Gastric proton pump localization. Application of triphosphatase and monophosphatase techniques

Potassium-dependent phosphatase activity can be demonstrated in unfixed frozen sections of mouse stomach using either adenosine triphosphate (ATP) or p-nitrophenyl phosphate (NPP) as substrate. In both cases the potassium-dependent reaction is confined to oxyntic cells, but with ATP, a strong, potassium-independent reaction occurs in the connective tissue of the lamina propria and elsewhere. In the NPP system potassium-independent reaction is very slight, and the oxyntic cell reaction shows responses to inhibitors that differentiate it from Na+, K+-ATPase and that are consistent with its identification with the dephosphorylation step of the proton pump enzyme H+, K+-ATPase, recognized as the active transport component in gastric acid secretion.     

Preparation of polyethylene glycol-modified streptokinase with disappearance of binding ability towards anti-serum and retention of activity

A 37 amino acid-peptide has been isolated from porcine jejuno-ileum on the basis of its glucagon-like activity in liver (interaction with glucagon-binding sites and activation of adenylate cyclase) using gel filtration, ion-exchange and high-performance liquid chromatography. Depending on the criteria chosen, this peptide is referred to as either ‘bioactive enteroglucagon’ (activity in liver), ‘oxyntomodulin’ (specific action in gastric oxyntic glands) or ‘glucagon-37’ (chemical structure).     

Histamine dependence of pentagastrin-stimulated gastric acid secretion in rats.

Does gastrin stimulate gastric acid secretion by direct action on oxyntic cells, by releasing histamine, or by being potentiated by histamine? Previous studies in the mouse pointed to gastrin-regulated histamine release. Guinea pig and rat are well known to vary in their sensitivity to histamine. Therefore, the effects of histamine and pentagastrin were compared quantitatively on isolated, lumen-perfused, stomach preparations from these species in the absence and presence of histamine H2-receptor blockade. The loss of potency of histamine in the rat was mirrored by a loss of potency of pentagastrin consistent with the idea that pentagastrin acts by releasing histamine. In the rat, a well-defined pentagastrin curve was obtained in the presence of histamine H2-receptor block as though pentagastrin acts both directly on the oxyntic cell and indirectly by releasing histamine. It was not necessary to invoke a potentiating interaction between histamine and pentagastrin at the oxyntic cell; the two effects appeared simply to add. Potentiation was observed, however, between other combinations of stimuli, for example, between vagal nerve and pentagastrin stimulation. The physiological consequences of these results are discussed.     

Distribution of intermediate filaments in amphibian oxyntic cells

Summary Intermediate filaments of toad oxyntic cells were isolated and analysed by SDS-PAGE. The major proteins of the residue were identified as actin and a 51,000 dalton polypeptide. Immunological crossreactivity between toad oxyntic cell intermediate filament components and anti-prekeratin, was shown by double immunodiffusion tests and indirect immunofluorescence. The immunofluorescent decoration of oxyntic cells and the electron microscope images are coincident in locating the intermediate filaments mainly at the cortical and perinuclear basal zones. Furthermore, the cortical zone appears especially rich in prekeratin-like material at its adluminal third. This results in a cup-like structure that encloses the cell portion occupied by the tubulovesicular system, which does not contain intermediate filaments. The translocation of membranes occurring during the secretory cycle of the oxyntic cell, has been attributed to a system of contractile proteins. The disposition of the prekeratin-like material suggests a role for intermediate filaments in the generation of movement, produced by actin and myosin interaction, by providing a fixed plane for the anchoring of actin microfilaments.     

Inhibition of aldolase A blocks biogenesis of ATP and attenuates Japanese encephalitis virus production

Viral replication depends on host proteins to supply energy and replication accessories for the sufficient production of viral progeny. In this study, we identified fructose-bisphosphate aldolase A as a binding partner of Japanese encephalitis virus (JEV) untranslated regions (UTRs) on the antigenome via RNA affinity capture and mass spectrometry. Direct interaction of aldolase A with JEV RNAs was confirmed by gel mobility shift assay and colocalization with active replication of double-stranded RNA in JEV-infected cells. Infection of JEV caused an increase in aldolase A expression of up to 33%. Knocking down aldolase A reduced viral translation, genome replication, and viral production significantly. Furthermore, JEV infection consumed 50% of cellular ATP, and the ATP level decreased by 70% in the aldolase A-knockdown cells. Overexpression of aldolase A in aldolase A-knockdown cells increased ATP levels significantly. Taken together, these results indicate that JEV replication requires aldolase A and consumes ATP. This is the first report of direct involvement of a host metabolic enzyme, aldolase A protein, in JEV replication.     

NH(4)Cl inhibition of acid secretion: possible involvement of an apical K(+) channel in bullfrog oxyntic cells

This study was undertaken to determine the mechanism by which ammonium chloride (NH(4)Cl) inhibits stimulated acid secretion in the bullfrog gastric mucosa. To this end, four possible pathways of inhibition were studied: 1) blockade of basolateral K(+) channel, 2) blockade of ion transport activity, 3) neutralization of secreted H(+) in the luminal solution, or 4) ATP depletion. Addition of nutrient 10 mM NH(4)Cl (calculated NH(3) concentration = 92.5 microM and NH(4)(+) concentration = 9.91 mM) inhibited acid secretion within 30 min. Inhibition of acid secretion did not occur by blockade of basolateral K(+) channel activity or ion transport activity or by neutralization of the luminal solution. Although ATP depletion occurred in the presence of NH(4)Cl, the magnitude of ATP depletion in 30 min was not sufficient to inhibit stimulated acid secretion. By comparing the effect of NH(4)Cl on the resistance of inhibited or stimulated tissues, we demonstrate that NH(4)Cl acts specifically on stimulated tissues. We propose that NH(4)Cl blocks activity of an apical K(+) channel present in stimulated oxyntic cells. Our data suggest that the activity of this channel is important for the regulation of acid secretion in bullfrog oxyntic cells.     

Ligand binding and enzymic catalysis coupled through subunits in tyrosyl-tRNA synthetase.

The interaction of the tyrosyl-tRNA synthetase from Bacillus stearothermophilus with its substrates in the aminoacyl adenylation reaction has been studied by stopped-flow fluorescence. The observed changes have been assigned to their chemical and physical processes by comparison with equilibrium dialysis, pyrophosphate exchange kinetics and rapid quenching and sampling techniques to give the rate constants for ligand binding, the formation of tyrosyl adenylate, and the reverse reaction. The stoichiometry of tyrosine and ATP binding in the catalytic process has been determined directly by equilibrium dialysis and equilibrium gel filtration under pyrophosphate exchange conditions, i.e., where a steady state has been set up in which the equilibrium position favors starting materials. It is shown that the rate-determining step in the formation of tyrosyl adenylate involves 1 mole each of tyrosine and ATP. A second mole of tyrosine and ATP bind to the aminoacyl adenylate complex stabilizing the high-energy intermediate. The enzyme tyrosyl adenylate complex that is isolated by gel filtration is in a different conformational state from that in the presence of tyrosine and ATP.     

A-like cells in the rat stomach contain ghrelin and do not operate under gastrin control

Ghrelin is a 28 a.a. gastric peptide, recently identified as a natural ligand of the growth hormone secretagogue receptor (orphan receptor distinct from the receptor for growth hormone releasing hormone). In the present study, radioimmunoassay demonstrated ghrelin-like material in the rat oxyntic mucosa with moderate amounts also in antrum and duodenum. Small amounts were found in the distal intestines and pancreas. Northern blot analysis revealed abundant ghrelin mRNA in the oxyntic mucosa. Immunocytochemistry demonstrated ghrelin-immunoreactivity in endocrine-like cells in the oxyntic mucosa. Such cells occurred in low numbers also in the antrum and duodenum. The rat oxyntic mucosa is rich in endocrine (chromogranin A/pancreastatin-immunoreactive) cells, such as the histamine-rich ECL cells (65-75% of the endocrine cells), the A-like cells (20-25%) and the D cells (somatostatin cells) (10%). The ghrelin-immunoreactive (IR) cells contained pancreastatin but differed from ECL cells and D cells by being devoid of histamine-forming enzyme (ECL cell constituent) and somatostatin (D cell constituent). Hence, ghrelin seems to occur in the A-like cells. The ghrelin-IR cells in the antrum were distinct from the gastrin cells, the serotonin-containing enterochromaffin cells and the D cells. Conceivably, ghrelin cells in the antrum and distally in the intestines also belong to the A-like cell population. The concentration of ghrelin in the circulation was lowered by about 80% following the surgical removal of the acid-producing part of the stomach in line with the view that the oxyntic mucosa is the major source of ghrelin. The serum ghrelin concentration was higher in fasted rats than in fed rats; it was reduced upon re-feeding and seemed unaffected by 1-week treatment with the proton pump inhibitor omeprazole, resulting in elevated serum gastrin concentration. Infusion of gastrin-17 for 2 days failed to raise the serum ghrelin concentration. Omeprazole treatment for 10 weeks raised the level of HDC mRNA but not that of ghrelin mRNA or somatostatin mRNA in the oxyntic mucosa. Hence, unlike the ECL cells, ghrelin-containing A-like cells do not seem to operate under gastrin control.     

Gastrin and somatostatin in the rat antrum. The effect of removal of acid-secreting mucosa

Female rats were subjected to operations aimed at reducing the amount of oxyntic gland mucosa draining its acid secretion to the antrum. The rats were provided either with Heidenhain or Pavlov pouches reducing the oxyntic mucosa draining its secretion to the antrum by about 50% or subjected to various degrees (75, 90 and 100%) of fundectomy. Ten weeks following surgery, plasma levels of gastrin and somatostatin were assayed. At the same time, antral mucosal content of gastrin and somatostatin was determined as well as the mucosal density of these hormone-producing cells. There was a relationship between the amount of acid-secreting mucosa removed and the ensuring plasma concentration of gastrin. Thus, a stepwise increase in plasma gastrin was found with the highest levels obtained in rats subjected to 90 or 100% fundectomy. The somatostatin concentration in plasma was reduced only in rats subjected to fundectomy with the most sustained decrease in animals in which all oxyntic gland mucosa had been removed. There was also a relationship between the amount of acid-secreting mucosa removed and the gastrin content of the antral mucosa. An inverse relationship seemed to exist between antral gastrin and somatostatin concentrations. However, a significant decrease in somatostatin concentration of the antral mucosa was seen only in rats subjected to a fundectomy. The number of gastrin cells in the antral mucosa was increased in fundectomized rats only, with the largest density seen in rats deprived of all oxyntic mucosa. A corresponding decrease in the number of somatostatin cells was noticed. Our results would suggest an apparent functional relationship between antral gastrin and somatostatin cells, where the antral acid load (or pH) appears to be the major factor of physiological significance.     

Protein conformational change and nucleotide binding involved in regulation of sigmaF in Bacillus subtilis.

We have studied the ability of three mutant forms of SpoIIAA, containing amino acid substitutions at the site of phosphorylation (serine 58), to interact with SpoIIAB. Native gel analysis revealed that SpoIIAAS58A could form a complex with SpoIIAB in the presence of ADP and more strongly in the presence of ATP. SpoIIAAS58N did not form a complex with SpoIIAB in the presence of ADP but displayed some interaction with SpoIIAB in the presence of ATP. SpoIIAAS58D was unable to form a complex with SpoIIAB in the presence of either ADP or ATP. Corresponding differences were found in the behavior of the three mutant proteins when studied by gel permeation with high-performance liquid chromatography and limited proteolysis. SpoIIAAS58A behaved like the wild-type SpoIIAA, SpoIIAAS58D like SpoIIAA-P, and SpoIIAAS58N in a way that was intermediate between the behaviors of SpoIIAA and SpoIIAA-P. Limited proteolysis was also used to show that on binding of ADP or ATP SpoIIAB undergoes a shift in conformation. The affinity of SpoIIAB for ADP and ATP was determined by limited proteolysis in the presence of a wide range of nucleotide concentrations. The results indicated that SpoIIAB has approximately equal affinity for ADP and for ATP.     

Quantitative subcellular study of apical pole membranes from chicken oxyntic cells in resting and HCl secretory state

Vertebrate oxyntic cells, responsible for gastric HCl production, undergo a remarkable morphological reorganization in relation to their secretory cycle. In resting state, the luminal surface of the cells is smooth; a peculiar system of endocellular membranes, the tubular system, occupies the luminal cytoplasm. Actin filaments frame a cortical network between the tubular system and the luminal plasma membrane. With the onset of HCl secretion, the tubular system becomes incorporated into the luminal plasma membrane. Villous processes containing microfilaments fill the secretory surface. This morphological reorganization of membranes and cytoskeletal matrix could regulate HCl secretion by translocation of membranes containing the proton pump from the endocellular compartment to the secretory surface. In this paper, we describe the isolation of membranes that selectively belong to the tubular system or to the cytoplasmic processes of the secretory surface of chicken oxyntic cells. Chicken oxyntic cells are the main cellular component of the proventricular glands. A resting state was obtained after cimetidine treatment, whereas the HCl- secretory state was induced by histamine. We present a comparative analysis of resting and stimulated chicken gastric glands by quantitative subcellular fractionation. The HCl secretory state was related to specific modifications in membrane fractions derived from the secretory pole of oxyntic cells. Morphological and functional reorganization of oxyntic cells was closely correlated with changes in: the sedimentation pattern of the marker enzyme of the apical pole membrane (K-NPPase), the total activity of K-NPPase and nonmitochondrial Mg-ATPase, the valinomycin dependence of K-ATPase, and polypeptides that cosediment in purified membrane fractions. Changes in the distribution pattern of K-NPPase after fractionation of histamine- stimulated glands were consistent with the replacement of the small vesicles typical of resting glands by dense membrane profiles, analogous to the luminal processes of stimulated oxyntic cells. SDS- PAGE showed that, in purified membrane fractions of stimulated glands, the concentration of 28-, 43-, and 200-kD polypeptides increased while that of 95- and 250-kD polypeptides decreased. The present results define the tubular system of oxyntic cells as an organelle with properties different from those of endoplasmic reticulum, mitochondria, and plasma membrane. The biochemical and physico-chemical properties of this membraneous system changed when the organization of the membranes and the cytoskeletal matrix of the apical pole was modified by the onset of HCl secretion.     

A comparison of ribosomal proteins from rabbit reticulocytes phosphorylated in situ and in vitro.

A comparison has been made between the ribosomal proteins phosphorylated in intact cells and proteins isolated from ribosomal subunits after modification in vitro by purified protein kinases and [gamma-32P]ATP. When intact reticulocytes were incubated for 2 h in a nutritional medium containing radioactive inorganic phosphate, one phosphorylated protein was identified as a 40S ribosomal component using two-dimensional polyacrylamide gel electrophoresis followed by electrophoresis in a third step containing sodium dodecyl sulfate. This protein, containing 99% of the total radioactivity associated with ribosomal proteins as observed by two-dimensional electrophoresis, is found in a nonphosphorylated form in addition to several phosphorylated states. These states differ by the number of phosphoryl group attached to the protein. The same 40S protein is modified in vitro by the three cAMP-regulated protein kinases from rabbit reticulocytes. Two additional proteins associated with the 40S subunit are phosphorylated in situ. These proteins migrate as a symmetrical doublet, and contain less than 1% of the radioactive phosphate in the 40S subunit. A number of phosphorylated proteins associated with 60S subunits are observed by disc gel electrophoresis after incubation of whole cells with labeled phosphate. These proteins do not migrate with previously identified ribosomal proteins and are not present in sufficient amounts to be identified as ribosomal structural proteins. Proteins in the large subunit are modified in vitro by cAMP-regulated protein kinases and ATP, and these modified proteins migrate with known ribosomal proteins. However, this phosphorylation has not been shown to occur in intact cells.     

Gastrin-releasing peptide: neuronal distribution and spatial relation to endocrine cells in the human upper gut

By using immunocytochemical techniques, we have studied the distribution of gastrin releasing peptide (GRP)-containing neurons as well as the spatial relationship between these neurons and the endocrine cells in the human stomach and duodenum. Moderate numbers of immunoreactive fibers were distributed in the smooth muscle and submucosa of the stomach; they were more rare in the duodenal wall. Numerous GRP-containing nerve fibers were found in the oxyntic mucosa, the antral mucosa harboured only few GRP immunoreactive nerve fibers. The mucosa of the proximal duodenum was found to be virtually devoid of such fibers. Only occasionally did we observe signs of a direct contact between GRP-containing nerve fibers and gastrin and somatostatin cells in the antral mucosa. In the oxyntic mucosa GRP-containing nerve fibers sometimes seemed to contact endocrine cells, including somatostatin cells as well as individual parietal cells. In conclusion, although GRP-containing nerve fibers were quite numerous in the wall of the human upper gastro-intestinal (GI)-tract, we observed a lack of intimate spatial relationship between these fibers and endocrine cells in the antral mucosa, suggesting additive mechanisms to a direct innervation of gastrin cells and somatostatin cells by GRP nerve fibers explaining the physiological effects on hormonal release.     

Gamma-aminobutyric acid immunoreactivity in the enterochromaffin cells of the rat stomach.

The present immunocytochemical study revealed gamma-aminobutyric acid (GABA) immunoreactivity in the oxyntic and pyloric mucosa of the rat stomach at light- and electron-microscopic levels. GABA-immunoreactive endocrine cells were numerously seen in the lower half portion of the pyloric mucosa but rarely in the oxyntic mucosa. These cells were round or oval in shape and sometimes had a short cytoplasmic process. Serotonin-immunoreactive enterochromaffin (EC) cells were also observed in the oxyntic and pyloric mucosa of the stomach. The distribution and shapes of the immunoreactive cells were similar to those of the GABA-immunoreactive cells. With a double immunolabeling technique using anti-GABA and antiserotonin serum, GABA-immunoreactive endocrine cells showed serotonin immunoreactivity and were identified as EC cells. At the electron-microscopic level the GABA-immunoreactive cells contained round or oval, spindle-like, pear-shaped granules in EC cells. The immunoreaction product in the EC cells was generally confined to the granular cores. These findings suggest that GABA may be synthesized in the EC cells and be released from the granules of the cells after adequate stimuli.     

Effects of ECL cell extracts and granule/vesicle-enriched fractions from rat oxyntic mucosa on cAMP and IP(3) in rat osteoblast-like cells

The existence of an osteotropic hormone (referred to as gastrocalcin) in the ECL cells of the gastric mucosa has been suggested. Both gastrin and an extract of the oxyntic mucosa lower blood Ca(2+) and stimulate Ca(2+) uptake into bone. The ECL cells are known to operate under gastrin control and, conceivably, gastrin lowers blood Ca(2+) indirectly by releasing the hypothetical ECL cell hormone. We have shown earlier that extracts of isolated ECL cells or of the granule/vesicle fraction of the oxyntic mucosa evoke a typical Ca(2+)-mediated second messenger response in osteoblastic cells. In the present study, we characterize this response further. An increase in intracellular inositol 1,4,5-trisphosphate (IP(3)) concentration was observed after treatment of UMR-106.01 osteoblast-like cells with extracts of ECL cells or granule/vesicle-enriched fractions from oxyntic mucosa. Intracellular cyclic adenosine monophosphate (cAMP) concentrations were not affected. Inhibition of phospholipase C (PLC) by U-73122 abolished the increase in [Ca(2+)](i). Preincubation of UMR-106.01 cells with pertussis toxin, which blocks many G-proteins, did not prevent the increases in IP(3) and [Ca(2+)](i). It was also found that the novel peptide hormone ghrelin, produced in the A-like cells of the oxyntic mucosa, did not evoke any Ca(2+) signal in osteoblastic cells. The results indicate that the extracts mediate their effects through a pertussis toxin-insensitive mechanism, and that binding to a receptor leads to activation of PLC and production of IP(3) resulting in increased [Ca(2+)](i). The putative osteotropic hormone is distinct from ghrelin.     

Ca2+ -dependent interaction of S100A1 with F1-ATPase leads to an increased ATP content in cardiomyocytes

S100A1, a Ca(2+)-sensing protein of the EF-hand family that is expressed predominantly in cardiac muscle, plays a pivotal role in cardiac contractility in vitro and in vivo. It has recently been demonstrated that by restoring Ca(2+) homeostasis, S100A1 was able to rescue contractile dysfunction in failing rat hearts. Myocardial contractility is regulated not only by Ca(2+) homeostasis but also by energy metabolism, in particular the production of ATP. Here, we report a novel interaction of S100A1 with mitochondrial F(1)-ATPase, which affects F(1)-ATPase activity and cellular ATP production. In particular, cardiomyocytes that overexpress S100A1 exhibited a higher ATP content than control cells, whereas knockdown of S100A1 expression decreased ATP levels. In pull-down experiments, we identified the alpha- and beta-chain of F(1)-ATPase to interact with S100A1 in a Ca(2+)-dependent manner. The interaction was confirmed by colocalization studies of S100A1 and F(1)-ATPase and the analysis of the S100A1-F(1)-ATPase complex by gel filtration chromatography. The functional impact of this association is highlighted by an S100A1-mediated increase of F(1)-ATPase activity. Consistently, ATP synthase activity is reduced in cardiomyocytes from S100A1 knockout mice. Our data indicate that S100A1 might play a key role in cardiac energy metabolism.     

ELECTRON MICROSCOPY OF THE OXYNTIC CELL IN THE GASTRIC GLANDS OF THE BULLFROG (RANA CATESBIANA) : I. The Non-Acid-Secreting Gastric Mucosa

The fine structure of the oxyntic cell from the gastric glands of the bullfrog was studied in lead hydroxide—stained sections of gastric mucosa fixed in buffered osmium tetroxide and embedded in n-butyl methacrylate. The oxyntic cell in non-acid-secreting stomachs (gastric juice pH, 7.4–7.8) is characterized by: (a) numerous closely packed smooth surfaced vesicular and tubular profiles disposed randomly in the cell; some of these elements show interconnections making it possible to identify this component with smooth surfaced endoplasmic reticula of certain other cell types, (b) a small percentage of rough surfaced profiles characteristic of endoplasmic reticula possessing RNP particles on the outer membrane surfaces, (c) a Golgi complex consisting of multiple isolated non-polarized arrays of smooth surfaced parallel elongated profiles and associated vesicular elements, (d) a sparse granular component (140 A) scattered freely in the cytoplasmic matrix, (e) numerous mitochondria with a dense matrix and containing an unusually large number of closely approximated cristae, (f) a number of zymogen granules consisting of either a dense body limited by a membrane or surrounded by a halo of less dense material which is in turn limited by a membrane, and (g) a number of granules (~260 A) containing several smaller granules (~80 A) identified presumably as glycogen. Intracellular canaliculi were not observed. Instead the free surface of the oxyntic cell facing the lumen of the gastric gland shows a complicated plication of the plasma membrane. Intercellular canaliculi are seen frequently between adjacent oxyntic cells. The walls of these canaliculi are made up of folded and ruffled cell membranes. The basal surface of the cell also exhibited this type of configuration. Occasional smooth surfaced profiles are seen communicating with the free surface, the wall of an intercellular canaliculus, or the basal surface of the cell. Although nerve endings were not found in association with oxyntic cells, unmyelinated nerves were observed in the vicinity of the gastric glands.     

Effect of reserpine on the generation of the chromogranin A-derived neuropeptide WE-14 in rat oxyntic mucosa

WE-14, a post-translational product of the neuroendocrine protein chromogranin A (CgA), is generated in distinct subpopulations of endocrine cells. The objective of this study was to investigate the generation of WE-14 in the endocrine cell types of the oxyntic mucosa of the stomach, after treatment with reserpine, an irreversible inhibitor of vesicular monoamine uptake 2 (VMAT2). Reserpine (10 mg/kg) was administered subcutaneously and tissue analysed 1, 3, 5 and 18 h following treatment. The oxyntic mucosa was analysed immunohistochemically employing a site-specific WE-14 antiserum, a region-specific CgA antiserum and an antiserum against histidine decarboxylase (HDC), a marker of the histamine-producing ECL cells in the oxyntic mucosa. The number of oxyntic endocrine cells exhibiting WE-14 immunostaining increased more than 100-fold 18 h after reserpine administration relative to vehicle treated controls. Double immunostaining with HDC revealed that most, but not all, of the WE-14 positive cells were ECL cells. These results suggest that reserpine has the ability to influence the post-translational processing of CgA to generate WE-14 in rat stomach ECL cells, presumably as a consequence of reduced VMAT2-driven accumulation of histamine.     

Immunohistochemical studies on glucagon, glicentin and pancreatic polypeptide in human stomach: normal and pathological conditions

Summary Endocrine-like cells containing glucagon, glicentin or pancreatic polypeptide immunoreactivity in human foetal and adult stomach, with or without disease, were studied with the indirect immunoperoxidase method and mirror sectioning technique. In foetal and neonatal oxyntic mucosae, there were endocrine-like cells with glucagon and glicentin immunoreactivities and argyrophilia. Cells containing glicentin immunoreactivity alone were detected earlier than glucagon cells during foetal development, and were also distributed throughout foetal to neonatal life. Bovine pancreatic polypeptide immunoreactivity coexisted in a subpopulation of the glucagon-glicentin cells. These cells were absent from normal oxyntic mucosa in the postneonatal period and from normal antral mucosa throughout life. Hamartomatous polyp in adult oxyntic mucosa, hyperplastic oxyntic mucosa in Menetrier's disease and atrophic oxyntic mucosa in a remnant stomach with cancer showed scattered glucagon-glicentin cells, but few or no cells containing bovine pancreatic polypeptide. Intestinalized mucosa showed plentiful glicentin cells with occasional glucagon and/or bovine pancreatic polypeptide immunoreactivity. Some gastric cancer cells of both diffuse and adenoplastic types contained immunoreactive glicentin and, less frequently, glucagon. Bovine pancreatic polypeptide immunoreactivity was detected in a few adenoplastic cancer cells, but not in diffuse type cells. Three different anti-pancreatic polypeptide sera against bovine, porcine or human pancreatic polypeptide detected basically the same cells mentioned above, but pancreatic polypeptide cells lacking human pancreatic polypeptide immunoreactivity were also present in foetal oxyntic mucosa. Immunoabsorption tests revealed that the bovine pancreatic polypeptide immunoreactivity was remote from peptide YY and neuropeptide Y.