A questionnaire survey on the management of Graves' orbitopathy in Europe.

OBJECTIVE: To determine management patterns among clinicians who treat patients with Graves' orbitopathy (GO) in Europe. DESIGN AND METHODS: Questionnaire survey including a case scenario of members of professional organisations representing endocrinologists, ophthalmologists and nuclear medicine physicians. RESULTS: A multidisciplinary approach to manage GO was valued by 96.3% of responders, although 31.5% did not participate or refer to a multidisciplinary team and 21.5% of patients with GO treated by responders were not managed in a multidisciplinary setting. Access to surgery for sight-threatening GO was available only within weeks or months according to 59.5% of responders. Reluctance to refer urgently to an ophthalmologist was noted by 32.7% of responders despite the presence of suspected optic neuropathy. The use of steroids was not influenced by the age of the patient, but fewer responders chose to use steroids in a diabetic patient (72.1 vs 90.5%, P<0.001). Development of cushingoid features resulted in a reduction in steroid use (90.5 vs 36.5%, P<0.001) and increase in the use of orbital irradiation (from 23.8% to 40.4%, P<0.05) and surgical decompression (from 20.9 to 52.9%, P<0.001). More ophthalmologists chose surgical decompression for patients with threatened vision due to optic neuropathy, who were intolerant to steroids than other specialists (70.3 vs 41.8%, P<0.01). CONCLUSION: Deficiencies in the management of patients with GO in Europe were identified by this survey. Further training of clinicians, easier access of patients to specialist multidisciplinary centres and the publication of practice guidelines may help improve the management of this condition in Europe.

Pathogenesis of thyroid-associated ophthalmopathy.

Thyroid-associated ophthalmopathy is a complex disease whose pathogenesis is thought to be autoimmune. Evidence has accumulated to implicate lymphocytes, antibodies, and cytokines targeting the orbital tissues whose structural and functional characteristics are unique. Epitope sharing between the orbit and the thyroid is the likely explanation for the close association of ophthalmopathy with autoimmune thyroid disease. Environmental and other nonautoimmune factors are identified that also play a part in the initiation or perpetuation of the disease process.

The stimulation of thyroid hormone secretion in vitro by thyroid-stimulating antibodies.

A new simple technique is described for the in vitro assay of thyroid-stimulating antibody (TSAb) in serum. Thyroid slices are incubated with serum in two-compartmental dialysis pot, followed by measurement of the thyroid hormone concentration in the dialysis fluid by RIA. Heat-inactivated normal serum applied directly to the tissue has no effect on the basal level of response. A dose-dependent relationship is demonstrated for both TSH and TSAb. Serum from patients with untreated Graves' hyperthyroidism is clearly shown to stimulate thyroid hormone secretion. Porcine thyroid tissue has been used and responds well to TSAb. This technique not only has the advantage of simple methodology, but also reflects the biological activity of TSAb. Therefore it will prove useful for further studies of the significance of TSAb and the pathogenesis of Graves' disease.

G protein gene mutations in patients with multiple endocrinopathies.

Point mutations of G protein genes that result in the constitutive activation of G proteins have been described. Such mutations have been shown to occur in a number of endocrine diseases. We have examined tissues from patients having more than one organ affected by an endocrine disorder and patients having separate distinct endocrine diseases for G protein gene mutations. G protein genes encoding for Gs alpha and Gi2 alpha were examined for activating mutations at codons 201 and 227 (Gs alpha) and codons 179 and 205 (Gi2 alpha) using site-directed oligonucleotide hybridization and direct sequencing of tissue DNA amplified by polymerase chain reaction. Tissues from six patients were examined. The only mutation that was identified was at codon 201 of Gs alpha (gsp), which encoded a change from arginine to cysteine. Patient 1 had the mutation in a corticotroph adenoma, a chemodectoma, and a nodular hyperplastic adrenal gland. patient 2 had the mutation in an extraadrenal pheochromocytoma, but an adrenal gland with medullary hyperplasia was wild-type. Patient 3 had an aggressive corticotroph adenoma and developed Nelson's syndrome after bilateral adrenalectomy. The corticotroph adenoma was wild-type, but both hyperplastic adrenal glands had the mutation. Patient 4 had the mutation in a parathyroid adenoma and in two hyperplastic parathyroid glands. Patient 5 had the mutation in both a primary and a metastatic pheochromocytoma. Patient 6 had the mutation in a parathyroid adenoma and also in histologically normal thyroid and parathyroid tissue. Leukocyte DNA was examined from five patients and was found to be wild-type in all cases. We conclude that G protein gene mutations occur in a wider range of endocrine conditions than has been recognized hereto. In addition, the presence of gsp mutations in different endocrine disorders in the same patient is suggestive of a common underlying etiology.

G protein and thyrotropin receptor mutations in thyroid neoplasia.

The cAMP pathway plays a central role in thyroid follicular cell growth and function. Mutations of the TSH receptor (TSHR) or G proteins (gsp) that activate adenylyl cyclase have been identified in autonomously functioning thyroid nodules. Gsp mutations have been identified also in other forms of thyroid neoplasia, but their reported prevalence has been extremely variable. We have studied the prevalence of gsp mutations and activating mutations of Gi2 alpha (gip) in a series of 66 benign and 34 malignant thyroid tumors. Thirty-six tumors were from Boston and 64 from the UK. In addition, we examined the 64 UK tumors for mutations of the TSHR gene. DNA extracted from fresh-frozen or paraffin-embedded tissue was amplified by PCR and examined for mutations using oligonucleotide-specific hybridization and single-strand conformation polymorphism analysis. No G protein gene mutations were identified in the Boston tumors. One gsp mutation, R201C, in a Hürthle cell adenoma and 1 gip mutation, R179C, in a follicular adenoma were demonstrated in tumors from the UK. Oligonucleotide-specific hybridization and single-strand conformation polymorphism analysis of the UK tumors did not demonstrate any mutations of the TSHR gene. Eleven normal thyroid tissue samples were wild-type for Gs alpha, Gi2 alpha, and the TSHR gene.

Concurrence of Pendred syndrome, autoimmune thyroiditis, and simple goiter in one family.

Pendred syndrome is the autosomal recessively transmitted association of familial goiter and congenital deafness. There is no specific biochemical marker of this disease, and the diagnosis depends upon the demonstration of the triad of congenital sensorineural hearing loss, goiter, and abnormal perchlorate discharge test. Pendred syndrome is caused by mutations within the putative ion transporter gene (PDS gene), located on chromosome 7q. A wide variation in the clinical presentation of this condition, and its well documented phenotypic overlap with other thyroid disorders (such as Hashimoto's thyroiditis), can lead to diagnostic difficulties. The potential for misdiagnosis increases when these disorders occur coincidentally in the same family. We describe a kindred in which Pendred syndrome, autoimmune thyroiditis, and simple goiter coexisted, to highlight these diagnostic pitfalls and to illustrate the use of mutational analysis in resolving diagnostic confusion.

Evidence for a Graves' disease susceptibility locus at chromosome Xp11 in a United Kingdom population.

Graves' disease (GD), which has a strong female preponderance (female/male ratio, >5:1), is inherited as a complex genetic trait. Loci for GD have started to be defined using genome-wide approaches for genetic linkage. To date, 3 loci have been confirmed in at least 2 cohorts of GD patients, the strongest effect being at the cytotoxic T lymphocyte antigen-4 (CTLA-4) locus on chromosome 2q33 in our population. Two other loci for GD have recently been proposed, but not confirmed, on chromosomes Xq21 (GD3) and 14q31 (GD1). We studied a cohort of 75 sibling pairs with GD from the United Kingdom for linkage to 12 markers over a 83-cM region of the X chromosome and for 8 markers over a 36-cM region of 14q31-q33. A peak multipoint nonparametric linkage score of 2.21 (P = 0.014) was found at marker DXS8083 on Xp11, which increased to a nonparametric linkage score of 3.18 (P = 0.001) in data that had been conditioned for allele sharing at the CTLA-4 locus under an epistatic model. There was no evidence to support linkage of GD to Xq21.33-q22 (GD3) or at the 14q31-q33 (GD1) region in our population. A locus with a moderate contribution to GD susceptibility (lambda(s) = 1.4) is likely to exist in the Xp11 region, but we are unable to confirm that the GD1 or the GD3 regions contain major susceptibility loci in our United Kingdom GD population.

Production of bioactive and immunoreactive interleukin-6 (IL-6) and expression of IL-6 messenger ribonucleic acid by human pituitary adenomas.

Preliminary studies have demonstrated that some pituitary adenomas secrete immunoreactive interleukin-6 (irIL-6) when cultured in vitro. We have extended these studies by investigating 100 pituitary adenomas of different types measuring immunoreactive and bioactive IL-6. Tumors were cultured either as explants without fetal calf serum or as dispersed cells with 10% total calf serum. Fifty-three of the 100 (53%) pituitary cultures were found to release irIL-6 and in 44 adenomas examined, 32 (72.7%) secreted bioactive IL-6. In 61 explant cultures, 30 adenomas released IL-6, indicating autonomous secretion. The amount of IL-6 released by adenomas in cell culture was generally higher, although the incidence was similar to explant cultures. IrIL-6 was released by 7 of 14 prolactinomas, 15 of 27 somatotrophinomas, 5 of 7 corticotrophinomas (including 2 Nelson's adenomas), 1 of 1 thyrotrophinomas, 2 of 2 gonadotrophinomas, and 23 of 49 clinically non-functioning adenomas. Periadenomatous tissue removed from a patient with a corticotrophinoma was found to secrete IL-6 but in much lower concentration than from the adenoma tissue. Tumor necrosis factor-alpha and -gamma-interferon were not detected in the conditioned media. Four IL-6-secreting adenomas were examined by in situ hybridization for IL-6 messenger RNA, and three of these were positive with fluorescence present throughout the tissue examined. We have provided evidence that over half of pituitary adenomas secrete IL-6 which is bioactive and that IL-6 is synthesized within the tumor by the adenoma cells.

Association analysis of the cytotoxic T lymphocyte antigen-4 (CTLA-4) and autoimmune regulator-1 (AIRE-1) genes in sporadic autoimmune Addison's disease.

Although autoimmune Addison's disease (AAD) may occur as a component of the monogenic autoimmune polyendocrinopathy type 1 syndrome (APS1), it is most commonly found as an isolated disorder or associated with the autoimmune polyendocrinopathy type 2 syndrome (APS2). It is likely that sporadic (non-APS1) AAD is inherited as a complex trait; however, apart from the major histocompatibility complex, the susceptibility genes remain unknown. We have examined polymorphisms at two non-major histocompatibility complex candidate susceptibility loci in sporadic (non-APS1) AAD: the cytotoxic T lymphocyte antigen-4 (CTLA-4) gene and the autoimmune regulator (AIRE-1) gene. DNA samples from AAD subjects (n = 90) and local controls (n = 144 for CTLA-4; n = 576 for AIRE-1) were analyzed for the CTLA-4A/G polymorphism in exon 1 of the CTLA-4 gene and for the common mutant AIRE-1 allele (964de113) in United Kingdom subjects with APS1, by using the restriction enzymes Bst7II and BsrBI, respectively. There was an association of the G allele at CTLA-4A/G in AAD subjects (P = 0.008 vs. controls), which was stronger in subjects with AAD as a component of APS2 than in subjects with isolated AAD. In contrast, the mutant AIRE-1 964del13 allele was carried in one each of the 576 (0.2%) control subjects and the 90 (1.1%) AAD subjects as a heterozygote (P = 0.254, not significant), suggesting that this common AIRE-1 gene abnormality does not have a major role in sporadic (non-APS1) AAD.

Specific epidermal growth factor receptors on porcine thyroid cell membranes.

Specific saturable receptors for epidermal growth factor (EGF) of high affinity (Ka 1.7 X 10(9) M-1) have been demonstrated on porcine thyroid membranes. Optimal conditions for EGF binding have been determined. TSH and other peptide hormones do not inhibit the binding of 125I-EGF and EGF does not inhibit 125I-TSH binding to thyroid membranes. The results suggest that EGF may be involved in the regulation of thyroid follicular cell growth and function.

The effect of growth hormone replacement therapy for up to 12 months on lipoprotein composition and lipoprotein(a) in growth hormone-deficient adults.

The effect of growth hormone replacement therapy in near physiological doses on lipoprotein composition and serum lipoprotein(a) concentrations was investigated in growth hormone-deficient subjects. A randomised, double-blind, placebo-controlled trial of recombinant growth hormone was undertaken for 6 months followed by an open extension for a further 6 months (0.125 IU/kg per week for the first 4 weeks of each 6 month period and thereafter 0.25 IU/kg per week). A total of 18 patients with isolated growth hormone deficiency or hypopituitarism were studied. Lipid concentrations were estimated in lipoprotein fractions and protein concentrations were measured in low density lipoprotein (LDL). Glucose and glycated haemoglobin in blood and insulin, cholesterol, triglyceride, apolipoproteins A-I and B and lipoprotein(a) concentrations were measured in serum. In the placebo-controlled phase fasting blood glucose concentrations increased with growth hormone treatment from 5.0 +/- 0.2 to 5.8 +/- 0.2 mmol/l (P = 0.02) (mean +/- S.E.M.), although no significant changes were seen in lipids or lipoproteins. In the group receiving active treatment total serum cholesterol decreased from 6.0 +/- 0.4 to 5.2 +/- 0.3 mmol/l (P = 0.002) after 6 months, due to reduced LDL cholesterol concentrations. Low density lipoprotein protein concentrations fell (0.8 +/- 0.1 versus 0.7 +/- 0.1 g/l) (P = 0.005), and LDL phospholipid levels decreased from 0.9 +/- 0.1 to 0.7 +/- 0.1 mmol/l (P = 0.007). Serum cholesterol and LDL composition reverted to pre-treatment values by 12 months. Fasting blood glucose remained above pre-treatment values (P = 0.036) and fasting insulin was significantly increased (P = 0.044). There was no effect of growth hormone therapy on serum triglyceride, apolipoprotein or lipoprotein(a) concentrations. In conclusion, growth hormone therapy with near physiological doses has no long term effects on serum lipoprotein(a) concentrations or lipoprotein composition.

Bromocriptine therapy for "nonfunctioning" pituitary tumors.

Bromocriptine therapy may cause regression of prolactinomas, but its effect on nonsecretory pituitary tumors is uncertain. Conventional treatment for such "functionless" tumors is surgery and/or radiotherapy, but recurrences pose a therapeutic dilemma. We describe a patient with such a tumor treated by surgery and radiotherapy who presented with recurrent disease 14 days later. On treatment with bromocriptine, 20 mg daily for 25 months, the intrasellar tumor recurrence had diminished in size and a suprasellar extension had almost disappeared. Bromocriptine therapy may therefore benefit some patients with nonsecretory pituitary tumors considered unsuitable for surgery or radiotherapy.

Hyperprolactinemia. Long-term effects of bromocriptine.

Patients with hyperprolactinemia may be managed by pituitary surgery or irradiation, bromocriptine treatment, or a combination of these methods, and some patients remain untreated. Little is known of the long-term consequences of some of these therapeutic regimens. Forty-six hyperprolactinemic patients (40 female and six male) managed solely with bromocriptine or no treatment over a period of 12 months to six years were therefore evaluated in this study. Nine patients with radiologically normal pituitary fossae were untreated and 10 received bromocriptine, 7.5 to 10 mg daily, while 20 patients with radiologic evidence of a pituitary tumor were treated with bromocriptine, generally 10 to 20 mg daily. Patients were assessed clinically, biochemically, and radiologically before treatment and at least six weeks after discontinuation of therapy. A further seven patients were similarly assessed before and after eight bromocriptine-induced pregnancies. Symptoms persisted in the untreated group of nine patients, although menstruation returned in four of the females with previous amenorrhea; serum prolactin levels remained elevated, other pituitary function did not change, and pituitary fossae remained normal radiologically. In all patients treated with bromocriptine, symptoms improved irrespective of radiologic findings on the pituitary, and were abolished in 67 percent during treatment associated with a decrease in serum prolactin levels in all, and a return of levels to within normal limits in 80 percent of patients. Persistent side effects were usually dose-related, but remained troublesome in 13 percent. Bromocriptine-induced tumor regression was evident radiologically in all patients with suprasellar tumor tissue and in some with purely intrasellar adenomas. This effect occurred rapidly and persisted or increased throughout follow-up. On discontinuation of treatment, prolactin levels remained significantly lower than before therapy (mean 2,934 versus 5,052 mU/liter, p less than 0.05) but were within the normal range in only two patients. Other pituitary function was unaltered, or improved in some patients with definite tumors. Bromocriptine-induced pregnancy produced no permanent change in clinical, biochemical, or radiologic status. Long-term bromocriptine treatment for hyperprolactinemia is thus highly effective in alleviating symptoms and suppressing prolactin secretion, and induces persistent tumor regression on treatment without deterioration of other pituitary function in patients with macroadenomas. On discontinuation of therapy, however, hyperprolactinemia usually recurs, and treatment may therefore need to be continued for years.

Effect of thyrotrophin on epidermal growth factor receptors in monolayer cultures of porcine thyroid cells.

Primary cultures of porcine thyroid cells, grown as monolayers, showed saturable binding of epidermal growth factor (EGF). Scatchard analysis resolved the binding to a high-affinity/low-capacity site (dissociation constant = 0.17 nmol/l, maximal binding capacity = 1.67 pmol/10(6) cells) and a low-affinity/high-capacity site. Preincubation of thyroid monolayers with TSH for 3 days caused an increase in binding of 125I-labelled EGF due to an increase in the number of receptors, with the binding affinity unchanged. This effect was dose-dependent within the range of TSH concentrations 0.01-100 mu/l. The same effect was seen with dibutyryl cyclic AMP (10-1000 mumol/l). When the protein synthesis inhibitor cycloheximide was included in the TSH preincubation, the increase in EGF binding was abolished. The TSH effect on EGF binding was not mediated by thyroid hormones, since neither thyroxine (T4) nor tri-iodothyronine (T3) at 0.1 nmol/l-10 mumol/l could mimic the effect of TSH, nor could antisera to T3 or T4 neutralize the effect of TSH. The concentration of extracellular iodide (10 nmol/l-10 mmol/l) had no effect on the binding of 125I-labelled EGF. The results demonstrate that TSH increases the number of receptor sites for binding of EGF to thyroid monolayers in vitro. This effect is dependent upon protein synthesis and is mediated by cyclic AMP but not by thyroid hormones or iodide. This effect on binding of EGF may contribute to the trophic action of TSH.

The pathogenesis of thyroid-associated ophthalmopathy.

TAO is an autoimmune disease the pathogenesis of which is unclear. Autoimmunity against orbital antigens has been demonstrated, although the precise immunological mechanisms responsible are unidentified. Cytokines probably play an important mediating role. Activation of orbital fibroblasts appears to be essential in bringing about interstitial oedema and enlargement of EOM. The challenge for the future lies in clarifying the role of cytokines, identifying the autoantigen and elucidating the immunological mechanisms underlying the disease.

Cross-reaction of eye muscle antibodies with thyroid tissue in thyroid-associated ophthalmopathy.

The nature of the association of ophthalmopathy with autoimmune thyroid disease is not understood. Serum autoantibodies to eye muscle have previously been identified and in this study we have explored the hypothesis that there may be shared antigenic determinants between orbital and thyroid tissues. Sera were obtained from patients in whom eye muscle antibodies (EMAb) had been detected by an enzyme-linked immunosorbent assay (ELISA); the sera were preincubated with membrane preparations of thyroid or eye muscle, hepatic membranes being used as control. Tissue-binding antibodies were removed from serum by centrifugation and the supernatant serum was analysed using an indirect ELISA and by immunoblotting. In the ELISA, all sera gave a positive response for EMAb. In one serum, the binding was entirely non-specific. All sera showed significant neutralization of EMAb by eye muscle. In six sera there was reduction of EMAb after exposure to thyroidal antigens, indicative of cross-reaction. Western blotting confirmed the non-specific nature of the binding in one serum. In five of the remaining nine sera, protein bands were identified which interacted specifically with eye muscle and, in two of these, the same determinants were neutralized by preincubation with thyroid tissue. The Western blots confirmed the findings in the ELISA. The determinants recognized by IgG were variable between patients and no common antigen could be identified. This study demonstrates that, in some cases of thyroid-associated ophthalmopathy, there is cross-reaction of EMAb with thyroidal antigens, but this is variable and not found in every case. This may explain the association of the disease with autoimmune thyroid disease, at least in some cases.

Demonstration of significant epidermal growth factor activity in murine thyroid tissue.

Epidermal growth factor (EGF) was first isolated from the mouse submaxillary gland (SMG) which remains its only known site of synthesis. Earlier work suggests that EGF may be implicated in the regulation of thyroid cell growth and proliferation. Moreover, thyroid hormones increase the EGF content of mouse SMG and mimic the maturational effects associated with EGF. The aim of this study was therefore to determine whether the mouse thyroid gland is a site of EGF production. Thyroid and submaxillary glands were homogenized in Tris-HCl, the EGF content was measured by a homologous radioimmunoassay (RIA) and receptor-binding activity assessed in a radioreceptor assay. Epidermal growth factor was readily detectable in each thyroid extract. Dilution curves were parallel to the standard curve. Values obtained for right and left lobes of thyroid were in excellent agreement (r = 0.997, P less than 0.001). The intrathyroidal EGF concentration obtained by RIA was 26.1 +/- 6.0 (S.E.M.) ng/mg protein (n = 40); values obtained by the receptor assay were slightly lower but correlated closely (r = 0.828, P less than 0.01). Assay of homogenates prepared at the same time from the submaxillary glands of these mice showed that thyroidal EGF was not correlated with EGF content in the SMG. These findings indicate that EGF, both biologically active and immunoreactive, is readily detectable in the mouse thyroid and that the thyroid gland is a probable site of EGF synthesis. The methodology provides a model for further studies of the regulation of EGF production and its significance in relation to thyroid disease.

Thyrotrophin-blocking antibodies in goitrous primary hypothyroidism: detection by a highly sensitive bioassay and determination of site of action.

Using a highly sensitive bioassay technique, the presence of antibodies capable of blocking the stimulation of thyrocyte function by TSH has been investigated in the sera of a group of 50 patients with primary hypothyroidism. TSH-blocking activity was detected in immunoglobulin (IgG) preparations from the sera of 13 patients (26%). All 13 IgG preparations blocked both TSH-stimulated iodide uptake and cyclic AMP generation (Spearman's rank correlation = 0.6, p = 0.02). However, only one IgG preparation blocked dibutyryl cyclic AMP-stimulated iodide uptake. The presence of TSH-blocking antibody activity was associated with goitrous (ten out of thirteen patients) as well as atrophic (two out of thirteen patients) primary hypothyroidism. Furthermore, TSH-blocking activity was not associated with other thyroid autoantibodies, as thyrotrophin-binding-inhibiting immunoglobulins and anti-TSH antibodies were undetectable in all cases and there was no correlation between TSH-blocking activity and the presence or titre of anti-thyroglobulin or anti-microsomal antibodies. This study indicates that TSH-blocking antibodies are present in the serum of some patients with primary hypothyroidism and are directed towards a site, presumably adjacent to or contiguous with the TSH receptor, that is not the binding site for TSH. The coexistence of TSH-blocking activity and goitre in the majority of these patients implies that these antibodies, although capable of blocking TSH-stimulated thyroid hormone biosynthesis, do not necessarily inhibit the mitogenic action of TSH in vivo.

Long-term in-vitro maintenance of growth hormone secretion from human somatotrophinomas with cortisol, and its effects on growth hormone-releasing factor.

The effect of cortisol (206 nmol/l) on GH secretion from enzymatically dispersed human somatotrophinoma cells in long-term culture was studied using adenomas from 13 patients with acromegaly. Basal GH secretion from cultures of three out of five tumours measured during periods of 4 hours declined to less than 10 mu u. GH/culture within 21 days. Secretion from two other tumours measured during 24 h also declined but GH concentrations were still readily detectable (greater than 100 mu u./culture) by 28 or 58 days after cell dispersal. The decline in GH secretion was reversed in all seven tumours when cultures were maintained in cortisol-supplemented medium (CM), 4-h secretion rates being increased or retained at greater than or equal to 100 mu u./culture for as long as the studies were continued (range 18-291 days), and 24-h secretion rates at greater than 1000 mu u./culture (range 28-58 days). GH secretion and content from cultures of two additional somatotrophinomas were increased by treatment with cortisol for 5 or 9 days but there was no concomitant effect on cell number. During long-term maintenance of cultures in CM (range 5-40 days), cortisol partially or completely blocked the 4-h GH response to a test dose of GH-releasing factor (GRF) (20 nmol/l) in five out of six tumours, and to a dose range of GRF (0.01-20 nmol/l) in two of them.(ABSTRACT TRUNCATED AT 250 WORDS)

Effects of insulin-like growth factor-II on growth hormone release from human somatotrophinoma cells in vitro.

Insulin-like growth factor-I (IGF-I) and IGF-II receptors have previously been demonstrated on membranes prepared from human somatotrophinomas. IGF-I has been shown to have a variable effect on GH secretion by these tumours in vitro. The effects of purified IGF-II on GH secretion have not been described. We have studied the direct actions of human recombinant IGF-II on GH release from eight somatotrophinomas cultured in vitro. Somatotrophinoma cells were cultured as monolayers at a density of 10(5) cells/0.5 ml. Treatment with IGF-II for 4 and 24 h resulted in discrete inhibitory effects on GH release from two tumours (tumour 5: 4 h, IGF-II 0.5 nmol/l; tumour 2; 24 h, IGF-II 1 nmol/l). Treatment with IGF-II for 24 h resulted in significant inhibitory effects on GH release from one tumour over a range of concentrations tested (IGF-II 0.5-10 nmol/l). Addition of human GH-releasing factor (hGRF) (1-44) (20 nmol/l) for 4 and 24 h resulted in stimulation of GH release by five tumours. Two tumours demonstrated significant inhibitory effects of IGF-II on GRF-stimulated GH release (tumour 2: 24 h, IGF-II 1-5 nmol/l; tumour 3; 4 h, IGF-II 5 nmol/l; 24 h, IGF-II 0.5-50 nmol/l). These data emphasize the heterogeneity of somatotrophinomas in terms of their response to modulators of GH secretion. IGF-II does not appear to have a modulatory role on GH release by most somatotrophinomas.