The importance of the glucose transporter isoform,. GLUT2, in the construction of glucose-sensitive surrogate insulin-secreting cells was evaluated using murine ...
75
Expression of GLUT2 in insulin-secreting AtT20 pituitary cells E L Davies, K I J Shennan1, K Docherty1 and C J Bailey Department of Pharmaceutical Sciences, Aston University, Aston Triangle, Birmingham B4 7ET, UK 1 Department of Molecular and Cell Biology, Institute of Medical Sciences, Foresterhill, University of Aberdeen, Aberdeen AB25 2ZD, UK (Requests for offprints should be addressed to C J Bailey)
ABSTRACT
The importance of the glucose transporter isoform, GLUT2, in the construction of glucose-sensitive surrogate insulin-secreting cells was evaluated using murine pituitary AtT20 cells. The cells were double transfected with cDNAs for human preproinsulin (hppI-1) driven by the cytomegalovirus promoter, and human GLUT2 driven by the â-actin promoter. The stably transfected clone, AtTinsGLUT2.36, which strongly expressed both the hppI-1 and GLUT2 genes, constitutively released 7·5 ng/106 cells/24 h of immunoreactive insulin-like material, 75% of which was fully processed mature human insulin. Increasing glucose concentrations in the subphysiological range up to 50 µM increased insulin release, but greater glucose concentrations did not further increase insulin release. Suppression of the low-Km glucosephosphorylating enzyme, hexokinase, with 2-deoxy-glucose increased glucose-stimulated insulin
release by two- to threefold in the presence of subphysiological and physiological glucose concentrations up to 10 mM. Physiological glucose concentrations increased the amount of GLUT2 mRNA, indicating that the â-actin promoter responds in a glucose-dependent manner. Implantation of 2#107 AtTinsGLUT2.36 cells intraperitoneally into streptozotocin-diabetic nude mice slowed the progression of hyperglycaemia. The implanted cells formed vascularised tumourlike cell aggregates attached to the peritoneum. The results demonstrate that the â-actin promoter is partially regulated by glucose. Expression of GLUT2 enables glucose to enter the cell at high Km, but high-Km glucose phosphorylation is also required to signal glucose-stimulated genes affecting insulin release.
INTRODUCTION
diabetic animals (Selden et al. 1987, Kawakami et al. 1992, Stewart et al. 1993, 1994). These cells release insulin constitutively, but it would be desirable to control insulin release in accordance with the blood glucose concentration (Bailey et al. 1997). To install a glucose-sensitive insulinsecretory mechanism similar to pancreatic â-cells presents a formidable challenge; however, the recently elucidated signalling pathway through which glucose controls insulin gene transcription provides a potential means to regulate insulin production, which in turn will govern insulin release (MacFarlane et al. 1994, 1997). The first step in this pathway is the transfer of glucose into the cell via the low-affinity glucose transporter isoform GLUT2. In the present study, we used murine pituitary AtT20 cells as a model to determine whether
Somatic cell gene therapy provides a means to create glucose-sensitive surrogate insulin-secreting cells that could be used in the treatment of diabetes mellitus (Docherty 1991). In principle somatic cells other than islet â-cells can be removed from a diabetic person, bioengineered to secrete insulin, grown and characterised in vitro, and implanted back into the donor with appropriate containment safeguards (Bailey & Docherty 1994). This method of insulin delivery avoids the main problems that have thwarted islet transplantation, notably the restricted availability of cadaveric tissue and immunological rejection (Ricordi 1996). We and others have demonstrated the feasibility of a gene therapy approach by creating insulinproducing somatic cells and implanting them into
Journal of Molecular Endocrinology (1998) 20, 75–82
Journal of Molecular Endocrinology (1998) 20, 75–82 ? 1998 Journal of Endocrinology Ltd Printed in Great Britain 0952–5041/98/020–075 $08.00/0
76
and others
· Expression of GLUT2 in insulin-secreting AtT20 pituitary cells
co-expression of transfected genes for human GLUT2 and preproinsulin would confer glucose sensitivity to the release of insulin. The GLUT2 gene is driven by the inducible â-actin promoter (Phillips et al. 1988), which we show herein to be regulated through glucose. MATERIALS AND METHODS
Cells, chemicals and animals AtT20d16v cells from Dr R E Mains, Johns Hopkins University, Baltimore, USA were maintained at 37 )C in 5% CO2 in air as described previously (Stewart et al. 1993). Cells were cultured in Dulbecco’s modified Eagle’s medium, 10% fetal calf serum, and 11 mM glucose, except where stated. Cell culture plastics were from Sarstedt (Leicester, UK) and reagents were from Gibco (Paisley, Strathclyde, UK). Distilled water was used throughout. 125I-Insulin (specific activity 2000 Ci/mmol) was from Amersham (Amersham, Bucks, UK). All other chemicals were of analytical grade from Sigma (Poole, Dorset, UK) and BDH (Poole, Dorset, UK). Adult male athymic nude mice were obtained from Bantin and Kingman (Hull, UK). The mice were maintained in an isolated environment with filtered air at 22 )C and 12 h light per day (lights on 0800–2000 h) and a standard pellet diet (Economy Rodent Diet, SDS, Essex, UK) with tap water available ad libitum. The mice were given 200 mg/kg streptozotocin (in citrate buffer pH 4·5) by intraperitoneal injection, to induce hyperglycaemia. Transfection of AtT20 cells with insulin and GLUT2 Plasmid pCB7–hppI-1 was prepared using a Hind III, Xba I fragment of the cDNA encoding the human preproinsulin gene (hppI-1). This was subcloned into the vector pCB7 immediately downstream of its cytomegalovirus (CMV) promoter. The pLK444–GLUT2 plasmid (gift of Dr G Gould, University of Glasgow, UK) contained the human GLUT2 cDNA subcloned downstream of the human â-actin promoter. The pLK444 vector also contained a neomycin resistance marker gene. The two plasmids were used in equal quantity for transfection of subconfluent monolayer cultures of AtT20 cells by the calcium phosphate coprecipitation method (Taylor & Docherty 1992). After selection in G418 sulphate at a concentration of 500 µg/ml in the culture medium, 45 stable transformants were isolated. The transformants were grown for three passages in selection media, Journal of Molecular Endocrinology (1998) 20, 75–82
and then analysed for expression of insulin and GLUT2. Analysis of insulin and GLUT2 expression For northern blot analysis, AtT20 cells were harvested from tissue culture and total RNA was extracted using the guanidinium thiocyanate method (Chomczynski & Sacchi 1987). RNA were resolved on formaldehyde gels and transferred to Hybond-N membrane filters. The blots were fixed using 1200J in a u.v. crosslinker and prehybridised in phosphate buffer with 7% SDS and 100 µg/ml denatured herring sperm DNA, at 65 )C for 3 h. Insulin or GLUT2 cDNA probes labelled with [32P]dCTP using a Pharmacia Oligolabelling Kit (Pharmacia Biotech, St Albans, UK) were added and the blots were hybridised for 20 h at 65 )C. Blots were then washed twice for 15 min in 2#SSC/0·1% SDS at room temperature, followed by a 30-min wash in 0·1#SSC/0·1% SDS at 50 )C. Blots were visualised by autoradiography, stripped with boiling 0·1% SDS, and reprobed with an 18S RNA probe to assess RNA loading on the blot. Characterisation of AtTinsGLUT2 cells in vitro The total amount of insulin-like immunoreactive material released into the culture medium was determined by radioimmunoassay with a broad specificity insulin antiserum that cross-reacted fully with mature rat and human insulins and partially processed proinsulin (des 31,32 proinsulin and des 64,65 proinsulin) (Sigma). An antiserum specific for mature human insulin (crossreactivity: proinsulin
