Gastric Amylin Expression: Cellular Identity and Lack of Requirement for the Homeobox Protein PDX-1. A Study in Normal and PDX-1-Deficient Animals with a Cautionary Note on Antiserum Evaluation

Abstract

The gene encoding amylin is implicated in the generation of amyloid in the islets of Langerhans of diabetics and is believed to be regulated by the homeodomain transcription factor PDX-1. Although gastric mucosa also produces amylin, studies on its cellular site of production have yielded highly divergent results, localizing this peptide to either gastrin, serotonin, or somatostatin cells or to combinations thereof. Using region-specific amylin antisera in combination with reverse transcriptase-polymerase chain reaction, we now document that the majority of cells expressing amylin correspond to somatostatin cells. Only a small subpopulation of gastrin cells contained immunoreactive amylin. Studies of PDX-1-deficient mice, which fail to develop gastrin cells while possessing normal numbers of somatostatin cells, revealed no detectable change in gastric amylin expression. These data show that neither normal gastrin cell development nor PDX-1 expression is needed for gastric amylin expression. (J Histochem Cytochem 47:973–980, 1999)

Keywords

serotonin PDX-1 amylin IAPP peptide hormones gastrin somatostatin

Amylin or islet amyloid polypeptide (IAPP) is a 37 amino-acid peptide that was originally identified as the chief constituent of amyloid present in an insulinoma and in the islets of patients with non-insulin-dependent diabetes mellitus (NIDDM) (Westermark et al. 1986; Cooper et al. 1987). The peptide shows structural similarities to calcitonin gene-related peptide (CGRP) and to adrenomedullin (Kitamura et al. 1994), and has been localized to the insulin cells of the islets of Langerhans (Westermark et al. 1987). In addition, amylin has been detected in islet somatostatin (D) cells (deVroede et al. 1992; Mulder et al. 1997).

Amylin possesses diabetogenic effects by virtue of its ability to inhibit pancreatic insulin release and to interfere with insulin action (Leighton and Cooper 1988; Ohsawa et al. 1989; Koopmans et al. 1991). These effects, together with its propensity to form amyloid fibrils in certain species and in humans, have been considered to represent potential contributing causes of NIDDM. In agreement with this, studies in pancreatic cancer patients as well as in transgenic animals support a diabetogenic role of human amylin (Permert et al. 1994; Janson et al. 1996)

Several studies have addressed mechanisms regulating amylin expression. Because amylin often, but not invariably, is co-expressed with insulin, many studies have focused on transactivating factors that can bind to specific motifs in both the insulin and the IAPP promoters. The most studied factor is the pancreatic duodenal homeobox 1 (PDX-1), which also is referred to as insulin promoter factor 1 (IPF-1), somatostatin transactivating factor 1 (STF-1), islet duodenal homeobox 1 (IDX-1), and insulin upstream factor 1 (IUF-1) (Leonard et al. 1993; Ohlsson et al. 1993; Miller et al. 1994; Stein et al. 1996). A/T-rich binding sites for PDX-1 have been detected in the insulin, amylin, and somatostatin promoters/enhancers. At least two such sites appear to be functionally involved in regulation of amylin gene expression (Bretherton-Watt et al. 1996; Watada et al. 1996; Carty et al. 1997). Moreover, stable transfection of PDX-1 into a glucagonoma cell line activated transcription of insulin and amylin in these cells (Serup et al. 1996). All these studies indicate that PDX-1 is an important factor for insulin and amylin expression. In contrast, studies employing antisense oligodeoxynucleotides have documented that diminishing PDX-1 expression by 80-90% does not have any appreciable effect on insulin or amylin expression (Kajimoto et al. 1997). However, these observations do not exclude the possibility that that the remaining PDX-1 molecules are sufficient to activate the insulin and amylin promoters. Mice deficient in PDX-1 have been produced by targeted gene deletion (Jonsson et al. 1994). Such mice fail to develop a pancreas and therefore cannot be directly used for studying β-cell expression of insulin and amylin. Insulin cell-targeted inactivation of PDX-1 by the CRE-Lox system has, however, shown that PDX-1 positively regulates amylin and insulin expression (Ahlgren et al. 1998). Amylin is also expressed in several extrapancreatic sites, with highest concentrations in antrum (Ferrier et al. 1989; Nakazato et al. 1989; Asai et al. 1990; Miyazato et al. 1991), and has effects on gastric emptying and gastric secretion (for a recent review see Guidobono 1998). Whereas there is general consensus that amylin is expressed in the stomach, the identity of the cells expressing it has been surprisingly controversial. Amylin immunoreactivity has been variously localized to gastrin cells (Ohtsuka et al. 1993), to somatostatin cells and to a few gastrin cells (Mulder et al. 1994, 1997), to neither gastrin nor somatostatin cells (Toshimori et al. 1990), and to serotonin cells (D'Este et al. 1995). Although these discrepancies cannot be ascribed to species differences (rats were included in all studies), they may relate to technical differences, unspecific absorption of immunoglobulins, and crossreactivity phenomena.

We decided to reinvestigate the cellular origin of amylin in rats and mice using an assortment of well-characterized antibodies, and to investigate the effects of PDX-1 deficiency on mouse gastric amylin expression.

Materials and Methods

Tissue Material

Mice deficient in PDX-1 (IPF-1) were generated by homologous recombination as previously described (Jonsson et al. 1994). Newborn, apancreatic mutant mice, and wild-type normal mice, as well as adult mice, were decapitated and the distal region of the stomach was quickly dissected out and frozen in liquid nitrogen. In addition, adult rats and mice were anesthesized with CO₂ and intracardially perfused with 2 ml of saline followed by either 4% paraformaldehyde (Merck; Darmstadt, Germany) in 0.1 M sodium phosphate buffer, pH 7.4, or Zamboni's fixative (Stefanini et al. 1967). Specimens from the antropyloric and oxyntic gland regions of the stomach, pancreas, and thyroid were subsequently immersion-fixed in the respective fixatives overnight and routinely embedded in paraffin, or were cryoprotected with 30% sucrose and prepared for cryostat sectioning.

Antisera

Two rabbit antisera, one raised against the synthetic N-terminal [1–13] (IHC 7324) and the other against the synthetic C-terminal [25–37] (IHC 7326) region of human amylin, were obtained from Peninsula Labs (Merseyside, UK). The [1–13] region of amylin is identical between human and rat, whereas the [25–37] region contains four substitutions localized at amino acids 25–26 and 28–29, respectively. Two additional rabbit antisera produced against the whole synthetic rat amylin molecule (IHC 7323) and against the synthetic rat amylin [11–37] region (“ODM”; Madsen et al. 1990) were also used. In addition, two rabbit antisera to synthetic rat αCGRP [IHC 6006 from Peninsula and CA 1134 from Affiniti Research (Nottingham, UK)], a guinea pig antiserum to synthetic human gastrin (Euro Diagnostica; Malmö, Sweden) and a monoclonal rat anti-somatostatin antibody (Chemicon International; Temecula, CA) were used. Second antibodies included fluorescein isothiocyanate-(FITC)-, Texas Red-, biotin-, and aminomethyl coumarin (AMCA)- conjugated species-specific donkey antisera to rabbit and guinea pig Ig, biotin-conjugated donkey anti-rat Ig antibody, and peroxidase- and AMCA-streptavidin conjugates (Jackson Labs; West Grove, PA). Alkaline phosphatase-conjugated swine anti-rabbit Ig (D306) was obtained from Dako (Glostrup, Denmark).

Immunocytochemical Models

Descending concentrations of synthetic rat CGRP, rat amylin, and rat adrenomedullin [1–50] were spotted on 0.22-μm nitrocellulose membranes (Hybri-slot precut filters PK/8; Gibco BRL, Gaithersburg, MD), dried, and vapor-fixed in paraformaldehyde fumes at room temperature (Larsson 1981, 1988). Then the membranes were blocked with 10% normal serum and stained with the four different amylin antisera diluted 1:2000. The site of the antigen-antibody reaction was revealed using alkaline phosphatase-conjugated anti-rabbit IgG, followed by development in bromochloroindolyl-phosphate-nitroblue tetrazolium (BCIP-NBT) medium (Larsson and Hougaard 1993).

Immunocytochemical Staining

Five-μm paraffin sections were dewaxed, hydrated, and either digested with 0.03% pepsin (Sigma; St Louis, MO) in 0.1 M HCl for 20 min or left unpretreated. Then the sections were treated with 1 % sodium borohydride (Sigma) in redistilled water for 15 min, preblocked in 10% horse serum for 30 min, and incubated with the different primary amylin and CGRP antisera diluted 1:1000-1:5000 overnight at 4C. The site of the antigen-antibody reaction was revealed by indirect immunofluorescence or immunoperoxidase staining followed by development in diaminobenzidine medium (Larsson 1988). Cell identification was determined by triple stainings using mixtures of the different rabbit amylin anti-sera with guinea pig anti-gastrin serum and monoclonal rat anti-somatostatin antibodies, followed by species-specific antibodies conjugated to different fluorochromes. Immunofluorescence specimens were examined in a Leica epillumination microscope equipped with selective filters for FITC, Texas Red, and AMCA. To obtain complete selectivity for AMCA, the commercial filter cube was supplemented with a 490-mm shortpass emission filter.

Controls included absorptions of the primary antisera with synthetic rat amylin [1–37], rat CGRP, human gastrin [1–17], somatostatin [1–14] (from Peninsula), bovine thyroglobulin, bovine serum albumin, and low molecular weight poly-L-lysine (from Sigma) as well as conventional staining controls, as previously detailed (Larsson 1988).

Reverse Transcriptase-Polymerase Chain Reaction

Antropyloric and pancreatic total RNA was prepared from adult mice and from 1-day-old wild-type and PDX-1-deficient mice by extraction of tissues with Trizol (Gibco BRL; Paisley, UK). One to 5 mg total RNA was reverse-transcribed using random hexamers and SuperScript RT according to the manufacturer's instructions (Gibco BRL) and amplified with primers specific for β-actin (5′CCACCACAGCTGAGAGGGAAAT3′) and (5′AGGGGCCGGACTCATCGTA3′) (Gene-bank accession number J00691:2394U22, gastrin (5′TCTTCTTCTTCCTCCATTCGTG3′) and (5′TGTGTACATGCTGGTCTTAGTGCT3′) (U58136:99U24), PDX-1, (5′GCTCACCTCCAC-CACCACCTT3′) and (5′GCAGTACGGGTCCTCTTGTTTTCC3′) (MMIPF1:520U21), IAPP (5′AACTGCCAGCTGTCCTCCTCATCC3′) and (5′G-CCGTGTTGCACT-TCCGTTTGTC3′) (M25389:22U24), and somatostatin (5′TGCTGGGTTCGAGTTGGCAGACCT3′) and (5′GTCCTGGCTTTGGGCGGTGTC3′) (X51468:941U21). The polymerase chain reactions were performed with the AmpliTaqGold kit (Perkin Elmer; Roche Molecular System, Branchburg, NJ) according to the manufacturer's instructions. The PCR was run for 35 cycles with each cycle consisting of 45 sec at 94C, 60 sec at 60C, and 60 sec at 72C. The primers all spanned an intron and the amplicons resulting from mRNA were 109 (amylin), 239 (gastrin and somatostatin), and 503 (actin) base pairs, respectively.

Figure 1

Mouse antropyloric mucosa. Serial sections immunoperoxidase-stained with amylin antiserum 7326 after preabsorption with poly-L-lysine (A), CGRP (B), or amylin (C), or with CGRP antiserum CA 1134 (D). Note that the amylin antiserum stains endocrine-like cells and that the staining is eliminated by preabsorption against amylin but not against CGRP or poly-L-lysine. In contrast, the CGRP antiserum stains many nerve fibers and terminals but fails to detect antropyloric endocrine cells. Bar = 50 μm.

Results

Immunocytochemistry

Paraffin sections of paraformaldehyde-fixed gastric mucosa required predigestion with pepsin to stain with the different amylin antisera. In contrast, sections from Zamboni-fixed, paraffin-embedded specimens did not require such predigestion. All amylin antisera stained many endocrine-like cells of the antropyloric mucosa (Figure 1). In addition, all amylin antisera stained few and scattered nerve fibers and terminals of the external muscle layer, mucosa, and submucosa. Staining with the CGRP antiserum CA 1134 revealed no endocrine-like cells but detected nerve fibers and terminals in the external muscle layer, mucosa, and submucosa (Figure 1). The number of nerve fibers detected by the CGRP antiserum vastly exceeded the number detected by the amylin antisera.

Triple stainings revealed that the vast majority of the amylin-immunoreactive endocrine-like cells corresponded to somatostatin cells, whereas only a few gastrin cells were amylin-immunoreactive (Figure 2). In addition, a few scarce amylin-immunoreactive cells were detected that were both gastrin- and somatostatin-negative. The same pattern emerged with all four amylin antisera. Absorption against synthetic CGPR eliminated all nerve fiber staining but did not affect the staining of the endocrine-like cells (Figure 1). Absorption against synthetic amylin [1–37] eliminated staining of both nerve fibers and endocrine-like cells (Figure 1), whereas absorption against poly-L-lysine, gastrin, or somatostatin was without effect.

Figure 2

Mouse antropyloric mucosa triple-stained for amylin (Texas Red immunofluorescence, A), somatostatin (blue AMCA immunofluorescence, B), and gastrin (green FITC immunofluorescence, C). Note in the triple exposure (D) that most somatostatin cells are amylin-immunoreactive (small arrows), whereas only one gastrin cell (arrowhead) is amylin-positive. Bar = 50 μm.

These results suggested that all four amylin antisera contained a subpopulation of antibodies that crossreacted with CGRP. Preabsorption of the antisera with CGRP eliminated this subpopulation and also eliminated neural staining but did not affect endocrine cell staining. Cytochemical models also confirmed the presence of CGRP-crossreactive antibodies in all four amylin antisera. Notably, whereas the amylin antisera detected down to 0.2 pmol spots of amylin, 100-fold higher (20 pmol) concentrations of CGRP were needed for detection. In contrast, neither of the four antisera crossreacted with adrenomedullin. Together, these results show that authentic amylin is present in a major population of somatostatin cells and in a smaller sub-population of other endocrine-like cells of the rat and mouse gastric mucosa. In addition, these results show that CGRP is not present in detectable amounts in antropyloric endocrine-like cells but is strictly a neural peptide in this tissue. The observation that the amylin antisera detected only few and scattered nerve fibers probably mirrors the fact that they needed 100-fold more CGRP than amylin for detection. Therefore, it is likely that the nerve fibers stained by the amylin antisera corresponded to nerve fibers containing particularly high CGRP concentrations. Supplementary studies with an additional CGRP antiserum (IHC 6006) were undertaken. However, in paraformaldehyde-fixed paraffin sections, this antiserum stained endocrine-like cells but did not detect any nerve fibers (Figure 3). In contrast, use of paraformaldehyde-fixed cryostat sections permitted detection of fibers and terminals in the mucosa, submucosa, and muscularis in numbers similar to those seen with the CGRP antiserum CA 1134. In addition, in cryostat sections antiserum IHC 6006, but not CA 1134, stained gastric endocrine-like cells (Figure 3). Absorption of IHC 6006 against synthetic CGRP eliminated all staining in both paraffin and cryostat sections (Figure 3). Triple staining revealed that all of the CGRP-immunoreactive cells detected by this antiserum corresponded to gastrin cells. Although gastrin possesses no overt structural similarity to CGRP, we also preabsorbed the IHC 6006 antiserum with synthetic human gastrin [1–17]. Such absorption completely eliminated staining of the endocrine-like cells but did not affect staining of the neural structures (Figure 3). To further evaluate the effects of gastrin preabsorption, thyroid tissue was also studied using the CGRP antisera. Whereas CGRP antiserum CA 1134 stained only parafollicular C-cells of the rat thyroid, antiserum IHC 6006, in addition, produced strong staining of colloid present in the thyroid follicles (Figure 4). This staining was unaffected by preabsorption with gastrin. Preabsorption against CGRP eliminated only the C-cell staining, whereas preabsorption against thyroglobulin eliminated the colloid but not the C-cell staining (Figure 4). Although the possible use of a conjugating protein for production of the IHC 6006 antiserum was never mentioned in the antiserum data sheet, it is possible that a CGRP-thyroglobulin conjugate had been used for immunization.

Figure 3

Rat antropyloric mucosa, paraffin sections (A-C) and cryostat section (D) immunocytochemically stained with CGRP antiserum 6006 preabsorbed with poly-L-lysine (A,D), CGRP (B), or gastrin (C). Note that in paraffin sections this antiserum detects only endocrine cells and that this staining is eliminated by preabsorbtion against either CGRP or gastrin. In cryostat sections, the antiserum detects both endocrine cells and nerve fibers. The number of endocrine cells stained in cryostat sections often appeared greater than the number stained in paraffin sections. Bar = 50 μm.

RT-PCR

Use of amylin-specific primers amplifying sequences present in exons 1 and 2 of the mouse amylin gene permitted the detection of an amplicon corresponding in size to that expected from amylin mRNA in mouse antropyloric mucosa. The absence of an amplicon corresponding in size to that expected from DNA excluded genomic contamination. This was further underscored by the negative controls (exclusion of template and exclusion of reverse transcriptase, respectively). RNA isolated from PDX-1-deficient mice and from wild-type controls showed amylin-specific amplicons of the same size and intensity (Figure 5). In contrast, amplification using specific primers confirmed absence of PDX-1 and gastrin mRNA in the PDX-1-deficient, but not in the wild-type animals (cf. Larsson et al. 1996).

Discussion

Our antisera characterizations and staining, as well as RT-PCR studies, strongly support the concept that authentic amylin is expressed in the majority of gastric somatostatin-producing D-cells of rat and mouse antropyloric mucosa. In addition, less expression occurs in gastrin cells and in an additional small population of unidentified endocrine-like cells. These data agree with findings published by Mulder et al. (1994, 1997) but disagree with findings obtained by D'Este et al. (1995), Ohtsuka et al. (1993), and Toshimori et al. (1990). There are many potential explanations for these discrepant results. Gastrin cells show a troublesome ability to absorb immunoglobulins unspecifically. Absorption controls using peptides containing basic and/or aromatic amino acid residues can become falsely negative with these cells. To correct for this, an extra control using poly-L-lysine-preabsorbed antiserum is required (Scopsi et al. 1986; Larsson 1988). Moreover, as also shown in our study, unsuspected crossreactivity phenomena may occur. One of the CGRP antisera used crossreacted with gastrin [1–17], although no overt similarity in primary amino acid sequence exists between these peptides. Interestingly, it has previously been described that antisera recognizing the C-terminus of gastrin and CCK may crossreact with CGRP (Ju et al. 1986; Arias et al. 1989). Moreover, other studies have documented an unexpected crossreactivity between serotonin and substance P antisera (Rawdon and Andrew 1994). Because such crossreactivity phenomena can hardly be foreseen, it is desirable to preabsorb antisera reacting with one type of cell with other molecules known to characterize that cell. This, however, is not always possible because many peptides (e.g., precursor peptides) are either unknown or are available only as cDNA sequences. A way to circumvent this is to use region-specific antisera recognizing different regions of the antigen to be studied (Larsson 1988). In the present study, amylin antisera produced against the [1–11], [11-37], and [25-37] regions of amylin were available. Because of the extended sequence similarity between CGRP and amylin, all three region-specific antisera contained antibody subpopulations that could be shown to crossreact with CGRP (but not with the more remotely related adrenomedullin). Preabsorption against CGRP, however, successfully removed these crossreactive antibodies while leaving the amylin-specific antibodies untouched. In this way we could map different regions of amylin to gastric somatostatin cells and to a smaller subpopulation of gastrin and unidentified endocrine cells.

Figure 4

Rat thyroid immunocytochemically stained with CGRP antiserum 6006 after preabsorption against poly-L-lysine (A) or thyroglobulin (B). Note that preabsorption against thyroglobulin abolishes staining of the thyroid colloid but not of the parafollicular C-cells. Bar = 50 μm.

Figure 5

RT-PCR of amylin, gastrin and β-actin mRNA in the antropyloric stomach of wild-type (WT) and PDX-1-deficient (KO) mice. Note the presence of amplicons corresponding in size to that expected for amylin and β-actin mRNA in both wild-type and PDX-1-deficient mice, whereas gastrin mRNA is detectable only in wild-type mice.

All peptide antisera must be carefully evaluated. Most peptides are of low immunogenicity and must be conjugated to larger carrier proteins before immunization. Ironically, although much effort goes into securing pure synthetic peptides for immunization, these are often conjugated to notoriously “dirty” proteins such as albumin or thyroglobulin. This practice most often leads to the production of contaminating antibodies that react with the conjugating proteins and with potential impurities contained in these. In most cases it is possible to absorb out these antibodies using excess quantities of the conjugating protein. However, this requires knowledge of the conjugating protein used. In the case of antiserum IHC 6006, it was only because of the chance observation that it stained thyroid colloid that we guessed that this thyroid peptide had been conjugated to thyroglobulin. This antiserum provides a “good bad example” of potential errors hidden in immunocytochemistry. Thus, when applied to routinely formaldehyde-fixed and paraffin-embedded specimens, antiserum IHC 6006 detected only the crossreactive (gastrin) antigen and not the correct (CGRP) antigen. Only after staining of formaldehyde-fixed cryostat sections was the correct antigen detected. Because the fixation was the same in both instances, the process of dehydration and paraffin embedding must have selectively masked the correct antigen. One possible explanation is that the crossreactive gastrin was present in higher concentrations than CGRP and was therefore the only antigen detected in tissues that underwent paraffin embedding. The dangers inherent in this type of unwanted staining are obvious.

At first sight, the fact that amylin is expressed both by a majority of somatostatin cells and by a subpopulation of gastrin cells may seem surprising. However, strong data suggesting that these cells evolve from a common precursor cell type have been presented (Larsson et al. 1996; Øster et al. 1998). Different transactivating factors, including PDX-1, may control amylin expression. Our data show that gastric amylin expression persists in mice deficient in PDX-1. Therefore, this homeobox factor is not needed for gastric amylin expression but may work by modulating levels of amylin expression. The role of amylin in antropyloric endocrine cells is unknown. However, many observations have shown that amylin affects gastric emptying and also inhibits gastric acid secretion (for review see Guidobono 1998). Interestingly, the amylin-induced inhibition of gastric acid secretion was recently found not to be dependent on somatostatin (Rossowski et al. 1998). Therefore, gastric D cells may simultaneously produce two inhibitors of gastric acid secretion that work through different receptors.

Footnotes

Acknowledgments

Supported by the Danish Biotechnology Program, the Danish MRC, Danish Cancer Foundation, and the Danish National Research Fund.

References

Ahlgren

Jonsson

Simu

Edlund

(1998) β-cell-specific inactivation of the mouse Ipf/Pdx1 gene results in loss of the β-cell phenotype and maturity onset diabetes. Genes Dev 12:1763–1768.

Arias

Scopsi

Fisher

Larsson

(1989) Light and electron-microscopical localization of calcitonin, calcitonin gene-related peptide, somatostatin and C-terminal gastrin/cholecytokinin immunoreactivities in rat thyroid. Histochemistry 91:265–272.

Asai

Nakazato

Miyazato

Kangawa

Matsuo

Matsukura

(1990) Regional distribution and molecular forms of rat islet amyloid polypeptide. Biochem Biophys Res Commun 169:788–795.

Bretherton-Watt

Gore

Boam

DSW

(1996) Insulin upstream factor 1 and a novel ubiquitous factor bind to the human islet amyloid polypeptide/amylin gene promoter. Biochem J 313:495–502.

Carty

Lillquist

Peshavari

Stein

Soeller

(1997) Identification of cis- and trans-active factors regulating human islet polypeptide gene expression in pancreatic β-cells. J Biol Chem 272:11986–11993.

Cooper

Willis

Clark

Turner

Sim

Reid

(1987) Purification and characterization of a peptide from amyloid-rich pancreases of type 2 diabetic patients. Proc Natl Acad Sci USA 84:8628–8632.

D'Este

Wimalawansa

Renda

(1995) Amylin immunoreactivity is costored in a serotonin cell subpopulation of the vertebrate stomach and duodenum. Arch Histol Cytol 58:537–547.

deVroede

Ferriers

Van de Winkel

Madsen

Pipeleers

(1992) Presence of islet amyloid polypeptide in rat islet B and D cells determines parallelism and dissociation between rat pancreatic islet amyloid polypeptide and insulin content. Biochem Biophys Res Commun 182:886–893.

Ferrier

GJM

Pierson

Jones

Bloom

Girgis

Legon

(1989) Expression of the rat amylin (IAPP/DAP) gene. J Mol Endocrinol 3:R1–4.

10.

Guidobono

(1998) Amylin and gastrointestinal activity. Gen Pharmacol 31:173–177.

11.

Janson

Soeller

Roche

Nelson

Torchia

Kreutter

Butler

(1996) Spontaneous diabetes mellitus in transgenic mice expression human islet amyloid polypeptide. Proc Natl Acad Sci USA 93:7283–7288.

12.

Jonsson

Carlsson

Edlund

(1994) Insulin-promoter-factor 1 is required for pancreas development in mice. Nature 371:606–609.

13.

Hökfelt

Fisher

Frey

Rehfeld

Dockray

(1986) Does cholecystokinin-like immunoreactivity in rat primary sensory neurons represent calcitonin-gene-related peptide?. Neurosci Lett 68:305–310.

14.

Kajimoto

Watada

Matsuoka

Kaneto

Fujitani

Miyazaki

Yamasaki

(1997): Suppression of transcription factor PDX-1/PF-1/STF-1/1DX-1 causes no decrease in insulin mRNA in MING cells. J Clin Invest 100:1840–1846.

15.

Kitamura

Kangawa

Kojima

Ichiki

Matsuo

Eto

(1994) Complete amino acid sequence of porcine adrenomedullin and cloning of cDNA encoding its precursor. FEBS Lett 338:306–310.

16.

Koopmans

van Mansfield

ADM

Jansz

Krans

HMJ

Radder

Frolich

de Boer

Kreutter

Andrews

Maassen

(1991) Amylin-induced in vitro insulin resistance in conscious rats: the liver is more sensitive to amylin than peripheral tissues. Diabetologia 34:218–224.

17.

Larsson

(1981) A novel immunocytochemical model system allowing extensive specificity and sensitivity screening. J Histochem Cytochem 29:408–410.

18.

Larsson

(1988) Immunocytochemistry, Theory and Practice. Boca Raton, FL, CRC Press.

19.

Larsson

Hougaard

(1993) Non-radioactive in situ mRNA hybridization using synthetic oligonucleotides: principles, combination with immunocytochemistry and quantitation. Neurosci Protocols 20:1–18.

20.

Larsson

Madsen

Serup

Jonsson

Edlund

(1996) Pancreatic-duodenal homeobox 1 —role in gastric endocrine patterning. Mech Dev 60:175–184.

21.

Leighton

Cooper

GJS

(1988) Pancreatic amylin and calcitonin gene related peptide cause resistance to insulin in skeletal muscle in vitro. Nature 335:632–635.

22.

Leonard

Peers

Johnsson

Ferreri

Lee

Montminy

(1993) Characterizations of somatostatin transactivating factor that stimulates somatostatin expression in pancreatic islet cells. Mol Endocrinol 7:1275–1283.

23.

Madsen

Nielsen

Petersen

Kofod

Dyrberg

(1990) Expression of islet amyloid polypeptide and insulin in the pancreatic islet cells. In: Shafir

, ed. Frontiers in Diabetes Research: Lessons from Animal Diabetes III. London, Smith-Gordon, 512–515.

24.

Miller

McGehee

Jr Habener

(1994) IDX-1: a new homeodomain transcription factor expressed in rat pancreatic islets and duodenum that transactivates the somatostatin gene. EMBO J 13:1145–1156.

25.

Miyazato

Nakazato

Shiomi

Aburaya

Toshimori

Kangawa

Matsuo

Matsukura

(1991) Identification and characterization of islet amyloid polypeptide in mammalian gastrointestinal tract. Biochem Biophys Res Commun 181:293–300.

26.

Mulder

Ekelund

Ekblad

Sundler

(1997) Islet amyloid polypeptide in the gut and pancreas: localization, ontogeny and gut motility effects. Peptides 18:771–783.

27.

Mulder

Lindh

Ekblad

Westermark

Sundler

(1994) Islet amyloid polypeptide is expressed in endocrine cells of the gastric mucosa in the rat and mouse. Gastroenterology 107:712–719.

28.

Nakazato

Asai

Kangawa

Matsukura

Matsko

(1989) Establishment of radioimmunoassay for human islet amyloid polypeptide and its tissue content and plasma concentration. Biochem Biophys Res Commun 164:394–399.

29.

Ohlsson

Karlsson

Edlund

(1993) IPF1, a homeodomain-containing transactivator of the insulin gene. EMBO J 12:4251–4259.

30.

Ohsawa

Kanatsuka

Yamaguchi

Makino

Yoshida

(1989) Islet amyloid polypeptide inhibits glucose-stimulated insulin secretion from isolated rat pancreatic islets. Biochem Biophys Res Commun 160:961–967.

31.

Ohtsuka

Iwanaga

Fujino

Fujita

(1993) Amylin-containing cells in the gastro-entero-pancreatic (GEP) endocrine system of the rat and humans—an immunohistochemical study. Arch Histochem Cytochem 26:405–414.

32.

Øster

Jensen

Edlund

Larsson

L-I

(1998) Homeobox gene product Nkx 6.1 immunoreactivity in nuclei of endocrine cells of rat and mouse stomach. J Histochem Cytochem 46:717–721.

33.

Permert

Larsson

Westermark

Herrington

Christmanson

Pour

Westermark

Adrian

(1994) Islet amyloid polypeptide in patients with pancreatic cancer and diabetes. N Engl J Med 330:313–318.

34.

Rawdon

Andrew

(1994) Distribution of serotonin immunoreactive gut endocrine cells in chicks at hatching—examination of possible co-localisations with peptides reveals unexpected cross-reactivity of substance P antiserum with serotonin. Histochemistry 102:93–100.

35.

Rossowski

Cheng

Jiang

Coy

(1998) Examination of somatostatin involvement in the inhibitory action of GIP, GLP-1, amylin and adrenomedullin on gastric acid release using a new SRIF antagonist analogue. Br J Pharmacol 125:1081–1087.

36.

Scopsi

Wang

Larsson

(1986) Nonspecific immunocytochemical reactions with certain neurohormonal peptides and basic peptide sequences. J Histochem Cytochem 34:1469–1475.

37.

Serup

Jensen

Andersen

Jørgensen

Blume

Hoist

Madsen

(1996) Induction of insulin and islet amyloid polypeptide production in pancreatic islet glucagonoma cells by insulin promoter factor 1. Proc Natl Acad Sci USA 93:9015–9020.

38.

Stefanini

de Martino

de Zamboni

(1967) Fixation of ejacuated spermatozoa for electron microscopy. Nature 216:173–174.

39.

Stein

Fritsch

Lemaire

Kessel

(1996) Checklist vertebrate homeobox genes. Mech Dev 55:91–108.

40.

Toshimori

Narita

Nakazato

Asai

Mitsukana

Kangawa

Matsuo

Matsukara

(1990) Islet amyloid polypeptide (IAPP) in the gastrointestinal tract and pancreas of man and rat. Cell Tissue Res 262:401–406.

41.

Watada

Kajimoto

Kaneto

Matsuoka

Fujitani

Miyazaki

Yamasaki

(1996) Involvement of the homedomain-containing transcription factor PDX-1 in islet amyloid polypeptide gene transcription. Biochem Biophys Res Commun 229:746–751.

42.

Westermark

Wernstedt

Wilander

Hayden

O'Brien

Johnsson

(1987) Amyloid fibrils in human insulinoma and islets of Langerhans of the diabetic cat are derived from a neuropeptide-like protein also present in normal islet cells. Proc Natl Acad Sci USA 84:3881–3885.

43.

Westermark

Wernstedt

Wilander

Sletten

(1986) A novel peptide in the calcitonin gene related peptide family as an amyloid fibril protein in the endocrine pancreas. Biochem Biophys Res Commun 140:827–831.