WO1995032740A1 - Transformed human hepatoma cell line which releases insulin - Google Patents

Abstract

A method of treating Type (I) diabetes in a subject comprising transfecting an hepatocyte with a gene encoding insulin, the gene being under control of an appropriate promoter, and introducing the transformed hepatocyte into the subject.

Description

TRANSFORMED HUMAN HEPATOMA CELL LINE WHICH RELEASES INSULIN The present invention relates to a method of treating Type I diabetes using a transformed hepatocyte. The present invention further relates to a transformed hepatocyte which secretes insulin.

In order to design a feasible somatic cell gene delivery system for the treatment of diabetes in which cells from the patient are removed, cultured In vitro, transfected with the insulin gene and reimplanted, several problems may be anticipated. Firstly, a suitable cell type needs to be determined. The cell type of choice for gene therapy of diabetes is not the beta cell. Islet cells are either greatly reduced or absent in patients suffering from Type I diabetes because of autoimmune destruction (Eisenbarth et al, 1986). The donor cell type must be accessible and capable of being engineered to synthesise, process, store and secrete insulin. Secondly, it is essential that the insulin output of such a modified cell type must be regulated appropriately. To achieve this insulin gene expression must be under the control of a suitable promoter.

The production of trans aryotic mice bearing an intact human insulin gene inserted into mouse Ltk fibroblast cells (Seldon et al, 1987) pioneered the field of somatic gene therapy in diabetes by proving that transfected fibroblasts can supply enough insulin in diabetic mice to initially produce healthy animals . However, as the fibroblasts had no facility to store insulin and no developed regulatory pathways to control its secretion the mice died of hypoglycaemia. It would appear that cell types worth foremost consideration are those already specifically adapted to protein secretion, which have well developed regulatory pathways capable of processing prohorraones . The AtT-20 cell line derived from the mouse anterior pituitary when transfected with cDNA for human proinsulin has been shown to be capable of processing proinsulin to insulin (Moore et al, 1983). The islet isoform of glucokinase is naturally expressed in AtT-20ins cells, but the cells lack ability to express the glucose transporter GLUT-2 and fail to respond to glucose (Hughes et al, 1991). Upon stable transfection with rat GLUT-2, the AtT-20ins cells exhibit increased intracellular storage of insulin, glucose potentiation of non-glucose secretogogues and a direct stimulation of insulin release by glucose, although the maximal effect was seen at non-physiological concentrations of glucose (Hughes et al, 1992). Work in this laboratory has generated stable transformants of the AtT-20 cell line containing insulin cDNA attached to an inducible promoter (Simpson et al, 1993). These clones secreted and stored insulin but were unresponsive to glucose.

Although the AtT-20 immortal cell line provides a useful place to start in developing a model system for gene therapy of diabetes, their primary equivalent is unlikely to be available from humans; therefore accessible primary cells must be considered. Of these the present inventors considered hepatocytes as possible target cells . Hepatocytes are known to play a crucial role in intermediary metabolism, synthesis and storage of proteins in the liver, exhibit glucokinase expression (similar to pancreatic islets) and are accessible. Introduction and stable expression of foreign genes into primary mammalian hepatocytes has been demonstrated recently using human liver tissue (Grossman et al, 1991). In the present study a cDNA for human insulin has been introduced by electroporation into a human hepatoma cell line HEP G2 that has retained most of the biosynthetic capabilities of normal liver cells (Knowles et al, 1980) . The results show that stable transfection of insulin cDNA into this liver cell line results in synthesis, storage and acute regulated insulin release. However, glucose responsiveness was not detected. However, when this cell line was further transfected with a GLUT-2 cDNA, regulated release of insulin upon glucose stimulation was observed.

Thus, liver cells which are readily accessible by biopsy may be engineered to secrete insulin acutely in response to glucose and possibly other physiological secretogogues. In a first aspect, the present invention consists in a method of treating Type I diabetes in a subject comprising transfecting an hepatocyte with a gene encoding insulin, the gene being under the control of an appropriate promoter, and introducing the transformed hepatocyte into the subject.

In a second aspect, the present invention consists in a transformed hepatocyte for use in treatment of Type I diabetes, the hepatocyte including a gene encoding insulin, the gene being under the control of an appropriate promoter.

Preferably, the hepatocyte used in the invention is a primary hepatocyte isolated from the subject and, following transfection with an insulin gene under the control of an appropriate promoter, is re-introduced to the subject. Such primary hepatocytes naturally express the glucose transporter GLUT-2, however, it may be desirable to increase the expression of GLUT-2 in the transfected hepatocyte by further transfecting with an homogeneous or heterogeneous GLUT-2 gene under the control of an appropriate promoter.

It is also possible to utilise hepatoma cell lines

(e.g., HEP G2) which may not express GLUT-2. Where a cell line of this kind is used, it is preferred to further transfeet the hepatocyte with a homogeneous or heterogeneous gene encoding GLUT-2 under the control of an appropriate promoter.

In a further preferred embodiment, the hepatocyte is further transfected with either an homogeneous or heterogeneous gene encoding the high capacity glucose phosphorylation enzyme glucokinase.

In yet a further preferred embodiment of the present invention, the hepatocyte includes either homogeneous or heterogeneous genes encoding the glucose transporter GLUT

2 and the high capacity glucose phosphorylation enzyme glucokinase.

In order that the nature of the present invention may be more clearly understood, preferred forms thereof will now be described with reference to the following examples and accompanying figures .

Brief Description of the Accompanying Figures;

Figure 1 provides light micrographs of a) HEP G2 cells

(negative controls) and b) HEP G2ins cells immunochemically stained for insulin and counterstained with Harris heamotoxylin and Scott bluing. Granular positive staining in cytoplasm of HEP G2ins cells is marked with arrows (X 495).

Figure 2a provides a transmission electron micrograph of part of the cytoplasm of a HEPG2ins cell. There are many membrane-bound vacuoles (v) in the cytoplasm, each containing electron dense material. Scale 1 μm.

Figure 2b provides a transmission electron micrograph of part of the nucleus and cytoplasm of a HEP G2 control cell. There is an abundance of rough endoplasmic reticulum (r) , mitochondria (m) and a Golgi complex (G) .

However, no membrane bound vacuoles are present. Scale,

1 urn.

Figure 3 shows insulin synthesis of HEP G2ins cells in response to varying concentrations of glucose (0-20 mM) .

Values are expressed as means + S.E. for 6 observations.

Figure 4 shows the effect of 20 mM glucose and 5 mM 8-Br- cAMP on a) the chronic release of insulin, and b) the insulin content of HEP G2ins cells. Values are expressed as means + S.E. for 3 observations. Figure 5 shows Northern blot analysis of Glut 2 in HEP G2 (untransfected cells, lane 1), HEP G2 control cells (transfected with CMV and Rep 4 vectors alone, lane 2) HEP G2ins cells (transfected with insulin cDNA alone, lane 3) and HEP G2ins/ g cells (lane 4.

Figure 6 shows stimulation by 20 mM glucose on the acute regulated release of immunoreactive insulin and proinsulin. HEP G2ins/ g cells were incubated in the basal medium for 2 consecutive 1 hr periods to stabilise the basal secretion of insulin. Monolayers were exposed to the stimulus for 1 hr. n: number of experiments, B: basal, S: stimulus. Values are express as means + S.E. Figure 7 shows stimulation by (a) 5 mM 8 Br cAMP; (b) 10 mM theophylline; (c) 20 mM arginine on the acute regulated release of immunoreactive (pro)insulin. HEP G2ins/ g cells were incubated in the basal medium for 2 consecutive 1 hr periods to stabilise the basal secretion of insulin. Monolayers were exposed to the stimulus for 1 hr. n: number of experiments, B: basal, S: stimulus. Values are expressed as means + S.E.

Figure 8 shows insulin secretion of HEP G2ins/ g cells in response to varying concentrations of glucose (0-20 mM. n=6, ± S.E. mean. Figure 9 shows perifusion of HEP G2ins/ g cells with 20 mM glucose.

Figure 10 shows insulin synthesis of HEP G2ins/ g cells in response to varying concentrations of glucose (0-20 mM) . n=4, + S.E. mean. EXAMPLE 1 METHODS

Reagents:

Insulin cDNA pC2 was kindly provided by Dr. M. Walker, Weizmann Institute, Israel. The expression vector pRcCMV was purchased from Invitrogen (San Diego, Cal, USA) and pSKII^"1"' from Statagene (La Jolla, Cal, USA). Eagles Minimal Essential Medium (MEM) and G418 antibiotic were purchased from Gibco Laboratories, Grand Island, NY, USA. Fetal calf serum (FCS) was supplied by Cytosystems Pty Ltd, Sydney, Australia. Restriction enzymes came from Boehringer Mannheim, Germany. Biosynthetic human proinsulin (hPI) and a polyclonal antibody to this peptide were kindly supplied by Lilly Research Laboratories, Indianapolis, USA. The radioimmunoassay for insulin was carried out with a human insulin kindly provided by Novo Nordisk, Sydney, Australia and guinea pig insulin antibody donated by D. Yue and ιl25 by J. Bryson, University of Sydney, Australia. ^H leucine was purchased from New England Biolabs, Ontario, Canada. Guinea pig anti-insulin antibody was supplied by Dako Corp. Ca, USA. 8-Br-cAMP was purchased from Sigma, St. Louis, Mo, USA. Millipore filters were purchased from Millipore, Bedford, MA, USA. Biorad Protein Assay Kit was obtained from Biorad, Richmond, CA, USA and pansorbin from Calbiochem, Behring Diagnostics, La Jolla, USA. Cell Culture; HEP G2 cells were grown in monolayers in Minimal

Essential Medium (MEM) supplemented with 10% fetal calf serum (FCS) in air at 37°C. DNA Transfection and Selection of Transformed Cells;

The full-length 0.6 Kb human insulin cDNA pC2 was ligated into the multi cloning site of the mammalian expression vector pSKII⁺ (EcoRl/BamHl) . The Xbal/Hindlll 0.6 Kb fragment was subsequently cloned into the multi- cloning site of pRcCMV which expresses resistance to the antibiotic neomycin/G418. HEP G2 cells were transfected with 20 μg of the recombinant plasmid and vector alone (these clones are used throughout the experiments as "control" cells) by electroporation at 200V and 960 μF in a Biorad gene pulser at a cell concentration of 5 x 10^ cells per cuvette, in MEM medium containing no FCS. 2.5 x 10^ cells were subsequently plated into culture dishes in MEM medium containing 10% FCS. To obtain stable transfectants of HEP G2 cells containing insulin cDNA, 48 hr later 1 mg/ml of G418 antibiotic was added to the culture medium. The antibiotic G418 had a purity of 46-49% (active drug, 460-490 μg/mg) . The concentration referred to above is the crude compound and not the actual drug. Medium plus drug was changed every 2-3 days. After 3-4 weeks of selection colonies were picked using cloning rings and screened for production of insulin by radioimmunoassay (RIA) . Selected clones were expanded into mass cultures and maintained in selective media until use. Cell Extraction and Radioimmunoassays;

Samples of culture medium were collected for estimation of insulin and human proinsulin by radioimmunoassays (RIA) . HPI was measured as described previously (Tuch et al, 1992). Iodinated hPI was prepared by the chloramine-T method. Specificity was established by showing <0.05% cross-reactivity with insulin. The standard RIA for insulin was carried out using guinea pig insulin antibody and 125ι_ι_a]-,_en_ecj insulin prepared by the chloramine-T method. Cross-reactivity with hPI was 73%; because of this, insulin values obtained have been appropriately reduced, and it is these corrected values that are reported in the results . Insulin content of the cells was determined following extraction overnight in acid-ethanol.

Albumin secretion of the cells was measured using a Kallestad QM 300 rate nephelometer. Immunohis ochemical Analysis; Immunohistochemical analysis was carried out on fixed samples of cultures that had been trypsinised, using guinea pig anti-insulin antibody and the streptavidin- diaminobenzidine chromogen complex. Electron Microscopy; After trypsinization a single cell suspension of 10^ control HEP G2 cells or insulin-secreting HEP G2ins cells were fixed for 45 minutes in a fixative modified after Ito and Karnovsky (1968). This fixative contained picric acid and was diluted 3:1 with distilled water. It contained 0.5% betaine hydrochloride, pH 7.8, osmolality 490 mOsM. After fixation, the cells were packed down by vacuum filtration onto a Millipore filter (pore size 0.45 μm) for ease of handling. Buffer washers were with 0.1 M cacodylate buffer containing 6.6% sucrose, pH 7.4. Post- fixation was in 2% osmium tetroxide in the same buffer, followed by 1% tannic acid in the same buffer after further buffer washes. After another buffer wash, samples were dehydrated for 1 hour in absolute ethanol changes, followed by 15 minutes in absolute acetone (to dissolve the Millipore filter) . The packed cells were embedded in Spurr's resin, sectioned grey on a MT-1 ultramicrotome, stained with uranyl acetate and lead citrate, and examined in a Jeol JEM-1010 transmission electron microscope at 80 kv. Static Stimulation of Insulin Secretion; Before stimulation, tissue culture plates were thoroughly washed with basal medium (Dulbecco's Phosphate Buffered Saline containing 1 mM CaCl2 and supplemented with 20 mM HEPES and 2 mg/ml bovine serum albumin (BSA) to remove FCS. Monolayers were incubated in the basal medium at pH 7.4 for three consecutive 1 hr periods to stabilise the basal secretion of insulin. Monolayers were exposed to stimuli for 1 hr periods. Glucose (20 mM) and 5 mM 8-Br-cAMP were dissolved in basal medium. 12-0-tetradecanoylphorbol-13-acetate (TPA.1.3 μM) was dissolved in DMSO and diluted in basal medium (DMSO final concentration 0.8%). Two controls were used: basal medium alone and medium with 0.8% DMSO. Insulin Synthesis; i. Dose response curve to gl ucose After trypsinization a single cell suspension of 10^

HEP G2ins cells was incubated at 37°C (5% C0 /95% 0₂) in vials containing 2 ml Krebs-Ringer bicarbonate buffer (Krebs, 1932) supplemented with 10 mM HEPES, 2 mg/ml BSA (KRB-BSA), 50 μCi/ l L-[4,5,-³H]leucine and 0-20 mM glucose. After 2 hr, the cells were washed in nonradioactive leucine solution and disrupted by sonication in 200 μl distilled water. The amount of labelled insulin was determined by an immunoprecipitation technique (Halban et al, 1980) with guinea pig anti- insulin antibody; guinea pig serum was used as a control. The antibody was precipitated by pansorbin. The total binding capacity of this system was 400 ng. The antibody bound and trichloroacetic acid precipitable radioactivity was estimated by liquid scintillation counting and a sample of the aqueous homogenate was analysed for DNA content (Labarea et al, 1980) . Specificity of the antibody was established by blocking measurement of ^E- insulin with an excess of cold insulin (7 mM) , a similar concentration having no effect. ii . Pulse/ chase labelling of cells After trypsinisation 10^ HEP G2ins cells per aliquot were washed and placed in 100 μl of KRB-BSA containing either 20 mM glucose or 5 mM 8-Br-cAMP as the stimulus or 2.8 mM glucose for control experiments. At time zero, the cells were pulse-labelled at 37°C for 5 min with 1.0 mCi L-[4,5-^H]leucine. Using data from the dose response curve (results) 20 mM glucose was used to stimulate insulin biosynthesis and thus get maximum [³H]-leucine incorporation into newly synthesised endogenous insulin during the 5 min radiolabelling period. Pulse labelling of the cells was stopped by washing in ice cold KRB-BSA and incubated in 2 ml KRB-BSA (containing 2.8, 20 mM glucose or 5 mM 8-Br-cAMP) for 175 min postlabel (chase) incubation at 37°C. The period of pulse labelling was considered as time zero to 5 min, and the chase period a further 175 min, for a total of 180 min. At the 15-, 30-, 45-, 60-, 90-, 120- and 180-min the chase period was ended and the cells centrifuged to form a pellet. The supernatant was removed for analysis and the pelleted cells were sonicated and the total immunoreactive insulin in the chase medium and the cell sonicates analysed by the immunoprecipitation technique outlined above.

Measurement of Glucokinase and Hexokinase Activity;

The glucose phosphorylation activity was determined by measuring the rate of glucose-6-phosphate formation in a modification of the method of Trus et al (1981). HEP G2 (control) and HEP G2ins cells were harvested with trypsin- EDTA, washed with pre-chilled Phosphate Buffered Saline twice to remove glucose and homogenized in ice-cold buffer (pH 7.7) containing 20 mM K2HPO4, 1 mM EDTA, 100 mM KCl and 5 mM dithiotreitol . The homogenate was centrifuged at 12,000 g for 10 min at 4°C, the supernatant was retained to measure glucose phosphorylation activity and protein. Protein was estimated by the Biorad protein assay kit.

The assay volume contained 4 μl of supernatant (containing 7-18.7 μg protein) in 100 μl of HEPES HCI 50 mM (pH 7.7), 100 mM KCl, 7.4 mM MgCl_2. 15 mM β-mercaptoethanol, 0.5 mM β-NAD⁺, 0.05% BSA, 2.5 mg/ml glucose-6-phosphate dehydrogenase from Leuconostoc mesenteroides, 5 mM ATP and glucose concentrations ranging from 0.03 to 100 mM. Glucose was added at 0.03, 0.06, 0.12, 0.25 and 0.5 mM to measure glucokinase activity.

The reaction was terminated after 1 hr at 30°C by adding 1 ml of 500 mM sodium bicarbonate buffer, pH 9.4. NADH fluoresence was then measured at 460 ran (excitation AT 340 n ) . The V_max and Km of the hexokinase and glucokinase activities were calculated in each experiment by the Eadie-Hofstee plots . Fluorescence of both reagent and tissue blanks were subtracted from the total fluorescence. Statistical Analysis; Data were expressed as means + SE and compared with

Student's t test for paired samples. RESULTS

Isolation of clones and production of (pro)insulin;

The levels of human (pro)insulin secreted into the culture supernatant from different stably selected clones of HEPG2 cells transfected with RcCMV-pC∑ differed by as much as 10-fold: 17.3 ± 1.0 pmol/10⁵ cells/day, n=5 (clone 11) to 2.0 + 0.2 (n=3) (clone 25). The HEP G2ins clone 11 was finally selected from 25 different clones isolated with cloning rings and five mixtures of clones to be used for further experimentation.

In other cell lines expressing foreign insulin genes (Laub & Rutter, 1981; Lomedico, 1982; Moore et al, 1983), the immunoreactive material secreted into the medium was proinsulin. To examine the form of insulin secreted by these cells insulin and proinsulin assays were carried out. It can be seen from Table 1 that significantly (P<0.001) more proinsulin than insulin was released daily by the transformed cell line HEP G2ins. Release of these peptides has been stable for greater than six months. Examination of acid/ethanol extracts of the cells however indicated that insulin was preferentially being stored (Table 1). HEP G2 cells transformed with pRcCMV vector alone (Table I) neither contained nor secreted (pro)insulin. Albumin secretion of the transformed cells (53.2±1.0 μg/10^ cells/day) was unaltered from untransformed cells (52.6 ± 0.4 μg/10^ cells/day) (n=3) . Immunohistochemical Analysis;

Immunohistochemical analysis using a polyclonal antibody confirmed that (pro)insulin was being stored (Fig. 1). Electron Microscopy;

Electron microscopy revealed large membrane-bound vacuoles, approximately 2.6 μm in diameter, containing electron-dense material in the HEP G2ins cells (Fig. 2a), while these were not seen in the control HEP G2 cells (Fig. 2b).

Acute Secretion from HEP G2ins Cells is Enhanced by Secretogogues; Eight times the amount of insulin (1.93 + 0.2 pmol/10⁵ cells) compared to proinsulin (0.33 + 0) (n=4) was acutely released following static stimulation with 5 mM 8-Br-cAMP. As would be expected in regulated release on removal of the stimulus insulin levels returned to the basal level of secretion: 0.16 + 0 pmol/10^ cells. Proinsulin levels remained unchanged throughout the experiment. However, the cells did not respond to 20 mM glucose or 1.3 μM TPA. Control cells and culture medium alone, as expected had no insulin present. (Pro)insulin Synthesis; i. Dose response curve to glucose

In response to varying concentrations of glucose from 0 to 20 mM, a dose response curve was obtained for insulin synthesis (Fig. 3) with half maximal synthesis occurring at 4.9 mM glucose and a V_max of 38.0 i 10^ cpm/ μg DNA (n=6) . ii . Pulse/ chase radiolabelling of HEP G2ins cells

Secretion of newly synthesised (pro)insulin that was labelled with [³H]leucine during a 5 minute pulse in response to 20 mM glucose and 5 mM 8-Br-cAMP is presented in Figure 4a. After a 60-min lag period, release of (pro)insulin was detected; this was stimulated significantly (P<0.05) in the presence of glucose and 8-Br-cAMP compared to control cells maintained in buffer containing 2.8 mM glucose (Fig. 4a). The [³H]

(pro)insulin retained in the cells (Fig. 4b) was first detectable at 60 min with no effect on its appearance by either stimuli before 120 min. From 120 min till the termination of the experiment there was a significant (P<0.05) decrease in the amount of [ ] (pro)insulin , retained in the cells in the presence of 20 mM glucose and 5 mM 8-Br-cAMP compared to control cells (Fig. 4b) . Kinetic Characteristics of Glucokinase and Hexokinase in Homogenates of Cells; As shown in Table 2, the activities and Km values of glucokinase in the HEP G2ins and control cells were not significantly different. However, the hexokinase activity of the transformed cells was significantly (P<0.05) lower. DISCUSSION The data presented in this example demonstrates that similar to what has been reported for other non-islet types transfected with the insulin gene (Laub & Rutter, 1993; Lomedico, 1982; Moore et al, 1983), the chronic insulin release of HEP G2ins cells was constitutive and 84% of the material secreted into the medium was proinsulin. Regulated release is that induced by exposure to a secretogogue, in this case 8-Br-cAMP. Following acute stimulation for periods of one hour 86% of the material secreted into the medium was mature insulin not proinsulin. Examination of acid/ethanol extracts of the cells also confirmed that it is insulin and not proinsulin that is preferentially being stored. Immunohistochemical analysis using a polyclonal insulin antibody confirmed that (pro)insulin was being stored. Thus it would appear that HEP G2ins cells have the ability to package insulin into some form of limiting membrane until regulated release occurs . The cells also appear to be functioning normally following transfection, as with regards to albumin secretion which was unaltered from untransfected cells.

The next logical step was to investigate where a hepatoma cell would store the processed mature insulin. Electron microscopy revealed large membrane-bound vacuoles containing electron-dense material in HEP G2ins cells, which were not seen in control cells. This provides morphological evidence that storage of insulin could occur here, but this will have to be determined in future work by immunoelectron microscopy.

It would also appear that HEP G2ins cells exhibit enhancement of insulin synthesis in response to both glucose and elevated levels of cAMP, similar to that seen in a pancreatic β cell (Maldonato et al, 1977). In a normal β cell insulin biosynthesis is primarily stimulated by glucose. The short term effects of glucose on preproinsulin synthesis are restricted to a stimulation of translation and this occurs within minutes of raising ambient glucose (Welsh et al, 1986). Over a longer period, glucose is thought to stimulate transcription and stabilise mRNA. A dose response curve for (pro)insulin biosynthesis of HEP G2ins cells in response to increasing concentrations of glucose was generated with a half V_max of 4.9 mM glucose, not significantly different to that recorded for adult rat islets (Shuit et al, 1988) and human fetal islets (Simpson, abstract, 1991). The sigmoidal appearance of the curve is also similar to the dose response curves generated by these tissues.

Five minute pulse/chase labelling of HEP G2ins cells with [³H]leucine further confirmed insulin synthesis is stimulated in the presence of 20 mM glucose and 5 mM 8-Br- cAMP. The results of these experiments were similar to normal pancreatic beta cells (Rhodes & Halban, 1987).

Newly synthesised (pro)insulin was only released from HEP G2ins cells after a 60 minute lag period. This is most likely a result of the time required from synthesis in the rough endoplasmic reticulum to passage in the Golgi complex, and thence exocytosis. A lag period of 45 minutes before the appearance of newly synthesised insulin was seen in a similar experiment using rat islets (Rhodes & Halban, 1987). The fact that insulin was released in response to elevated levels of glucose and 8-Br-cAMP could imply release via a regulated pathway or simply the fact that translation is stimulated. The [^H] (pro)insulin retained in the cells increased during the early points of the chase, at latter times there was a significant decrease in [³H] (pro)insulin in the presence of elevated glucose and the cAMP analogue 8-Br-cAMP, most likely as a consequence of its release. These results for insulin synthesis indicate that translation ± transcription of the insulin gene is stimulated by glucose and cAMP in HEP G2ins cells.

In general, two pathways for cellular secretion have been defined (Kelly, 1985). Firstly, the constitutive pathway seen in most body cells which involves transport of products from the trans-Golgi followed by fusion with the plasma membrane. This process takes approximately 10 minutes. In specialised secretory cells such as the pancreatic beta cell, in addition to this, a regulated pathway is present which packages products of protein synthesis into secretory granules, conversion of proprotein to mature protein and appropriately releases products by exocytosis in response to a stimulus. All major hormones of the islets of Langerhans (insulin, glucagon, somatostatin and pancreatic polypeptide) are handled by such a regulatory pathway; in contrast to the constitutive pathway products can take hours or days in transition between the Golgi complex and the plasma membrane. The fundamental issue in designing a somatic cell gene delivery system for Type I diabetes is to engineer an appropriate non-islet cell that has the ability on transfection of the insulin gene to regulate insulin expression in response to physiological stimuli. The death from hypoglycaemia of diabetic mice transplanted with constitutively expressed insulin from fibroblast cells (Seldon et al, 1987) underlies the magnitude of the problem.

Like transfected AtT-20 cells (Moore et al, 1983, Simpson et al, 1993) HEP G2ins cells differ from pancreatic beta cells in at least one important way - they do not respond to an acute glucose stimulus . As there is no defect in signal tranduction concerning insulin synthesis it would appear that the glucose insensitivity of HEP G2ins cells is confined to the secretory process. The key elements of the "glucose sensing system" which regulates insulin release from pancreatic β cells in response to small external nutrient changes are the high capacity glucose transporter GLUT 2 (Kasanicki et al, 1990; Thorens, et al, 1990; Newgard et al, 1990; Johnson et al, 1990) and the high capacity glucose phosphorylation enzyme glucokinase (Weinhouse, 1976). Both of these are known to function similarly to pancreatic β cells and liver cells. However, like the AtT-20ins cells (Hughes et al, 1992), the HEP G2ins hepatoma cell lines lack of responsiveness to glucose is probably linked at least in part to failure to express the high K_m islet-liver transporter GLUT 2. The parent HEP G2 cells express the erythrocyte/brain glucose transporter (Permutt et al, 1989) which has 55% sequence homology to GLUT 2 (thorens, 1988), and it is reasonable assumption that the HEP G2ins cells do likewise. Human GLUT 2 may be of paramount importance in augmenting the glucose- sensitive insulin release from any given cell type, as illustrated in work by Ferber et al (1993), who have stably transfected rat GLUT 2 and glucokinase into

RIN 1046-38 cells which lose glucose stimulated insulin secretion with continual passaging in culture. GLUT 2 conferred glucose sensitivity on these cells which exhibited a 4.2-fold increase in activity relative to untransformed cells at >5 mM glucose and a 2.3-fold increase in glucokinase activity. Thus for insulin secretion to occur in response to glucose, evidence suggests that the high K_m glucose transporter isoform in the plasma membrane of beta cells allows the rate of glucose uptake to increase proportionately to the extracellular glucose concentration over a range of 5- 15 mM, therefore the presence of a high-K_m glucokinase is also paramount for normal glucose sensing. Whilst expression of glucokinase activity in HEP G2ins cells is present, the K^ for stimulation over a range of glucose concentrations was only 1.46 mM compared to 7 mM

(Hattersley and Turner, 1993) for islets and 9 mM for liver cells (Meglasson & Matchinsky, 1986). By comparison, the K_m for hexokinase which probably provides an alternative means of controlling glucose phosphorylation in the presence of stimulants such as acetylcholine (Trus et al, 1981), was very similar to that reported for mouse islets using the same technique of measurement as used in this study (Trus et al, 1981). This lack of glucokinase activity in the same range as that for pancreatic β cells most likely also contributes to the lack of glucose responsiveness.

In summary, the results show that the introduction of insulin cDNA into a liver cell line results in synthesis, storage and acute regulated insulin release. However, chronic insulin release was constitutive and the cells did not secrete insulin in response to glucose.

In order to use the transfected liver cell described in this example in the treatment of diabetic humans it appears to be necessary to include an additional gene encoding for glucose transport. This is in order to render it responsive to the same physiological stimuli as the normal beta cell. Of course, utilisation of these cells in humans could only be carried out if they were placed inside immunoprotective capsules. Rejection of these cells would occur without this. Clearly this would not be required if the liver cells were taken from the patient with diabetes, infected with the insulin gene (using, for example, a retrovirus), and these beta cytes then transplanted back into the patient. Rejection should not be a problem as the recipient and donor are the same. Alternatively it may be possible to inject the insulin gene and retrovirus into the blood stream so that it travels to the liver and alters the cells in vivo without the need to remove cells from the patient.

EXAMPLE 2 METHODS

Cell Culture;

HEP G2ins cells were grown in monolayers in Eagles minimal essential medium (MEM) supplemented with 10% fetal calf serum (FCS) and 1 mg/ml G418 antibiotic (presence of G418 maintains presence of insulin cDNA/RcCMV construct through subculture) in air at 37 C.

DNA transfection and selection of transfectants;

The full length human GLUT 2 cDNA was ligated into the multi cloning site of the vector pREP4 (Groger et al . , 1989) which expresses resistance to the eucaryocidal antibiotic hygromycin.

HEP G2ins cells were transfected with 40 μg of the recombinant plasmid and vector. Cells transfected with the vector alone have been used as controls. Transfection was accomplished by electroporation at 200 V and 960 μF in a Biorad gene pulser at a cell concentration of 5 x 10 cells per cuvette, in MEM medium containing no FCS.

Subsequently, 2.5 x 10 were plated into culture dishes in

MEM containing 10% FCS. To obtain stable transfectants of HEP G2ins cells containing GLUT 2 cDNA, 48 hr later 500 μg of the hygromycin antibiotic was added to the culture medium. Medium plus drug was changed every 2-3 days.

After 5 weeks of selection, colonies were picked using cloning rings (HEP G2ins/g cells). Cell extraction and radioimmunoassays (RIA) ;

Samples of culture medium were collected for estimation of insulin and human proinsulin (hPI) by RIA as described in Example 1. Insulin content of the cells was determined after overnight extraction in acid-ethanol. Albumin secretion of the cells was measured using a

Kallestad QM 300 rate nephelometer. RNA preparation and Northern (RNA) blot;

Total cellular RNA was isolated by the method of Wilkinson et al . , 1988,. Northern blot analysis was carried out as described in Simpson et al . , 1995. Immunohistochemical analysis;

Immunohistochemical analysis was carried out on formalin-fixed samples of cultures that had been trypsinised, using Dako LSAB kit and guinea pig anti- insulin (first antibody), a second incubation with anti- rabbit antibody and a final incubation with rabbit anti- guinea pig antibody and the streptavidin-diaminobenzidine chromogen complex. Static simulation of insulin secretion;

Before stimulation, tissue culture plates were thoroughly washed with basal medium [Dulbecco's phosphate buffered saline (PBS) containing 1 mM CaCl₂ and supplemented with 20 mM HEPES and 2 mg/ml bovine serum albumin (BSA)] to remove culture medium and FCS. Monolayers were incubated in the basal medium at pH 7.4 for three consecutive 1 hr periods to stabilise the basal secretion of insulin. Monolayers were then exposed to stimuli for 1 hr. Glucose (20 mM) , 5 mM 8-Br-cAMP and 10 mM theophylline were dissolved in basal medium, 12-0- tetradecanolyphorbol-13-acetate (TPA: 1.3 μm) was dissolved in DMSO and diluted in basal medium (DMSO final concentration 0.8%). Calcium (10 mM) was dissolved in basal medium without phosphate. Three controls were used - basal medium alone, medium with 0.8% DMSO and basal medium without phosphate. The dose response curve to insulin was measured in a similar fashion. The basal level of insulin was established and monolayers were exposed to increasing concentrations of glucose from 0-20 mM. Insulin secretion of cells perifused with 20 mM glucose; Cells were trypsinised and attached to cytodex beads and cultured in MEM for a period of 5 days. The cells, attached to the beads were then transferred to a perifusion apparatus attached to a fraction collector and basal medium was pumped round the cells . A basal level of insulin secretion was established over a 30 min period. This was followed by a 30 min stimulus with 20 mM glucose followed by a further 30 min basal period. This system was designed to more closely monitor insulin secretion on a minute-to-minute basis.

Insulin Synthesis; Dose response curve to glucose - after trypsinisation 10

HEP G2ins/g cells were incubated at 37°C (5% C0₂/95% 0₂) in vials containing 2 ml Krebs-Ringer bicarbonate buffer supplemented with 10 mM HEPES, 2 g ml BSA (KRB/BSA) ,

3 50 μCi/ml L-[4,5- H] leucine and 20 mM glucose. After 2 hr, the cells were washed with KRB-BSA containing 0.05% nonradioactive leucine and disrupted by sonication in 200 μl distilled water. The amount of labelled insulin was determined by the immunoprecipitation technique as described in Simpson et al . , 1995. Statistical Analysis;

Data were expressed as means + S.E., paired sample means were compared with Student's t-test. RESULTS; Northern Blot analysis of cellular RNA; (Pro)insulin mRNA was clearly abundant in the RNA from HEP G2ins/g cells (Fig. 5), but completely absent from HEP G2 (untransfected cells), HEP G2 control cells (transfected with CMV and REP 4 vectors alone) and HEP G2ins cells (transfected with insulin cDNA alone) . Chronic Secretion and storage of insulin;

In other cell lines expressing foreign insulin genes, the immunoreactive material secreted into the medium was proinsulin and/or its split products . To examine the form of insulin secreted by our cells insulin and proinsulin assays were carried out. It can be seen from Table III, that three times more proinsulin than insulin was released daily by the transformed HEP G2ins/g cells. By comparison the parental HEP G2ins cell line (does not possess GLUT 2) chronically secreted a greater proportion of proinsulin (6X) compared to mature insulin. Transfection of HEP G2 ins cells with the GLUT 2 cDNA (HEP G2ins/g) also resulted in a 6-fold increase in intracellular insulin content (Table III) compared to ells lacking GLUT 2. The proportion of insulin : proinsulin was also greater (14.6 : 1) in the HEP G2ins/g cells compared to 3.6 : 1 in the HEP G2ins line.

Immunohistochemical analysis using a polyclonal antibody further confirmed that (pro)insulin was being stored.

Albumin secretion of the transformed cells (52.5 +_0.3 μg/10 cells) was unaltered from the untransfected cells (52.6 ± 0.4). Acute insulin secretion;

Nine times the amount of insulin compared with proinsulin was released acutely following static stimulation with 20 mM glucose (Fig. 6). As would be expected in regulated release, on removal of the stimulus, insulin levels returned to basal. Proinsulin levels remained unchanged throughout the experiment. Similar results were seen when arginine, the cAMP analogue 8-Br- cAMP and the phosphodiesterase inhibitor, theophylline were applied as a stimulus (Fig. 7). However, the cells did not respond to TPA or calcium.

In response to varying concentrations of glucose from 0-20 mM, a dose response curve for insulin secretion was generated (Fig. 8). While glucose responsiveness commenced at a lower concentration than normal islets a secretion curve approaching normal physiological conditions was generated. Perifusion; Results of perifusion experiments are illustrated in

Fig. 9. It can be seen from this figure that there is a definite increase in insulin secretion at 30 min on the addition of 20 mM glucose, which returns to basal levels on removal of the stimulus . DISCUSSION; This example examined whether transfection of GLUT 2 cDNA, which is absent from HEP G2ins cells, will allow glucose to exert an effect on insulin secretion. It was found that the cell line established from this transfection (HEP G2ins/g cells) were capable, following synthesis, to process proinsulin to insulin and store it until regulated insulin release occurs .

The HEP G2ins/g cells also secrete and store significantly greater amounts of (pro)insulin than the parental line (HEP G2ins) .

Thus, for the first time, regulated release of insulin to the physiological stimulus glucose has been demonstrated in a liver cell line. While glucose responsiveness commenced at a lower concentration than normal islets, a secretion curve approaching normal physiological conditions was generated.

These observations indicate that primary liver cells, which naturally express GLUT2, may serve as useful vehicles for the gene therapy of diabetes.

REFERENCES

1. Eisenbarth, G.S. 1986. Type I diabetes mellitus: a chronic autoimmune disease. N. Engl . J. Med. 314:1360-1368.

2. Seldon, R.F., Skoskiewicz, M.J., Russell, P.S., and Goodman, H.M. 1987. Regulation of insulin gene expression: Implications for gene therapy. N. Engl. J. Med. 317:1087-1076. 3. Moore, H-P.H., Walker, M.D., Lee, F., and Kelly,

R.B. 1983. Expressing a human proinsulin cDNA in a mouse ACTH-secreting cell. Intracellular storage, proteolytic processing, and secretion on stimulation. Cell. 35:531-538. 4. Hughes, S.D., Quaade, C, Milburn, J.L.,

Cassidy, L., and Newgard, C.B. 1991. Expression of normal and novel glucokinase mRNAs in anterior pituitary and islet cells. J. Biol. Chem. 266:4521-4530.

5. Hughes, S.D., Johnson, J.H., Quaade, C, and Newgard, C.B. 1992. Engineering of glucose-stimulated insulin secretion and biosynthesis in non-islet cells. Proc. Natl. Acad. Sci. USA. 89:688-692.

6. Simpson, A.M., Hand, H., Clarke, R. , Burn, A., and Tuch, B.E. 1993. Transformation of pituitary and fibroblast cell lines using insulin cDNA and a dexamethasone-inducible promoter. Trans. Proc. 25:2915-2916.

7. Grossman, M. , Raper, S.E., and Wilson, J.M. 1991. Towards liver-directed gene therapy: retrovirus-mediated gene transfer into human hepatocytes. Som. Cell. Mol. Gen. 17:601-607.

8. Knowles, B.B., Howe, CC, and Aden, D.P. 1980. Human hepatocellular carcinoma cell lines secrete the major plasma proteins and hepatitis B surface antigen. Science 209:497-499. 9. Tuch, B.E., Roberts, E.C, and Darby, K.B. 1992. Release of proinsulin from the human fetal β cell.

J. Endo. 132:159-167.

10. Ito, S., and Karnovsky, K. 1968. Formaldehyde- glutaraldehyde fixatives containing trinitro compounds . J. Cell. Biol., 39:168 (Abstr) .

11. Krebs, H.A. , and Henseleit, K. 1932. Untersuchungen uber die Harnstoffbildung im Tiekorper. Hoppe Seyler Z. Physiol. Chem. 210:33-36. 12. Labarea, C, and Paigen, K. 1980. A simple, rapid and sensitive DNA assay procedure. Anal. Biochem. 102:344-352.

13. Laub, 0., and Rutter, W.J. 1983. Expression of the human insulin gene and cDNA in a heterologous mammalian system. J. Biol. Chem. 258:6043-6050.

14. Lomedico, P.T. 1982. Use of recombinant DNA technology to program eucaryotic cells to synthesise rat proinsulin - a rapid expression assay for cloned genes. Prog. Nat. Acad. Sci. USA. 79:5798-5803. 15. Maldonato, A., Renold, A.E., Sharp, G.W.G. , and

Cerasi, E. 1977. Glucose-induced proinsulin giosynthesis . Role of islet cyclic AMP. Diabetes 26, 538-545.

16. Welsh, M. , Scherberg, N., Gilmore, R. , and Steiner, D.F. 1986. Translation control of insulin biosynthesis: Evidence for regulation of elongation, initiation and signal-recognition-particle-mediated translational arrest by glucose. Biochem. J. 235:459.

17. Schuit, F.C, IN 'T Veld, P.A., and Pipeleers, D.G. 1988. Glucose stimulates proinsulin biosynthesis by a dose-dependent recruitment of pancreatic beta cells. Proc. Natl. Acad. Sci. USA. 85,3865-3869.

18. Simpson, A.M., Tuch, B.E., and Vincent, P.C 1991. Control of insulin biosynthesis in the human fetal beta cell. Annual Meeting of the Australian Diabetes Society, Sydney, 21 (Abstr). 19. Rhodes, C.J., and Halban, P.A. 1987. Newly synthesised proinsulin/insulin and stored insulin are released from pancreatic β cells predominantly via a regulated, rather than a constituitive pathway. J. Cell. Biol. 105:145-153.

20. Kelly, R.B. 1985. Pathways of protein secretion in eukaryotes . Science 230:25-

21. Kasaniciki, M.A. , and Pilch, P.F. 1990. Regulation of glucose-transporter function. Diabetes Care 13:219-227.

22. Thorens, B., Charron, M.J., and Lodish, H.F. 1990. Molecular Physiology of glucose transporters. Diabetes Care 13:209-218.

23. Newgard, C.B., Quaade, C, Hughes, S.D., and Milburn, J.L. 1990. Glucokinase and glucose transporter expression in liver and islets. Implications for control of glucose homeostasis. Biochem. Soc. Trans. 18:851-853.

24. Johnson, J.H., Ogawa, A., Chen, 1., Orci, L., Newgard, C.B., Alam, T., and Unger, R.H. 1990. Underexpression of β cell high Km Glucose transporter in noninsulin-dependent diabetes. Science 250:546-549.

25. Weinhouse, S. 1976. Regulation of glucokinase in liver. In Current Topics in Cellular Regulation. B.L. Horecker, and E.R. Stadtman, editors. Academic Press/ New York, San Francisco, London. Vol. 11:1-50.

26. Permutt, M.A. , Koranyi, L. , Keller, K. , Lacy, P.E., Sharp, D.W., and Mueckler, M. 1989. Cloning and functional expression of a human pancreatic islet glucose- transporter cDNA. Proc. Natl. Acad. Sci. USA. 86:8688-8692.

27. Thorens, B., Sarkar, H.K. , Kaback, H.R., and Lodish, H.F. 1988. Cloning and functional expression in bacteria of a novel glucose transporter present in liver, intestine, kidney and β-pancreatic islet cells. Cell 55:281-290. 28. Ferber, S., Johnson, J., Beltrandelrio, H., Hughes, S., Clark, S. Chick, and Newgard, C.B. 1993. Glucose sensing in GLUT-2 and glucokinase transfected RIN cells. Diabetes Suppl . 1:31. 29. Meglasson, M.D., and Matschinsky, F.M. 1986.

Pancreatic islet glucose metabolism and regulation of insulin secretion. Diabetes Metab. Rev. 2:163-214.

30. Trus, M.D., Zawalich, W.S., Burch, P.T., Berner, D.K. , Weill, V.A. , and Matchinsky, F.M. 1981. Regulation of glucose metabolism in pancreatic islets. Diabetes 30:911-922.

31. Hattersley, A.T. and Turner, R.C, 1993. Mutations of the glucokinase gene & Type 2 diabetes. Quart. J. Med. 86:227-232 32. Groger, R.K. , Morrow, D.M. , Tykocinski, M.I., 1989. Directional antisense and sense cDNA cloning using Epstein-Barr virus episomal expression vectors. Gene 81:285.

33. Wilkinson, M. , 1988. RNA isolation: a mini-prep method. Nucleic Acids Res 16: 10933.

34. Simpson, A.M., Tuch, B.E., Swan, M.A. , Tu, J., Marshall, G.M. 1995. Functional expression of the human insulin gene in a human hepatoma cell line (HEP G2) . Gene Therapy 2:1-9.

TABLE I

Detection of Insulin and Proinsulin in

Culture Supematants and Cell Extracts from HEP G2ins and HEP G2 Cells

Culture Supernatant Cell Extract

Cell Insulin Proinsulin Insulin Proinsulin

HEP G2ins 13.8 + 3.1 78.2 + 4.6* 29.4 ±5.2 7.7 ±1.6*

Control: HEP G2 0 0 0 0

Insulin and proinsulin from the culture supernatant

(pmol/lθ5 cells/day) was compared with that isolated from the cell extracts (pmol/10⁵ cells) in transformed

HEP G2ins and control HEP G2 cells. Results expressed as mean ± S.E. for five observations.

* P<0.001 (proinsulin vs. insulin).

TABLE II

Kinetic Characteristics of Glucokinase and Hexokinase in HEP G2 and HEG dins Cells

Gluco inase Hexo cinase

Cell Vmax Km(mM) Vmax Km(mM)

HEPG2 110.0 + 30.0 1.2 + 0 208.0 ± 11.5 0.05 + 0

HEP G2 INS 90 ±2.3 1.5 ±0.3 139.0 ±7.5* 0.06 + 0

The V_max values for glucokinase and hexokinase

(nmol. min~l.mg protein-^) were calculated from V vs. V/S plots at different glucose levels (see methods) and 5 mM

ATP. Values are means + S.E. for 3 observations.

* (P=0.0474) vs. control HEP G2 cells. TABLE HI

Culture Supernatant Cell Extract (pmol/10⁵ cells/day) (pmol/10⁵ cells/day)

Cell Insulin Proinsulin Insulin Proinsulin

HEP G2ins/ g 15.5 + 3.9 50.1 + 2.4 19.0 ± 2.2 1.3 + 0.2

HEP G2ins 1.4 + 0.3 7.8 + 0.5 2.9 ± 0.5 0.8 ± 0.2

Control: HEP G2 0 0 0 0

Immunoreactive insulin and proinsulin from culture supematants was compared with that isolated from the cell extracts in transfected HEP G2ins/ g, Hep G2ins and HEP G2 (control: transfected with CMV and REP 4 vectors alone) cells.

Results expressed as mean + for 6 experiments. P < 0.001

It will be appreciated by persons skilled in the art that numerous variations and/or modifications may be made to the invention as shown in the specific embodiments without departing from the spirit or scope of the invention as broadly described. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive.

Claims

CLAIMS : -

1. A method of treating Type I diabetes in a subject comprising transfecting an hepatocyte with a gene encoding insulin, the gene being under control of an appropriate promoter, and introducing the transformed hepatocyte into the subject.

2. A method according to claim 1, wherein the hepatocyte is a primary hepatocyte isolated from said subject.

3. A method according to claim 1 or 2, wherein the hepatocyte is further transfected with a homogeneous or heterogeneous gene encoding glucose transporter GLUT 2.

4. A method according to any one of the preceding claims, wherein the hepatocyte is further transfected with a homogeneous or heterogeneous gene encoding the high capacity glucose phosphorylation enzyme glucokinase.

5. A method according to claim 1 or 2, wherein the hepatocyte is further transfected with either homogeneous or heterogeneous genes encoding the glucose transporter GLUT 2 and the high capacity glucose phosphorylation enzyme glucokinase.

6. A method according to any one of the preceding claims, wherein the subject is a mammal.

7. A method according to any one of claims 1 and 3 to 6, wherein the hepatocyte is a hepatoma cell line.

8. A method according to claim 7, wherein the hepatoma cell line is HEP G2.

9. A method according to claim 7 or 8, wherein the transfected hepatoma cell line is introduced to the subject through the use of immunoprotective capsules.

10. A transformed hepatocyte for use in treatment of Type I diabetes, the hepatocyte including a gene encoding insulin, the gene being under the control of an appropriate promoter.

11. A transformed hepatocyte according to claim 10, further including either an homogeneous or heterogeneous gene encoding glucose transporter GLUT 2.

12. A transformed hepatocyte according to claim 10, further including either an homogeneous or heterogeneous gene encoding the high capacity glucose phosphorylation enzyme glucokinase.

13. A transformed hepatocyte according to claim 10, further including either homogeneous or heterogeneous genes encoding the glucose transporter GLUT 2 and the high capacity glucose phosphorylation enzyme glucokinase.

14. A mammalian transformed hepatocyte according to any one of claims 10 to 13.