Erica Penman

Development and Validation of a Specific Radioimmunoassay for Somatostatin in Human Plasma

Little is known about the factors controlling somatostatin secretion in man, and data are not available on the changes in circulating levels in various human physiological or pathophysiological states. This is mainly a consequence of the technical difficulties involved in measuring somatostatin in plasma. In the presence of plasma, binding of somatostatin tracer to antibody was consistently decreased by about 20%, and this could not be abolished by the addition of EDTA and aprotinin or by the use of specially prepared somatostatin-free plasma. Furthermore, in the presence of plasma, endogenous somatostatin does not dilute in parallel with synthetic cyclic somatostatin standard. We have, therefore, developed and validated a radioimmunoassay for somatostatin using prior extraction of the peptide onto leached silica glass. Tyrosine-11 somatostatin was iodinated using lactoperoxidase and purified on ODS silica. This method is superior to iodination using chloramine-T with CMC cellulose purification, and gives a highly purified preparation with a shelf-life of at least eight weeks. Using this tracer and a specific antiserum, the limit of sensitivity of the assay was 10 pg/ml, with an intra-assay coefficient of variation of 12% (n = 16) and inter-assay coefficient of variation of 15% (n = 10). Parallelism has been demonstrated between standard synthetic cyclic somatostatin and all extracted plasma samples. The mean recovery of exogenous somatostatin from plasma was 78%. The fasting level of immunoreactive somatostatin at 0900 hours in 40 normal subjects ranged from 17 to 81 pg/ml. Care is needed, however, when comparing these values with those obtained from other laboratories since standard preparations of somatostatin vary considerably in their immunopotency.

Distribution and characterisation of immunoreactive somatostatin in human gastrointestinal tract

Immunoreactive somatostatin (IRS) was measured in acid extracts of human gastrointestinal tissue. The highest levels were found in the duodenum, pancreas, jejunum and stomach with lower levels in the ileum and colon. In the antrum, pylorus, duodenum and pancreas the main peak of IRS (1.6K IRS) coeluted with synthetic somatostatin-14 on both gel filtration chromatography and HPLC. In the body of stomach, jejunum, ileum and colon, a large peak coeluting with synthetic somatostatin-28 (3.5K IRS) on both chromatographic systems was also identified, while minor peaks of IRS assigned molecular weights of 6000 (6K) and greater than 15 000 (15K) were seen in some extracts. The total IRS content and pattern of molecular forms were similar in tissues obtained from adults at surgery or rapid post mortem, and in tissue taken from human fetuses after prostaglandin termination of pregnancy. When tissues were divided into mucosal and muscle layers, greater than 90% of the IRS was in the mucosa with less than 10% in the muscle layer. In the muscle layer the IRS was almost entirely the 1.6K form in all tissues. Immunohistochemical studies showed the IRS in the mucosa to be localised in endocrine-type cells, while in the muscle layer the IRS is present in nerve fibres and neurones of the myenteric plexus. It is suggested that (1) different mechanisms may control the biosynthesis of somatostatin-14 and somatostatin-28 in mucosal cells in different parts of the gut, (2) different biosynthetic controls may operate in endocrine-like and neuronal cells in the same region of the gut.

Circulating somatostatin after food and glucose in man.

Using a recently validated radioimmunoassay, changes in circulating somatostatin have been measured in normal subjects after food (a standard breakfast), and oral and intravenous glucose. After the standard breakfast, a clear and sustained rise in plasma somatostatin was seen in all subjects from a mean value (± 1 SE) of 28 ± 7 pg/ml to a mean peak value, at 60 min of 57 ± 11 pg/ml. When glucose was taken by mouth a significant but smaller rise was seen, but intravenous glucose caused no significant change in plasma somatostatin. A rise in circulating somatostatin after feeding has not previously been demonstrated in normal man and it is suggested that somatostatin may have an important endocrine role in the gut.

Bombesin-like immunoreactivity in human gastrointestinal tract

Abstract In the present study the distribution and molecular characteristics of bombesin-like immunoreactivity (BLI) were studied in acid extracts of human gastrointestinal tract. The highest levels were found in the fundus, antrum, pylorus and pancreas with lower levels in the duodenum, jejunum, terminal ileum and colon. BLI was also detected in both the muscle and mucosal layers of the antrum and colon. Sephadex G-50 gel chromatography under acid dissociating conditions revealed two peaks of immunoreactivity, one in the position of synthetic porcine gastrin releasing peptide (GRP) and the second eluting with synthetic amphibian bombesin. Variations in the proportions of the two molecular forms were seen in different regions of the gut. In the stomach and pancreas > 70% of the BLI eluted with the GRP marker while in pylorus, jejunum and terminal ileum only 20% was present in this form. Reverse-phase ODS silica HPLC of the major antral BLI peak, utilising a methanol/trifluoroacetic acid gradient indicated that this peptide was similar to porcine GRP. We have therefore (1) demonstrated the presence and heterogeneity of bombesin-like immunoreactivity throughout the human gastrointestinal tract and (2) shown for the first time that a proportion of this BLI closely resembles porcine GRP.

Evidence for direct production of somatostatin-14 from a larger precursor than somatostatin-28 in a phaeochromocytoma ☆

Gel-filtration chromatography of an acid-extract of a phaeochromocytoma, under dissociating conditions, revealed 4 peaks of immunoreactive somatostatin (IRS) of approx. 8-10 kilodaltons (K), 6K, 3.5K and 1.6K as detected by an antiserum (R9) directed against the central region of tetradecapeptide somatostatin (S14). The 3.5K and 1.6K forms of IRS co-eluted with synthetic cyclic S28 and S14 respectively on reversed phase HPLC. Using another radioimmunoassay for the 1-14 sequence of S28 (N-peptide) a peak of immunoreactive N-peptide (IRN) with a molecular weight of approx. 4500 was observed. The antiserum (N3) used in the N-peptide assay was raised against N-Tyr N-peptide and cross-reacts less than 5% with synthetic S28. Two peaks were further characterised by partial tryptic digestion and gel-filtration chromatography. The 3.5K IRS peak was partially converted to a 1.6K IRS form together with an approximately equimolar amount of IRN with apparent molecular weight of 2500. This 2.5K IRN co-eluted both with N-Tyr N-peptide and with the IRN generated by tryptic digestion of synthetic cyclic S28. No IRN peak of this size was observed in the original extract. Tryptic digestion of the 6K IRS peak generated 3.5K and 1.6K IRS and 2.5K IRN. These results suggested that (1) this human phaeochromocytoma contains IRS very similar to the known structure of ovine and porcine S28 and S14. (2) The 6K IRS is composed of an unknown peptide sequence attached via trypsin-susceptible bond to the N-terminus of S28. (3) In this tumour S14 is being generated directly from 6K IRS and not via S28.

MOLECULAR FORMS OF SOMATOSTATIN IN NORMAL SUBJECTS AND IN PATIENTS WITH PANCREATIC SOMATOSTATINOMA

Two patients with somatostatin‐secreting pancreatic tumours are described, one presenting with hypoglycaemia due to hyperinsulinism, and the other with Cushings syndrome due to ectopic ACTH production. When plasma from these patients was subjected to gel chromatography under conditions designed to prevent somatostatin binding to larger proteins, a peak of monomeric immunoreactive somatostatin was observed as well as several large molecular weight forms. These larger forms of somatostatin could be dissociated into monomeric somatostatin by dithiothreitol. Similar studies on plasma obtained from normal subjects also showed heterogeneity of circulating somatostatin. Extracts of tumour tissue from both patients contained predominantly monomeric somatostatin, but only small amounts of high molecular weight somatostatin which differed from the profile seen in plasma. The site(s) of origin of the large molecular weight forms of somatostatin seen in plasma and their relative biological activities remain to be established.

Characterisation of bombesin-like immunoreactivity in human fetal lung.

A sensitive radioimmunoassay for bombesin-like immunoreactivity (BLI) was developed and utilised in conjunction with G50 gel chromatography and reverse-phase HPLC, to study the content and molecular characteristics of bombesin-like peptides in acid extracts of human fetal lung. The antiserum, (B5), is directed towards the C-terminal region of the bombesin molecule and cross-reacts 70% with synthetic porcine GRP and the synthetic GRP fragment, GRP (14-27). Specimens of lung were collected from fetuses of gestational ages 15-22 weeks, following prostaglandin termination of pregnancy. The tissue was extracted into 0.1 N HCl at 90 degrees C. The mean BLI content was 50.2 pg/mg wet weight of tissue (range 15.5-136 pg/mg; n = 13). No correlation between gestational age and BLI content could be established. G50 gel chromatography of acid extracts, under dissociating conditions, revealed two peaks of BLI, one in the position of synthetic porcine GRP and the second, constituting greater than 90% of the immunoreactivity, eluting with synthetic amphibian bombesin. Reverse-phase ODS silica HPLC of this major G50 peak, utilising a methanol/trifluoroacetic acid gradient, indicated that this peptide was similar to the GRP C-terminal fragment, GRP (14-27). We have therefore (1) confirmed the presence and heterogeneity of BLI in human fetal lung, and (2) shown, for the first time, that the majority of this BLI more closely resembles a fragment of GRP than amphibian bombesin itself.

IMMUNOREACTIVE SOMATOSTATIN CHANGES DURING INSULIN‐INDUCED HYPOGLYCAEMIA AND OPERATIVE STRESS IN MAN

Little is currently known about the factors controlling somatostatin secretion. A radioimmunoassay has been developed that is sufficiently specific and sensitive to be used for physiological studies of circulating levels in man. During insulin‐induced hypoglycaemia a rise in plasma somatostatin was seen in each of ten subjects studied. Although this paralleled the rise in circulating glucagon and growth hormone, no individual relationships were found either between these variables or to any change in cortisol or insulin C‐peptide. In contrast no rise in somatostatin was seen during surgical stress. Thus, contrary to expectation, circulating somatostatin levels can be altered by metabolic stimuli. It seems likely that this peptide may serve an endocrine as well as a paracrine role since its modulating effects may occur not only near to but also at a distance from the site of secretion. It is not yet clear whether the somatostatin measured comes from the hypothalamus, any other part of the central nervous system or the gastrointestinal tract.

PANCREATIC SOMATOSTATINOMA PRESENTING WITH HYPOGLYCAEMIA

A 33‐year‐old woman with a 2 month history of vague ill health was admitted to hospital in hypoglycaemic coma. Preoperative investigation suggested malignant insulinoma as the probable cause of illness, but immunohistological examination of the tumour showed it to consist mainly of somatostatin‐containing cells but sparse insulin‐secreting cells were also present. Plasma immuno‐reactive somatostatin levels were from fifty to 200 times the upper limit of normal and rose in response to arginine and fell during diazoxide infusion. The hypoglycaemia was unusually sensitive to the hyperglycaemic effects of diazoxide and chlorothiazide and, with 5‐fluorouracil as the only specific anti‐tumour agent, there has been clinical, biochemical and radiological evidence of tumour regression.

SOMATOSTATIN SECRETION BY LUNG AND THYMIC TUMOURS

Ectopic secretion of somatostatin in one patient with a thymic tumour and in two patients with lung tumours is described. All three patients also had ectopic ACTH secretion. Gel filtration chromatography under dissociating conditions showed the two lung tumour extracts to contain predominantly 1600 MW somatostatin monomer while the thymic tumour contained predominantly a 3000–3500 MW form of somatostatin, 77% of which was not converted to 1600 MW somatostatin by dithiothreitol (an agent which reduces disulphide bridges). This form may therefore represent a covalently‐bound precursor of somatostatin rather than a dimer of two somatostatin monomers. Plasma from all three tumour patients and from normal subjects contained both 1600 and 3000–3500 MW somatostatin. It is suggested that somatostatin secretion may frequently be associated with multiple hormone producing tumours.