469 Culture of human insulinoma cells: development of a neuroendocrine tumor cell- and human pancreatic islet cell-specific monoclonal antibody

L Wagner, E Templ, G Reining, W Base, M Weissel, P Nowotny, K Kaserer1 and W Waldha¨usl Department of Medicine III, and 1Department of Clinical Pathology, University of Vienna, Vienna, Austria (Requests for offprints should be addressed to L Wagner, Department of Medicine III, Division of Clinical Endocrinology and Metabolism, University of Vienna, Wa¨hringer Gu¨ rtel 18-20, A-1090 Vienna, Austria)

Abstract We report on the culture of human insulinoma cells neuroendocrine tumor cells. D24 mAb also reacted derived from a 32-year-old male patient with hyper- immunohistochemically with normal pancreatic â-cells insulinism due to an insulinoma of the pancreas. A and tumors such as vipoma, gastrinoma and carcinoid. single-cell suspension was made by passing insulinoma Insulinoma cell clusters separated from fibroblasts by fragments through a fine-gauge stainless-steel mesh. micromanipulation and plated into multiwell culture Cluster-forming insulinoma cells resembling pancreatic dishes exhibited an insulin-secretion rate of 30 U/100 islets grew in the presence of fibroblasts. The insulinoma cells per 24 h with no insulin-secretory response to cell clusters could be differentiated from fibroblasts by elevated glucose concentration. Purified insulinoma cells using in situ pan optic staining and specific immunocyto- incubated with 1 ng/ml human nerve growth factor chemical staining (anti-human insulin and anti-human expressed neurofilament and neurite extension. These insulinoma monoclonal antibody (mAb) D24). mAb D24 findings together with earlier observations in animal was generated using insulinoma cells as antigen for immu- models suggest that human pancreatic â-cells share nization of a Balb/C mouse and cell fusion by the some properties with neurons and are related to other hybridoma cell technique. The anti-insulinoma cell mAb neuroendocrine cells in the gastrointestinal tract. recognized a 32 kDa protein on immunoblot analysis of Journal of Endocrinology (1998) 156, 469–476

Introduction In this study we describe an attempt to grow insulinoma cells derived from a benign insulinoma from a young adult. A number of â-cell lines have been generated from The cells were tested for insulin secretion at glucose insulinomas and hyperplastic islets arising in mice that concentrations of 5 and 25 mM. Using the insulinoma cells express a transgene encoding the SV-40Tag oncogene we were able to develop a monoclonal antibody (mAb) under the control of the insulin promoter (Efrat et al. 1988, that reacted with pancreatic â-cells, insulinoma cells 1993, D’Ambra et al. 1990, Miyazaki et al. 1990, and also other neuroendocrine tumors such as vipoma, Hamaguchi et al. 1991, Radvanyi et al. 1993). Two of gastrinoma and carcinoid. We also present evidence these insulinoma cell lines that maintain correct glucose that isolated insulinoma cells express neurofilament and responsiveness have been cloned (Efrat et al. 1993, Knaack undergo neurite extension on incubation with human et al. 1994). Metabolic studies (Fehmann et al. 1994, nerve growth factor (hNGF). Ferber et al. 1994) have been performed using animal- derived insulinoma cell lines. Because of the lack of human Materials and Methods insulinoma cell lines, rat and mouse cell lines have been used as antigens for detection of pancreatic â-cell-specific autoantibodies from humans with insulin-dependent Insulinoma tissue diabetes mellitus (Karounos & Thomas 1990, Thomas et al. A 32-year-old male patient presented with hyper- 1990). No human insulinoma cell line is available. For insulinism resulting from an insulinoma of 17 mm establishment of cDNA libraries, isolated human islets diameter. This was surgically removed from the lateral part of Langerhans have been used (Permutt et al. 1989, Rabin of the pancreas corpus. Individual clusters of tumor cells et al. 1992). were surrounded by connective tissue. The tumor did not

Journal of Endocrinology (1998) 156, 469–476  1998 Journal of Endocrinology Ltd Printed in Great Britain 0022–0795/98/0156–0469 $08.00/0

Downloaded from Bioscientifica.com at 10/01/2021 05:44:03AM via free access 470 L WAGNER and others · Pancreatic islet cell- and neuroendocrine tumor cell-specific monoclonal antibody

show any signs of malignancy. Parts of the tumor appeared microscopic observation in order to avoid unwanted fibro- to extend into the normal pancreatic parenchyma, but blast transfer. Clusters were plated into 24-multiwell the pancreatic margin of the resection was free of tumor. dishes for measurement of insulin release. Cells adhered Tumor cells were absent from the conjointly removed within 24 h. lymph node. Immunohistochemical analysis was per- formed on tissue cryostat sections and revealed positive staining for neuron-specific enolase, chromogranin A and Phenotype analysis insulin. Negative results were obtained for glucagon, pancreatic peptide, gastrin, serotonin, somatostatin and Immunofluorescence After removal of cell supernatant, á-subunit of human choriogonadotropin. adherent cells were washed twice with PBS. The air-dried cell layer was fixed with ice-cold methanol for 10 min. Fixed cells were wetted with PBS and then stained in situ Insulinoma cell culture with fluorescein isothiocyanate (FITC)-conjugated sheep A part of the tumor was minced and a single-cell anti-human insulin serum (ICN Immunobiologicals, Lisle, suspension was made by passing insulinoma fragments Illinois, USA). After staining, dishes were washed with through a fine-gauge stainless-steel mesh. Cells were PBS (3#5 min) and a drop of 1% n-propyl gallate (Sigma, washed in RPMI medium supplemented with 10% fetal St Louis, MO, USA) mounting fluid (8:2 mixture of calf serum. A 3 ml sample of insulinoma cell suspension glycerol and Tris–HCl, pH 8·4) was used to reduce (1#106 cells/ml) was seeded on to Costar Petri dishes fluorescence photobleaching (Ge et al. 1994). Specimens (5 cm). Two different cell culture media were tested: (1) were examined under a fluorescence microscope. RPMI 1640 containing 25 mM Hepes buffer, 5 mM glucose and 580 µg/ml -glutamine with the addition of Immunocytochemistry Briefly, slides fixed in acetone 100 U/ml penicillin and 100 µg/ml streptomycin and (5 min) or culture dishes containing fixed cells (fixation supplemented with 10% fetal calf serum and 5% human procedure described above) were sequentially incubated serum; (2) Dulbecco’s modified Eagle’s medium with primary mouse mAb followed by rabbit anti-mouse (DMEM), 5 mM glucose, 25 mM Hepes with sodium immunoglobulin, then with alkaline phosphatase– pyruvate with the addition of 100 U/ml penicillin and anti-alkaline phosphatase complex (APAAP) and finally 100 µg/ml streptomycin and supplemented with 10% fetal with fast red/naphthol ASBI phosphate as the chromo- calf serum and 5% human serum. After 5 days of genic substrate. The cell layer was then counterstained incubation at 37 )Cin5%CO2/humidified air, non- in haematoxylin, and a coverslip was applied using adherent cells were removed and cells adhering to the Glycergel as mounting medium (Dako, Glostrup, Petri dish were fed every third day. Denmark). GLUT1 expression was detected by using anti-GLUT1 serum followed by affinity-isolated alkaline phosphatase-conjugated swine anti-rabbit immuno- Cell harvesting globulin. After removal of tissue culture media, cells were washed twice with RPMI (protein-free) and then incubated with Flow cytometry Cells were harvested using trypsin as 0·05% trypsin and 0·053 mM EDTA at 37 )Cfor7 min. described above. After being washed with culture When all cells were detached from the plastic dish, the cell medium, D24 mAb or an isotype control was added to suspension was washed twice in tissue culture medium (1) cells and incubated for 30 min on ice. After two washes, and either seeded again after dilution 1:3 in tissue culture the samples were incubated with the FITC conjugate of medium (1) or frozen for later use for mouse immunization the F(ab*)2 fragment of rabbit anti-mouse immuno- or hybridoma selection. globulin (Dako) for 30 min on ice. Cells were then washed twice, resuspended in ice-cold medium containing sodium azide, and analyzed immediately on a Becton Dickinson Micromanipulation FACStarPlus flow cytometer using LYSIS II software In order to obtain slow trypsin activity, dishes containing (Becton Dickinson, Erembodegem, Belgium) to generate fibroblasts and insulinoma cells were incubated at room logarithmic fluorescence histograms. temperature. At a stage when cells had started to detach, insulinoma cells could easily be differentiated from fibro- blasts because of their cluster morphology and picked Cytospin preparation out using a micropipette under continuous inverted- microscope observation. The harvested insulinoma cells Insulinoma cell and fibroblast (5#105/ml) suspensions were expelled into a second Petri dish containing fresh were obtained from tissue culture harvest or from the tissue culture medium only. The separated cells were initial tissue processing. Aliquots (100 µl) of cell suspension selected a second time using a new pipette tip again under were placed into each well of a Hettich cytocentrifuge and

Journal of Endocrinology (1998) 156, 469–476

Downloaded from Bioscientifica.com at 10/01/2021 05:44:03AM via free access Pancreatic islet cell- and neuroendocrine tumor cell-specific monoclonal antibody · L WAGNER and others 471 spun on to the slide at 700 r.p.m. for 10 min. All slides cells/well and 10 U/ml interleukin-6 by micromanipu- were allowed to dry (4–12 h) before fixation and staining. lation. The clones were screened again by immuno- cytochemistry, and the best producers were chosen for bulk growth. Insulin secretion of isolated insulinoma cell clusters Clusters (1–10 per well) of 3-month-grown insulinoma cells were plated into 24-well culture dishes. Insulinoma Extraction of cellular antigen and detection by Western blot cells were free of fibroblasts. Assays were performed using the insulinoma cell-specific mAb D24 14 days after plating. Cells were incubated with fresh Insulinoma cells and other control cell lines were medium containing 5 mM glucose 24 h before the solubilized by adding 100 µl lysis buffer containing secretion assay was started. On the day of the secretion 100 mM NaCl, 50 mM Tris–HCl, pH 7·4, 10 mM assay, new tissue culture medium (1) containing 5 or EDTA, 1% Nonidet P40, 0·1% SDS, 100 µM phenyl- 25 mM glucose was added. methylsulphonyl fluoride and 1 mM diisopropyl fluoro- phosphate to 3#106 cells. The vials were immediately Addition of hNGFâ, pan optic staining and snap-frozen in liquid nitrogen. The sample was thawed immunocytochemistry for neurofilament expression and centrifuged at 12 000 g for 5 min at 4 )C. An aliquot of the clarified supernatant was loaded on to SDS– Isolated insulinoma clusters were exposed to 1 ng/ml polyacrylamide gels (12% polyacrylamide) and run under hNGFâ for 10 days. Cells were then fixed in situ as reducing and non-reducing conditions. described above and stained by pan optic staining as Nitrocellulose filters containing whole cell extract from described by Romanovsky (Worman 1989) or by the insulinoma cells, HepG2 human hepatoma cells or post anti-neurofilament mAb (Dako). The anti-neurofilament mortem human brain extract were obtained from the mAb labels neurons, neuronal processes, peripheral nerve SDS–polyacrylamide gels run under reducing and non- cells and sympathetic ganglion cells. Stimulated and reducing conditions. Hybridoma supernatant from mAb control cells were photographed in the culture dish using D24 was used as primary antibody which was followed a Leitz Aristoplan microscope. by monospecific biotinylated goat anti-mouse immuno- globulin G, and an avidin-biotinylated horseradish Reagents peroxidase complex (Vectastain ABC Kit, Vector Laboratories, Burlingame, CA, USA) and developed hNGFâ was obtained from Boehringer-Mannheim, with 4-chloro-1-naphthol (Sigma, Deisenhofen, Germany) Vienna, Austria. RPMI 1640, DMEM and fetal bovine and hydrogen peroxide according to the manufacturer’s serum were obtained from Gibco, Vienna, Austria. Insulin directions. and amylin were measured by RIA kits (insulin RIA 100) obtained from Pharmacia, Uppsala, Sweden and Peninsula Laboratories, Belmont, CA, USA respectively. The Results fibroblast-specific mAb FibAS01 (Saalbach et al. 1996) was used for identification of fibroblasts. Anti-GLUT1 serum Insulinoma cell culture was obtained from Chemicon, Temecula, CA, USA. Affinity-isolated alkaline phosphatase-conjugated swine In order to investigate whether insulinoma cells grow anti-rabbit immunoglobulin was purchased from Dako. in vitro, a fraction of minced tumor cells was incubated under the tissue culture conditions described above. After 6–12 h incubation, a number of insulinoma cells started to Development of the insulinoma cell-specific mAb D24 adhere to the RPMI culture medium (culture medium Mouse mAb D24 was obtained by immunizing a Balb/C (1)). Dead cells were removed after 6 days and cells were mouse with insulinoma cells and fusing the immune re-fed every fourth day. After 10 days, growth of fibro- mouse spleen cells with the myeloma cell line 63A8·653 blasts between the adherent insulinoma cell clusters could using polyethylene glycol 1500 (Sigma Chemical Co.) by be observed. In contrast, insulinoma cells did not grow in standard techniques. Wells were screened for isulinoma DMEM cultures (culture medium (2)). Cells from culture cell-specific mAb production by immunocytochemistry medium (1) were subcultured at 1:3 dilution the first time using insulinoma cytospin preparations. Positive wells 3 weeks after seeding. By the fourth passage, fibroblasts were analysed by using cryosections of post mortem human already represented about half of the cell population. pancreatic tissue as well. Only hybridomas found to be Insulinoma cells grew in clusters of different size mostly in immunocytochemically positive for both insulinoma cells several cell layers (Fig. 1A). After 77 days of culture, the and human pancreatic islets were processed further. rate of insulin secretion into the culture supernatant was Selected hybridomas were cloned (one cell per well) 723 U/ml per 24 h. The rate of amylin secretion was into 96-well microcultures containing 1#105 spleen 32 pg/ml per 24 h. From each passage, some Petri dishes

Journal of Endocrinology (1998) 156, 469–476

Downloaded from Bioscientifica.com at 10/01/2021 05:44:03AM via free access 472 L WAGNER and others · Pancreatic islet cell- and neuroendocrine tumor cell-specific monoclonal antibody

Figure 1 (A) Pan optic staining of an insulinoma cell cluster in co-culture with fibroblasts in a Petri dish. An insulinoma cell cluster is shown surrounded by fibroblasts after 3 months of co-culture. Nuclei of insulinoma cells are smaller and cytoplasmic staining is more intense than in fibroblasts. Magnification #600. (B) Pan optic staining of an isolated insulinoma cell cluster after incubation with hNGFâ. Isolated insulinoma cells incubated with 1 ng/ml hNGFâ produce neurite-like projections. In situ pan optic staining of isolated insulinoma cells with distinctive neurite-like projections after 8 days of culture in the presence of NGF. Magnification #400.

were fixed 24 h after subculture and stained in situ by pan cell clusters were maintained for 3 months in culture optic staining (Fig. 1A) and by immunocytochemistry medium (1). During this time, isolated insulinoma cells using either the fibroblast-specific mAb FibAS01 or the developed cytoplasmic extensions, which, upon addition anti-insulinoma mAb D24 (Fig. 2F). The D24 mAb of hNGFâ (1 ng/ml) tended to become neurite-like clearly differentiated fibroblasts from insulinoma cells. projections (Fig. 1B). After 6 days of stimulation, cells Insulinoma cells were much smaller than fibroblasts and started to form distinct processes. Immunocytochemically, had ovoid nuclei (Fig. 1A). On pan optic staining, neurofilament protein could be detected with a insulinoma cells had blue cytoplasm whereas fibroblasts neurofilament-specific mAb in more than 90% of the had a large nucleus and pale cytoplasm (Fig. 1A and induced insulinoma cells (data not shown). This was Fig. 2F). Cluster formation was unique to insulinoma accompanied by detachment of some cells from the cells. To ascertain the morphology of insulin-secreting original cluster and formation of axon-like projections insulinoma cells further, FITC-conjugated sheep anti- which maintained contact with the original cluster (Fig. human insulin serum was used to localize the insulin- 1B). Insulin secretion still occurred (20 µU/ml per 24 h producing insulinoma cells (Fig. 2D). per cluster). Individual clusters contained between 30 and When cells were further subcultured for more than 60 insulinoma cells. 3 months, fibroblast growth was found to be faster than insulinoma cell growth. Insulinoma cell number therefore decreased from subculture to subculture in parallel with Development and specificity of anti-insulinoma mAb D24 decreasing insulin content in the supernatant. As a part of our continuing effort to characterize pancreatic â-cell-specific antigens, we generated a novel mAb using human insulinoma cells as immunogen which recognizes a Culture of isolated insulinoma cells in the absence of fibroblasts 32 kDa protein in human neuroendocrine tumor cells and In a second attempt to obtain a pure insulinoma cell pancreatic â-cells. This mAb D24 has been successfully culture, insulinoma clusters were separated from the Petri utilized for immunoblotting (Fig. 3) and immunocyto- dish by micromanipulation as described above. Insulin chemical procedures. Using this mAb, we have been able secretion from the isolated insulinoma cells was measured to demonstrate that a further three human insulinomas in the presence of 5 and 25 mM glucose. Insulin release (Fig. 2A) and tumors relating to the neuroendocrine was dependent on cell number but not on glucose con- system such as carcinoid (Fig. 2B), vipoma (Fig. 2C) and centration (Table 1) as correct glucose responsiveness gastrinoma also express this antigen. The antibody also was not present. To understand the impaired glucose recognizes pancreatic â-cells of post mortem cryostat tissue responsiveness further, the original tumor tissue was im- sections (Fig. 2E). There was broad variety in staining munostained for GLUT1 expression, which was positive intensity of different pancreatic islets. Other tissues such as not only in this insulinoma but also in two further ones liver, exocrine pancreas, brain, connective tissue, endo- used for testing mAb D24 specificity. Isolated insulinoma thelial cells, epithelium of the bile duct and peripheral

Journal of Endocrinology (1998) 156, 469–476

Downloaded from Bioscientifica.com at 10/01/2021 05:44:03AM via free access Pancreatic islet cell- and neuroendocrine tumor cell-specific monoclonal antibody · L WAGNER and others 473

Figure 2 (A) Immunocytochemical staining of permeabilized insulinoma cells. A cytospin preparation of insulinoma cells was stained with D24 mAb using the APAAP technique. The mAb labeled more than 95% in a granular staining pattern. Magnification #1220. (B) Immunocytochemical staining of permeabilized cells derived from a liver metastasis of a carcinoid tumor. A cytospin preparation of carcinoid cells after fixation in acetone was stained with D24 mAb (APAAP method). More than 40% of cells were labeled all over the cytoplasm. Magnification #940. (C) Immunocytochemical staining of a cryostat section of a pancreatic vipoma. A cryostat section of a pancreatic vipoma after fixation in acetone was stained with D24 mAb (APAAP method). All of the tumor cells were labeled over all the cytoplasm. Magnification #560. (D) In situ immunofluorescence staining of an insulinoma cell cluster growing in a Petri dish surrounded by fibroblasts. With FITC-conjugated sheep anti-human insulin serum, only insulinoma cells were stained. Fibroblasts surrounding the insulinoma islet show low background stain. Magnification #280. (E) Immunocytchemical staining of a cryostat section of post mortem pancreatic tissue. A cryostat section of pancreatic tissue after fixation in acetone was stained with D24 mAb (APAAP method). More than 95% of the cells of the islets of Langerhans were labeled over all the cytoplasm. Magnification #940. (F) In situ immunocytochemical staining of an insulinoma cell cluster surrounded by fibroblasts. The cell layer was stained after fixation with ice-cold methanol using D24 mAb and the APAAP technique. Only insulinoma cells were stained. Magnification #140.

Journal of Endocrinology (1998) 156, 469–476

Downloaded from Bioscientifica.com at 10/01/2021 05:44:03AM via free access 474 L WAGNER and others · Pancreatic islet cell- and neuroendocrine tumor cell-specific monoclonal antibody

Table 1 Insulin secretion measured in tissue culture supernatant Discussion obtained from 24 h incubation of isolated insulinoma cell clusters exposed to 5 or 25 mM glucose In order to obtain an insulin-secreting â-cell-resembling Insulin secretion cell line as a source of antigen for the development of (ìU/ml supernatant) islet cell-specific mAbs, we attempted to culture benign human insulinoma cells, which has not been previously 5 mM glucose 25 mM glucose reported. Insulinoma cells were shown to grow in co- Number of cells culture with fibroblasts at a low rate and could be separated 0 4·4 3·7 from the fibroblasts. The cultured insulinoma cells were 87 26·3 13·2 shown to be devoid of any insulin-secretory response to 215 61·2 33·0 256 84·2 46·6 elevated glucose concentration. This finding is in line with abnormal insulin-secretory responses to glucose described for rat- (Praz et al. 1983) and mouse- (Miyazaki et al. 1990) derived insulinomas. Such non-physiological insulin secretion in insulinoma cells is due to abnormal expression of glucose transporter (GLUT1) (Boden et al. 1994). GLUT1 in insulinomas represents a low-Km glucose transporter protein which is not normally present in pancreatic â-cells. However, when the GLUT2 gene was transferred into insulinoma cells, low- and high-affinity glucose-stimulated insulin release could be demonstrated (Ferber et al. 1994). In order to evaluate the potential of cultured insulinoma cells for further differentiation, hNGFâ was added to isolated insulinoma cell clusters. Evidence was thereby provided that hNGFâ is able to induce neurite formation, as previously also shown for the animal insulinoma cell line RINm5F (Polak et al. 1993). Although insulinoma cells undergo some morphological changes when isolated from fibroblasts, the neurite extension is caused by NGF. Figure 3 Immunoblotting of human insulinoma cells using mAb Neurite extension of murine pancreatic insulin cells on D24. Human insulinoma cell extract was loaded on to SDS–12% dissociation of cells from the pancreatic islets along polyacrylamide gels and run under reducing conditions. Separated with neurofilament expression has also been reported proteins were transferred to nitrocellulose and detected using D24 mAb following monospecific biotinylated goat anti-mouse (Teitelman 1990). The demonstration of similar behavior immunoglobulin G and avidin-biotinylated peroxidase complex for human insulinoma cells in this study supports the and developed with 4-chloro-1-naphthol. suggestion that human insulinoma cells also share charac- teristics with neurons. In this context, it is of note that NGF receptors are present in pancreatic â-cell lines blood leukocytes tested negative. Negative results (Scharfmann et al. 1993) and that NGF itself can be were obtained with HepG2 human hepatoma cells and produced by fibroblasts (Yoshida & Gage 1992) and can U937 human histiocytoma cells using both methods exert profound regulatory stimulation in receptor-bearing (immunoblotting and immunocytochemistry). No positive target cell lines (Greenberg et al. 1985). reaction was found with â-cells of rodents and a dog D24 mAb recognizes a molecule that is strongly insulinoma. expressed not only in insulinoma cells but also in tumor In order to investigate whether the D24 mAb- cells (carcinoid, vipoma and gastrinoma) as well as being recognized protein represents a cell surface expression moderately expressed in post mortem human pancreatic molecule, flow cytometric analysis of isolated insulinoma islet tissue. The somewhat surprising expression of the cells or insulinoma cells obtained from tissue culture D24 mAb-recognized protein in neuroendocrine-derived was performed (Fig. 4). Surface expression of the tumors other than insulinoma can be explained by the D24 mAb-reactive protein was observed on insulinoma observation (Alpert et al. 1988) that all four main islet cell cells, but not on fibroblasts. After morphological differ- types arise from common precursors. The hormone- entiation of insulinoma cells with multiple cytoplasmic specific genes of these precursors become activated projections, D24 mAb staining was still present but sequentially during development (Teitelman et al. decreased in intensity (immunocytochemistry) when 1993). The most primitive islet cells are often found compared with that of freshly isolated insulinoma cells co-expressing two, three or even four islet hormones (data not shown). (Lukinius et al. 1992). Similarly, animal insulinoma cell

Journal of Endocrinology (1998) 156, 469–476

Downloaded from Bioscientifica.com at 10/01/2021 05:44:03AM via free access Pancreatic islet cell- and neuroendocrine tumor cell-specific monoclonal antibody · L WAGNER and others 475

Figure 4 Flow cytometry of insulinoma cells and fibroblasts with mAb D24. Insulinoma cells were incubated with D24 mAb (including an isotype control) followed by the FITC-conjugated F(ab*)2 fragment of rabbit anti-mouse immunoglobulin. The mAb labeled 70% of insulinoma cells. Using the same mAb, no cell surface staining was observed on fibroblasts. lines have been found co-expressing several peptide (D24) that specifically recognizes pancreatic â-cells hormones (Madsen et al. 1986, Philippe et al. 1987). and cells derived from gastrointestinal neuroendocrine Therefore it has to be assumed that these different tumors such as insulinoma, vipoma, gastrinoma and neuroendocrine cells share protein molecules that are carcinoid. These observations together with earlier find- involved in hormone packaging, secretion or signal trans- ings in animal models support the hypothesis that human duction. One of these shared molecules may be the protein pancreatic â-cells share properties with neurons and are recognized by D24 mAb. related to other neuroendocrine cells in the gastrointestinal In conclusion, we have been able to grow insulinoma tract. cells in co-culture with fibroblasts. These insulinoma cells, which grow in clusters, show inadequate insulin-secretory response to glucose because of their expression of GLUT1. Acknowledgements When isolated clusters of insulinoma cells were incubated in the absence of fibroblasts in the presence of hNGFâ, This work was supported by the Fonds zur Foerderung der neurofilament protein transcription and neurite extension wissenschaftlichen Forschung Oesterreich grant P11782- were observed. Using the insulinoma cells as an antigen for Med. We thank Professor Niederle for support, and Britt immunization of Balb/C mice, we have generated a mAb Schier for photographic work.

Journal of Endocrinology (1998) 156, 469–476

Downloaded from Bioscientifica.com at 10/01/2021 05:44:03AM via free access 476 L WAGNER and others · Pancreatic islet cell- and neuroendocrine tumor cell-specific monoclonal antibody

References line that retains glucose-inducible insulin secretion: special reference to expression of glucose transporter isoforms. Endocrinology 127 Alpert S, Hanahan D & Teitelman G 1988 Hybrid insulin genes 126–132. reveal a developmental lineage for pancreatic endocrine cells and Permutt MA, Koranyi L, Keller K, Lacy PE, Scharp DW & Mueckler imply a relationship with neurons. Cell 53 295–308. M 1989 Cloning and functional expression of a human pancreatic Boden G, MurerE&MozzoliM1994 Glucose transporter proteins in islet glucose-transporter cDNA. Proceedings of the National Academy of 86 human insulinoma. Annals of Internal Medicine 121 109–112. Sciences of the USA 8688–8692. D’Ambra D, Surana M, Efrat S, Starr RG & Fleischer N 1990 Philippe J, Chick WL & Habener JF 1987 Multipotential phenotypic Regulation of insulin secretion from beta-cell lines derived from expression of genes encoding peptide hormones in rat insulinoma 7 transgenic mice insulinomas resembles that of normal beta-cells. cell lines. Journal of Clinical Investigation 351–358. Endocrinology 126 2815–2822. Polak M, Scharfmann R, Seilheimer B, Eisenbarth G, Dressler D, Efrat S, Linde S, Kofod H, Spector D, Delannoy M, Grant S, Verma IM & Potter H 1993 Nerve growth factor induces neuron- Hanahan D & Baekkeskov S 1988 Beta-cell lines derived from like differentiation of an insulin-secreting pancreatic beta cell line. transgenic mice expressing a hybrid insulin gene-oncogene. Proceedings of the National Academy of Sciences of the USA 90 Proceedings of the National Academy of Sciences of the USA 85 5781–5785. 9037–9041. Praz GA, Halban PA, Wollheim CB, Blondel B, Strauss AJ & Renold Efrat S, Leiser M, Surana M, Tal M, Fusco-DeMane D & Fleischer N AE 1983 Regulation of immunoreactive-insulin release from a rat 1993 Murine insulinoma cell line with normal glucose-regulated cell line (RINm5F). Biochemical Journal 210 345–352. insulin secretion. Diabetes 42 901–907. Rabin DU, Pleasic SM, Palmer-Crocker R & Shapiro JA 1992 Fehmann HC, StrowskiM&GokeB1994 Interaction of glucagon- Cloning and expression of IDDM-specific human autoantigens. like peptide-I (7–37) and somatostatin-14 on signal transduction and Diabetes 41 183–186. proinsulin gene expression in beta TC-1 cells. Metabolism 43 Radvanyi F, Christgau S, Baekkeskov S, Jolicoeur C & Hanahan D 787–792. 1993 Pancreatic beta-cells cultured from individual preneoplastic Ferber S, Beltrandel-Rio H, Johnson JH, Noel RJ, Cassidy LE, Clark foci in a multistage tumorigenesis pathway: a potentially general S, Becker TC, Hughes SD & Newgard CB 1994 GLUT-2 gene technique for isolating physiologically representative cell lines. transfer into insulinoma cells confers both low and high affinity Molecular Cell Biology 13 4223–4232. glucose-stimulated insulin release. Relationship to glucokinase Saalbach A, Anderegg U, Bruns M, Schnabel E, Herrmann K & activity. Journal of Biological Chemistry 269 11523–11529. Haustein U 1996 Novel fibroblast-specfic monoclonal antibodies: Ge Y, Ezzell RM, Tompkins RG & Warren HS 1994 Cellular properties and specificities. Journal of Investigative Dermatology 106 distribution of endotoxin after injection of chemically purified 1314–1319. lipopolysaccharide differs from that after injection of live bacteria. Scharfmann R, Tazi A, Polak M, Kanaka C & Czernichow P 1993 Journal of Infectious Diseases 169 95–104. Expression of functional nerve growth factor receptors in pancreatic Greenberg ME, Greene LE & Ziff EB 1985 Nerve growth factor and beta-cell lines and fetal rat islets in primary culture. Diabetes 42 epidermal growth factor induce rapid transient changes in proto- 1829–1836. oncogene transcription in PC 12 cells. Journal of Biological Chemistry Teitelman G 1990 Insulin cells of pancreas extend neurites but do not 260 14101–14110. arise from the neuroectoderm. Developmental Biology 142 368–379. Hamaguchi K, Gaskins HR & Leiter EH 1991 NIT-1, a pancreatic Teitelman G, Alpert S, Polak JM, Martinez A & Hanahan D 1993 beta-cell line established from a transgenic NOD/Lt mouse. Precursor cells of mouse endocrine pancreas coexpress insulin, Diabetes 40 842–849. glucagon and the neuronal proteins tyrosine hydroxylase and Karounos D & Thomas JW 1990 Recognition of common islet antigen neuropeptide Y, but not pancreatic polypeptide. Development 118 by autoantibodies from NOD mice and humans with IDDM. 1031–1039. Diabetes 39 1085–1090. Thomas NM, Ginsberg-Fellner F & McEvoy R 1990 Strong Knaack D, Fiore DM, Surana M, Leiser M, Laurance M, Fusco- association between diabetes and displacement of mouse anti-rat DeMane D, Hegre OD, Fleischer N & Efrat S 1994 Clonal insulinoma cell monoclonal antibody by human serum in vivo. insulinoma cell line that stably maintains correct glucose Diabetes 39 1203–1211. responsiveness. Diabetes 43 1413–1417. Worman CP 1989 Morphology and cytochemistry. In Laboratory Lukinius A, Ericsson JLE, GrimeliusL&Korsgren O 1992 Haematology, edn 1, pp 169–176. Ed I Chanarin. London: Churchill Ultrastructural studies on the ontogeny of fetal human and porcine Livingstone. endocrine pancreas, with special reference to colocalization of four Yoshida K & Gage FH 1992 Cooperative regulation of nerve growth major islet hormones. Developmental Biology 153 376–385. factor synthesis and secretion in fibroblasts and astrocytes by Madsen OD, Larssen LI, Rehfeld JF, Schwartz TW, Lemmark A, fibroblast growth factor and other cytokines. Brain Research 569 Labreque AD & Steiner DF 1986 Cloned cell lines from a 14–25. transplantable islet cell tumor are heterogenous and express cholecystokinin in addition to islet hormones. Journal of Cell Biology 103 2025–2034. Received 6 May 1997 Miyazaki J-I, Araki K, Yamato E, Ikegami H, Asano T, Shibasaki Y, Revised manuscript received 26 August 1997 Oka Y & Yamaura K-I 1990 Establishment of a pancreatic beta-cell Accepted 24 September 1997

Journal of Endocrinology (1998) 156, 469–476

Downloaded from Bioscientifica.com at 10/01/2021 05:44:03AM via free access