Immunochemical characterisation of somatostatin in ruminants

Following development and validation of a radioimmunoassay for somatostatin, the immunoreactivity of this peptide in the plasma of ruminants was measured and the levels in sheep were 9-31 pM (mean 18 +/- 7 pM, n = 48), in lambs 10-54 pM (mean 25 +/- 10 pM, n = 18) and in calves 5-35 pM (mean 12 +/- 6 pM, n = 22). Somatostatin-like immunoreactivity was present in sheep in high concentrations in the antrum (2342 +/- 280 pmol/g wet weight), duodenum (446 +/- 73 pmol/g) and pancreas (832 +/- 208 pmol/g). Lower concentrations (6-150 pmol/g) were found in other regions of the gastrointestinal tract. Molecular sieve chromatography on Bio-Gel P-10 showed that while most of the somatostatin in the antrum was somatostatin-14, in the duodenum about 30% of the total immunoreactivity was somatostatin-28.     

The intestinal mucosa preferentially releases somatostatin-28 in pigs

We studied the molecular forms of somatostatin-like immunoreactivity (SLI), newly released from isolated perfused preparations of the porcine antrum, stomach, pancreas and upper small intestine: Perfusion effluents were concentrated by Sep-Pak C-18 adsorption, eluted with ethanol, dessicated, and subjected to gel filtration with subsequent radioimmunoassays for somatostatin-14 and N-terminal somatostatin-28 immunoreactivity. All the SLI newly released from the stomach and antrum eluted at the position of somatostatin-14, and such was also the case for more than 95% of the SLI newly released from the pancreas, while 68 -/+ 7% and 75 -/+ 8% of the SLI newly released from the isolated perfused jejunum and ileum, respectively, corresponded to somatostatin-28. By reverse phase HPLC the identity of these peptides with synhetic somatostatin-14 and -28 was established.     

Immunochemical study of galanin in the cat digestive tract and autonomic ganglia

The presence of galanin was examined in the cat gut and related autonomic nervous structures using radioimmunoassay (RIA) and high performance liquid chromatography (HPLC). In the gut wall, the concentration of galanin-like immunoreactivity (GAL-LI) was assayed separately in the muscular layers with the nervous plexuses and in the mucosa and ranged from 0.35 to 0.55 pmol/g wet tissue. In the autonomic nervous structures, GAL-LI concentrations ranged from 0.22 (thoracic spinal ganglia) to 0.81 (inferior mesenteric ganglion) pmol/g wet tissue. The presence of galanin was checked by HPLC in the antrum, intestine, and colon. HPLC of extractable material revealed a major peak coeluting with the synthetic porcine peptide and minor earlier peaks representing likely different molecular forms of galanin. Our study strengthens the notion that galanin acts in nervous control of the cat gut functions.     

PACAP 1-38 as neurotransmitter in the porcine antrum

The concentration of PACAP 1-38 in porcine antrum amounted to 15.4+/-7.9 and 20.3+/-8 pmol/g tissue in the mucosal and muscular layers. PACAP immunoreactive (IR) fibres innervated the muscular (co-localised with VIP) and submucosal/mucosal layers (some co-storing VIP and CGRP) including myenteric and submucosal plexus and blood vessels. Only myenteric nerve cell bodies contained PACAP-IR (co-storing VIP). In isolated perfused antrum, vagus nerve stimulation (8 Hz) and capsaicin (10(-5) M) increased PACAP 1-38 release. PACAP 1-38 (10(-9) M) increased substance P (SP), gastrin releasing peptide (GRP) and VIP release. PACAP 1-38 (10(-8) M) inhibited gastrin secretion and stimulated somatostatin secretion and motility dose-dependently. PACAP-induced motility was strongly inhibited by the antagonist PACAP 6-38 but also by atropine and substance P-antagonists (CP99994/SR48968) but PACAP 6-38 had no effect on vagus-induced secretion or motility. Conclusion: PACAP 1-38 may be involved in antral motility and secretion by interacting with cholinergic, SP-ergic, GRP-ergic and/or VIP-ergic neurones, and may also be involved in afferent reflex pathways.     

Identification by specific radioimmunoassay of two novel peptides derived from the C-terminus of porcine preprogastrin

A radioimmunoassay has been developed using antibodies to a synthetic analogue of the C-terminal hexapeptide sequence of the porcine gastrin precursor. Boiling water extracts of porcine antral mucosa contained immunoreactive material that diluted in parallel with standard peptide. Concentrations of immunoreactivity were 5.5 +/- 0.8 nmol X g-1 (mean +/- S.E.M.) in antral mucosa and were closely similar to those of C-terminal heptadecapeptide gastrin immunoreactivity (5.0 +/- 0.6 nmol X g-1). Approximately 30-fold lower concentrations were found in porcine duodenum. A similar distribution was found in ferret, but human, rat and chicken antrum did not contain significant quantities of immunoreactivity. Gel filtration of porcine antral extracts on Sephadex G-50 revealed a single peak of immunoreactivity eluting in a similar position to G17, but on anion-exchange chromatography two peaks of immunoreactive material were separated. These also differed in their retention time on reverse phase HPLC. Both peptides are probably derived by tryptic cleavage at the C-terminus of porcine preprogastrin. No evidence was found to suggest that there are significant quantities of unprocessed preprogastrin in hog antral mucosa. The precise chemical difference between the two immunoreactive peptides identified here remains to be established; together, however, they provide specific markers for progastrin synthesis.     

The distribution of GAWK-like immunoreactivity in neuroendocrine cells of the human gut, pancreas, adrenal and pituitary glands and its co-localisation with chromogranin B

GAWK is a recently discovered peptide isolated from extracts of human pituitary gland and subsequently shown to be identical to sequence 420-493 of human chromogranin B. The distribution of this peptide was studied in human gut, pancreas, adrenal and pituitary glands using antisera to two portions of the 74 amino acid peptide (sequences 1-17 and 20-38). In addition, the co-existence of GAWK immunoreactivity with other peptides and chromogranin B was investigated using comparative immunocytochemistry. In the gut, GAWK was localised mainly to serotonin-containing cells of the mucosal epithelium, where electron microscopy showed it to be stored in typical electron-dense (250 nm diameter) granules, and to a moderate population of nerve fibres in the gut wall. Considerable quantities of GAWK-like immunoreactivity were measured in the gut, up to 36.3 +/- 18 pmol GAWK 1-17/g wet weight of tissue (mean +/- SEM) and 12.4 +/- 2.9 pmol GAWK 20-38/g. Chromatography of gut extracts revealed several GAWK-like immunoreactive peaks. GAWK-like immunoreactivity was also detected in endocrine cells of pancreas, pituitary gland and adrenal medulla, where the highest concentrations of GAWK-like immunoreactivity were measured (GAWK 1-17 2071.8 +/- 873.2 and GAWK 20-38 1292.7 +/- 542.7 pmol/g). Endocrine cells containing GAWK-like immunoreactivity were found also to be immunoreactive for chromogranin B. Our results define a discrete distribution of GAWK immunoreactivity in human endocrine cells and nerves and provide morphological support for the postulated precursor-product relationship between chromogranin B and GAWK. Details of the functions of this peptide are awaited.     

Neuropeptide Y- and somatostatin-like immunoreactivities in ganglioneuroma, ganglioneuroblastoma and neuroblastoma

Neuropeptide Y (NPY)- and somatostatin (SS)-like immunoreactivities (LI) were investigated in tumor tissues of one ganglioneuroma (GN), 3 ganglioneuroblastomas (GNB) and one neuroblastoma (NB) by radioimmunoassay. NPY-LI was detected from all 5 tumor tissues (16.4-1247 pmol/g wet tissue). Sephadex G-50 column chromatography and reverse phase high performance liquid chromatography (HPLC) revealed that most of the NPY-LI in tumor extracts was eluted in an identical position to synthetic human NPY except one GNB (case 2). In this case, most of the NPY-LI was eluted in a higher molecular weight region than synthetic human NPY in Sephadex G-50 column chromatography and in a more hydrophobic position in HPLC. SS-LI was detected from 4 tumor extracts except one GNB (case 2) (21.3-787 pmol/g wet tissue). Sephadex G-25 column chromatography and reverse phase HPLC revealed that SS-LI in tumor extracts was eluted just after the void volume and then in the same positions as SS-28 and SS-14. These results suggest that NPY, SS-14 and SS-28 exist in tumor tissues of GN, GNB and NB, and most of the NPY-LI in one GNB was a higher molecular and more hydrophobic form of NPY-LI.     

Progastrin-like immunoreactivity in porcine antrum: identification and characterization with region-specific antisera

In an effort to identify and characterize precursors of gastrin in tissues, we generated region-specific antisera against a synthetic progastrin peptide, Try-Gly-Trp-Met-Asp-Phe-Gly-Arg-Arg (GL9), as deduced from the nucleotide sequence of gastrin mRNA. This antisera did not cross-react with gastrin or progastrin peptides with shorter carboxyl-terminal extensions. Progastrin-like immunoreactivity (PGLI) was measured in porcine antrum at a concentration of 6.8 +/- 1.2 pmol/g wet weight (mean +/- SE, n = 5), or roughly 0.2% of that of gastrin. On Sephadex G50 chromatography, a major peak of PGLI was eluted as a slightly larger molecule than gastrin heptadecapeptide (G17) but possessed the same N-terminal immunoreactivity. These findings suggest that G17 may be formed by processing of a carboxyl-terminally extended precursor as an alternative to cleavage of big gastrin (G34).     

Plasma gastrin and somatostatin in newborn infants and their relationship to catecholamines

We investigated the relationship between gastrin and somatostatin, and catecholamine concentrations in the cord blood of newborn infants. We also measured the levels of the two peptides during the first postnatal hours in the infants and furthermore characterized their molecular pattern. Twenty-two healthy infants who had been born at term were studied. Blood samples were collected from the umbilical cord and from the infants 0.5 h and 3.5 h after delivery. Peptides were measured with radioimmunoassay and further characterized by HPLC. Catecholamines were analysed by HPLC. We found that gastrin and somatostatin concentration in the umbilical cord blood was 106 +/- 40 pmol/l and 29 +/- 17 pmol/l, respectively. A significant relationship between the concentrations of somatostatin and noradrenaline in cord blood was found, (r = 0.7, n = 11, P less than 0.01). No such relation was found for gastrin. No change occurred in gastrin concentrations postnatally. Somatostatin concentration in the blood collected from the infant 0.5 h and 3.5 h after delivery was 19 +/- 11 pmol/l and 16 +/- 7 pmol/l, respectively. These concentrations were significantly lower (P less than 0.01) compared to the level measured in cord blood. Circulating gastrin was found to correspond to non-sulphated gastrin-34 and somatostatin to both somatostatin-28 and somatostatin-14. The proportion of somatostatin-28 was 30-40% and of somatostatin-14, 60-70%. We conclude that the somatostatin level, but not the gastrin level is influenced by the degree of fetal stress during labour, as evidenced by the relationship with noradrenaline. The gastrin level remained unchanged during the 3.5 h following delivery, whereas the somatostatin level decreased significantly during the same time.     

Hypersomatostatinemia in chronic renal failure

Plasma concentrations of somatostatin-like immunoreactivity (SLI) were determined in uremic patients on maintenance hemodialysis. Plasma SLI levels were significantly (p less than 0.001) elevated in 26 diabetic uremic patients (67.1 +/- 6.8 pg/ml, mean +/- SE) and in 24 non-diabetic uremic patients (43.5 +/- 7.2 pg/ml), when compared with 60 healthy subjects (5.0 +/- 0.7 pg/ml). Paired pooled plasma from uremic patients before and after hemodialysis was subjected to a reverse-phase octadecasilyl-silica (C-18) cartridge and then the extract was gel filtered on a Sephadex G-25 column (1.6 X 90 cm). Both elution profiles showed two peaks of SLI which coeluted with synthetic somatostatin (SS)-28 and SS-14 markers, respectively. The SS-28-like immunoreactivity (LI) peak, which was estimated by using SS-14 as a reference standard, was 3-fold larger than that for SS-14 LI. On the basis of immunoequivalency of the two components in the present assay, SS-28 LI constitutes approximately 75% of circulating somatostatin. In conclusion, plasma SLI is substantially high in uremic patients of both diabetic and non-diabetic etiology and the SS-28 is a predominant form of circulating SLI in these patients, probably, in part, for a lower clearance of this molecule.     

Conversion of somatostatin-28 to somatostatin-14 during maturation of epithelial cells in the porcine jejunum

Fractions of isolated epithelial cells were harvested from a segment of porcine jejunum by ten successive incubations with a chelating buffer. The cell fractions showed a progressive decrease in the activity of the brush-border enzymes, alkaline phosphatase and sucrase, with increasing incubation number but a progressive increase in the ability to incorporate labelled thymidine into DNA. Fractions enriched in cells from the crypt region (fractions 9 and 10) contained higher concentrations per mg protein of somatostatin-like immunoreactivity (1.8-fold), glucagon-like immunoreactivity (5.3-fold) and serotonin (3.0-fold) than fractions enriched in cells from the villus tip (fractions 1 and 2). Analysis of extracts of the fractions by gel filtration/radioimmunoassay showed that somatostatin-28 represented the predominant molecular form of somatostatin-like immunoreactivity in all cell fractions but the relative proportion of somatostatin-14 (and related metabolites) to somatostatin-28 was significantly higher (P less than 0.05) in fractions enriched in villus cells (fraction 1 and 2) than in fractions enriched in crypt cells (fractions 5-10). This result suggests that metabolism of somatostatin-28 to somatostatin-14 takes place during migration of the D cell from the crypt base to the villus tip. Heterogeneity in the somatostatin-14 region of the chromatograms indicates that the peptide may be further metabolized by the action of aminopeptidases.     

Identification and characterization of glycine-extended post-translational processing intermediates of progastrin in porcine stomach

We developed a radioimmunoassay specific for glycine-extended progastrin processing intermediates (G-Gly) using antisera generated against the synthetic peptide Tyr-Gly-Trp-Met-Asp-Phe-Gly. Distribution of immunoreactivity in the porcine gastrointestinal tract obtained with this antibody paralleled that of gastrin with the mucosa containing the highest quantity, 116 +/- 22 pmol/g, wet weight (mean +/- S.E., n = 5), or roughly 4% of gastrin concentration. This immunoreactivity was localized specifically to antral mucosal G-cells by immunohistochemistry. On Sephadex G-50 column chromatography of porcine antral mucosal extracts glycine-extended progastrin processing intermediates were separated into three principal molecular forms, each corresponding to known molecular forms of gastrin, component I, tetratriacontagastrin (G34) and heptadecagastrin (G17). Following purification by antibody-coupled affinity chromatography, one molecular form corresponding to G17 in size was shown to have an amino terminus identical to that of G17. Another molecular form corresponding to G34 in size could be converted to the molecular form corresponding to G17 by tryptic digestion. Our findings indicate that glycine-extended progastrin processing intermediates may serve as immediate precursors for each molecular form of gastrin, thus suggesting an alternative pathway for gastrin biosynthesis more complex than that previously conceived.     

Distribution and heterogeneity of immunoreactive cholecystokinin (CCK) in the mucosa of the porcine gastrointestinal tract

The concentration and molecular nature of cholecystokinin-like immunoreactivity (CCK-LI) in extracts of porcine intestinal mucosa were determined using sequence-specific radioimmunoassays. Highest CCK concentrations were measured in duodenal mucosa (258 +/- 60 pmol/g in the distal duodenum) followed by jejunal mucosa (204 +/- 36 pmol/g in the proximal jejunum) and pylorus (51 +/- 9 pmol/g). All other gastrointestinal regions proximal to the pylorus and distal to the jejunum contained less than 20 pmol/g. Pancreas contained less than 1 pmol/g. Gel chromatography in 6 M urea revealed four immunoreactive forms and this was confirmed by reverse-phase high-pressure liquid chromatography (HPLC). The predominant molecular form in acid extracts of duodenal mucosa resembled CCK-33 although high concentrations of the larger CCK form ('CCK-58') and of the form intermediate in size between CCK-33 and CCK-8 were measured. A molecular form resembling CCK-8 was the principal form in neutral extracts of the duodenum.     

Endocrine cells with gastrin/cholecystokinin-like immunoreactivity in the pancreas of the spiny dogfish,Squalus acanthias

Endocrine cells containing gastrin/cholecystokinin (CCK)-like immunoreactivity were localized to the islet tissue in the pancreas of the spiny dogfish. Most of these cells were located in the 'intestinal' lobe of the pancreas; only occasional cells were observed in the 'splenic' lobe. The gastrin/CCK-like immunoreactive cells were often co-localized with the 'classical' pancreas hormones (insulin, glucagon and somatostatin). Radioimmunoassay of water extracts with a C-terminally directed antiserum revealed high levels of immunoreactive material in the intestinal part (48.6 +/- 19.9 pmol/g) and lower levels (4.5 +/- 0.6 pmol/g) in the splenic part. Acetic acid extracts of the intestinal lobe contained low levels (6.8 +/- 3.3 pmol/g) of gastrin/CCK-like immunoreactivity, whereas corresponding extracts of the splenic part showed no immunoreactivity. When the extracts were subjected to DEAE ion-exchange chromatography the gastrin/CCK-like peptides eluted as a major peak. After Sephadex gel filtration, pooled immunoreactive material from the main DEAE chromatographic peak eluted at a position close to that of CCK4. Further characterization by ion-exchange and reversed-phase HPLC showed that, in general, the immunoreactive material behaved like the shorter forms of the gastrin/CCK family (CCK4/G5 and CCK8/Cae 3-10).     

Isolation of a porcine intestinal peptide with C-terminal somatostatin

A peptide showing somatostatin-like immunoreactivity has been isolated from acid extracts of porcine upper small intestine. Purification was followed by radioimmunoassay, high pressure liquid chromatography, thin layer chromatography and isotachophoresis. Preliminary chemical characterization shows that it has N-terminal serine and that it is composed of somatostatin extended from its N-terminus by an additional peptide.     

Pancreastatin-like immunoreactivity in human carcinoid disease

Pancreastatin-like immunoreactivity has been demonstrated in human carcinoid tumors by immunohistochemistry and radioimmunoassay, employing antisera raised to a synthetic C-terminal fragment of porcine pancreastatin. Immunohistochemistry revealed intense immunoreactivity in all tumors. By radioimmunoassay, high concentrations of pancreastatin-like immunoreactivity were measured in carcinoid tumors arising from the fore-gut (mean +/- S.D. and range: 369 +/- 955 and 9.4-3670 pmol g-1, respectively, n = 14), mid-gut (mean +/- S.D. and range: 1354 +/- 1538 and 337-3978 pmol g-1, respectively, n = 5) and in metastases associated with mid-gut tumors (mean +/- S.D. and range: 684 +/- 739 and 31-2255 pmol g-1, respectively, n = 7), compared to corresponding normal tissues (less than 1.4 pmol g-1). Individuals with hepatic metastases and carcinoid syndrome had elevated circulating levels of pancreastatin-like immunoreactivity (mean +/- S.D. and range: 770 +/- 1249 and 42-4120 pmol l-1; n = 12), significantly above the normal, fasting range (mean +/- S.D. and range: 14.9 +/- 7.5 and 4-37.5 pmol l-1, respectively, n = 42). However, patients with non-metastatic carcinoid tumors (n = 4), who had been clinically cured after primary tumor resection, had plasma levels within the normal range. Chromatographic analysis of extracts of primary lung and ileal tumors, hepatic metastases from ileal tumors and plasma from individuals with carcinoid syndrome revealed molecular heterogeneity of pancreastatin-like immunoreactivity.     

Somatostatin Precursor in the Rat Striatum: Changes After Local Injection of Kainic Acid

The molecular forms of somatostatin contained in the rat striatum were separated by size-exclusion HPLC. Three major peaks of somatostatin-like immunoreactivity (SLI) were resolved. Two peaks cochromatographed with synthetic somatostatin-14 (SS-14) and somatostatin-28 (SS-28), respectively. One peak exhibited a higher molecular weight (about 10,000) and may contain a proform of somatostatin. Local injection of the neurotoxin kainic acid (1 microgram) into the left striatum resulted in a persistent decrease (65-85%) of all three forms of somatostatin. In the contralateral--not injected--striatum a decrease of SLI was also observed which was maximal (45%) after 2 days and was largely abolished after 7 days. This decrease of SLI in the contralateral striatum, however, was due mainly to a decrease of SS-14 and SS-28 but not of the putative proform. Our data suggest that kainic acid causes a destruction of somatostatin-containing perikarya in the injected striatum, whereas in the contralateral striatum increased release with subsequent inactivation of SS-14 and SS-28 takes place. The putative somatostatin proform may serve as neurochemical marker for somatostatin-containing perikarya in the striatum.     

Processing of synthetic somatostatin-28 and a related endogenous rat hypothalamic somatostatin-like molecule by hypothalamic enzymes

Synthetic somatostatin-28 (S-28) as well as a related endogenous rat hypothalamic somatostatin-like compound (3K SLI) were incubated with hypothalamic extracts from which endogenous somatostatin-like immunoreactivity (SLI) had been removed by immunoabsorption. The reaction products were analyzed by gel chromatography, HPLC as well as two different radioimmunoassays for tetradecapeptide somatostatin (S-14) in which S-28 crossreacted either 100% (RIA R149) or < 0.001% (RIA S39). The results indicate that incubation of S-28 with SLI free hypothalamic extracts results in a rapid decrease of total immunoreactivity measured with RIA R149 ($\text{t}{}_{\mathrm{<mtext>1</mtext><mtext>2</mtext>}}\text{= 14}\text{min}$). By contrast, with RIA S39 a rise from zero to a peak value at 8 min was measured suggesting the formation of S-14. This was confirmed by subsequent analysis by gel chromatography and HPLC. Using endogenous 3K SLI a decrease of total R149-immunoreactivity with a similar time course ($\text{t}{}_{\mathrm{<mtext>1</mtext><mtext>2</mtext>}}\text{= 17}\text{min}$) was observed simultaneously with the emergence of material that corresponded to S-14. This converting activity seems to be specific for SLI-containing tissues since similar rates of conversion were observed with extracts from cerebral cortex and cerebellum but not with lung and liver extracts.It is concluded that (1) S-28 is converted to S-14 by hypothalamic enzymes; (2) the processing of 3K SLI is similar, suggesting the two molecules are closely related, if not identical, and (3) the regulation of S-28 to S-14 conversion could represent an important mechanism for controlling the functional activity of somatostatinergic cells.     

Plasma somatostatin-like immunoreactivity responses to a mixed meal and the heterogeneity in healthy and non-insulin-dependent (NIDDM) diabetics

Peripheral plasma somatostatin-like immunoreactivity (SLI) was estimated in non-extracted plasma using a specific somatostatin-14 (SS-14) antiserum. The basal plasma SLI level in healthy subjects (n = 18) was 43 +/- 2.9 pg/ml (mean +/- SE) and rose significantly to 8.3 +/- 2.7, 7.3 +/- 1.1 and 5.8 +/- 2.1 pg/ml above the mean basal level 20, 30, and 40 min after a mixed meal, respectively (P less than 0.05). Basal plasma SLI levels in diet (n = 8), sulfonyl urea (n = 8), and insulin groups (n = 8) of non-insulin-dependent maturity onset diabetics (NIDDM) were 50 +/- 1.6, 59 +/- 4.5, and 74 +/- 5.8 pg/ml, respectively. The basal levels for patients with NIDDM were significantly higher than those for healthy subjects (P less than 0.05). No significant increases in plasma SLI were observed after a mixed meal in any group of NIDDM subjects. Elevated plasma SLI levels are considered to be closely related to the severity of the diabetes. The ratios of SS-14 and SS-28 to the total amount of basal plasma SLI were analyzed using high pressure liquid chromatography (HPLC). The ratio of SS-14 to the total SLI was 71-80% in healthy subjects. The ratio of SS-28 to the total SLI increased from 26-30% in the diet group to 50-55% in the group on insulin. These findings suggest a possible pathophysiological role for gastrointestinal somatostatin in NIDDM.     

Somatostatin-like immunoreactivity (SLI) in pancreatic and intestinal tissues of the duck. A possible origin for the portal circulating SLI

Substances with Somatostatin-Like Immunoreactivity (SLI) were extracted using 2 N acetic acid, from the three pancreatic lobes and the intestine of the duck. The concentration of SLI was found to be very high in the pancreas (4.2 micrograms/g wet weight), the splenic lobe containing 80% of pancreatic SLI compared with 10% for the dorsal and 10% for the ventral lobes. SLI was equally distributed between duodenum, jejunum and ileum and between their mucosal and muscular layers. Chromatography of pancreatic extracts, using a Sephadex G-25 column, showed mainly the tetradecapeptide form (somatostatin-14, S-14) with a small amount of big somatostatin. Chromatography of intestinal extracts revealed three peaks with SLI: big somatostatin, somatostatin-28 (S-28) and S-14. The substance represented by the predominant peak was co-eluted with that of synthetic S-28. In normal ducks, portal plasma SLI corresponded to big somatostatin S-28 and S-14. After total pancreatectomy the S-14 form disappeared from portal plasma, whereas, when the intestinal blood vessels were ligatured, the S-28 form disappeared. We therefore hypothesize that in portal blood, S-14 has a mainly pancreatic origin, and S-28 a mainly intestinal origin.