Abstract
Xenin is a 25-amino-acid peptide extractable from mammalian tissue. This peptide is biologically active. It stimulates exocrine pancreatic secretion and intestinal motility and inhibits gastric secretion of acid and food intake. Xenin circulates in the human plasma after meals. In this study, the cellular origin of xenin in the gastro-entero-pancreatic system of humans, Rhesus monkeys, and dogs was investigated by immunohistochemistry and immunoelectron microscopy. Sequence-specific antibodies against xenin detected specific endocrine cells in the duodenal and jejunal mucosa of all three species. These xenin-immunoreactive cells were distinct from enterochromaffin, somatostatin, motilin, cholecystokinin, neurotensin, and secretin cells, and comprised 8.8% of the chromogranin A-positive cells in the dog duodenum and 4.6% of the chromogranin A-positive cells in human duodenum. In all three species, co-localization of xenin was found with a subpopulation of gastric inhibitory polypeptide (GIP)-immunoreactive cells. Immunoelectron microscopy in the canine duodenal mucosa demonstrated accumulation of gold particles in round, homogeneous, and osmiophilic secretory granules with a closely adhering membrane of 187 ± 19 nm diameter (mean ± SEM). This cell type was found to be identical to the previously described canine GIP cell. Immunocytochemical expression of the peptide xenin in a subpopulation of chromogranin A-positive cells as well as the localization of xenin immunoreactivity in ultrastructurally characterized secretory granules permitted the identification of a novel endocrine cell type as the cellular source of circulating xenin.
Keywords
I
Identification of C-terminally extended xenin in canine pancreas (proxenin) and database sequence alignment have revealed complete homology of xenin and proxenin with the mammalian N-terminus of a 140-kD cytosolic coat protein (α-COP) (Hamscher et al. 1996). This protein, together with six other cytosolic coat proteins, forms a hetero-oligomeric protein complex, the coatomer, that is involved in the formation of COP I Golgi-derived transport vesicles (Wieland and Harter 1999). It has, however, been considered highly unlikely that the N-terminus of a cytosolic coat protein could enter the secretory pathway and could appear as an endocrine regulatory peptide (Feurle 1998). Detailed extraction experiments have revealed that xenin can be extracted from all mammalian tissues in concentrations of approximately 100 pmol/g wet tissue, provided that the extraction medium contains the aspartarte protease pepsin and is acidified with 2% trifluoroacetic acid (Hamscher et al. 1995). The release of xenin into the extraction medium under these circumstances is due to proteolysis of the ubiquitously present α-COP at an enzymatic cleavage site between Leu and Thr in coat protein α. Xenin concentrations in a similar range extractable from acidified gastric mucosa without prior pepsin addition have been attributed to proteolytic cleavage of α-COP by endogeneous pepsin of the gastric mucosa (Feurle 1998). From the duodenal mucosa of humans and dogs, however, xenin can also be extracted in pmol concentrations but without addition of pepsin (Hamscher et al. 1995). Duodenal mucosa contains no pepsin but an aspartic protease, gastricsin (EC 3.4.23.3, PG II) (Samloff and Liebman 1973; Szecsi 1992). This observation suggested the duodenal mucosa as possible cellular source of xenin and led us to investigate the mammalian duodenal mucosa with region-specific antisera against xenin and proxenin, using immunohistochemistry and immunoelectron microscopy.
Materials and Methods
Tissue Preparation
Samples of stomach, pancreas, and small and large bowel were collected from tissue specimens resected during abdominal surgery of 12 patients and from tissue specimens of three beagle dogs and three Rhesus monkeys obtained during experiments in other laboratories.
Tissues were rapidly removed and fixed for 48 hr in Bouin–Hollande fixative. The tissues from Rhesus monkey were perfused with 4% formaldehyde/PBS before postfixation in Bouin–Hollande for 24–48 hr as described (Rausch et al. 1994). After dehydration in a graded series of 70% 2-propanol solution, the tissues were embedded in Paraplast Plus (Merck; Darmstadt, Germany). Adjacent sections (3 or 7 μm thick) were cut and deparaffinized. Antigen retrieval to increase the sensitivity of immunodetection was performed by heating the sections at 92–95C for 15 min in 0.01 M citrate buffer (pH 6) according to the instructions of DAKO (Hamburg, Germany). Nonspecific binding sites were blocked with 5% bovine serum albumin (BSA; Serva, Heidelberg, Germany) in PBS, followed by an avidin–biotin blocking step (avidin-biotin blocking kit; Boehringer Ingelheim, Germany).
For postembedding electron microscopic immunocytochemistry, fixation was performed for 12 hr in 1% glutaraldehyde and 4% paraformaldehyde. After rinsing in 70% ethyl alcohol over 3 days, tissues were infiltrated with LR White (medium grade; Plano W. Plannet, Wetzlar, Germany) and polymerized in gelatin capsules over 24 hr at 55C.
Ultrathin sections were sampled on formvar-coated nickel grids (100-mesh). After drying for 30 min, the sections were preincubated in PBS containing 1% BSA at pH 7.6.
Antibodies
Rabbits were immunized by repeated footpad injections of various synthetic peptide sequences: N-terminal xenin 1–9, xenin 1–25, C-terminal xenin 17–25, proxenin 17–35, and proxenin 26–35 (Figure 1) coupled to BSA by glutaraldehyde and emulsified with Freund's adjuvant. The antisera against xenin 1–25 (AS 9/4), proxenin 17–35 (AS 2519/2), and against xenin 17–25 (AS 2815/3) were useful in specific radioimmunoassays (Feurle et al. 1992; Stoschus et al. 1998). All antisera used in the present study are listed in Table 1.
Immunocytochemistry
Tissue sections were incubated with the primary antibodies (diluted as shown in Table 1) overnight at 18C and further incubated for 2 hr at 37C. After washing in distilled water and in 50 mM PBS, the sections were incubated with species-specific biotinylated secondary antibodies (Dianova; Hamburg, Germany) for 45 min at 37C, washed several times, and incubated for 30 min with the ABC reagents (Vectastatin Elite ABC kit; Boehringer, Ingelheim, Germany). Immunoreactions were visualized with 3′3-diaminobenzidine (DAB; Sigma, Deisenhofen, Germany) enhanced by 0.08% ammonium nickel sulfate (Fluka; Buchs, Switzerland), resulting in a dark blue staining. No binding was detected in the absence of the primary antibody. The same patterns of immunostaining for xenin and the other antigens examined were seen without heating the sections, although at antibody concentrations that were approximately twice as great as those indicated in Table 1 (data not shown).
Region-specific antibodies against xenin and proxenin used in this study.
List of primary antibodies
Control Experiments
There was no immunostaining with preimmunization sera. For preabsorption, the antisera were incubated at 4C overnight with 25 μmol of the respective peptide sequence and then used for immunocytochemistry. Heterologous preabsorption controls were performed for antisera against xenin with neurotensin and for antisera against neurotensin with xenin.
Co-localization Studies and Double Immunostaining
To study the co-localization of xenin immunoreactivity with other peptides and monoamines, four strategies were used: (a) alternate staining of adjacent sections; (b) the two-color immunoperoxidase technique (Hancock 1984); (c) combination of enzyme- and fluorochrome-enhanced immunohistochemistry; and (d) double-fluorescence labeling.
For adjacent section analysis 3-μm slides were stained as described above. For the two-color peroxidase technique and for combined enzyme/fluorochrome-enhanced immunohistochemistry, the first primary antibody was visualized with the nickel-enhanced DAB procedure. After dehydration through a graded series of 2-propanol and one passage through xylene, sections were rehydrated in a gradient series of 2-propanol and treated with BSA and the avidin–biotin reagents to block potential nonspecific binding of the second avidin–biotin–peroxidase complex. The second primary antibody was then visualized by the DAB/peroxidase reaction without nickel enhancement, resulting in a brown staining product, or by CY3 fluorochrome labeling (Dianova). In control sections, primary antibodies were omitted. No false-positive immunostaining of the different antibodies of the detection system was found.
Double immunofluorescence detection of xenin and chromogranin A or 5-hydroxytryptamine (5HT) was performed by covering the sections with a mixture of the two different primary antibodies in appropriate dilutions (see Table 1) and by subsequent labeling with the species-specific secondary antibodies bearing the fluorochrome CY3 or DTAF (Dianova). Sections were analyzed and photographed with the AX70 microscope (Olympus; Hamburg, Germany).
Quantitative Analysis of Xenin-positive Cells
To determine the number of xenin-immunoreactive cells as percentage of CgA-positive cells, randomly selected sections (n = 3) of dog (n = 2) and human (n = 2) duodenum were first stained with xenin antibody 2815/3 and immunoreactions were visualized with the nickel-enhanced DAB-peroxidase reaction as described above. The same sections were then stained with the monoclonal CgA antibody LK210 (human duodenum) or the polyconal antibody against the WE14 sequence of CgA (dog duodenum). CgA immunoreactivities were visualized with Cy3 labeled species-specific secondary antibodies (Dianova). In each section of dog and human duodenum, at least 400 CgA-positive cells were counted in a given area. In the same area, the number of xenin-positive cells (which consistently co-stained for CgA) was counted. For each species, the number of xenin-positive cells (co-positive for CgA) was expressed as the mean of the percentage of CgA-positive cells.
Immunoelectron Microscopy
Ultrathin sections were incubated overnight at room temperature with the polyclonal antibody AS 2815/3 (1:1000). After six 5-min rinses in PBS, pH 7.6, anti-rabbit IgG from goat conjugated with 10-nm gold particles (Dianova) was applied at 1:30 dilution. After a further incubation period of 90 min, the sections were rinsed with distilled water and contrasted with uranyl acetate and lead citrate (Weihe et al. 1991). Preabsorption controls were performed as described. Sections were analyzed with the Philipps EM 400 electron microscope (Eindhoven; The Netherlands).
Ethics
The procurement of human material during surgery was approved by the Ethics Committee of the Medical Faculty of the University of Bonn. Oral informed consent was obtained from each patient before surgery. Animal tissue specimens were collected in accordance with German Federal Law of Animal Welfare (Tierschutzgesetz). Rhesus monkey tissues, generously supplied by Dr. Lee E. Eiden (NIMH; Bethesda, MD), were obtained in accordance with NIH/NIMH governmental rules.
Results
Localization Studies and Control Experiments
Analysis of xenin expression consistently revealed specific xenin immunoreactivity in a subpopulation of endocrine cells in the duodenal epithelium, Brunner's glands, and in the jejunal mucosa of all three species examined (Figure 2).
A cytoplasmic granular pattern with peri- and infranuclear distribution of immunoreactivity was found, suggesting the endocrine nature of these cells (Figure 2, insets). The greatest density of xenin-immunoreactive cells was observed in Brunner's glands of humans and monkeys and in the epithelium of the crypts of Lieberkühn in the dog. In the jejunal mucosa, only a few xenin-immunoreactive cells were present. The specifity of the xenin immunoreactivity was confirmed by homologous preabsorption of the xenin antibody with xenin peptide (Figure 2). Adjacent sections revealed that the antibodies against the C-terminus (AS 2815/3), the N-terminus (AS 28/2), and the complete sequence of xenin 1–25 (AS 9/4) stained identical endocrine cells (Figure 3). Immunocytochemistry with antibodies against the C-terminally extended xenin (AS 2519/2 and AS 2509/2) did not reveal any immunostaining (data not shown).
Immunoreactivity with antiserum 2815/3 against the C-terminus of xenin (XN) in the crypts of Lieberkühn and Brunner's glands of dog, human, and monkey. Immunoreactive xenin cells (arrows and asterisks in
Use of sequence-specific antisera against xenin in the canine duodenal mucosa. Adjacent sections reveal identical immunoreactive cells labeled by antibody XN 2815/3 against the C-terminus of xenin 17–25 (arrows in
Adjacent section analysis and double-staining experiments demonstrated absence of neurotensin immunoreactivity in xenin-positive duodenal cells (Figure 2). Sparse neurotensin cells were present in the duodenum of human and monkey (data not shown), but none were found in the duodenal mucosa of dog (Figure 2).
Antibodies against the entire xenin molecule (AS 9/4) and against the C-terminus (AS 2815/3) stained xenin- and neurotensin-immunoreactive cells. In accordance with these results, control studies revealed that preabsorption of antibodies AS 9/4 and AS 2815/3 with xenin or with neurotensin abolished the immunoreaction. The N-terminally directed antibody AS 28/2, however, did not stain neurotensin cells, and its reaction with xenin cells was not abolished after preabsorption with neurotensin but only after preabsorption with xenin. The immunostaining of the antibody against neurotensin was abolished after preincubation of the antibody with neurotensin but not with xenin. The neurotensin antibody did not react with xenin-immunoreactive cells. Antibody AS 2815/3 against the C-terminus of xenin diffusely labeled the gastric chief cells and, to a lesser degree, pancreatic islets and the perikarya of intrinsic neurons. None of these reactions was abolished after homologous preabsorption with xenin. These findings were therefore characterized as unspecific (data not shown).
Identification of Cell Phenotype
In the three species investigated, analysis of adjacent sections and double immunostaining consistently revealed expression of xenin in a subpopulation of CgA-positive endocrine cells in the duodenal mucosa (Figure 4). Xenin-immunoreactive cells comprised 8.8% of CgA-positive cells in canine duodenal mucosa and 4.6% of CgA-positive cells in human duodenal mucosa. Xenin immunoreactivity was absent from enterochromaffin cells (EC cells) identified by their staining for 5HT and the vesicular monoamine transporter 1 (VMAT1), which has been recently localized to EC cells (Weihe et al. 1994; Erickson et al. 1996) (Figure 6). Xenin immunoreactivity was found in approximately 50% of GIP-immunoreactive cells, both in the duodenal epithelium and Brunner glands (Figures 4 and 5), whereas xenin-immunoreactive cells were different from endocrine cells in the duodenal mucosa staining for somatostatin, CCK, gastrin, neurotensin, motilin, or secretin (Figure 6; and data not shown).
Co-expression of xenin in a subpopulation of chromogranin A (CgA) cells and gastric inhibitory polypeptide (GIP) cells in the canine duodenal mucosa (AS 2815/3). Adjacent sections reveal co-localization of xenin immunoreactivity (XN, arrows in
Immunoelectron Microscopy
Immunoelectron microscopy was performed in the duodenal mucosa of two dogs with the C-terminal antibody AS 2815/3. One endocrine cell type with the ultrastrucural features of the gastric inhibitory polypeptide (GIP) cell (Usellini et al. 1984) consistently accumulated the gold particles (Figure 7). Xenin immunogold labeling was localized to round and homogeneously osmiophilic secretory granules, with a closely adhering membrane and a mean diameter of 187 nm ± 19 nm (Figure 7). Immunogold staining was not seen in other endocrine cell types identified in the same section. The immunoreaction of the antibody 2815/3 with the secretory granules was abolished after preabsorption of this antibody with 25 μmol xenin peptide (data not shown). No immunoreaction and back-ground staining were present when the secondary antiserum was omitted.
Co-expression of xenin (XN) and GIP in two endocrine cells of human Brunner's glands, demonstrated in high-power magnification on adjacent sections (AS 2815/3). Bar = 10 μm.
Discussion
The use of three different region-specific antibodies against xenin and consistent preabsorption findings permitted the identification of specific xenin-immunoreactive cells in the duodenal and jejunal mucosa of humans, monkeys, and dogs. The xenin cell was characterized at the microscopic as well as the ultrastructural level.
Xenin cells represent a subpopulation of CgA-immunoreactive cells. CgA is an established marker for neuroendocrine cells and a pan-neuronal marker (Facer et al. 1985; Siegel et al. 1988; Schäfer et al. 1994). It is clear that CgA is localized at the soluble cores of peptide hormone storage vesicles (O'Connor 1983). CgA transcription is activated in neuroendocrine cells that possess a regulated secretory pathway (Huttner et al. 1991). Our findings that all xenin-immunoreactive cells were CgA-positive, that 8.8% of CgA-positive cells in the dog reacted with antibodies to xenin, and that 4.6% of CgA positive cells were positive for xenin in human underscore the conclusion that the xenin cell is a member of the neuroendocrine cell family. The ultrastructural findings, moreover, demonstrate that xenin is localized to secretory vesicles, indicating that xenin is transported to the cell surface via the regulated, signal-dependent secretory pathway. The observation that xenin-immunoreactive plasma concentrations rise after food intake (Feurle et al. 1992) and after pharmacological stimulation (Stoschus et al. 1998) is consistent with these morphological findings.
After identification of the xenin cell in the duodenal and jejunal mucosa, we searched for co-localization with other peptidergic and monoaminergic endocrine cell types known to be present in the small intestine. To our surprise, in all three species examined we observed partial co-localization of xenin-immunoreactive cells with gastric inhibitory polypeptide (GIP)-positive cells. Approximately 50% of the GIP-immunoreactive cells were also positive for xenin. We found no xenin cells that did not also stain with the antibody to GIP.
We also identified the xenin-immunoreactive cells at the ultrastructural level. We distinguished an endocrine cell type with round, homogeneous, and osmiophilic granules, with a mean diameter of 187 ± 19 nm surrounded by a closely adhering membrane that accumulated the xenin-immunoreactive gold particles. When this cell was compared with the ultrastructurally defined endocrine cell types of the canine duodenal mucosa from the literature, it became apparent that the xenin cells we describe are identical to the GIP-producing endocrine cells. Usellini et al. (1984) described a cell with uniformly electron-dense secretory granules with a closely adherent membrane and with a mean diameter of 188 ± 34 nm recognized as the I-cell of the endocrine cell classification in dogs. These findings at the ultrastructural level corroborate our conclusions from the immunhistochemical investigation. Both the xenin cells and the GIP cells are confined to the duodenal and jejunal mucosa, and both represent subsets of CgA-positive endocrine cells. GIP is known for its ability to inhibit gastric acid secretion and to stimulate the release of insulin from the pancreatic β-cell after luminal carbohydrate and triglyceride ingestion (Pederson 1994). At the peptide sequence level, there is no structural relationship of GIP to xenin. In the human GIP gene, there are no xenin-specific nucleotide sequences (Pederson 1994). No data are available on possible co-secretion of GIP with xenin.
Segregation of xenin immunoreactivity from EC cells and various monoaminergic endocrine phenotypes in the canine duodenal mucosa (AS 2815/3). Xenin-immunoreactive cells (XN, arrowheads in
Ultrastructural demonstration of xenin-immunoreactive secretory granules in an endocrine cell of the canine duodenal mucosa (AS 2815/3). EM immunogold cytochemistry demonstrates the subcellular distribution of xenin immunoreactivity in secretory granules of an endocrine cell at low-power magnification (
Previous extraction experiments are consistent with our present immunomorphological findings. The duodenal mucosa and, to a lesser degree, the jejunal mucosa were the only sites containing no endogenous pepsin from which xenin was extractable without prior pepsinization (Hamscher et al. 1995). The unspecific immunostaining with the xenin antibody in the gastric mucosa indicates cleavage of the xenin sequence from ubiquitous α-COP under acidic conditions in the presence of endogenous pepsin (Feurle 1998). The abundant generation of xenin from α-COP in the stomach precluded a reliable immunocytochemical study in the gastric mucosa. The lack of immunocytochemical reaction in the duodenal mucosa with our antibodies against C-terminally extended xenin sequences (AS 2519/2 and AS 2509/2) indicates that free and not C-terminally extended xenin 1–25 is localized in the secretory granules of the xenin cell. Because proxenin sequences were not detectable with our antibodies directed to these sequences, the xenin immunoreactivity we observed in specific endocrine duodenal cells does not appear to be part of coat protein α.
It is not surprising that we found xenin-immunoreactive cells in the duodenal and jejunal epithelium as well as in Brunner's glands. Enteroendocrine cells, enterocytes, Paneth cells, and goblet cells arise from a common totipotent stem cell located in the mid-portion of the intestinal gland (Rindi et al. 1999). The co-localization of xenin and GIP in one cell type of the duodenal mucosa indicates that the xenin/GIP cell belongs to the developmentally related group of small intestinal endocrine cells (Rindi et al. 1999). It remains to be determined whether the gene for xenin is independent of the α-COP gene or whether the selective expression of the peptide xenin in duodenal endocrine cells is due to specific transcriptional and/or translational processing.
Our immunomorphological findings, together with previous biochemical observations, support the concept that xenin represents a further endocrine regulatory peptide, with its source in a specific endocrine cell of the duodenal and jejunal mucosa.
Footnotes
Acknowledgements
Supported by grants from the Deutsche Forschungsgemeinschaft Fe 127/10-1 and We 910/7-1.
We thank Petra Sack, Elke Rodenberg-Frank, and Ursula Egner for excellent technical assistance, and Heidemarie Schneider for brilliant photo documentation.