Neuroendocrine-Specific and Gastrin-Dependent Expression of a Chromogranin A-Luciferase Fusion Gene in Transgenic Mice

GASTROENTEROLOGY 2001;121:43–55

Neuroendocrine-Specific and Gastrin-Dependent Expression of a Chromogranin A–Luciferase Fusion Gene in Transgenic Mice ¨ CKER,* THORSTEN CRAMER,* DANIEL T. O’CONNOR,‡ STEFAN ROSEWICZ,* MICHAEL HO BERTRAM WIEDENMANN,* and TIMOTHY C. WANG§ *Medizinische Klink mit Schwerpunkt Hepatologie und Gastroenterologie, Universita¨tsklinikum Charite´, Campus Virchow-Klinikum, Humboldt Universita¨t Berlin, Berlin, Germany; ‡Department of Medicine and Center for Molecular Genetics, University of California, San Diego, California; and §Gastrointestinal Unit and Department of Internal Medicine, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts

Background & Aims: Chromogranin A (CgA) is a multifunctional acidic protein specifically expressed in neuroendocrine cells. In the stomach, CgA is found predominantly in enterochromaffin-like (ECL) cells, where it is regulated by gastrin. We investigated the ability of a promoter fragment comprising 4.8 kb of 5ⴕflanking DNA of the mouse CgA (mCgA) gene to direct cell-specific expression as well as gastrin responsiveness in the gastroenteropancreatic neuroendocrine system. Methods: Two independent lines of mCgA 4.8 kb–luc transgenic mice were created. Transgene expression was assessed by determination of luciferase activity and reverse-transcription polymerase chain reaction analysis of luciferase messenger RNA. Cell specificity of transgene expression was investigated by immunohistochemical analysis. The influence of hypergastrinemia on transgene expression was determined after repeated omeprazole injections. Results: In both transgenic lines, mCgA 4.8 kb–luc expression paralleled the expression pattern of the endogenous CgA gene. ECL cells were identified as the major gastric cell population expressing the transgene. Omeprazole treatment stimulated expression of the transgene and the endogenous CgA gene selectivity in the gastric corpus (3– 4-fold). Conclusions: mCgA 5ⴕflanking DNA (4.8 kb) contain the major cis-regulatory element(s) required for cell-specific CgA expression in the neuroendocrine system and gastrin-responsiveness in the gastric corpus. Further analysis of the CgA promoter in transgenic studies may elucidate the general molecular mechanisms underlying cell-specific gene expression in the gastroenteropancreatic neuroendocrine system.

hromogranin A (CgA) is a 48-kilodalton acidic protein specifically expressed in neuroendocrine and neuronal cells. In the gastroenteropancreatic (GEP) system, CgA is expressed by neuroendocrine cells of stom-

C

ach, pancreas, and intestine,1–7 whereas outside the gastrointestinal tract, high levels of CgA can be found in the brain and adrenal glands.8,9 CgA was initially identified as the major soluble protein co-stored and -released with neurotransmitters and secretory peptides of neuroendocrine secretory vesicles, and it is thought to influence granule stability, prohormone processing, and peptide sorting into the regulated secretory pathway.1–3,10 Cleavage products of CgA, such as pancreastatin, are able to influence endocrine and exocrine secretory functions as well as cellular adhesion processes and vascular tension.11–14 In the stomach, enterochromaffin-like (ECL) cells of the corpus mucosa have been identified as the main source for CgA expression.15,16 Additionally, antral G cells have often been shown to express the CgA gene, whereas somatostatin-positive D cells are mostly CgA negative.5,6,17,18 In vivo experiments in rodents and observations in hypergastrinemic patients suffering from gastrinoma or atrophic gastritis show that gastric CgA expression is under control of circulating gastrin levels.15,16,19 –22 Furthermore, infusion of synthetic gastrin, omeprazole treatment, and refeeding of fasted rats have all been shown to increase gastric CgA protein and messenger RNA (mRNA) levels, suggesting that gastrin influences CgA expression in the stomach through transcriptional mechanisms.15,19 Gastrin as a prerequisite for appropriate gastric CgA expression was further high-

Abbreviations used in this paper: bp, base pairs; CCK, cholecystokinin; CgA, chromogranin A; CRE, cyclic adenosine monophosphate– responsive element; dCTP, deoxycytidine triphosphate; ECL, enterochromaffin-like; GEP, gastroenteropancreatic; mCgA, mouse chromogranin A; mRNA, messenger RNA; PBS, phosphate-buffered saline; PCR, polymerase chain reaction; RT, reverse transcription. © 2001 by the American Gastroenterological Association 0016-5085/01/$35.00 doi:10.1053/gast.2001.25526

44

¨ CKER ET AL. HO

lighted by observations made in cholecystokinin (CCK) B/gastrin receptor– or gastrin– deficient mice, in which targeted disruption of either gene abrogated basal and stimulated CgA expression.23–26 In a recent study, we characterized the effect of gastrin on the mouse CgA (mCgA) promoter in vitro and mapped the gastrin-responsive element of the mCgA gene to ⫺92 to ⫺62 base pairs (bp) of the proximal mCgA core promoter.27 We found that the functional integrity of 2 cis-acting elements within this region was indispensable for gastrin responsiveness: an Egr-1/Sp1 motif located at ⫺88 to ⫺77 bp and a cyclic adenosine monophosphate–responsive element (CRE)–like element at ⫺71 to ⫺64 bp. In addition to its essential role for gastrin responsiveness of the mCgA promoter, in vitro studies suggest that the proximal CgA CRE site, which is highly conserved among mammalian CgA genes, is also critical for determination of cell type–specific CgA expression in neuroendocrine cells.28 –30 Nevertheless, additional elements located at more distal positions of the 5⬘-flanking region may contribute to the cell-specific expression pattern of the CgA gene.31,32 Although these in vitro studies added considerably to our understanding of the mechanisms controlling tissue-specific CgA gene expression, the structural elements responsible for the neuroendocrine-specific expression pattern of the CgA gene in vivo have not been analyzed yet. Therefore, we decided to investigate the ability of a 4.8-kb mCgA promoter fragment to confer neuroendocrine-specific expression and gastrin responsiveness in a transgenic mouse model. Two independent mouse lines, transgenic for a fusion gene in which expression of the luciferase reporter gene is under control of 4.8 kb of 5⬘-flanking mCgA DNA, were analyzed for transgene expression. We found that this 5⬘-flanking mCgA sequence is sufficient to provide tissue- and cell-specific expression in the neuroendocrine system as well as gastrin responsiveness in vivo and therefore contains the major cis-regulatory element(s) required for cell-specific CgA expression in the neuroendocrine system and gastrin-dependent regulation in vivo.

Materials and Methods Generation of Transgenic Mice The mCgA 4.8 kb–luc transgene was created by releasing a DNA fragment comprising a fusion gene consisting of nucleotides mCgA ⫺4.800 to ⫹42 bp linked to the firefly luciferase gene from the vector pXP1 by PvUII/BamH1 digestion. The pXP1-based mCgA 4.8 kb–luc reporter construct has previously been used in transient transfection studies.27,28 The mCgA 4.8 kb–luc transgene also comprised a SV40 polyadenylation signal as well as SV40 3⬘-untranslated se-

GASTROENTEROLOGY Vol. 121, No. 1

Figure 1. Schematic diagram showing the structure of the transgene and probes used in this study. The mCgA 4.8 kb–luc transgene contained nucleotides ⫺4.800 to ⫹42 bp of the mouse CgA gene linked to the firefly luciferase gene and was released from the vector PXP1 by PvuI/BamHI digestion. The transgene also comprised a SV40 poly adenylation signal as well as SV40 3⬘-untranslated sequence. The random-labeled probe for tail DNA Southern blot analysis comprised the most 5⬘-located 248 nucleotides of the firefly luciferase gene. Primers used to amplify luciferase sequence in mouse tail DNA were sense 5⬘-CAGAGGACCTATGATTATGTCCGG-3⬘ (corresponding to nucleotides ⫹1227 to ⫹1250 of the firefly luciferase gene) and antisense 5⬘-CGGTACTTCGTCCACAAACACAAC-3⬘ (corresponding to nucleotides ⫹1597 to ⫹1620 of the firefly luciferase gene).

quence (Figure 1). The injected DNA fragment was prepared by restriction digest, followed by sepharose gel separation, CsCl2 density gradient centrifugation, and dialysis against injection buffer as previously described.33 The transgene was microinjected into the male pronucleus of fertilized eggs from FVB inbred (Taconic Farms Inc., Germantown, NY) matings according to standard protocols.33 Progeny were tested at age 3 weeks for the presence of the transgene by Southern blot or polymerase chain reaction (PCR) analysis of restriction enzyme– digested mouse tail DNA. The [32P]-labeled Southern blot probe comprised the most 3⬘-located 248 bp of the firefly luciferase gene. The primers used for PCR analysis of mouse tail DNA were primer 1, sense 5⬘-CAGAGGACCTATGATTATGTCCGG-3⬘ (corresponding to ⫹1227 to ⫹1250 bp of the firefly luciferase gene), and primer 2, antisense 5⬘CGGTACTTC-GTCCACAAACACAAC-3⬘ (corresponding to ⫹1597 to ⫹1620 bp of the firefly luciferase gene). Use of this primer combination resulted in amplification of a PCR product of approximately 400 nucleotides, which corresponds well with the predicted size of 391 nucleotides, based on the reported sequence of the intronless firefly luciferase gene.34 Transgenic mice were generated in the Transgenic Core for the Center for the Study of Inflammatory Bowel Disease at Massachusetts General Hospital.

RT-PCR Analysis of Transgene Expression in Different Tissues Primers used for reverse-transcription polymerase chain reaction (RT-PCR) amplification of luciferase mRNA in various mouse tissues were identical to primers used for mouse tail DNA analysis: primer 1, sense 5⬘-1227-CAGAGGACCTATGATTATGTCCGG-1250-3⬘, and primer 2, antisense 5⬘-1597-CGGTACTTCGT-CCACAAACACAAC-1620-3⬘. Primer 1 was used for first-strand complementary DNA (cDNA) transcription using 1.0 ␮g of total RNA and Mo-MuL reverse transcriptase (RT; Gene Amp RNA PCR Kit; Perkin

July 2001

Elmer Cetus). Resulting cDNAs were used for PCR amplification in a reaction mixture containing 10 pmol of both primers in a 50-␮L volume of 1⫻ PCR buffer (10 mmol/L Tris-HCl, pH 8.3, at 25°C, 50 mmol/L KCl2 , 2 mmol/L MgCl2 , and 0.001% gelatine) containing 250 ␮mol/L of each deoxynucleotide triphosphate and 1 U of AmpliTaq DNA polymerase (Perkin Elmer Cetus). Reactions were subjected to 35 cycles in a Perkin Elmer Cetus thermal cycler at 72°C for 1 minute, 94°C for 30 seconds, and 52°C for 30 seconds, followed by extension at 72°C for 5 minutes. Negative controls underwent identical treatment without addition of RT in the cDNA transcription step. Ten-microliter aliquots of reaction products were electrophoresed in 2.0% agarose gels and visualized by ethidium bromide staining. Using RNA samples obtained from wild-type mice in parallel determinations, no amplification of luciferase transcripts could be detected (data not shown).

Luciferase Measurements in Tissue Extracts Tissue samples excised from transgenic mice were snap-frozen and sonicated in 500 ␮L of lysis buffer containing 25 mmol/L glycylglycine, 15 mmol/L MgSO4 , 1 mmol/L dithiothreitol, and 1% Triton X-100. Lysates were microfuged at 12,000 rpm and supernatants were used for luciferase assays. Luciferase assays of tissue lysates were performed using luciferin, adenosine triphosphate (ATP), and coenzyme A (Promega system) with a monolight Luminometer (Analytic Luminescence Laboratory) as previously described.27,35 For each determination, 20 ␮L of lysate was incubated with 100 ␮L of assay buffer and light emission was measured for 20 seconds. Values for mCgA luc activity were expressed as arbitrary light units and normalized for tissue protein content, as determined by the Bradford protocol (Biorad). In addition, luciferase measurement of tissue samples from nontransgenic mice (C57B6/ SJL F1) served as control. All experiments were approved by the Massachusetts General Hospital Subcommittee on Research Animal Care.

Immunohistochemical Analysis of Transgene Expression The polyclonal rabbit anti-human CgA antibody used in this study is commercially available (Dako, Hamburg, Germany) and has been described before.36 The rabbit antiluciferase antibody was kindly provided by W. Just, Institut fu¨r Biochemie, Universita¨t Heidelberg, Germany.37 The rabbit anti-CgA antibody was used at a 1:200 dilution, whereas the antiluciferase antibody was used at a 1:3000 dilution. Specific antibodies for gastrin and somatostatin were obtained from Dako and used at final dilutions between 1:300 and 1:500.38,39 After incubation of sections with the primary antibody, detection of antigen–antibody complex was performed with the avidin/biotin method according to the manufacturer’s instructions (Vectastain ABC kit; Vector Laboratories, Burlingame, CA). In brief, after deparaffinization with xylol, sections were washed with decreasing concentrations of alcohol and finally

CHROMOGRANIN A–LUCIFERASE TRANSGENE EXPRESSION

45

washed for 5 minutes with ddH2O. After a 5-minute damp chamber treatment, sections were microwaved for 2 minutes, followed by several washes in phosphate-buffered saline (PBS). Thereafter, sections were treated with 0.2% Triton/PBS solution for 10 minutes and subsequently incubated with primary antibodies at appropriate dilutions for 24 hours. After removal of the antibody-containing solution and washing in PBS, detection of primary antibodies was performed with the Vectastain ABC kit. After removal of the secondary antibody complex with ddH2O, counterstaining with Hemalaun solution according to the method of Mayer was performed for 4 minutes at room temperature. Thereafter, slides were washed with tap water and covered with glycerol gelatine. Microphotographs were taken using a camera-equipped Zeiss Axiophot microscope with Zeiss objective. Exposures were obtained with the Ektachrome 64T daylight film (Kodak, Berlin, Germany). In all sets of stainings with a given antibody, a nonrelated antibody targeting a neuroendocrine-specific (anti synaptophysin) as well as a non–neuroendocrine-specific epitope (antiH⫹,K⫹-ATPase) was included. Furthermore, secondary antibodies used were also investigated without addition of a specific primary antibody. Staining patterns that showed similarity to the pattern obtained with the specific antibody were not obtained with these control antibodies in any of our studies.

Omeprazole Treatment of mCgA 4.8 kb–luc Transgenic Mice To determine gastrin-dependent regulation of the mCgA 4.8 kb–luc transgene in vivo, we injected fasted 7– 8week-old mCgA 4.8 kb–luc mice with the proton pump inhibitor omeprazole. Omeprazole was dissolved in 0.25% methylcellulose and 0.9% saline at a concentration of 80 ␮mol/mL (27.8 mg/mL) and animals received intraperitoneal injections of 400 ␮mol/kg (139 mg/kg) body weight or an appropriate concentration of methylcellulose (control) every 12 hours for 2 days. This treatment has previously been shown to result in 3– 4-fold elevation of endogenous gastrin plasma levels in rodents.19,40 Animals were killed 48 hours after the first injection. The gastric antrum and corpus were separated, and samples of each were snap-frozen for luciferase activity and processed for RNA. RNA was prepared from homogenized tissue extracts from using the Trizol method (Gibco BRL, Gaithersburg, MD). RNA blots were performed by standard techniques using 10 ␮g of RNA per lane. Blots were hybridized with a [32P]deoxycytidine triphosphate (dCTP)-labeled mCgA probe. This was a 279-bp murine cDNA probe generated by RT-PCR that has been previously described.23 Blots were then reprobed using a glyceraldehyde3-phosphate dehydrogenase random-primed probe.23 Probes were labeled using the Megaprime labeling kit (Amersham, Arlington Heights, IL). These experiments were approved by the Massachusetts General Hospital Subcommittee on Research Animal Care.

46

¨ CKER ET AL. HO

GASTROENTEROLOGY Vol. 121, No. 1

Results Generation of Transgenic Lines and Tissue Distribution of Transgene Expression Using Southern blot and PCR analysis of tail DNA prepared from offspring, we identified 7 founder animals. Germline transmission of the mCgA 4.8 kb–luc fusion gene to offspring was found for 5 transgenic lines. All transgenic animals developed normally and did not show alterations of the normal phenotype by introduction of the transgene. To determine the tissue distribution pattern of transgene expression, we performed RTPCR analysis of mCgA mRNA using primers amplifying a 400-bp fragment of the firefly luciferase gene. This approach showed that in these 2 positive lines (mCgA 4.8 kb–luc1 and mCgA 4.8 kb–luc2), the transgene was expressed only in tissues known to express the CgA gene endogenously,4 –9 showing the presence of luciferase mRNA in adrenal, brain, stomach, pancreas, and different regions of small and large bowel (Figure 2).To confirm the data obtained by RT-PCR analysis and to determine transgene expression quantitatively, we additionally assessed mCgA 4.8 kb–luc expression by luciferase measurements in tissue lysates prepared from various adult mouse organs. The 2 positive mCgA 4.8 kb–luc transgenic lines showed an expression pattern of the transgene that basically matched the results of RTPCR analysis (Figure 3A and B). High luciferase levels were found in stomach, pancreas, adrenal, and brain, whereas lower levels of luciferase activity were detected in the small and large intestines. In contrast, luciferase expression was basically undetectable in tissues known to be negative for CgA such as heart, liver, and kidney (Figure 3). Although the luciferase expression levels differed between the 2 transgenic lines, the relative differences between luc-positive tissues were similar in animals of both lines (Figure 3A and B). Immunohistochemical Analysis of mCgA 4.8 kb–luc Transgene Expression in Gastrointestinal Tissues To analyze the cell-specific expression of the mCgA 4.8 kb–luc transgene, we performed an immunohistochemical analysis using antibodies specific for luciferase and CgA. In the gastric corpus, CgA and luciferase immunoreactivity was confined to cells located at the basal third of the glands, the typical localization for ECL cells (Figure 4A and B). No luciferase immunoreactivity could be detected in mesenchymal structures or parietal cells, whereas occasionally neuronal structures of the submucosa showed CgA and luciferase immunoreactivity (not shown). Although ECL cells represent the most

Figure 2. Analysis of mCgA 4.8 kb–luc transgene expression by RTPCR. Transgene-specific primers amplifying the sequence amplifying a 391-bp fragment spanning the luciferase sequence ⫹1597 to ⫹1620 bp were used for RT-PCR amplification of luciferase RNA in various mouse tissues. First-strand cDNA was generated by transcription of 1.0 ␮g of total RNA obtained from different tissues of transgenic animals with primer 1. Reaction products were electrophoresed in a 1.5% agarose gel and visualized by ethidium bromide staining. Lane 1 shows PhiX174 size markers. Data shown are representative for results obtained in transgenic lines mCgA 4.8 kb–luc1 and mCgA 4.8 kb–luc2.

abundant neuroendocrine cell type in the corpus mucosa of rats and mice, cells of the A-like type as well as somatostatin-producing D cells and rarely EC cells can be found in the corpus.41 In contrast, in the murine antrum D cells, EC cells, and gastrin-producing cells of the G type are frequently found, whereas ECL cells are essentially absent.41 To characterize the CgA- and luciferase-expressing cells, we stained consecutive sections of the stomach of transgenic animals with specific antibodies directed against CgA, luciferase, somatostatin, and gastrin, respectively. In the corpus, we found numerous cells staining positively for both luciferase and CgA (Figure 5, upper panel). Few cells were positive for somatostatin (Figure 5, lower panel), and most of these were also positive for CgA and luciferase (Figure 5, circled areas). Because gastrin-positive cells were absent in the

July 2001

CHROMOGRANIN A–LUCIFERASE TRANSGENE EXPRESSION

47

cells (Figure 6, upper panel). Although gastrin cells were absent in the corpus, this cell type represented the most abundant neuroendocrine cell population in the antrum (Figure 6, lower panel). Some of the gastrin-positive cells were also positive for luciferase and CgA, and few gastrin-positive cells stained positively for only luciferase or CgA. Taking into account that 4 serial sections were investigated, it is reasonable that a cell present on the first section might be absent on the last section of a series, resulting in negative staining for that particular marker. Compared with gastrin-positive cells, cells positive for somatostatin were less abundant (Figure 6, lower panel). Similar to the corpus, most of the somatostatinpositive cells were also positive for CgA and luciferase (Figure 6, lower panel). In the pancreas, luciferase immunoreactivity was restricted to the islets of Langerhans matching the distribution of CgA immunoreactivity (Figure 7A and B).

Figure 3. Expression of the mCgA 4.8 kb–luc transgene in different tissues of transgenic animals. Tissue samples were obtained from different organs of transgenic animals, snap-frozen, homogenized in luciferase lysis buffer, and analyzed for luciferase activity. Results were expressed as arbitrary light units normalized to tissue protein content. (A ) Data obtained from line mCgA 4.8 kb–luc1; (B) results from line mCgA 4.8 kb–luc2. Data represent means ⫾ SEM obtained in a given line.

corpus (Figure 5, lower panel), it can be concluded that the predominant neuroendocrine cell type in the corpus expressing the mCgA 4.8 kb–luc transgene are ECL cells and to a lesser extent D cells. In the antrum of transgenic mice, we found results similar to those obtained in the corpus with a staining pattern for luciferase that was almost superimposable to the pattern of CgA-positive

Figure 4. Immunohistochemical analysis mCgA 4.8 kb–luc transgene expression in the mouse corpus. Paraffin-embedded sections from mCgA 4.8 kb–luc mice were stained with specific antibodies for (A ) chromogranin A and (B) firefly luciferase (for details, see Materials and Methods). Bound primary antibodies were visualized by avidin/ biotin detection. Results shown are representative for both transgenic lines (original magnification ⫻200).

Figure 5.

Figure 6.

July 2001

Essentially all islet cells were stained with ␣-luciferase and ␣-CgA antibodies, although differences in staining intensity were occasionally observed among different islet cell types (Figure 7A and B). CgA and luciferase immunoreactivity were absent in exocrine and mesenchymal cells of the mouse pancreas. Along the small and large bowel, CgA immunoreactivity and mCgA 4.8 kb– luc immunoreactivity was less abundant than in the stomach but clearly restricted to cells with typical features of neuroendocrine cells: elongated, bottle-shaped cell structure with apical extensions often reaching the intestinal lumen. CgA- and luciferase-positive cells of the intestine were located mostly at the base of the crypts and were less frequent along the villi (Figures 7C–H). Based on morphologic criteria, no mCgA 4.8 kb–luc immunoreactivity could be detected in nonneuroendocrine cells of the gut, such as paneth cells, goblet cells, or enterocytes (Figures 7C–H). By use of the same set of antibodies on tissues that were negative for expression of the mCgA 4.8 kb–luc transgene by RT-PCR analysis or luciferase measurements such as liver, spleen, heart, and skin, no immunoreactivity for CgA or luciferase could be detected (data not shown). Similarly, sections obtained from nontransgenic mice did not show mCgA 4.8 kb–luc immunoreactivity (not shown). Immunohistochemical findings for mCgA 4.8 kb–luc transgene and endogenous CgA expression in non-GEP tissues were similar to the results in the gastrointestinal tract, in the sense that identical cells stained positively for both gene products. In the adrenal, in 1 line (mCgA 4.8 kb–luc1), CgA and luciferase immunoreactivity were detected in chromaffin cells of the adrenal medulla, whereas no staining was found in the adrenal cortex, confirming previous observations.39,40 The other line (mCgA 4.8 kb–luc2) differed from these results in the sense that only the outer portion of the adrenal medulla showed mCgA 4.8 kb–luc immunoreactivity, whereas CgA immunoreactivity could be found in the entire adrenal medulla. In the brain, CgA and luc immunoreactivity were present in neuronal cells of the cortex and basal ganglia (not shown).

CHROMOGRANIN A–LUCIFERASE TRANSGENE EXPRESSION

49

Omeprazole Treatment Stimulates Expression of the mCgA 4.8 kb–luc Transgene and Endogenous CgA RNA Levels in the Gastric Corpus To determine whether the mCgA 4.8-kb promoter fragment was able to confer gastrin responsiveness in vivo, we injected fasted mCgA 4.8 kb–luc transgenic mice by intraperitoneal injection of omeprazole (400 ␮mol/kg body wt) every 12 hours for 2 days. This approach has previously been shown to result in 3– 4-fold elevation of serum gastrin concentrations in rodents.19,27 Transgene activity was assessed by measuring luciferase activity in various gastric and nongastric mouse tissues. We found that in the corpus, omeprazole treatment selectively elevated transgene expression approximately 4-fold compared with control animals, whereas no changes in transgene expression were observed in the antrum (Figure 8A). Similarly, omeprazole treatment had no influence on transgene expression in extragastric tissues such as brain, adrenal, or pancreas (data not shown). In parallel to luciferase measurements, expression of endogenous CgA RNA levels in antrum and corpus were determined by Northern blot analysis. Without omeprazole treatment, CgA RNA levels in the corpus were clearly higher than in the antrum (Figure 8B and C). This result is in accordance with those of previous studies in rodents19,27 and also fits well with the differences in mCgA 4.8 kb–luc transgene expression between antrum and corpus as determined by luciferase measurements (Figure 8B and C). Similar to the selective increase in transgene activity in the corpus after omeprazole treatment, injection of the proton pump inhibitor elevated endogenous CgA RNA levels in the corpus, but no changes in CgA RNA abundance could be observed in the antrum (Figure 8B and C).

Discussion In the present study, we demonstrate the ability of 4.8 kb of 5⬘-flanking DNA of the mCgA gene to direct tissue- and cell-specific gene expression to the

Š Figure 5. Immunohistochemical analysis of mCgA 4.8 kb–luc transgene expression in the mouse corpus. Paraffin-embedded consecutive sections from mCgA 4.8 kb–luc mice were stained with specific antibodies for CgA, firefly luciferase, somatostatin, or gastrin (for details, see Materials and Methods). Bound primary antibodies were visualized by avidin/biotin detection. Results shown are representative for both transgenic lines. Arrows indicate CgA-, luciferase (LUC)-, gastrin (GAS)-, or somatostatin (SMS)-positive cells. Circles indicate somatostatinpositive cells that also stain positively for luciferase and CgA (original magnification ⫻400). Figure 6. Immunohistochemical analysis of mCgA 4.8 kb–luc transgene expression in the mouse antrum. Paraffin-embedded consecutive sections from mCgA 4.8 kb–luc mice were stained for CgA, firefly luciferase, somatostatin, or gastrin using the same set of antibodies used for the stainings shown in Figure 4 (for details, see Materials and Methods). Bound primary antibodies were visualized by avidin/biotin detection. Results shown are representative for both transgenic lines. Arrows indicate CgA-, luciferase (LUC)-, gastrin (GAS)-, or somatostatin (SMS)-positive cells. Results shown are representative for both transgenic lines (original magnification ⫻630).

Figure 7. Immunohistochemical analysis of mCgA 4.8 kb–luc transgene expression in various mouse tissues. Paraffin-embedded sections from the pancreas (upper panel), small intestine (middle panel), and colon (bottom panel) of mCgA 4.8 kb–luc mice were stained for the expression of either (A, D, G) CgA or (B, C, E, F, H) firefly luciferase. (A, B, D, E, G, H) Stainings of consecutive sections; (C, F) representative stainings of the (C ) small bowel and (F ) colon. The antibodies used were a polyclonal rat ␣-chromogranin A and a monoclonal rabbit ␣-luciferase antibody. Results shown are representative for both transgenic lines; (A, B, D, E ), original magnification ⫻430; (C, E), original magnification ⫻200.

July 2001

CHROMOGRANIN A–LUCIFERASE TRANSGENE EXPRESSION

51

Š Figure 8. Omeprazole treatment selectively stimulates mCgA 4.8 kb–luc transgene activity and CgA RNA expression in the gastric corpus. Fasted animals were treated with repeated intraperitoneal injections of either methylcellulose or omeprazole/methylcellulose suspension (400 ␮g/kg body wt) every 12 hours for 2 days. In previous studies, this treatment has been shown to elevate endogenous plasma levels in rodents 3– 4-fold. Tissue samples from the corpus and antrum were taken 48 hours after the first injection and analyzed for (A ) transgene expression and (B) CgA RNA abundance by Northern blot analysis. Luciferase activity was expressed as fold increase of results measured in tissues obtained from vehicle-treated control animals. Luciferase data shown represent the mean ⫾ SEM obtained from 3 different animals of one transgenic line. CgA Northern blots were performed using a 279-bp [32P]deoxycytidine triphosphate– labeled murine cDNA probe as outlined in Materials and Methods. Data shown represent hybridization signals from 3 different animals that received either methylcellulose (animals 1–3) alone (vehicle) or methylcellulose– omeprazole (animals 4 – 6) suspension. To correct for loading differences between individual lanes, blots were stripped and reprobed with a glyceraldehyde-3-phosphate dehydrogenase–specific cDNA probe (data not shown). (C ) Summary of densitometry results of the data shown in B; densitometry results obtained for the CgA Northern blot were corrected for differences in RNA loading on the basis of densitometry of the glyceraldehyde-3-phosphate dehydrogenase Northern blot. Subsequently, results in each group were pooled and analyzed statistically.

neuroendocrine system of transgenic mice. We found that in mice harboring the mCgA 4.8 kb–luc construct, transgene expression closely resembled the expression pattern of the endogenous CgA gene, with selective expression in cells of the neuroendocrine system. In all tissues investigated, no ectopic transgene expression in nonneuroendocrine epithelial or mesenchymal cells could be detected. Therefore, the 4.8-kb fragment of 5⬘-flanking mCgA DNA appears to contain all the necessary genetic information for neuroendocrine-specific expression and gastrin responsiveness of the CgA gene. In previous studies, endogenous CgA expression has been shown in both gastrointestinal and nongastrointes-

tinal tissues. Within the GEP neuroendocrine system, which comprises at least 15 different cell types, CgA is widely expressed.4 –7 In the stomach, the major cell type expressing CgA is ECL cells of the acid-producing mucosa, in which the release and production of CgA are controlled by circulating gastrin levels.19 –22 Whereas along the small and large intestines, CgA expression is confined to the different types of peptide-producing neuroendocrine cells scattered throughout the mucosa, pancreatic CgA expression is restricted to the islets of Langerhans.4 –7,42,43 In addition to its presence in intestinal neuroendocrine epithelial cells, CgA expression has also been detected in neuronal elements of the gut.44 Outside the gastrointestinal tract, hormone-producing glands such as pituitary and adrenal glands as well as neuronal structures of the brain have been identified as tissues with high CgA levels.8,9 Using a combination of RT-PCR determinations, luciferase measurements, and immunohistochemistry, the current study showed that the tissue- and cell-specific expression pattern of the mCgA 4.8 kb–luc transgene closely matched the expression pattern of the endogenous CgA gene. In the stomach, high levels of luciferase activity were found in the gastric corpus, and lower levels were present in the gastric antrum. Immunohistochemically, expression of CgA and luciferase in the oxyntic mucosa was restricted to cells located at the basal third of the glands, which showed typical morphologic features of ECL cells. In mammals, the gastric neuroendocrine

52

¨ CKER ET AL. HO

cell population comprises ECL cells, G cells, D cells, and cells of the A-like type.41 In the mouse corpus, ECL cells represent the most abundant neuroendocrine cell type, and somatostatin-producing D cells and neuroendocrine cells of the A-like type are less frequently found.41 G cells, whose primary secretory product is gastrin, are usually absent in the corpus, whereas they are numerous in the antrum.41 5-Hydroxytryptamine–producing EC cells are rarely found in the stomachs of mice, in contrast to those of other rodents or humans, wherein this cell type is frequently found.41 Because G cells were absent in the corpus and D cells were identified as a much less abundant cell population, our data show that the predominant neuroendocrine cell type of the corpus mucosa expressing the mCgA 4.8 kb–luc transgene must be ECL cells. Similar to the corpus, the antral expression pattern of the transgene was superimposable with the endogenous expression of the CgA gene, although the number of CgA/luciferase-positive cells was clearly smaller than in the corpus (Figure 6). Using staining of serial sections, coexpression of CgA and the mCgA 4.8 kb–luc transgene was localized to antral G cells and also D cells (Figure 6). Our studies show that in contrast to observations in the human antrum, where D cells mostly do not stain positively for CgA, in the mouse antrum most of the D cells showed positive CgA immunoreactivity and also expressed the mCgA 4.8 kb–luc transgene. Because remarkable differences exist between different species regarding expression levels of neuroendocrine peptides in a particular neuroendocrine cell type, these observations can most likely be attributed to differences in the CgA expression pattern between mouse and human neuroendocrine systems. Overall, these data are in accordance with previous studies showing that ECL cells are the predominant neuroendocrine cell population expressing the CgA gene in the corpus of various species, whereas in the antrum G cells and also D cells represent cellular sources for CgA.4 – 6,17,41 In the pancreas, transgene expression was restricted to islet cells, and virtually all islet cells showed positive immunoreactivity for both CgA and luciferase. These data are in accordance with previous studies showing CgA expression in ␤, ␣, pancreatic polypeptide, and D cells of the endocrine pancreas.5 There was no evidence of mCgA 4.8 kb–luc expression in the exocrine part of the gland (i.e., pancreatic acinar cells) or in mesenchymal structures. Expression of the transgene was also detected in the small and large bowel, where it was restricted to CgApositive neuroendocrine cells. The intestinal mucosa is known to harbor at least 15 different neuroendocrine cell

GASTROENTEROLOGY Vol. 121, No. 1

types,5 which have been shown to express multiple peptides, including neurotensin, gastric inhibitory polypeptide, CCK, secretin, substance P, gastrin, peptide tyrosine tyrosine, and enteroglucagon, as well as possibly unidentified peptide hormones.4 –7,17,18 Because expression of the transgene in the intestine was restricted to CgA-positive cells, it can reasonably be concluded that the aforementioned neuroendocrine cell types were likely targeted by the mCgA 4.8 kb–luc construct. Outside the GEP neuroendocrine system, CgA is known to be present in high amounts in the adrenal medulla and in the brain, where it has been found in neuronal structures of the cortex and basal ganglia.8,9 Similarly, we found luciferase immunoreactivity in the corresponding areas of the mouse brain in both transgenic lines, and in the adrenal glands in one line. In the other line, transgene expression was seen only in a distinct population of medullary cells located at the outer margin of the medulla, whereas on serial sections CgA immunoreactivity was detected in the entire medulla. Integration-specific effects could account for these discrepancies between transgenic lines. In gastric ECL cells, CgA secretion and gene expression have been found to be under tight control of circulating gastrin levels.19 –22 In addition, selective activation of CgA gene expression in the corpus mucosa in response to gastrin has been previously described by Dimaline et al.19 These investigators found that refeeding of fasted rats resulted in parallel elevation of plasma gastrin levels and CgA mRNA expression in the gastric corpus and that this feeding effect could be abolished by application of a specific CCK-B/gastrin receptor antagonist or immunoneutralization using a specific antigastrin antibody.19 However, these studies did not show transcriptional regulation of the CgA gene by gastrin. In a recent study, we showed that the mCgA promoter could be regulated by gastrin in a gastric cell line in vitro through a proximal promoter sequence mCgA ⫺92 to ⫺62 bp.27 Furthermore, this study showed that the Egr-1/Sp1 motif located at ⫺88 to ⫺77 bp and the CRE-like element at ⫺71 to ⫺64 bp were both necessary for gastrindependent mCgA promoter activation.27 To extend these earlier in vitro studies, we have now shown that the 4.8-kb mCgA promoter fragment is sufficient to confer transcriptional responses to gastrin in vivo. Omeprazolemediated hypergastrinemia in our mCgA 4.8 kb–luc transgenic mice resulted in 4-fold increases in promoter activity in the gastric corpus, whereas no changes in promoter activity could be detected in the gastric antrum or nongastric tissues (Figure 8A). Similarly, analysis of gastric CgA RNA expression in mCgA 4.8 kb–luc trans-

July 2001

genic mice showed that expression of the endogenous CgA gene was very similar to the expression of the mCgA 4.8 kb–luc transgene. Under unstimulated conditions, CgA RNA expression in the corpus was clearly more pronounced compared to the antrum, while omeprazole treatment selectively increased CgA RNA abundance in the corpus without having an effect on antral CgA gene expression (Figure 8B). These data show that basal and omeprazole-stimulated expression of the mCgA 4.8 kb–luc transgene parallels the expression pattern of the endogenous CgA gene, indicating that the mCgA 4.8-kb promoter fragment comprises the genetic information required for appropriate expression and gastrindependent regulation of the CgA gene in the mouse stomach. Previous studies in rodents have also found that omeprazole-induced hypergastrinemia has no effect on CgA mRNA abundance in the antrum.19 A possible explanation for this phenomenon has been provided by autoradiographic studies analyzing gastric CCK-B/gastrin receptor expression in humans and dogs.45,46 Although high CCK-B/gastrin receptor abundance was found in the ECL cell compartment of the corpus, antral expression of the receptor was much less pronounced and confined mostly to neuronal or muscular structures.45,46 Therefore, gastrin-dependent activation of the mCgA 4.8 kb–luc transgene in the corpus of transgenic mice is most likely determined by the cell-specific expression pattern of the CCK-B/gastrin receptor. Several transgenic models in which intestinal and/or pancreatic neuroendocrine cells were targeted have been described previously. Islet cell expression has been achieved using a variety of cell-specific promoters, including those from the rat glucagon, insulin, secretin, or peptide tyrosine tyrosine genes.47 Although most of these transgenes target pancreatic ␤ and pancreatic polypeptide cells, rat insulin promoter– controlled transgenes were also expressed in intestinal secretin-producing cells.47 Similarly, a 1.6-kb rat secretin promoter fragment was able to direct expression to both fetal islet cells and intestinal neuroendocrine cells.48 Intestinal neuroendocrine expression has also been reported for transgenes using different portions of the 5⬘-flanking region of the gene encoding rat liver fatty acid– binding protein, a gene that is normally expressed in enterocytes, hepatocytes, and a subpopulation of serotonin-positive enteroendocrine cells.49,50 However, the transgenic constructs used in these previous studies were essentially not expressed in the stomach and targeted only subpopulations of enteropancreatic neuroendocrine cells. In contrast, the mCgA 4.8 kb–luc transgene analyzed in our current study displays a broad cell-specific expression

CHROMOGRANIN A–LUCIFERASE TRANSGENE EXPRESSION

53

pattern in the neuroendocrine system and allows simultaneous targeting of most GEP neuroendocrine cells. In addition, the mCgA 4.8 kb–luc construct represents the first transgene that is able to target gastric ECL cells in vivo. Future transgenes based on the mCgA 4.8-kb promoter fragment could be used to target gene expression to ECL cells in the mouse and begin to address the biology and pathobiology of this neuroendocrine cell lineage. In addition, site-specific mutants derived from the 4.8-kb mCgA promoter can be used in future transgenic studies to identify the underlying cis-acting element(s) responsible for tissue-specific expression and may be useful in comparative investigations of transcriptional mechanisms in neuroendocrine cells of the GEP system. In conclusion, our study shows that in a transgenic mouse model, the 4.8-kb mouse CgA promoter specifically targets gastrointestinal and extragastrointestinal neuroendocrine cells in vivo. In addition, the expression level of the transgene in the gastric corpus can be selectively induced by omeprazole treatment, confirming in vivo the transcriptional regulation of the mCgA gene by gastrin. Therefore, the mouse 4.8-kb chromogranin A promoter comprises the genetic information required for tissue- and cell-specific expression of the CgA gene in vivo and gastrin responsiveness in the stomach. This CgA promoter fragment represents a powerful tool for further investigations of the biology and pathobiology of neuroendocrine cells, both inside and outside of the gastrointestinal tract.

References 1. Deftos LJ. Chromogranin A: its role in endocrine function and as an endocrine and neuroendocrine tumor marker. Endocrine Rev 1991;12:181–187. 2. Bauerfeind R, Ohashi M, Hutttner WB. Biogenesis of secretory granules and synaptic vesicles. Ann N Y Acad Sci 1994;733: 385–390. 3. Iacangelo A, Eiden L. Chromogranin A: current status as a precursor for bioactive peptides and a granulogenic/sorting factor in the regulated secretory pathway. Regul Pept 1995;58:65– 88. 4. Portela-Gomes GM, Stridsberg M, Johansson H, Grimelius L. Complex co-localization of chromogranins and neurohormones in the human gastrointestinal tract. J Histochem Cytochem 1997; 45:815– 822. 5. Wiedenmann B, Huttner W. Synaptophysin and chromogranins/ secretogranins—widespread constituents of distinct types of neuroendocrine vesicles and new tools in tumor diagnostic. Virchows Archiv B Cell Pathol 1989;58:95–121. 6. Wiedenmann, B, Waldherr R, Buhr H, Hille A, Rosa P, Huttner WB. Identification of neuroendocrine cells in normal and neoplastic human tissue with antibodies against synaptophysin, chromogranin A, secretogranin I (chromogranin B), and secretogranin II. Gastroenterology 1988;95:1364 –1374. 7. Facer P, Bishop AE, Lloyd RV, Wilson BS, Hennesy RJ, Polak JM. Chromogranin: a newly recognized marker for endocrine cells of the human gastrointestinal tract. Gastroenterology 1985;89: 1366 –1373. 8. Weiler R, Marksteiner G, Wohlfarter T, Schober M, Fischer-Colbrie

54

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

¨ CKER ET AL. HO

R, Sperk G, Winkler H. Chromogranins in rat brain: characterization, topographical distribution and regulation of synthesis. Brain Res 1990;532:87–94. Scha¨fer MK-H, Nohr D, Romeo H, Eiden LE, Weihe E. Pan-neuronal expression of chromogranin A in rat nervous system. Peptides 1994;15:263–279. Yoo SH. pH-and Ca2⫹-dependent aggregation property of secretory vesicle matrix proteins and the potential role of chromogranins A and B in secretory vesicle biogenesis. J Biol Chem 1997; 271:1558 –1565. Tatemoto K, Efendic S, Mutt V, Makk G, Feistner GJ, Barchas JD. Pancreastatin, a novel pancreatic peptide that inhibits insulin secretion. Nature 1986;324:476 – 478. Herzig K-H, Louie DS, Tatemoto K, Owyang C. Pancreastatin inhibits pancreatic enzyme secretion by presynaptic modulation of acetylcholine release. Am J Physiol 1992;G113–G117. Angeletti RH, Aardal S, Serck-Hanssen G, Gee P, Helle KB. Vasoinhibitory activity of synthetic peptides from the amino terminus of chromogranin A. Acta Physiol Scand 1994;152:11–19. Gasparri A, Sidoli A, Perez Sanchez L, Longhi R, Siccardi AG, Marchisto PC, Corti A. Chromogranin A fragments modulate cell adhesion. J Biol Chem 1997; 272:20835–20843. Dockray GJ, Varro A, Dimaline R. Gastric endocrine cells: gene expression, processing and targeting of active products. Physiol Rev 1996;76:767–798. Hakanson R, Xi-Quin D, Norle´n P, Chen D. Circulating pancreastatin is a marker for enterochromaffin-like cells of the rat stomach. Gastroenterology 1995;108:1445–1452. Cetin Y, Grube D. Immunoreactivities for chromogranin A and B, and secretogranin II in the guinea pig entero-endocrine system: cellular distributions and intracellular heterogeneities. Cell Tissue Res 1991;264:231–241. Cetin Y, Grube D. Immunoreactivities for chromogranin A and B, and secretogranin II in the guinea pig endocrine pancreas. Histochemistry 1990;264:231–241. Dimaline R, Evans D, Forster ER, Sandvik AK, Dockray GJ. Control of gastric corpus chromogranin A messenger RNA abundance in the rat. Am J Physiol 1993;264:G583–G588. Watkinson A, Dockray GJ. Functional control of chromogranin A and B concentrations in the body of rat stomach. Regul Pept 1992;40:51– 61. Syversen U, Mignon M, Bonfils S, Kristensen A, Waldum HL. Chromogranin A and pancreastatin-like immunoreactivity in serum of gastrinoma patients. Acta Oncol 1993;32:161–165. Borch K, Strindsberg M, Burman P, Rehfeld JF. Basal chromogranin A and gastrin concentrations in circulation correlate to endocrine proliferation in type A gastritis. Scand J Gastroenterol 1997;32:198 –202. Koh TJ, Goldenring JR, Ito S, Mashimo H, Kopin AS, Varro A, Dockray GJ, Wang TC. Gastrin deficiency results in altered gastric differentiation and decreased colonic proliferation. Gastroenterology 1997;113:1015–1025. Langhans N, Rindi G, Chiu M, Rehfeld JF, Ardman B, Beinborn M, Kopin AS. Abnormal gastric histology and decreased acid production in cholecystokinin-B/gastrin receptor-deficient mice. Gastroenterology 1997;90:391–399. Friis-Hansen L, Sundler F, Li Y, Gillespie PJ, Saunders TL, Greenson JK, Owyang C, Rehfeld JF, Samuelson LC. Impaired gastric acid secretion in gastrin-deficient mice. Am J Physiol 1998;274: G561–G568. Nagata A, Ito M, Iwata M, Kuno J, Takano H, Minova O, Chihara K, Matsui T, Noda T. G protein-coupled cholecystokinin-B/gastrin receptors are responsible for physiological cell growth of the stomach mucosa in vivo. Proc Natl Acad Sci U S A 1997;93: 11825–11830. Ho¨cker M, Raychowhury R, Plath T, Wu H, O’Connor DT, Wiedenmann B, Rosewicz S, Wang TC. Sp1 and CREB mediate gastrin-

GASTROENTEROLOGY Vol. 121, No. 1

28.

29.

30. 31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

dependent regulation of chromogranin A promoter activity in gastric carcinoma cells. J Biol Chem 1998,273:34000 –34007. Wu H, Mahata SK, Rozansky DJ, Webster NJG, Parmer RJ, O’Connor DT. A functional cyclic AMP response element plays a crucial role in neuroendocrine cell-type specific expression of the secretory granule protein chromogranin A. J Clin Invest 1995;96:568 –578. Wu H, Rozansky DJ, Webster NJG, Parmer RJ, O’Connor DT. Cell-type specific expression in the neuro-endocrine system: a neuroendocrine-specific regulatory element in the promoter of chromogranin A, a ubiquitous secretory granule protein. J Clin Invest 1994;94:118 –129. Mouland AJ, Bevan S, White JH, Hendy GN. Human chromogranin A gene. J Biol Chem 1994;269:6918 – 6926. Nolan EM, Cheung TC, Burton DW, Deftos LJ. Transcriptional regulation of the human chromogranin A gene by ist 5⬘ distal regulatory element: novel effects of orientation, structure, flanking sequences, and position on expression. Mol Cell Endocrinol 1996;124:51– 62. Canaff L, Bevan S, Wheeler, DG, Mouland AJ, Rehfuss RP, White JH, Hendy GN. Analysis of molecular mechanisms controlling neuroendocrine cell specific transcription of the chromogranin A gene. Endocrinology 1998;139:1184 –1196. Wang TC, Bonner-Weir S, Oates PS, Chulak M, Simon B, Merlino GT, Schmidt EV, Brand SJ. Pancreatic gastrin stimulates islet differentiation of transforming growth factor ␤–induced ductular precursor cells. J Clin Invest 1993;92:1349 –1356. De Wet JR, Wood KV, DeLuca M, Helinski DR, Subramani S. Firefly luciferase gene: structure and expression in mammalian cells. Mol Cell Biol 1987;2:725–737. Ho¨cker M, Zhang Z, Henihan R, Rosewicz S, Riecken E-O, Koh TJ, Wang TC. Gastrin and PMA regulate the human histidine decarboxylase promoter through Raf-dependent activation of ERK-related signaling pathways in gastric cancer cells. J Biol Chem 1997;272:27015–27024. Wiedenmann B, Franke WW. Identification and localization of synaptophysin, an integral membrane glycoprotein of Mr 38,000 characteristic of presynaptic vesicles. Cell 1985;41: 1017–1028. Soto U, Pepperkok R, Ansorge W, Just WW. Import of firefly luciferase into mammalian peroxisomes in vivo requires nucleoside triphosphates. Exp Cell Res 1993;205:66 –75. Rehfeld JF, Larsson L-I. Pituitary gastrins. Different processing in corticotrophs and melanotrophs. J Biol Chem 1981;256:10426 – 10429. Dubois MP. Immunoreactive somatostatin is present in discrete cells of the endocrine pancreas. Proc Natl Acad Sci U S A 1975;72:1340 –1345. Ho¨cker M, Zhang Z, Fenstermacher DA, Tagerud S, Chulak M, Joseph D, Wang TC. The rat histidine decarboxylase promoter is regulated by gastrin through a protein kinase C pathway. Am J Physiol 1996;270:G619 –G633. Sundler F, Hakanson R. Gastric endocrine cell typing at the light microscopic level. In: Sundler F, Hakanson R, eds. The stomach as an endocrine organ, Volume 15. Amsterdam: Elsevier, 1991:9 –26. O’Connor DT, Burton D, Leonard MS, Deftos L J. Chromogranin A: immunohistology reveals 1st universal occurrence in normal polypeptide-producing endocrine glands. Life Sci 1983;33: 1657–1663. Iacangelo A, Okayama H, Eiden LE. Primary structure of rat chromogranin A and distribution of its mRNA. FEBS Lett 1988; 227:115–121. Shen Z, Ekelund M, Sundler F. Immunohistochemical evidence for chromogranin A and B in neuronal elements in human gut. Regul Pept 1994;51:229 –236. Manthy CR, Pappas TN, Vigna SR. Localization of cholecystokin A and cholecytokinin B/gastrin receptors in the canine upper gastrointestinal tract. Gastroenterology 1994;107:1019 –1030.

July 2001

46. Reubi JC, Wasser B, Laderach U, Stettler C, Friess H, Halter F, Schmassmann A. Localization of cholecystokin A and cholecytokinin B-gastrin receptors in the human stomach. Gastroenterology 1997;112:1197–1205. 47. Rindi G. Characterization of neuroendocrine tumors in transgenic mice. Digestion 1994;55(suppl 3):24 –30. 48. Lopez MJ, Upchurch BH, Rindi G, Leiter AB. Studies in transgenic mice reveal potential relationships between secretin-producing cells and other endocrine cell types. J Biol Chem 1995;270:885– 891. 49. Roth KA, Kim S, Gordon JI. Immunocytochemical studies suggest two pathways for enteroendocrine cell differentiation in the colon. Am J Physiol 1992;263:G174 –G180. 50. Roth KA, Hertz JM, Gordon JI. Mapping enteroendocrine populations in transgenic mice reveals an unexpected degree of complexity in cellular differentiation within the gastrointestinal tract. J Cell Biol 1990;110:1791–1801.

CHROMOGRANIN A–LUCIFERASE TRANSGENE EXPRESSION

55

Received June 10, 2000. Accepted April 13, 2001. Address requests for reprints to: Timothy C. Wang, M.D., Gastroenterology Division, University of Massachusetts Medical Center, Biotech Two Suite 210, 373 Plantation Street, Worchester, Massachusetts 016052377. e-mail: [email protected]; fax: (508) 856-4770. Supported by a grant from the National Institutes of Health (2RO I DK 48077) (to T.C.W.); by a grant of the Deutsche Forschungsgemeinschaft (DFG), Bonn (Ho 1288/6-1) (to M.H.); and by grants from the DFG and the Mildred Scheel Stiftung (to B.W. and S.R.). The authors acknowledge the excellent technical assistance by Hucheng Bei, Ines Eichhorn, and Michael Niesar. The ␣-luciferase antibody was generously provided by W. Just, Institut fu ¨r Biochemie, Universita ¨t Wu ¨rzburg, Germany. Transgenic mice were generated at the Transgenic Core at the Center for the Study of Inflammatory Bowel Diseases at Massachusetts General Hospital.