A 338-bp proximal fragment of the glucose transporter type 2 (GLUT2) promoter drives reporter gene expression in the pancreatic islets of transgenic mice

ELSEVIER Molecular and Cellular Endocrinology 114 0995) 205-215

~lecular and

C ellular f%docrhology

A 33%bp proximal fragment of the glucose transporter type 2 (GLUT2) promoter drives reporter gene expression in the

pancreatic islets of transgenic mice

Gerard Waeber * a , Thierry Pedrazzinib, Olivier Bonny”, Christophe Bonny”, Myriam Steinmanna, Pascal Nicod”, Jacques-Antoine Haefliger”

aDepaltment of Internal Medicine B, Laboratory of Molecular Biology 19-533, Centre Hospitalier Uniwrsitaire Vaudois, CHW-1011 Luusanne, Switzerland

bDi&sion of Hypertension, Lhiuersity Hospital of Lausanne, Lausanne, Switzerland

Received 31 May 1995; accepted 7 August 1995

Abstract

The high K, glucose transporter GLUT2 is a membrane protein expressed in tissues involved in maintaining glucose homeostasis, and in cells where glucose-sensing is necessaq. In many experimental models of diabetes, GLUT2 gene expression is decreased in pancreatic p-cells, which could lead to a loss of glucose-induced insulin secretion. In order to identify factors involved in pancreatic p-cell specific expression of GLUT2, we have recently cloned the murine GLUT2 promoter and identified &-elements within the 33%bp of the proximal promoter capable of binding islet-specific tins-acting factors. Furthermore, in transient transfection studies, this 338-bp fragment could efficiently drive the expression of the chloramphenicol acetyl transferase (CAT) gene in cell lines derived from the endocrine pancreas, but displayed no promoter activity in non-pancreatic cells. In this report, we tested the cell-specific expression of a CAT reporter gene driven by a short (338 bp) and a larger (1311 bp) fragment of the GLUT2 promoter in transgenic mice. We generated ten transgenic lines that integrated one of the constructs. CAT mRNA expression in transgenic tissues was assessed using the RNAse protection assay and the quantitative reverse transcribed polymerase chain reaction CRT-PCR). Overall CAT mRNA expression for both constructs was low compared to endogenous GLUT2 mRNA levels but the reporter transcript could be detected in all animals in the pancreatic islets and the liver, and in a few transgenic lines in the kidney and the small intestine. The CAT protein was also present in Langerhans islets and in the liver for both constructs by immunocytochemistry. These findings suggest that the proximal 338 bp of the murine GLUT2 promoter contain c&elements required for the islet-specific expression of GLUTZ

Keywords: Glucose transporter; Gene expression; Promoter; Transgenic; Diabetes mellitus

lhtroduction

The high K, glucose transporter GLUT2 is a membrane protein that facilitates glucose diffusion through plasma membranes of mammalian cells and therefore plays a key role in carbohydrate homeostasis (Carruthers, 1990; Unger, 1991; Thorens, 1992a).

*Corresponding author, Tel.: 21 314 11 11 Ext. 740968, Fax: 21 314 0630.

The GLUT2 gene is expressed in the liver, the smal1 intestine, the kidney, and in the pancreatic P-cells (Thorens et al., 1988; Thorens, 1992a). Cell-specific expression of GLUT2 is also modulated in vivo and in vitro by changes in glucose concentration. Hyper- glycemia has been shown to increase GLUT2 gene expression, whereas hypoglycemia decreases GLUT2 mRNA accumulation (Chen et al., 1990; Kahn et al., 1990; Yasuda et al., 1992; Ferrer et al., 1993; Postic et al., 1993; Purrello et al., 1993; Waeber et al., 1994). In various rodent models of diabetes such as the Zucker

0303-7207/95/$09.50 0 1995 Elsevier Science Ireland Ltd. All rights reserved. SSDI 0303-7207(95)3662-Q

206 G. Waeber et al. /Molecular and Cellular Endocrinology 114 (1995) 205-215

fa/fa rat, the neonatal low-dose STZ-induced diabetic rat, the GK rat, the BB/W rat or in the db/db mouse, the immunodetectable GLUT2 expression is markedly decreased in pancreatic p-cells while it re- mains unchanged or is slightly elevated in the liver, the intestine and the kidneys (Johnson et al., 1990; Orci et al., 1990a,b; Thorens et al., 1990, 1992b; Unger, 1991; Ohneda et al., 1993). The p-cell-specific reduction of GLUT2 expression has been associated with a 50-75% decrease of GLUT2 mRNA (Johnson et al., 1990; Ohneda et al., 1993). One could conclude that hyperglycemia of diabetic animals is responsible for the observed decrease in expression of GLUT2 in the endocrine pancreas. However, experiments in which development of hyperglycemia was prevented by ascarbose treatment in Zucker diabetic rats showed that even in the presence of normoglycemia, GLUT2 expression steadily decreased over time (Orci et al., 1990a). Thus, the above in vivo and in vitro observa- tions suggest that high glucose level normally stimu- lates GLUT2 gene expression; however, in diabetes, there is an unidentified factor which causes a decrease in GLUT2 gene expression in the pancreatic p-cells, despite ambient hyperglycemia. It has been suggested that this decrease in GLUT2 expression could further limit glucose-induced insulin secretion, a hallmark of the diabetic state (Unger, 1991). This hypothesis has recently been supported by experiments with transgenic mice expressing GLUT2 antisense RNA under the control of the rat insulin promoter: these animals showed an 80% reduction in GLUT2 in p-cells, an impaired glucose-induced insulin secretion and developed diabetes (Valera et al., 1994).

The mechanisms of tissue- and cell-specific transcriptional control involve the binding of regulatory proteins (trans-acting factors) to specific DNA sequences (&elements). The variety of binding sites for truns-acting factors and their combinatorial and spatial arrangements suggest that regulation of gene transcription results from complex interactions between multiple trans-acting factors (M&night et al., 1982). In order to identify the &elements and the puns-acting factors involved in pancreatic p-cell- specific regulation of GLUT2 expression, we have recently cloned the murine 5’-flanking region of the GLUT2 gene (Waeber et al., 1994). The mouse and the human GLUT2 promoters share regions of nucleic acid sequence identity and have similar transcriptional glucose-responsiveness, as tested by the transfection of fusion constructs of each promoter with a chloramphenicol acetyltransferase (CAT) reporter gene into a differentiated insulin-secreting cell line. Furthermore, transient expression studies show that a 338bp fragment of the murine GLUT2 promoter can efficiently drive the expression of a CAT

reporter gene only in cell lines derived from the endocrine pancreas, but that the promoter activity has no activity in non-pancreatic cells. Several islet-specific trans-acting factors have been identified which are able to bind &s-elements in the proximal 338 bp of the murine GLUT2 promoter (Bonny et al., 1995). These data support the concept that 338 bp of this promoter contain the required regulatory information for correct spatial expression of the GLUT2 gene.

The major goal of this study was to test the cell- specific expression of a short (338 bp) and a larger (1311 bp) fragment of the GLUT2 promoter fused to a CAT reporter gene in a transgenic mouse model. Based on the transfection studies using insulinoma cells described above, one could predict that the short (338 bp) promoter would drive reporter gene expression in pancreatic cells of the transgenic mice.

This manuscript describes the generation of several transgenic lines that have integrated either - 1311/+ 49 or -338/+ 49 bp of the murine GLUT2 promoter linked to a CAT reporter gene. In the transgenic animals, the CAT mRNA expression was assessed using RNAse protection assays and quantitative reverse transcribed polymerase chain re- actions (RT-PCR). CAT mRNA expression was low compared to endogenous GLUT2 mRNAs, but was present in all tested animals in the pancreatic islets and the liver. CAT expression assessed by comparison with GLUT2 or p-actin mRNAs was similar in the 1311 and 338 bp transgenic pancreatic islets suggesting that the small promoter region can drive expression of the reporter transgene in the endocrine pancreas as efficiently as does the larger one. In addition, the CAT transcripts are translated into immunodetectable CAT in the Langerhans islets and the liver of the transgenic lines. This is consistent with the hypothesis that our promoter fragments contain most required &elements for their transcriptional control in normally expressing GLUT2 tissues such as the liver and the endocrine pancreas.

2. Methods

2.1. Fusion gene constructs The cloning of the murine upstream regulatory

region of the GLUT2 gene was described previously (Waeber et al., 1994). The promoter - 1311 to + 49 bp was subcloned into the SaZI/xbaI sites of the promoterless pCATbasic vector (Promega). For microinjection studies, two GLUE/CAT fragments were purified. A SalI/BamHI and PstI/BamHI which contain - 1311/ + 49 and - 338/ + 49 bp, respectively, of the murine GLUT2 promoter fused to the CAT gene and the SV40 polyadenylation sequence. Before microinjection, the DNAs were purified by agarose gel electrophoresis, electroeluted, pre-

G. Waeber et al. /Molecular and Cellular Endocrinology 114 (1995) 205-215 207

cipitated and resuspended at a concentration of 3 pg/ml in 5 mM Tris-HCl, pH 7.4, 0.1 mM EDTA.

2.2. Generation of transgenic mice and identification of the posit& lines

Fertilized eggs, obtained from superovulated 6- week-old NMRI female mice (IFFA CREDO, L’Arbresles, France), were injected into the male pronuclei at a constant pressure of 2 psi. Viable embryos were reimplanted in the oviduct of pseu- dopregnant NMRI mice. At 3 weeks of age, 2 cm of the tail were cut from the offspring and digested overnight at 55°C in 0.7 ml of Tris-HCl, pH 8.0, 100 mM EDTA, 0.5% SDS containing 500 pg/ml pro- teinase K. DNA was purified by phenol/chloroform extraction and ethanol precipitated. Transgenic mice were identified by dot blot analysis, Southern blot analysis or PCR. For the dot blot analysis, the tail DNA was denatured in 0.4 M NaOH, 100 mM EDTA at 100°C for 10 min and transferred to a Zeta-probe nitrocellulose GT blotting membrane (Bio-Rad) using a Bio-Dot SF blotting apparatus (Schleicher and Schuell minifold apparatus). The membrane was prehybridized for 30 min at 65°C in presence of 0.5M NA,HPO,, pH 7.2,0.1 mM EDTA and 7% SDS and hybridized overnight at 65°C in the same buffer. Southern blot analysis was done by an overnight digestion by &a1 of 10 pg of the tested genomic DNA followed by electophoresis to separate the digested DNA on a 0.8% agarose gel. The Southern gel was depurinated by soaking the gel in 0.25 N HCl for 15 min and then denatured in 0.5 N NaOH for 30 min. The DNA was transferred by diffusion blotting (10 x

SSC) to a Zeta-Probe GT membrane (Bio-Rad). The filters were then baked for 2 h at 80°C prehybridized and hybridized as described with the dot blot screening. A 1644-bp XbaI/BamHI fragment containing the CAT gene sequences was isolated from the pCAT basic vector and used as a probe. The insert was labelled with [a-32 P]dCTP (Amersham) using a ran- dom primed labelling kit (Boehringer). PCR screening of the putative positive lines was done using CAT specific oligonucleotides (A = 5’- CCTAAACGCCTGGTGCTACGCCTG-3’, B = 5’- CATAATITTCTI’GTATAGCAGTGCA-33 with 1 pg of genomic DNA amplified for 30 cycles with the following conditions: 94°C for 1 min, 58°C for 1 min and 72°C for 1 min.

2.3. Transgenic RNA isolation Mouse tissues were homogenized in 9 ml of 4 M

guanidium isothiocyanate buffer with a Kinametic polytron blender (Kriens, Switzerland) and layered onto a 4 ml 5.7 M CsCl cushion. RNAs were pelleted at 33 000 rev./min for 17 h in a 50 Ti rotor, RNA was isolated from mouse islets by the method of Gotoh et

al. (1987). Before analysis, RNAs were DNAse I treated (Pharmacia) followed by a phenol/chloroform extraction and ethanol precipitation to eliminate trace amounts of genomic DNA. All RNAs were obtained from either normal or transgenic NMRI mice at age lo-14 weeks. All procedures involving husbandry, production and analysis of transgene mice were ap- proved by the local Purdue Animal Care and Use committee (authorization VD 909).

2.4. Ribonuclease (RNAse) protection assay A probe was constructed which contained a 262 bp

XbaI/EcoRI fragment of the CAT cDNA together with a PCR-generated murine GLUT2 cDNA fragment (Sal1 and xhol digested) that includes bp 202-450 of the GLUT2 cDNA. Both fragments were subcloned into the pBluescript KS + vector (Strata- gene) and linearized with XbaI. The antisense RNA probe was synthesized using T3 RNA polymerase (Promega). The CAT antisense probe was generated using the same 262-bp XbaI/EcoRI fragment of the CAT cDNA subcloned into the pBluescript Ks + vector. The RNAse protection assay was carried out according to the protocol of the RPA II kit (Ambion) using 30 pg of total RNA of the different transgenic RNAs. The product of the RNAse protection assays were separated on a 6% polyacrylamide/urea sequencing gel. A sequencing reaction was used as a sizing marker.

2.5. Rewrse-transcription polymeruse chain reaction CRT-PCR)

One pg of DNAse-treated islet RNA was reverse transcribed using an oligo dT primer in 50 mM KCl, 10 mM Tris-HCl, pH 8.3, 1.5 mM MgCl,, 0.01% gelatin and 1 mM of each dNTPs. The reverse transcribed products were amplified by a PCR reaction in the presence of the CAT, p-actin and GLUT2 specific oligonucleotides and a trace amount of ‘*P- or 33P- labelled OdCTP with the following conditions: 94°C for 1 min, 58°C for 1 min and 72°C for 1 min. Forty cycles of amplification were used and 5 ~1 aliquots were removed every five to 10 cycles. The PCR generated products were analyzed on a 6% sequencing polyacrylamide gel with a sequencing reaction as size marker. To generate the PCR fragments, specific oligonucleotides were designed. For the p-actin: primer 1 sequence is 5’-TGGCACCACACCTTC- TACAATGAG-3’; and primer 2 sequence is 5’-GCT- TCTCTITGATGTCACGCACG-3’. The GLUT2 specific oligonucleotides are: primer (1) 5’-TC- CACTGTTCCACTGGATG-3’ and primer (2) 5’- AGTITGTCCCCGAGCCAC-3’. For the CAT expression the same oligonucleotides were used as for the identification of the transgenic mice. Quantifica-

208 G. Waeber et al. /Molecular and Cellular Endocrinology 114 (1995) 205-215

tion of the PCR-generated fragments was done by electronic autoradiography with an Instant Imager 2024 (Packard Instrument Co.) or using a Molecular Dynamics scanner (Sunnyvale, CA).

2.6. Immunocytochemisby Adult mice under deep anesthesia were perfused

with phosphate-buffered saline (PBS) and pancreata were quickly removed, incubated for 6 h in 4% paraformaldehyde and processed for paraffin embed- ding. Sections of 8 pm were used for immunocytochemistry. In brief, sections were incubated in the presence of 5% H,O, for 30 min at room temperature, washed in PBS and incubated overnight in the presence of 5% bovine serum albumin dissolved in PBS buffer containing 0.5% Triton X-100 and 2% goat serum. The sections were then incubated for 20 h at 4°C with the primary antibody (rabbit polyclonal anti-CAT from SPrime3Prime, Inc. Boulder, CO; diluted lo-fold in phophate-buffered saline containing 0.5% Triton X-100 and 2% goat serum), then for 4 h at room temperature with the secondary antibody (biotinylated goat anti-rabbit IgG from Vector/Re-

actolab S.A. diluted 200-fold) and for 2 h at room temperature with avidin-biotinylated peroxidase complex (ABC, Vector Reactolab SA, diluted lOOO-fold). The peroxidase reaction was finally visualized with 3,3’-diaminobenzidine tetrahydrochloride dihydrate (FLUKA) in PBS containing 0.2% H,O,. Pho- tographs were taken with a Leica microscope using transmitted light optics and Kodak Ektachrome EPJ 320 films.

3. Results

3.1. Generation of transgenic mice carrying - 1311/ + 49 or -338/ + 49 bp of the murine GLUT2 promoter fused to the CAT reporter gene

- 1311 to + 49 and - 338 to + 49 bp of the murine promoter region of GLUT2 were individually cloned into the promoterless pCAT-basic vector. A SalI- BamHI and a PstI-BamHI fragment, respectively, were used for microinjection into the male pronuclei of fertilized mouse eggs. The two fragments comprise 1311 or 338 bp of the 5’-flanking region of the murine

1

CAT } I

I Transgene (-1311/+49 GLUTSCAT)

R R R

Transgene (-338/+49 GLUTPCAT)

Digested genomic DNA (bp)

Probe

Fig. 1. Identification of the founders by Southern blot analysis. A typical Southern blot analysis of RraI-digested mice genomic DNA is shown

in panel A. Using a CAT probe, the RFaI digestion detects a 538- and 662-bp fragment when the transgene is integrated in the mouse line

(panel B). The Southern blot shows an example of identification of founder 13 and three positive offspring (lanes I,4 and 5) and two negative

offsprings (lanes 2 and 3).

G. Waeber et al. /Molecular and Cellular Endocrinolqy 114 (1995) 205-215 209

Table 1 Characterization of the transgenic lines

A. - 1311/ + 49 GLUIXAT Founders Sex

1 M 5 F 8 F

13 M 16 F 68 F 71 M 72 M

Integrated copies(n) Z-10 l-2 l-2 l-2 l-2 l-2

10-20 ?

Transmission + + + + + + + -

B. - 338/ + 49 GLUTXAT 41 F l-2 + 46 M Z-10 +

GLUT2 gene, the main transcription start site for GLUT2, the CAT gene and the SV40 polyadenylation sequence. Identification of the founders was done by Southern blot analysis of the genomic DNA. A typical blot is shown in Fig. 1 where identification of founder 13 and 3 offspring was done using a CAT probe which detects two RFaI fragments of 662 and 538 bp of positive transgenic genomic DNA. The transmission

of the transgene to the progeny was followed by dot blot hybridization or PCR of tail genomic DNA. Table 1 summarizes the different mice lines obtained with the two constructs. Eight positive founders that integrated multiple copies of the - 1311/ + 49 CAT constructs were identified and only one mouse line, num- ber 72, was unable to generate positive offspring suggesting a non-germinal integration of the trans-

T7 T3

probe t sense RNA

pos. control (274 bp)

CAT mRNA (262 bp)

Fig. 2. Transgene mRNA analysis by RNAse protection assay. (Al A typical RNAse protection assay of DNAse treated RNAs extracted from a wild type. (WT) or a transgenic line (1311CATl. The CAT transcript is detected only in the transgenic line. The spatial expression of the transgene shows CATmRNAs in the liver and in RNA obtained from a stabIy transfected insulin-producing ceh line (hneoCAT-INSl) with the - 1311 to +49 bp of the GLUT2 promoter Iinked to the CAT reporter gene (Waeber et al., 1994). (B) The RNAse assay, using a cRNA probe that includes a fragment of the CATcDNA, is shown schematically. The sixes of the protected fragments after RNAse A and Tl digestion are shown for the CAT mRNA and the positive control.

210 G. Waeberet al. /Molecular and Cellular Endocrinology 114 (1995) 205-215

gene. Two mice lines were generated that integrated the - 338/ + 49 bp CAT construct. Other than founder 72, all other transgenic lines followed an autoso- ma1 inheritance pattern. No obvious phenotypic dif- ference could be observed in the transgenic popula- tion and the development appeared normal.

3.2. Analysis of CAT expression by RNAse protection assays

As CAT activity measured in the various transgenic tissues was low and could not be used as a quantitative assay, we designed a ribonuclease protection assay using a radiolabelled RNA probe which contains a fragment of the CAT gene (262 bp) to determine the pattern of CAT mRNA expression. As schematically represented in Fig. 2B, the probe detects, after RNAse A and Tl digestion, a 262-bp fragment when CAT mRNA is present in the assay. As a positive control, we generated a cold sense RNA complementary to

the CAT probe which is 12 bp larger than the expressed CAT mRNA due to a small part of the multiple cloning site. RNA obtained from a stably transfected insulin-producing cell line with the - 1311/+ 49-bp GLUT2 promoter CAT gene (Waeber et al., 1994) was also used as positive control and size marker. Fig. 2A demonstrates the presence of a protected transcript corresponding to CAT mRNA in the transgenic liver RNA and the RNA extracted from the stably transformed cell line (hneoCAT INS) but no CAT transcript could be detected in all other tested organs. This particular experiment was done with a - 1311 CAT transgenic line corresponding to founder 1 which shows no CAT expression in the kidney or intestine. While CAT expression was present in the liver of all transgenic lines, the transgene was also detected in the kidney and the small intestine of only two transgenic lines: 46 and 68.

A

probe * sense RNA

I pos. control (274 bp) I CAT mRNA (262 bp)

I I GLUT2 mRNA (248 bp)

Fig. 3. Quantitative assessment of the transgene expression in comparison to the endogenous GLUT2 mRNA. (A) An RNAse protection assay of RNAs extracted from a wild type (WT) or a transgenic line (1311CAT, line 68). The CAT transcript is detected only in the transgenic line in the liver, to a lesser extent, in the intestine and the kidney. Transgenic lines 68 and 46 are the only animals where CAT expression was found to be present in the kidney and the intestine beside its presence in the liver. Endogenous GLUT2 mRNA is present in high level in the liver, the intestine, the kidney and faintly in the whole pancreas of all RNA tested but absent in the heart. CAT expression is low in comparison to the endogenous GLUE?. (B) The RNAse assay, using a cRNA probe that includes a fragment of the GLUT2 cDNA and a fragment of the CATcDNA, is schematically shown. The sizes of the protected fragments after RNAse A and Tl digestion are shown for the endogenous GLUT2 mRNA, CAT mRNA and the positive control.

G. Waeber et al. /Molecular and Cellular Endocrinology 114 (1995) 205-215 211

In order to quantitate CAT expression relative to the endogenous GLUT2, we generated a ribonuclease protection assay using a unique radiolabelled RNA probe which contains both a fragment of the murine GLUT2 gene (248 bp) and the CAT gene (262 bp). As depicted in Fig. 3B, the probe detects, after RNAse digestion, a 24$-bp fragment corresponding to the endogenous GLUT2 mRNA and a 262-bp fragment corresponding to CAT mRNA. Fig. 3A shows the presence of a protected transcript corresponding to CAT mRNA in a transgenic animal (line 68) in organs where CAT expression exhibits the same tissue- specific expression as GLUT2 (liver, kidney and small intestine). GLUT2 mRNA is not expressed in the heart in the wild type animals, nor is the transgene GLUT2/CAT expressed in the transgenic heart. The endogenous GLUT2 mRNA could be detected at low level when analyzing total pancreatic RNA as the Langerhans islets represent a small percentage of the organ (Fig. 3A). Even though the CAT transgene is accurately expressed, the level of expression of the CAT mRNA is low when compared to GLUT2 mRNA.

The CAT transgene was also detected using the same RNAse assay described in Fig. 2B in the 33%bp transgenic lines. As shown in Fig. 4, liver and kidney

RNAs obtained from a 33%bp transgenic line (founder 46) contain the CAT transcripts, whereas all other tested organs were negative. Quantitative assessment of CAT expression in comparison to the endogenous GLUT2 mRNA did not show any signitlcant differ- ences between the 1311 or the 338 bp CAT transgenic lines.

3.3. Immunodetection of CAT in transgenic liver sections Immunocytochemical detection of the transgene

product in the liver was done using a commercially available anti-CAT antibody. Fig. 5 shows a hemalun-etythrosin-stained liver section, and a wild type and transgenic liver section immunostained with the anti-CAT antibody. No CAT protein could be detected in the wild type animal, whereas immunos- taining was positive in the transgenic hepatocytes.

3.4. CAT mRNA and its translated product are detected in pancreatic islets of the - 1311- and - 338-bp CAT transgenic lines

As the RNAse protection assay could not be used as a sensitive assay to quantitate GLUT2 or CAT expression using total pancreatic RNA (see Fig. 3), we isolated Langerhans islets of several transgenic lines

338CAT

CAT

Fig. 4. Transgene mRNA analysis by RNAse protection assay in - 338CAT line. The figure shows an RNAse protection assay, using a similar probe to that described in Fig. 2B, of DNAse-treated RNAS extracted from a transgenic line (338CAT) The CAT transcript is detected in the transgenic liver and the kidney RNAs.

212 G. Waeber et al. /Molecular and Cellular Endoctinology 114 (1995) 205-215

Fig. 5. Imrnunocytochemistry of transgenic (46) and wild type (W) liver sections using an anti-CAT antibody. (A) A liver section stained with hemalun and erythrosin demonstrates the normal morphology of a transgenic liver (line 41) (X 60). The hepatocytes are surrounding a central veines. (B) Liver sections stained with anti-CAT antibodies in transgenic line 41 and a wild type animal (C). CAT protein immunolabeling is present only in the hepatoqtes of the transgenic animals (B) and is absent in the wild type &&al (C).

ia) -131 l/+49 CAT -338/+49 CAT

Ifi

PCR cycles

CAT

B actin

GLUT2

G. Waeber et al. /Molecular and Cellular Endocrinology 114 (199.5) 205-215 213

and the RNA was extracted followed by DNAse treatment. Detection and quantification of CAT mRNAs in Langerhans islets of transgenic animals were analyzed by quantitative RT-PCR (Cross, 1995). This approach was done using either - 1311- or -338-bp transgenic lines to compare the relative transcriptional activity of the two tested promoters. We designed specific oligonucleotides flanking the 66-bp small-t intron of the Simian Virus 40 localized 3’ to the reporter CAT gene. The reverse transcribed RNAs were amplified in presence of trace amounts of 32P- or 33P-labelled crdCTP, with the appropriate CAT, p-actin or GLUT2 specific oligonucleotides. Aliquots of the PCR reaction were sampled every five cycles and separated on a sequencing gel. The sampling of the PCR amplification during the process of amplification allowed us to determine a CAT/&actin or CAT/GLUT2 ratio during the linear phase of amplification. The PCR products of the CAT, GLUT2 and p-actin mRNAs of the two transgenic islets are dis-

lb)

Fig. 6. Transgene expression in endocrine pancreas. (A) Typical RT-PCR of DNAse-treated RNAs obtained from - 1311- or - 33% bp GLUECAT transgenic pancreatic islets. There is a cycle-de- pendent increase of CAT, GLUT2 and @-actin transcript, and quantitative autoradiographic scanning was done during the linear phase of amplification. Sampling of the CAT aliquots was done at cycles 20, 25, 30, 35, 30 and 45. The GLUT2 and pactin sampling was done at cycles 10, 15, 20, 25, 30 and 35. (B) Quantitative assessment of GLUT2/P-actin or CAT/j%actin expression for the two tested transgenic mice lines (see text).

played in Fig. 6A. In this particular experiment, the CAT samplings were done at 20,25,30,35,40 and 45 cycles. The p-actin and GLUT2 samplings were done in the same experiment at 10, 15, 20, 25, 30 and 35 cycles. Quantitative assessment, by autoradiographic scanning, of the CAT/p-actin or GLUE/p-actin in both transgenic lines is shown in Fig. 6B. The CAT, GLUT2 and p-actin quantification was done at 30,25 and 20 cycles of amplification, respectively, in the linear phase of amplification for the different PCR products. The CAT/P-actin ratio were of 43 and 42% for the - 131 l- and - 338-bp transgenic lines, respectively, and the GLUT2/P-actin ratio were of 77 and 72% for the - 1311- and -338-bp transgenic lines, respectively.

Immunocytochemical detection of the transgene product in Langerhans islet was performed and is shown in Fig. 7. A wild type and transgenic pancreatic islet (line 46 that integrated a -338-bp CAT transgene) were immunostained with the anti-CAT antibody. No CAT protein could be detected in the wild type animal, whereas a clear signal for the protein was present in the Langerhans islet of the transgenic animal.

4. Discussion

In the present study, we demonstrate that both 1311 or 338 bp of the GLUT2 promoter are able to drive expression of the transgene in the pancreatic islet and in the liver, as demonstrated by quantitative RT-PCR analysis, RNAse protection assays and immunodetection of the CAT protein. CAT expression was also detected in the kidney and the intestine of two transgenic lines. Our results corroborate our re- cent data obtained from transient transfection studies using several constructs of the murine GLUT2 promoter fused to the reporter CAT gene (Bonny et al., 1995). In this model, 338 bp of the murine GLUT2 proximal promoter could efficiently drive the expression of a CAT reporter gene in insulin-producing cells, but displayed no promoter activity in non-pancreatic derived cells. Furthermore, several cis-regulatory elements have been characterized by DNAse I footprinting, gel mobility and SouthWestern analysis and localized within this proximal region of the promoter. Two &-elements appear to bind proteins expressed in all tissues, whereas a third element binds trans-acting factors present only in p-cells (Bonny et al., 1995). Taken together, these data show that 338 bp of the proximal promoter contain several regulatory elements sufficient to drive in vivo and in vitro islet-specific expression of the GLUT2 gene. As the overall CAT expression is low in comparison to the endogenous GLUT2 in our transgenic lines, the iden-

214 G. Waeber et al. /A4ohmdar and Cellular Endocrinology 114 (1995) 205-215

tifled promoter used in this study may be suitable to study its tissue-specific expression but may lack some positive enhancer(s) susceptible to increase gene expression to the transgene. Further work is currently in progress to identify a larger regulatory region of the GLUT2 gene that may contain enhancer sequence(s) in order to facilitate the in vivo study of the transcriptional gene regulation of GLUT2.

The mechanism by which tissue- and cell-specific transcriptional control is achieved is complex and results from interactions between various Puns-acting factors expressed in a temporal and cellular-specific manner and c&regulatory elements localized on pro-

moter regions (M&night et al., 1982). In the endocrine pancreas, the insulin gene is probably the most widely studied in regard to its cell-specific expression. In vivo experiments have shown that 660 base pairs located upstream in the rat insulin II gene conferred pancreatic p-cell specific expression in transgenic mice (Hanahan, 1985). For the rat insulin I gene, the -346 to - 103 fragment of the promoter appears to be sufficient to drive transgene expression in the endocrine pancreas and to a lesser extent in the brain (Dandoy-Dron et al., 1991). To our knowledge, this represents the shortest fragment described that is able to confer tissue-specific expression

Fig. 7. Immunocytochemistry of transgenic (46) and wild type (WT) pancreatic sections using an anti-CAT antibody. (A) Pancreatic islet stained with hemalun and erythrosin demonstrates the normal morphology of a transgenic pancreas (line 46) ( X 60). (B) Pancreatic sections stained with anti-CAT antibodies in transgenic line 46 and a wild type animal (C). Specific CAT protein immunolabeling is observed only in the endocrine pancreas of the transgenic animals (B) and is absent in the wild type animal (0.

G. Wueber et al. /Molecular and Cellular Endocrinology 114 (199s) 205-215 215

in p-cells. T.L. Jetton and co-authors have shown, using a transgenic mouse model, that 294 bp of the upstream glucokinase promoter, described as the islet-specific promoter, can drive the expression of a reporter gene into pancreatic and non-pancreatic neuroendocrine cells present in jejunal enterocytes, thyroid, lungs, pituitary and medial hypothalamus (Jetton et al., 1994). We have added similar observa- tions for the two transgenic GLUT2 promoter constructs described in this manuscript as GLUT2 expression studied with these transgenes is restricted to a few tissues.

PI 131

[41 [51

Kl

[71

In every animal model of diabetes studied such as the diabetic Zucker fa/fa rat, the neonatal low-dose STZ-induced diabetic rat, the GK rat, the db/db mouse or the BB/W rat, there is strong reduction in GLUT2 expression which is restricted to pancreatic P-cells, whereas GLUT2 expression is unaltered or increased in the liver and the kidney (Johnson et al., 1990; Orci et al., 1990a,b; Thorens et al., 1990, 1992b; Unger, 1991; Ohneda et al., 1993). The present trans- genie model may allow the identification of regulatory elements of the GLUT2 promoter involved in the P-cell specific decrease in GLUT2 expression seen in diabetes. This approach may therefore contribute to the understanding of the molecular events involved in the pathogenesis of diabetes.

Bl 191

[lOI

[ill WI

[131

[141

1151

Ml

Acknowledgements [171

We are grateful to Bernard Thorens for helpful discussions and for isolating the pancreatic islets and to Nancy Thompson for critical comments on the manuscript. G.W. is supported by a career award from the Swiss National Science Foundation (32-31915.91 and 32-29317.91) and this work was supported by a Juvenile Diabetes Foundation grant (194183). J.-A.H is recipient of a grant from the Max Cliietta Founda- tion and supported by a grant from the Swiss National Science Foundation (31-37393.93).

[181

[191

DOI La

t221 1231

References [241

D51 [ll Bonny, C., Thompson, N., Nicod, P. and Waeber, G. (1995)

Mol. Endocrinol. (in press).

Carruthers, A. (1990) Physiol. Rev. 70, 1135-1176.

Chen, L., Alam, T., Johnson, J.H., Hughes, S., Newgard, C.B. and Unger, R.H. (1990) Proc. Natl. Acad. Sci. USA 87, 4088-4092. Cross, N.C.P. (1995) Br. J. Haematol. 89, 693-697. Dandoy-Dron, F., Monthioux, E., Jami, J. and Bucchini, D. (1991) Nucleic Acids Res. 19, 4925-4930.

Ferrer, J., Gomis, R., Femandez-Alvarez, J., Casamitjana, R. and Vilardell, E. (1993) Diabetes 42, 1273-1280. Gotoh, M., Maki, T., Satomi, T., Porter, J., Bonner-Weir, S., O’Hara, C.J. and Monaco, A.P. (1987) Transplantation 43, 725-730. Hanahan, D. (1985) Nature 315, 115-122.

Jetton, T.L., Liang, Y., Pettepher, C.C., Zimmerman, E.C., Cox, F.G., Horvath, K., Matschinsky, F.M. and Magnuson, M.A. (1994) J. Biol. Chem. 269, 3641-365.,

Johnson, J.H., Ogawa, A., Chen, L., Orci, L., Newgard, C.B., Alam, T. and Unger, R.H. (1990) Science 250, 546-548. Kahn, B.B. and Flier, J.S. (1990) Diabetes Care 13, 548-564. M&night, S.L. and Kingsbury, R. (1982) Science 217,316-32.

Ohneda, M., Johnson, J.H., Inman, L.R., Chen, L., Suzuki, K.I., Goto, Y., Alam, T., Ravazzola, M., Orci, L. and Unger, R.H. (1993) Diabetes 42, 1065-1072.

Orci, L., Ravazzola, M., Baetens, D., Inman, L., Amherdt, M., Peterson, R.G., Newgard, C.B., Johnson, J.H. and Unger, R.H. (1990a) Proc. Natl. Acad. Sci. USA 87, 9953-9957.

Orci, L., Unger, R.H., Ravazzola, M., Ogawa, A., Komiyia, I., Baetens, D., Lodish, H.F. and Thorens, B. (199Ob) J. Clin. Invest. 86, 1615-162.

Postic, C., Burcelin, R., Rencurel, F., Pegorier, J.-P., Loizeau, M., Girard, J. and Leturque, A. (1993) Biochem. J. 293, 119-124.

Purrello, F., Buscema, M., Rabuazzo, A.M., Cahabiano, V., Forte, F., Vinci, C., Vetri, M. and Vigneri, R. (1993) Diabetes 42, 199-205.

Thorens, B., Sarkar, H.K., Kaback, H.R. and Lodish, H.F. (1988) Cell 55, 281-290.

Thorens, B., Weir, CC., Leahy, J.L., Lodish, H.F. and Bon- ner-Weir, S. (1990) Proc. Natl. Acad. Sci. USA 87,6492-6496. Thorens, B. (1992a) Int. Rev. Cytol. 137A, 209-237.

Thorens, B., Wu, Y.-J., Leahy, J.L. and Weir, G.C. (1992b) J. Clin. Invest. 90, 77-85. Unger, R.H. (1991) Science 251, 1200-1205.

Valera, A., Solanes, G., Femandez-Alvarez, J., Pujol, A., Ferrer, J., Asins, G., Gomis, R. and Bosch, F. (1994) J. Biol. Chem. 269, 28543-28546.

Waeber, G., Thompson, N., Haefliger, J.-A. and Nicod, P. (1994) J. Biol. Chem. 269, 26912-26919.

Yasuda, K., Yamada, Y., Inagaki, N., Yano, H., Okamoto, Y., Tsuji, K., Fukumoto, H., Imura, H., Seino, S. and Seino, Y. (1992) Diabetes 41, 76-78.

Documents

A 338-bp proximal fragment of the glucose transporter type 2 (GLUT2) promoter drives reporter gene expression in the pancreatic islets of transgenic mice