in different alleles of the hLHR. Constructs of mutated hLHR (hLHR-delta Exon8, hLHR-872/1847, hLHR-1847, and hLHR-872) were used to transfect 293 cells, and the properties of the hLHR expressed were examined. Ligand-binding assays failed to detect the expression of hLHR-delta Exon8. Transfectants expressing hLHR-872/1847 demonstrated greatly reduced ligand binding and ligand-induced cAMP accumulation in comparison to those expressing wild type hLHR. Similar reduction in cAMP accumulation was observed in transfectants expressing hLHR-1847, but not hLHR-872 alone. These findings suggest that, in addition to the 7-TM helices, the polypeptide encoded by exon 8 plays an important role in LHR expression and signal transduction. On the other hand, glycosylation of Asn291 may not be critical for these activities. These results also establish that LCH can result from impaired signal transduction due to compound heterozygous mutations. Implications of these mutations on structure-function relationship of the hLHR and the genotype-phenotype correlation in LCH are discussed.

Inactivation of the luteinizing hormone/chorionic gonadotropin receptor by an insertional mutation in Leydig cell hypoplasia.

We previously identified a nonsense mutation (Cys545Stop) in the paternal human LH/CG receptor (hLHR) allele in a family with two 46,XY children afflicted with Leydig cell hypoplasia. This mutation abolished the signal transduction capability of the affected hLHR. We have now examined all coding exons and the transcript of both alleles of the hLHR gene of the affected children. A 33-bp in-frame insertion was found in the maternal hLHR allele. This insertion occurred between nucleotide 54 and 55 and might be the result of a partial gene duplication. Genomic DNA-PCR showed that this defective maternal hLHR allele was inherited by the two affected children. However, examination of the inheritance of the 935-A/G polymorphism of the hLHR by genomic- and RT-PCR indicated that the maternal hLHR allele was not expressed in cultured fibroblasts of the patients. The effect of the in-frame insertion on the biological activity of the hLHR was examined by expressing the mutated hLHR construct, generated by site-directed mutagenesis, in HEK 293 cells. The expression of the mRNA for the mutant hLHR in HEK 293 cells was not affected. Response of cells expressing the mutated hLHR to hCG stimulation was impaired as demonstrated by reduced intracellular cAMP biosynthesis. This change in signal transduction was the result of a profound reduction in hormone binding at the cell surface due to altered expression and processing of the mutated receptor. We conclude that Leydig cell hypoplasia in this family is the result of compound heterozygous loss-of-function mutations of the hLHR gene.

Effects of fasting on the growth plate: systemic and local mechanisms.

Inadequate caloric intake inhibits longitudinal bone growth. This study was designed to investigate the mechanisms responsible for this suppression of growth plate function, focusing on the roles of systemic and local insulin-like growth factor 1 (IGF-1). Five week-old male rabbits were fasted for 48 h. Fasting significantly decreased proximal tibial growth velocity and growth plate width (both proliferative and hypertrophic zones). During the fast, systemic IGF-1 production was down-regulated. Serum IGF-1 levels and hepatic IGF-1 messenger RNA (mRNA) levels decreased despite increased GH levels. Serum levels of GH binding protein (a circulating fragment of the GH receptor) and hepatic GH receptor mRNA levels were not significantly changed. In contrast, the local, growth plate IGF-1 system appeared to be up-regulated. Growth plate GH receptor mRNA and IGF-1 mRNA levels were both increased during fasting. We conclude that, in the rabbit, fasting induces a rapid depletion of growth plate chondrocytes and inhibition of longitudinal bone growth. These effects appear to be mediated by systemic endocrine mechanisms; circulating IGF-1 levels are diminished because of hepatic resistance to GH. In contrast, the local, paracrine IGF-1 system in growth plate does not appear to contribute to the growth inhibition but instead appears to be up-regulated by fasting.

Treatment of familial male precocious puberty with spironolactone, testolactone, and deslorelin.

Combined antiandrogen (spironolactone) and aromatase inhibitor (testolactone) are effective for the short term treatment of familial male precocious puberty. During this therapy, plasma testosterone levels remain in the adult range, since spironolactone blocks the testosterone receptor without significantly affecting plasma testosterone levels. After our initial 18-month pilot study, we continued to treat eight boys with the combined therapy for 2.0-4.2 yr. During this time all boys exhibited a pubertal rise in gonadotropin secretion and a diminishing response to treatment, which was manifested by the recurrence of clinical features of puberty and an increase in the bone maturation rate (P < 0.05). Addition of the LHRH agonist deslorelin (4 micrograms/kg.day, sc) to the combined therapy decreased peak LH, plasma testosterone, bone maturation rate, and growth velocity (P < 0.05) over the next year. We conclude that the rise in gonadotropin levels during central activation of hypothalamic LHRH secretion in boys with familial male precocious puberty causes a partial escape from the combined effect of spironolactone and testolactone. The addition of deslorelin to the combined therapy appears to restore the control of puberty in this setting.

Effect of chronic treatment with the glucocorticoid antagonist RU 486 in man: toxicity, immunological, and hormonal aspects.

Suppression of immune function was traditionally thought to occur only with pharmacological levels of glucocorticoids. However, recent studies in rodents have suggested that glucocorticoids exert tonic antiinflammatory/immunosuppressive effects even at basal nonstress concentrations. To examine whether basal glucocorticoid secretion modulates immune function in man we employed the specific glucocorticoid receptor antagonist RU 486. If a tonic level of inhibition of the immune system by basal glucocorticoid levels was present, then a potentiation or enhancement of immune function might evolve in the absence of glucocorticoid action. To examine this hypothesis, we studied 11 healthy male normal volunteers who received RU 486 (10 mg/kg.day) or placebo vehicle, divided into 2 daily oral doses, for 7-14 days. Blood samples were collected every 2 days for measurement of plasma ACTH and cortisol concentrations along with 24-h urine samples for measurement of 17-hydroxysteroid and free cortisol excretion. Complete and differential blood counts, erythrocyte sedimentation rates, C-reactive protein, antinuclear antibodies, rheumatoid factor, and quantitative immunoglobulins were also determined at 2-day intervals. Leukocytes were obtained by leukopheresis for phenotypic characterization and functional analysis before and 7 days after the initiation of RU 486 or placebo therapy. Blockade of cortisol receptors with RU 486 was associated with marked compensatory elevations of plasma ACTH and cortisol and increases in 24-h urinary excretion of 17-hydroxysteroids and free cortisol. Unexpectedly, 8 of the 11 subjects developed generalized exanthem after 9 days of RU 486 treatment. One subject developed symptoms and signs consistent with the diagnosis of adrenal insufficiency. Total white blood cell counts, absolute lymphocyte, neutrophil and eosinophil counts, erythrocyte sedimentation rate, and quantitative immunoglobulins did not change with RU 486 therapy. Similarly, T-, B-, and natural killer cell subsets did not change during RU 486 treatment. Furthermore, functional evaluation of lymphocyte cytotoxicity and proliferation revealed no changes. We conclude that administration of high doses of RU 486 to normal volunteers does not result in measurable enhancement of immune function. This suggests that in man, glucocorticoids may not exert a tonic inhibitory effect on the immune system as they appear to do in rodents. Alternatively, the compensatory increase in endogenous cortisol may obviate any effect of the glucocorticoid antagonist on the immune system.(ABSTRACT TRUNCATED AT 400 WORDS)

The antiglucocorticoid and antiprogestin steroid RU 486 suppresses the adrenocorticotropin response to ovine corticotropin releasing hormone in man.

The glucocorticoid and progesterone antagonist RU 486 normalizes the clinical and biochemical features of hypercortisolism in patients with nonpituitary Cushing's syndrome, presumably by antagonizing the action(s) of cortisol. Since RU 486 has progesterone agonist activity in addition to its progesterone antagonist action, the possibility that it might have some glucocorticoid agonist action did not seem unreasonable. To test this hypothesis we examined the effects of RU 486 on pituitary ACTH secretion in 10 patients with primary adrenal insufficiency in whom glucocorticoid replacement was withheld for 36 h. Each patient received, in randomized sequence 3-7 days apart, an oral dose of placebo, RU 486 (20 mg/kg), cortisol (0.1 mg/kg), or a combination of RU 486 and cortisol at 1800 h. Two hours later, an iv bolus dose of ovine CRH (1 microgram/kg) was administered, and plasma ACTH levels were measured serially for 3 h. RU 486 suppressed ovine CRH-stimulated ACTH secretion, albeit less than cortisol. Its glucocorticoid agonist effect was calculated to be approximately 1/250th that of cortisol on a weight basis. Additionally, RU 486 partially antagonized cortisol-induced suppression of ACTH secretion. These findings suggest that RU 486 is a partial glucocorticoid agonist and offer some insight as to its action in patients with Cushing's syndrome. Whether this degree of glucocorticoid agonist activity is adequate to support life, however, is not known.

PIP: The purpose of this study was to assess the glucocorticoid agonist activity of the antiprogestin steroid RU-486 by examining its ability to exert a glucocorticoid-like negative feedback effect on pituitary adrenocorticotropic hormone secretion. 10 patients with nonpituitary Cushing's syndrome, from whom cortisol therapy had been withheld for 36 hours, were given an oral dose of either a placebo, cortisol, or 20mg/kg of RU-480. 2 hours later, they were given an intravenous injection of lug/Kg of ovine corticotropin-releasing hormone. Blood samples were taken 15 minutes before the injection, at the time of the injection, and 15, 30, 60, 90, 120, and 180 minutes afterwards to measure adrenocorticotropic hormone and cortisol levels. Blood samples from the patients who received the RU-486 showed that RU-486 had suppressed the ovine corticotropin-releasing hormone-stimulated secretion of adrenocorticotropic hormone. However, the RU-486-induced suppression of the secretion of adrenocorticotropic hormone was only 80% that of the suppression induced by .1 mg/kg of cortisol, i.e., RU-486 had only 1/250 the glucocorticoid agonist effect of cortisol. RU-486, therefore, is a partial glucocorticoid agonist, but the effect is not of sufficient magnitude to prevent adrenal insufficiency in patients with nonpituitary Cushing's syndrome.

The antiglucocorticoid and antiprogestin steroid RU 486: its glucocorticoid agonist effect is inadequate to prevent adrenal insufficiency in primates.

The glucocorticoid receptor antagonist RU 486 has been used to treat the hypercortisolism of patients with nonpituitary Cushing's syndrome. Since endogenous cortisol production fluctuates in many patients with either the ectopic ACTH syndrome or adrenocortical tumors, treatment of these patients with a fixed dose of RU 486 introduces the risk of adrenal insufficiency. While RU 486 possesses some glucocorticoid agonist activity in addition to its potent antagonist effects, it is not known whether this intrinsic agonist activity is of sufficient magnitude to prevent adrenal insufficiency and sustain life. To answer this question three groups of bilaterally adrenalectomized cynomolgus monkeys (n = 5/group) were randomized to receive a daily injection of RU 486 (5 mg/kg.day), cortisol (1.25 mg/kg.day), or saline (placebo). All adrenalectomized monkeys received weekly im injections of deoxycorticosterone pivalate (1 mg) to prevent mineralocorticoid deficiency. Five sham-adrenalectomized monkeys served as controls and received im injections of saline (placebo). Blood was collected before adrenalectomy or sham operation and every 3 days postoperatively for measurement of serum electrolytes, blood urea nitrogen, and creatinine; plasma ACTH concentrations; and complete blood and differential cell counts. All sham-operated and cortisol-replaced adrenalectomized monkeys survived, and none developed overt biochemical evidence of adrenal insufficiency. All placebo and RU 486-replaced adrenalectomized monkeys expired within 33 days after adrenalectomy, presumably from adrenal insufficiency. These findings suggest that while RU 486 is a partial glucocorticoid agonist, its degree of glucocorticoid agonism is inadequate to prevent adrenal insufficiency and support life in adrenalectomized primates.

The effect of mevinolin on steroidogenesis in patients with defects in the low density lipoprotein receptor pathway.

Adrenal steroidogenesis is dependent upon cholesterol derived from both de novo biosynthesis and uptake of plasma lipoproteins through the low density lipoprotein (LDL) receptor pathway. Recent studies have demonstrated that patients homozygous for familial hypercholesterolemia have a mild impairment in cortisol secretion during maximal ACTH stimulation. Mevinolin, a competitive inhibitor of 3-hydroxy-3-methylglutaryl coenzyme A reductase, has been used clinically to inhibit de novo cholesterol synthesis in patients with familial hypercholesterolemia. In this study we examined hypothalamic-pituitary-adrenal function in seven patients with defects in the LDL receptor pathway, both before and during treatment with oral mevinolin (20 mg, twice a day), to assess whether inhibition of cholesterol synthesis influences steroidogenesis under basal conditions and in response to ovine CRH and exogenous ACTH. Two months after initiation of therapy, high density lipoprotein cholesterol levels were significantly elevated, and LDL cholesterol levels were reduced, although not normalized. Basal and ovine CRH-stimulated adrenocortical function were normal in all patients both before and during therapy. Plateau plasma cortisol concentrations achieved during maximal ACTH stimulation were lower than those in control subjects in all patients both before and during therapy. All patients, however, had an approximately 3-fold increase over basal values. These results suggest that treatment of patients with defects in the LDL receptor pathway with mevinolin improves the plasma lipid profile and does not result in adrenal dysfunction or further exacerbation of the mild impairment of adrenal function during maximal ACTH stimulation.

Adrenal androgen secretion in postadolescent acne: increased adrenocortical function without hypersensitivity to adrenocorticotropin.

Basal and ACTH-stimulated plasma levels of cortisol, delta 4-androstenedione, and dehydroepiandrosterone (DHEA) were measured in a group of 11 female patients with postadolescent acne resistant to or relapsing after conventional therapy and in a group of 10 normal women without acne or hirsutism. Each patient received, in a blinded random fashion, a series of 5 1-h ACTH tests. For each test a different dose of ACTH-(1-24) was administered, ranging from 0-1 microgram/kg, given as an iv bolus. Blood samples were collected 0, 10, 30, and 60 min after ACTH bolus injection. Patients with acne had slightly higher concentrations of basal cortisol, delta 4-androstenedione, and DHEA than normal controls (P less than 0.05). After ACTH-(1-24) stimulation, the same patients had greater peak and time-integrated DHEA concentrations (P less than 0.03). The ED50 values of the cortisol dose-response curves were similar in patients and normal women (P less than 0.05), suggesting that there are no differences in the sensitivity of the adrenal cortex to ACTH between the acne patients and the controls studied. The ratio of DHEA to cortisol response was significantly elevated in women with acne compared to that in control women, suggesting some preponderance of the delta 5 pathway of steroidogenesis in acne (P less than 0.05). These findings of basal and ACTH-stimulated hypersecretion of delta 5-androgens in patients with postaldolescent acne are consistent with an increased volume of androgen-secreting tissue, rather than hypersensitivity of the adrenal zona reticularis to ACTH.

A preliminary study of flutamide, testolactone, and reduced hydrocortisone dose in the treatment of congenital adrenal hyperplasia.

Treatment outcome in congenital adrenal hyperplasia is often suboptimal due to hyperandrogenism, treatment-induced hypercortisolism, or both. As a new approach, we hypothesized that the effects of androgen could be blocked by an antiandrogen (flutamide) and an inhibitor androgen to estrogen conversion (testolactone), thus allowing the hydrocortisone dose to be reduced. We conducted a short term pilot study in 12 children with congenital adrenal hyperplasia in a randomised cross-over open design to determine whether flutamide, testolactone, reduced hydrocortisone dose, and fludrocortisone are more effective than hydrocortisone and fludrocortisone treatment in normalizing linear growth, weight gain, and bone maturation. Each regimen was administered for 6 months, with a 3-month washout period, consisting of hydrocortisone and fludrocortisone treatment, between regimens. Compared to hydrocortisone and fludrocortisone treatment, the regimen of flutamide, testolactone, reduced hydrocortisone dose (from 12.9 to 7.9 mg/m2 day), and fludrocortisone produced an increase in plasma 17-hydroxyprogesterone levels (P < 0.05) and a decline in urinary cortisol (P < 0.01), linear growth rate (-0.9 +/- 0.5 vs. 1.4 +/- 0.6 SD U; P = 0.003), weight velocity (-0.80 +/- 4.0 vs 0.6 +/- 0.4 SD U; P = 0.01), and bone maturation (0.6 +/- 0.6 vs. 1.4 +/- 0.9 yr bone age/yr chronological age; P = 0.02). Although no important adverse effects were observed, the known potential for flutamide-induced hepatotoxicity made frequent monitoring essential. We conclude that the regimen of flutamide, testolactone, reduced hydrocortisone does, and fludrocortisone improve the short term control of growth and bone maturation in children with congenital adrenal hyperplasia. Long term studies are required to determine whether this approach can improve these children's growth and development.

Evidence for impaired activation of the hypothalamic-pituitary-adrenal axis in patients with chronic fatigue syndrome.

Chronic fatigue syndrome is characterized by persistent or relapsing debilitating fatigue for at least 6 months in the absence of a medical diagnosis that would explain the clinical presentation. Because primary glucocorticoid deficiency states and affective disorders putatively associated with a deficiency of the arousal-producing neuropeptide CRH can be associated with similar symptoms, we report here a study of the functional integrity of the various components of the hypothalamic-pituitary-adrenal axis in patients meeting research case criteria for chronic fatigue syndrome. Thirty patients and 72 normal volunteers were studied. Basal activity of the hypothalamic-pituitary-adrenal axis was estimated by determinations of 24-h urinary free cortisol-excretion, evening basal plasma total and free cortisol concentrations, and the cortisol binding globulin-binding capacity. The adrenal cortex was evaluated indirectly by cortisol responses during ovine CRH (oCRH) stimulation testing and directly by cortisol responses to graded submaximal doses of ACTH. Plasma ACTH and cortisol responses to oCRH were employed as a direct measure of the functional integrity of the pituitary corticotroph cell. Central CRH secretion was assessed by measuring its level in cerebrospinal fluid. Compared to normal subjects, patients demonstrated significantly reduced basal evening glucocorticoid levels (89.0 +/- 8.7 vs. 148.4 +/- 20.3 nmol/L; P less than 0.01) and low 24-h urinary free cortisol excretion (122.7 +/- 8.9 vs. 203.1 +/- 10.7 nmol/24 h; P less than 0.0002), but elevated basal evening ACTH concentrations. There was increased adrenocortical sensitivity to ACTH, but a reduced maximal response [F(3.26, 65.16) = 5.50; P = 0.0015). Patients showed attenuated net integrated ACTH responses to oCRH (128.0 +/- 26.4 vs. 225.4 +/- 34.5 pmol/L.min, P less than 0.04). Cerebrospinal fluid CRH levels in patients were no different from control values (8.4 +/- 0.6 vs. 7.7 +/- 0.5 pmol/L; P = NS). Although we cannot definitively account for the etiology of the mild glucocorticoid deficiency seen in chronic fatigue syndrome patients, the enhanced adrenocortical sensitivity to exogenous ACTH and blunted ACTH responses to oCRH are incompatible with a primary adrenal insufficiency. A pituitary source is also unlikely, since basal evening plasma ACTH concentrations were elevated. Hence, the data are most compatible with a mild central adrenal insufficiency secondary to either a deficiency of CRH or some other central stimulus to the pituitary-adrenal axis. Whether a mild glucocorticoid deficiency or a putative deficiency of an arousal-producing neuropeptide such as CRH is related to the clinical symptomatology of the chronic fatigue syndrome remains to be determined.

Sex differences in sensitivity of the hypothalamic-pituitary-adrenal axis.

Two studies examined sex differences in responsiveness of the hypothalamic-pituitary-adrenal cortical axis, a major component of the stress response. The first measured pituitary-adrenal responses to ovine corticotropin-releasing hormone (oCRH) in 24 health men and 19 healthy women. Plasma adrenocorticotropin hormone (ACTH) response to oCRH were significantly greater among women than among men. In contrast, cortisol concentrations were similar in both groups, though elevations were more prolonged in women. Differences in corticotropin-releasing activity between men and women may help account for these findings; such differences in central components of the stress response might play a role in the known epidemiological differences in diseases of stress system dysregulation between men and women.

Ligand-independent hormone secretion.

The secretion of many hormones is regulated by extracellular signals, such as hormones, growth factors, neurotransmitters, and ions, that mediate signal transduction via a G protein-coupled pathway. Three components comprise the G protein-coupled pathway: the G protein-coupled receptor, the G protein, and the effector. G protein-coupled receptors allow cells to respond to external stimuli and comprise a large superfamily with hundreds of members. G proteins function as signal transducers between ligand-bound receptors and intracellular effectors. G protein-regulated effectors include enzymes of second messenger metabolism, such as adenylyl cyclase, phospholipase C, cyclic GMP phosphodiesterase, and ion channels. Abnormalities in any of these three components alter signal transduction and can lead to human disease. For example, mutations of G protein-coupled receptors that promote G protein activation in the absence of an agonist cause retinitis pigmentosa, hyperthyroidism due to hyperfunctioning thyroid adenomas and thyroid hyperplasia, male-limited precocious puberty, and hypocalcemia. Human disorders attributed to constitutively activating mutations of the alpha subunit of Gs include the McCune-Albright syndrome, adrenocorticotropic hormone-independent Cushing's syndrome, and functional endocrine tumors.

Growth and neuroendocrine dysfunction in children with acquired immunodeficiency syndrome.

To assess whether neuroendocrine dysfunction is present in children with acquired immunodeficiency syndrome (AIDS) and growth failure, we evaluated the thyroid, adrenal, and growth hormone-insulin-like growth factor I (IGF-1) axes in nine children with AIDS and failure to thrive. Basal thyroid-stimulating hormone, free thyroxine, and triiodothyronine levels were normal in eight of the nine children and indicated primary hypothyroidism in one child; thyroxine levels were elevated in four and normal in five children. Thyroxine-binding globulin levels were elevated in all children. Serial measurements of thyroid-stimulating hormone, made hourly from 2 to 6 pm and from 10 pm to 2 am, revealed a flat diurnal rhythm of thyroid-stimulating hormone in six children, which may indicate early central hypothyroidism, and a normal nocturnal rise in the remaining three children. Basal plasma corticotropin and aldosterone levels were normal in all children, plasma renin levels were normal in three and elevated in six children, and cortisol levels were normal or elevated in all children. Corticotropin-stimulated cortisol levels exceeded 500 nmol/L (18 micrograms/dl) in all children except one, who was receiving treatment with ketoconazole. Thus adrenocortical function appeared to be grossly intact. The peak growth hormone responses to provocative testing was normal (greater than 7 ng/ml) in eight children and low in one child. The plasma level of insulin-like growth factor I was normal in eight of the nine children and low in one child. We conclude that growth failure in children with AIDS does not usually result from a recognized endocrine cause and that adrenal function is usually normal. However, endocrine deficiency may contribute to morbidity in some children with AIDS.