Vitamin D receptor gene polymorphisms are associated with breast cancer risk in a UK Caucasian population.

There is increasing evidence that vitamin D can protect against breast cancer. The actions of vitamin D are mediated via the vitamin D receptor (VDR). We have investigated whether polymorphisms in the VDR gene are associated with altered breast cancer risk in a UK Caucasian population. We recruited 241 women following a negative screening mammogram and 181 women with known breast cancer. The VDR polymorphism Bsm I, an intronic 3' gene variant, was significantly associated with increased breast cancer risk: odds ratio bb vs BB genotype = 2.32 (95% CI, 1.23-4.39). The Bsm I polymorphism was in linkage disequilibrium with a candidate translational control site, the variable length poly (A) sequence in the 3' untranslated region. Thus, the 'L' poly (A) variant was also associated with a similar breast cancer risk. A 5' VDR gene variant, Fok I, was not associated with breast cancer risk. Further investigations into the mechanisms of interactions of the VDR with other environmental and/or genetic influences to alter breast cancer risk may lead to a new understanding of the role of vitamin D in the control of cellular and developmental pathways.

Insulin upstream factor 1 and a novel ubiquitous factor bind to the human islet amyloid polypeptide/amylin gene promoter.

The islet amyloid polypeptide (IAPP) gene is expressed primarily in the islet beta-cell and the peptide is co-secreted with insulin. To investigate mechanisms important in its regulation, we have used the electrophoretic mobility-shift assay and methylation interference to determine systematically sites of DNA-protein interactions in the human IAPP promoter. We identified beta-cell-specific DNA-protein complexes at three sites, each of which contained a consensus binding site for insulin upstream factor I (IUF-I). This complex was displaced with an antiserum to IUF-1, confirming that IUF-1 binds to the human IAPP promoter in vitro. We have also identified a DNA-protein complex within the region -220/-250 in both beta- and non-beta-cell lines. This region contains a motif with partial identity with the binding site for the ubiquitous transcription factor upstream stimulatory factor (USF), which binds to the human insulin promoter. However, purified USF was not able to bind to this putative site in the IAPP promoter and an oligonucleotide containing a functional USF-binding site was unable to displace binding from the IAPP oligonucleotide. Methylation interference revealed that the DNA-protein complex binds to a sequence that overlaps the USE-like sequence, and may therefore be a novel helix-loop-helix protein. These results suggest that, although both IAPP and insulin are beta-cell peptides, IAPP contains regulatory regions both common to and distinct from insulin.

Islet amyloid polypeptide-like immunoreactivity in human tissue and endocrine tumors.

We have detected islet amyloid polypeptide (IAPP)-like immunoreactivity (-LI) in human pancreas and in a range of endocrine tumors, including oat cell carcinoma of the lung and pancreatic tumours producing insulin, gastrin, glucagon, and vasoactive intestinal peptide. Gel permeation chromatography of the IAPP-LI revealed that, except in the carcinoid, more than 80% coeluted with synthetic human IAPP. The remaining immunoreactivity consisted of variable amounts of larger and smaller molecular forms. The concentration of IAPP-LI in the circulation of patients with diagnosed pancreatic endocrine tumors was not significantly elevated above normal fasting levels. IAPP is, therefore, produced by a range of endocrine tumors and may relate to the deposition of endocrine amyloid.

Molecular form of islet amyloid polypeptide (amylin) released from isolated rat islets of Langerhans.

Islet amyloid polypeptide (IAPP) is a 37-amino acid residue polypeptide, originally isolated from the pancreatic amyloid deposits of patients with type II diabetes mellitus. Subsequently, IAPP was found to be colocalised with insulin in beta-cell secretory granules of the normal mammalian pancreas. Recently, IAPP has been reported to inhibit glucose-stimulated insulin release from isolated rat islets and to be released in response to glucose and arginine. To investigate further the regulation of IAPP release from the islet, we used a previously developed specific radioimmunoassay for IAPP and measured IAPP secretion from isolated rat islets of Langerhans. Release of IAPP-like immunoreactivity (-LI) was stimulated by glucose: 3.3 +/- 0.3, 3.9 +/- 0.3, and 11.1 +/- 1.5 (n = 5, mean +/- SEM) fmol/islet/60 min at 2, 7, and 20 mM, respectively. Carbachol (0.1 mM) increased the release of IAPP-LI at the lower glucose concentrations: 8.1 +/- 0.9, 8.7 +/- 0.6, and 11.7 +/- 1.8 fmol/islet/60 m in at 2, 7, and 20 mM glucose. Somatostatin (1 microM) suppressed glucose-stimulated IAPP-LI release (17.5 +/- 1.5 vs. 5.1 +/- 0.5 fmol/islet/60 min). Chromatographic characterisation of the IAPP-LI released into the incubation medium revealed two immunoreactive forms: The major peak (74% of the total IAPP-LI) corresponded to synthetic IAPP-37, while a smaller form, comprising 26% IAPP-LI, eluted later. In acid extracts of islets, all (> 95%) immunoreactivity co-eluted with the synthetic IAPP.

The physiology of calcitonin gene-related peptide in the islet compared with that of islet amyloid polypeptide (amylin).

Following the discovery of a second gene containing a CGRP-like sequence, we demonstrated that "beta-CGRP" was indeed translated as a 37-amino acid peptide in vivo and was the predominant form of CGRP produced by the enteric nervous system. The presence of CGRP in the islet has been reported by several groups. We now show that beta-CGRP is again the major form. Another 37-amino acid peptide was recently isolated from islet amyloid deposits and found to have approximately 50% amino acid sequence homology with CGRP. Islet amyloid polypeptide, or amylin, is co-localized with insulin to the beta-cell secretory granule and is synthesized and released in parallel with insulin in response to a range of physiological and pharmacological stimuli. IAPP was subsequently shown, like CGRP, to inhibit the release of insulin pharmacologically. Interestingly, it was also shown to decrease the uptake of glucose by striated muscle, though it was considerably less potent than CGRP. This led to the suggestion that IAPP might be a circulating hormone regulating peripheral insulin sensitivity. Infusion of IAPP in human volunteers to produce plasma concentrations more than 100-fold higher than those seen physiologically, however, failed to alter peripheral glucose disposal. We conclude that beta-CGRP and IAPP are likely to play a role in local paracrine control of the islet.

Very high concentrations of islet amyloid polypeptide are necessary to alter the insulin response to intravenous glucose in man.

Islet amyloid polypeptide (IAPP) is a beta-cell peptide that can oppose insulin action in animal systems, but has not been shown to have any action in man; previously, we failed to show an effect of infused IAPP on iv glucose tolerance in human volunteers. We have reexamined its effects at even higher concentrations in six volunteers who received iv glucose (0.5 g/kg) during infusions of IAPP (25 and 50 pmol/kg.min) or normal saline. IAPP rose from a mean basal of 14.7 +/- 5.3 pmol/L to peak levels of 1,420 +/- 110, 2,240 +/- 140, and 27.7 +/- 9 pmol/L, respectively. IAPP at 25 pmol/kg.min had no effect on the plasma glucose disposal rate or the total incremental insulin response, but, in contrast, at 50 pmol/kg.min decreased the insulin response to glucose compared to the saline infusion (incremental area under the curve, 11,276 +/- 2,353 vs. 17,549 +/- 2,687 U; mean +/- SEM; P less than 0.02). This decrease was observed both during the first phase (0-10 min postglucose) insulin response (3,210 +/- 985 vs. 4,382 +/- 815 U; P less than 0.05) and the second phase response (11-90 min, 8,520 +/- 1,719 vs. 13,679 +/- 2,326 U; P less than 0.03). Glucose disposal rate, however, was unaffected (2.0 +/- 0.2 vs. 1.9 +/- 0.2). Thus, circulating IAPP concentrations greater than 90 times normal postprandial peaks were necessary to affect the insulin response to glucose. IAPP appears unlikely to be a circulating hormone influencing carbohydrate metabolism in man.

Islet amyloid polypeptide: production by an osteoblast cell line and possible role as a paracrine regulator of osteoclast function in man.

1. The recently discovered peptide islet amyloid polypeptide shows considerable sequence homology with calcitonin-gene-related peptide, itself an alternative product of the calcitonin gene. The possibility that islet amyloid polypeptide might affect calcium homoeostasis and bone cell function was investigated. 2. Islet amyloid polypeptide messenger RNA was found to be expressed by human HTb 96 osteoblast-like cells in culture, and islet amyloid polypeptide immunoreactivity was present in the cell culture medium. 3. Infusion of islet amyloid polypeptide (150 pmol min-1 kg-1) caused a fall in serum calcium and phosphate concentrations in five patients with Paget's disease of the bone. This was similar to that caused by infusion of calcitonin (50 pmol min-1 kg-1). 4. These findings raise the possibility that islet amyloid polypeptide may act as a local factor within bone, produced by osteoblasts and regulating osteoclast function. The possibility of an action of islet amyloid polypeptide on the renal handling of calcium seems unlikely but is not totally excluded.

Islet amyloid polypeptide: the cause of type-2 diabetes?

Islet amyloid polypeptide (IAPP or amylin), first identified as the peptide deposited as amyloid in type-2 diabetic pancreas and insulinoma, turns out to be a peptide produced in the pancreatic beta-cell secretory granule that is costored and coreleased with insulin. Experimental evidence suggests that, under certain conditions, IAPP can counter insulin action in peripheral tissue and inhibit insulin release from the pancreas. IAPP therefore appears to respond to the same physiologic stimuli as insulin, but has opposing biologic actions. The role of IAPP, both in normal physiology and in pathology, remains unclear, but current evidence suggests against a role as a circulating hormone in favor of a paracrine or autocrine modulator of insulin secretion.

Depletion of islet amyloid polypeptide in the spontaneously diabetic (BB) Wistar rat.

Islet amyloid polypeptide (IAPP) in the pancreas of the spontaneously diabetic (BB) Wistar rat was examined by radioimmunoassay, and IAPP mRNA levels were determined by Northern blotting. IAPP-like immunoreactivity in the diabetic rat pancreas was found to be significantly depleted compared with control (non-diabetic) BB rats (85.9 +/- 5 pmol/g in control rats, n = 8, vs 8.97 +/- 0.9 pmol/g in diabetic rats, n = 5; mean +/- S.E.M.). A similar change in insulin concentrations was found, although insulin was present in approximately 100-fold greater amounts than IAPP. Chromatography of the IAPP immunoreactivity revealed a single molecular form, corresponding to synthetic IAPP. Northern blot analysis of pancreatic RNA (n = 4) revealed that IAPP mRNA in the diabetic group was depleted to 22% of the signal intensity in the control group. Insulin mRNA was dramatically reduced to only 4% of the control group and, in contrast, somatostatin was relatively unaffected, with the diabetic group retaining 86% of signal compared with the controls. This animal model of insulin-dependent diabetes results from severe autoimmune destruction of the beta cell. The extremely low levels of both insulin and its messenger RNA are in agreement with this. These results demonstrate that this pathological state is also associated with a loss of IAPP from the pancreas. Insulin-dependent diabetes is associated with a range of metabolic disturbances. It is possible that the concomitant depletion of IAPP may be a contributory factor in exacerbating the condition.

Islet amyloid polypeptide-like immunoreactivity (amylin) in rats treated with dexamethasone and streptozotocin.

Islet amyloid polypeptide (IAPP) is a 37 amino acid peptide present in pancreatic beta-cells. Pancreatic and circulating IAPP concentrations in rat models of diabetes were measured using a specific radioimmunoassay. Pancreatic IAPP-like immunoreactivity (IAPP-IR) in dexamethasone-treated rats was twice that of the control rats (1571 +/- 137 vs 657 +/- 176 (S.E.M.; n = 6) pmol/g), and this was reflected by similar changes in the plasma IAPP-IR (272 +/- 17 vs 102 +/- 10 pmol/l). In streptozotocin-treated rats, pancreatic IAPP-IR (200 +/- 90 pmol/g) was reduced compared with controls. There was a significant positive correlation between pancreatic and plasma IAPP-IR and insulin, with r values of 0.82 and 0.91 for the plasma and pancreas respectively. Characterization of pancreatic immunoreactivity, using gel chromatography, revealed two peaks of IAPP-IR, one which coeluted with synthetic human amidated IAPP and another peak, presumably a fragment or breakdown product, which eluted later. Chromatography of the plasma IAPP-IR revealed that greater than 90% of the IAPP-IR eluted in the void volume, although the remaining IR coeluted with the synthetic IAPP standard. These results are not straightforwardly compatible with the suggested role for IAPP as a hormonal, paracrine or autocrine inhibitory regulator of insulin secretion in the maintenance of carbohydrate homeostasis.

The distribution of galanin message-associated peptide-like immunoreactivity in the pig.

Using a radioimmunoassay (RIA) developed to the N-terminal part of the predicted sequence of porcine galanin message-associated peptide (GMAP), we have confirmed the existence of GMAP-like immunoreactivity (-LI) in normal porcine tissues. GMAP-LI was found to parallel the distribution of galanin-immunoreactivity (-IR), although consistently the concentrations detected were, on a molar ratio, significantly less than those measured for galanin throughout the gastrointestinal tract, brain, spinal cord, adrenal and pituitary gland. As cleavage of the prohormone would be expected to produce galanin and GMAP on an equimolar basis, it is possible that the endogenous, intact GMAP peptide does not fully cross-react with the antibody raised to the N-terminal GMAP sequence. Gel chromatography of tissue extracts revealed a single molecular form of galanin-IR in the gut and four distinct molecular forms in the adrenal gland. GMAP-LI eluted as a single immunoreactive component in the gut, and in the adrenal gland there were two major molecular forms, one of which was apparently also detected by the galanin assay, and a small amount of N-terminal fragment. This molecular heterogeneity seems likely to be a result of the various possible prohormone cleavage products and/or posttranslational processing modifications. Further analysis of the galanin gene products needs to be undertaken in order to confirm this.

Amylin and amylin-amide lack an acute effect on blood glucose and insulin.

Amylin-amide has been implicated in the pathogenesis of type II diabetes due to its proposed inhibitory effect on insulin release from beta cells of the pancreatic islets, and on glucose uptake by the skeletal muscle. In experiments with rats and rabbits we failed to demonstrate these anti-insulin actions of amylin and amylin-amide. A single bolus dose of the two peptides (500 pmol) administered i.v. failed to suppress plasma insulin levels or to elevate blood glucose levels. The continuous infusion of amylin-amide into rabbits also failed to suppress the release of insulin in response to hyperglycaemia produced by an i.v. bolus injection of glucose. These in vivo observations imply that the amylin peptides may not have a primary physiological role in carbohydrate metabolism, but in view of our previous findings, we speculate that the peptide has a more prominent role in calcium homeostasis.

Failure to establish islet amyloid polypeptide (amylin) as a circulating beta cell inhibiting hormone in man.

The presence of islet amyloid polypeptide in amyloid within pancreatic islet cells in Type 2 (non-insulin-dependent) diabetes, and its reported inhibition of glucose uptake by skeletal muscle in vitro, has prompted speculation concerning its role in the pathogenesis of diabetes. We investigated the effect of infused synthetic amidated human islet amyloid polypeptide (mol. wt. 3904, confirmed by mass spectroscopy) on intravenous glucose tolerance. Seven healthy, non-obese volunteers (age +/- SD, 27 +/- 4 years) were infused over 50 min with normal (0.9%) saline or islet amyloid polypeptide at 50 pmol.kg-1.min-1. After 20 min, a bolus of 0.5 g/kg glucose was given within 1 min and blood sampling continued for up to 60 min. Circulating concentrations of islet amyloid polypeptide reached at steady state were 1130 +/- 90 pmol/l. The calculated half-life was 11.8 +/- 0.9 min, metabolic clearance rate 5.7 +/- 0.6 ml.kg-1.min-1 and apparent distribution space therefore 94 +/- 12 ml/kg. However, islet amyloid polypeptide was found to have no effect on the peak value reached, or the total area under the curve for plasma glucose, insulin or glucagon following intravenous glucose. This study suggests circulating islet amyloid polypeptide may not be an important influence on intravenous glucose tolerance in man.

Altered islet amyloid polypeptide (amylin) gene expression in rat models of diabetes.

The response of the islet amyloid polypeptide gene to chronic dexamethasone treatment in adult rats was investigated. After 12 daily injections, rats were severely underweight and fasting blood glucose levels were elevated. When pancreatic mRNA was analysed, a 16-fold elevation in islet amyloid polypeptide mRNA was observed with only a four-fold increase in insulin mRNA levels. Pancreatic islet amyloid polypeptide and insulin mRNA levels were also determined 12 days after streptozotocin treatment. In these rats, which were not severely diabetic, the reduction in islet amyloid polypeptide mRNA levels was sixfold less than the reduction in insulin mRNA levels. In both these models of diabetes the ratio of islet amyloid polypeptide to insulin mRNA levels was raised. This would not be expected if the physiological role of islet amyloid polypeptide is as a simple hyperglycaemic agent opposing insulin action or release.

Novel peptide pancreastatin: its occurrence and codistribution with chromogranin A in the central nervous system of the pig.

The distribution of pancreastatin immunoreactivity was investigated in porcine brain, spinal cord, dorsal root ganglia, and pituitary. In the brain, immunoreactive cell bodies were present in many areas including the cortex, basal ganglia, hippocampus, thalamus, hypothalamus, mesencephalic reticular formation, cerebellum, and medulla oblongata. Immunoreactive fibres were most abundant in the globus pallidus, stria terminalis, entopeduncular nucleus, hippocampus, and in the substantia nigra. In the spinal cord, immunoreactive cells were found in laminae IV-IX. Immunoreactive fibres were concentrated in the dorsal horn. Pancreastatin immunoreactivity was localised to fibres and small cells (5-10% of the total) in the dorsal root ganglia. In the posterior pituitary, many immunoreactive fibres were present and in the anterior lobe subsets of gonadotrophs and thyrotrophs were pancreastatin-immunoreactive. The localisation of pancreastatin showed a parallel distribution with chromogranin A. Coexistence of pancreastatin with calcitonin gene-related peptide (CGRP) immunoreactivity in cell bodies in the spinal cord, including motoneurones, and with CGRP or galanin immunoreactivities in dorsal root ganglion cells was also noted. The differential pattern of pancreastatin immunostaining was reflected in the extractable levels of peptide with highest concentrations in the cortex (55.8 +/- 6.0 pmol/g wet weight, mean +/- S.E.M.), thalamus (60.0 +/- 5.0 pmol/g), hypothalamus (54.4 +/- 6.5 pmol/g), and anterior pituitary (2,714 +/- 380 pmol/g). Characterisation of pancreastatin immunoreactivity in the hypothalamus and pituitary by gel permeation and high-pressure liquid chromatography revealed multiple molecular forms, one of which was indistinguishable from natural porcine pancreastatin. The widespread distribution of pancreastatin immunoreactivity suggests this peptide may play a part in several neuroendocrine, autonomic, somatic, and sensory functions, and its colocalisation with chromogranin A is consistent with a precursor-product relationship.

Formation of endothelin by cultured airway epithelial cells.

Immunoreactivity to endothelin was detected in conditioned culture medium from both canine and porcine tracheal epithelial cells. Gel permeation chromatography and fast protein liquid chromatography were used to confirm the identity of the endothelin. The two peaks demonstrated on fast protein liquid chromatography co-eluted with endothelin 1 and endothelin 3 respectively.

Expression of tachykinins by ileal and lung carcinoid tumors assessed by combined in situ hybridization, immunocytochemistry, and radioimmunoassay.

Mid-gut carcinoid tumors have been shown to produce substance P, a tachykinin. A recent addition to this family of peptides is neurokinin A which is cleaved from the same precursor as substance P; beta-pre-pro-tachykinin. The authors have examined mid-gut and pulmonary carcinoid tumors for the presence of the two tachykinins, using immunocytochemical study and radioimmunoassay, and have applied the techniques of in situ hybridization and Northern blot analysis to investigate the expression of mRNA for beta-pre-pro-tachykinin. All gut tumors (n = 8) and three of the six lung tumors examined were found by immunocytochemical study to contain both tachykinins or neurokinin A alone. Chromatographic analysis of tumor extracts suggests that this peptide is being detected as a separate molecule and/or as the C-terminal portion of a larger, uncleaved form. Three of the cases positive for tachykinins showed no detectable serotonin immunoreactivity. Strong hybridization signals for beta-pre-pro-tachykinin mRNA were seen in all but one of the cases studied which contained tachykinin immunoreactivity. Intact mRNA and positive hybridization was found by Northern blot analysis in two mid-gut tumors. Concentrations of tachykinins were found by radioimmunoassay to be higher in mid-gut tumors (substance P 27.2 +/- 19.7 pmol/g; neurokinin A 31.8 +/- 24.2 pmol/g; mean +/- SEM, n = 5) than in lung cases (substance P mean 0.8, range 0.5-1.0 pmol/g; neurokinin A mean 11.0, range 10.0-12.0 pmol/g; n = 3). These results show that mid-gut and pulmonary carcinoid tumors produce tachykinins, which are detected, in some cases, where no serotonin immunoreactivity can be found, possibly because of a high rate of amine secretion. Screening for tachykinins may prove to be a useful diagnostic adjunct for these tumors.

Increase in diazepam binding inhibitor-like immunoreactivity (51-70) in Huntington's disease.

In Huntington's disease (HD), reduction in striatal GABA is one of the most striking abnormalities and alterations in benzodiazepine receptors, which are allosterically linked to the GABAA receptor, have also been reported. Diazepam binding inhibitor (DBI), recently isolated from rat and human brain, has been proposed as an endogenous ligand at the benzodiazepine receptor. The content of DBI-like immunoreactivity(51-70) (DBI-IR(51-70), has therefore been compared in control postmortem human brains and in HD brains (matched for age, sex and post-mortem delay), using a specific radioimmunoassay. DBI-IR(51-70) was more than 1.5-fold increased in the putamen, caudate, globus pallidus and nucleus accumbens of HD brains compared to the control group (P less than 0.001). Gel filtration chromatography showed similar elution profiles of the peptide in both control and HD extracts, thus providing no evidence for a change in the nature of the peptide itself.

Pancreastatin distribution and plasma levels in the pig.

Pancreastatin is a peptide isolated from porcine pancreas which has insulin-suppressive actions in vitro and sequence homology with chromogranin A. Using radioimmunoassay and immunocytochemistry we investigated whether pancreastatin has a more widespread distribution and a possible endocrine role in the pig. Pancreastatin immunoreactivity was found in plasma, adrenal gland, pancreas, anterior pituitary and throughout the gastrointestinal tract. The immunoreactivity was colocalized with chromogranin immunoreactivity in endocrine cells and ultrastructurally (in the pancreas) to storage granules. Characterization of pancreastatin-like immunoreactivity, using gel permeation and high performance liquid chromatography, separated 3 different pancreastatin-like immunoreactive forms: one molecular form, indistinguishable from synthetic pancreastatin 1-49, was predominant in pancreas and thyroid and released into the circulation postprandially. However, a high dose (greater than 1 nmol/l) infusion of pancreastatin 33-49 (the biologically active moiety in vitro) into conscious pigs had no effect on either basal or glucose-stimulated insulin secretion.

The occurrence of pancreastatin in tumours of the diffuse neuroendocrine system.

We have reported previously the localization of the 49 amino acid peptide pancreastatin to all identifiable endocrine cells of porcine gut, pancreas and adrenal, thyroid and pituitary glands. In this study, we have investigated the occurrence of pancreastatin in a series of human neuroendocrine tumours using an antibody to whole synthetic porcine pancreastatin. The most consistent immunostaining for pancreastatin was found in carcinoid tumours of ileum (four out of six), rectum (four out of six), ovary (two out of two) and lung (nine out of 10). Radioimmunoassay of tumour extracts showed that the concentrations of pancreastatin in ileal carcinoids were very high (mean 71.6, range 31.0-184.0 pmol g-1). The high rate of positivity in lung carcinoids contrasted sharply with the results of 10 pulmonary small cell carcinomas which displayed no immunoreactivity and contained minimal concentrations of pancreastatin (mean 2.0, range 0-6.0 pmol g-1). Extra-adrenal paragangliomas also contained pancreastatin (seven out of 10), but although radioimmunoassay detected peptide in phaeochromocytomas (mean 29.8, range 8.0-69.0 pmol g-1), immunocytochemistry did not. Porcine pancreastatin shows structural homology with bovine chromogranin A, an observation which has led to suggestions that chromogranin is a precursor for the peptide. More recently, a sequence homologous to porcine pancreastatin has been identified in the human chromogranin A molecule. In this study, immunostaining with an antiserum to human chromogranin gave positive results in most cases of each tumour type except the small cell carcinomas. The lack of consistent relationships between chromogranin and pancreastatin immunoreactivities may reflect the fact that the antiserum to pancreastatin was raised against the porcine peptide. When antibodies to human pancreastatin become available, the peptide may prove to be a more consistent marker for neuroendocrine tumours.