Proc. Natl. Acad. Sci. USA Vol. 91, pp. 2051-2055, March 1994 Biochemistry Ribozyme-mediated attenuation of pancreatic ,8-cell glucokinase expression in transgenic mice results in impaired glucose-induced insulin secretion (antisense RNA/3-celi lines/diabetes/glucose phosphorylation/glucose sensing) SHIMON EFRAT*t, MARGARITA LEISER*, Y.-JIAN Wu§, DAVID FUSCO-DEMANE*, OBAIDULLAH A. EMRAN*, MANJU SURANAt, THOMAS L. JETTONI, MARK A. MAGNUSON¶, GORDON WEIR§, AND NORMAN FLEISCHER* Departments of *Molecular Pharmacology and WMedicine, Albert Einstein College of Medicine, Bronx, NY 10461; §Joslin Diabetes Center, Boston, MA 02215; and IDepartment of Molecular Physiology and Biophysics, Vanderbilt University School of Medicine, Nashville, TN 37232 Communicated by Frank Lilly, November 22, 1993 ABSTRACT Phosphorylation ofglucose to glucose 6-phos- mutant proteins manifest drastically reduced activities (8, 9). phate by glucokinase (GK; EC 2.7.1.2) serves as a glucose- Since GK expression in 3 cells is not transcriptionally sensing mechanism for regulating insulin secretion in 13 cells. regulated, the wild-type allele cannot compensate for this Recent rindings ofheterozygous GK gene mutations in patients reduction. The decreased GK activity may be sufficient to with maturity-onset diabetes of the young (MODY), a form of shift the threshold for glucose sensing, thereby resulting in type H (non-insulin-dependent) diabetes characterized by au- impaired insulin secretion at physiological glucose levels. tosomal dominant inheritance, have raised the possibility that However, it remains unclear whether abnormal glucose up- a decrease in 3-cell GK activity may impair the insulin take by the liver contributes to the disease. secretory response of these cells to glucose. To generate an Understanding of the human disease can benefit from an animal model for MODY we have expressed in transgenic mice animal model, which will allow detailed biochemical and a GK antisense RNA with a ribozyme element under control of physiological studies. We sought to generate a mouse model the insulin promoter. Mice in two independent lineages had for MODY by specifically reducing GK activity in (3 cells, about 30% of the normal islet GK activity. Insulin release in without affecting its function in the liver. To this end a GK response to glucose from in situ-perfused pancreas was im- ribozyme was expressed in (3 cells in transgenic mice. Ri- paired; however, the plasma glucose and insulin levels of the bozymes are RNA molecules that possess catalytic RNA- mice remained normal. These mice are likely to be predisposed cleavage activity (10, 11). By flanking the ribozyme catalytic to type II diabetes and may manifest increased susceptibility to genetic and environmental diabetogenic factors. They provide element with two gene-specific fiagments in antisense orien- an animal model for studying the interaction of such factors tation the ribozyme activity can be targeted against a unique with the reduced islet GK RNA sequence, thus reducing target mRNA levels and activity. activity. The transgenic mice expressing the GK ribozyme The mechanisms by which pancreatic P cells sense and manifest only a third of the normal islet GK activity. Insulin respond to physiological changes in blood glucose have been secretion in response to glucose is impaired; however, the the subject of extensive investigation. Glucose-induced in- mice remain euglycemic. These findings suggest that GK sulin secretion requires the metabolism of glucose in ( cells deficiency in the liver, in addition to that in the islets, may (1). The phosphorylation of glucose to glucose 6-phosphate, play a role in the induction of diabetes in MODY patients. which determines the rate ofglycolysis, has been proposed to These mice provide an animal model for studying the inter- constitute a key glucose-sensing mechanism for regulating action ofgenetic and environmental diabetogenic factors with insulin secretion (2). (3 cells and hepatocytes express a the reduced islet GK activity. high-Km member ofthe hexokinase family, glucokinase (GK; EC 2.7.1.2), which is responsible for the majority of glucose MATERIALS AND METHODS phosphorylation activity in (3 cells (2). While in the liver transcription of the GK gene is induced by insulin (3, 4), GK Plasmid Constructs. Two oligonucleotides were synthe- expression in ( cells is primarily regulated by glucose at the sized containing two 12-base fragments derived from mouse translational and post-translational levels (3, 5). GK gene exon 3 sequence (12) that flank a hammerhead Recent DNA polymorphism studies have established a ribozyme catalytic element (10, 11). Annealing of the sense linkage between the GK locus and diabetes in patients with oligonucleotide 5'-GATCCTCTCCCACTTTCTGATGAG- a non-insulin-dependent diabetes mellitus form termed ma- TCCGTGAGGACGAAACCAGCATCACCGGTAC-3' and turity-onset diabetes of the young (MODY) (6, 7). This the antisense oligonucleotide 5'-CGGTGATGCTGGTT- disease is characterized by an early age of onset and an TCGTCCTCACGGACTCATCAGAAAGTGGGAGAG-3' autosomal dominant inheritance. Sequencing ofthe GK gene created a fragment (GKRZ) with BamHI and Kpn I protrud- from MODY patients has detected a number ofnonsense and ing ends, which was ligated downstream of a hybrid intron missense mutations which are associated with regions of the element (13) and upstream of the simian virus 40 late poly- enzyme molecule involved in glucose and ATP binding (6, 7). adenylylation site in- pMLSIS.CAT (13). The combined The molecular mechanism which makes these mutations 675-bp fragment was inserted into the Xba I and Sal I sites of dominant remains unknown. The inheritance pattern of the pRIP-Tag (14) downstream ofthe rat insulin II gene promoter disease suggests that the patients' (3 cells contain normal to form pRIP-GKRZ. A DNA fragment containing the neo- enzyme molecules encoded by the wild-type allele. The mycin-resistance gene under control ofthe pgk promoter (15) The publication costs ofthis article were defrayed in part by page charge Abbreviations: GK, glucokinase; MODY, maturity-onset diabetes of payment. This article must therefore be hereby marked "advertisement" the young; U, unit(s). in accordance with 18 U.S.C. §1734 solely to indicate this fact. tTo whom reprint requests should be addressed. 2051 Downloaded by guest on September 29, 2021 2052 Biochemistry: Efrat et al. Proc. Natl. Acad. Sci. USA 91 (1994) was inserted into the Sal I site of pRIP-GKRZ to form A GKRZ pRIP-GKRZ/neo. - + p E :'.. Generation of Transgenic Mice. Linearized pRIP-GKRZ - 3'lI''1.e DNA was microinjected into C3HeB/FeJ mouse embryos, and transgenic mice were generated and bred according to established procedures (16). GKR RNA Analysis. Five million fTC6 cells (17) were electro- B porated with 30 pg of pRIP-GKRZ/neo DNA in 1 ml of Dulbecco's modified Eagle's medium (DMEM) at 800 ,uF and -28S 250 V. RNA was extracted from G418-resistant and untrans- fected cells by using RNA Stat-60 (Tel-Test, Friendswood, 18S TX). Five micrograms of RNA was reverse-transcribed, and 1% of the reaction volume was used in 35 cycles of PCR amplification with sense (5'-CTCTAGAATTCGCTGTCT- rc - tub No _i GCGAGG-3') and antisense (5'-CACTGCATTCTAGTT- GTGGTTTGTCC-3') oligonucleotides flanking the intron. The amplified fragments were fractionated on a 1.25% aga- C GKRZ rose gel, visualized with ethidium bromide, and photo- graphed under UV light. Poly(A)+ mRNA was selected with -106 the Poly(A) Quik kit (Stratagene). Ten micrograms ofmRNA was fractionated on a 1% agarose/0.22 M formaldehyde -80 gel, blotted onto nitrocellulose filter, hybridized with a [32P]dCTP-labeled rat GK cDNA probe, and autoradio- -49 graphed on an x-ray film. The blot was stripped and rehy- bridized with an a-tubulin probe to normalize for mRNA loading. The autoradiographs were scanned with an LKB Ultroscan XL densitometer. FIG. 2. Analysis of pRIP-GKRZ effect on GK expression in transfected .3TC cells. (A) Reverse transcriptase-PCR analysis of Immunoblotting. Proteins were extracted from G418- GKRZ expression in the transfected cells. Expression is manifested resistant and untransfected cells as described (18). Twenty by the 200-bp band that results from spliced transcripts. Lane p micrograms of protein was fractionated on a 12% polyacryl- contains an unspliced fragment amplified from pRIP-GKRZ DNA. amide/SDS gel, electroblotted onto an Immobilon-P filter Size markers are in bp. (B) Northern blotting analysis ofGK mRNA. (Millipore), and probed as described (19) with a sheep an- Ribosomal RNA bands serve as size markers. Rehybridization with ti-GK serum (19). The bound antibody was visualized with a an a-tubulin probe is shown at the bottom. (C) Immunoblotting horseradish peroxidase-conjugated second antibody and a analysis of GK protein. Size markers are in kDa. chemiluminescent substrate (ECL, Amersham) by exposure samples in separate lanes were stained with Coomassie blue, to an x-ray film and quantitated by densitometry. Protein and the dried gel was quantitated by densitometry to nor- A malize for protein loading. Immunohistochemistry. Pancreas samples were sectioned 86 87 88 89 90 91 92 93 and analyzed with a sheep anti-GK serum as described (19). aa V M L V K V G E Glucose Phosphorylation. Islets were isolated from the pancreas by collagenase infusion through the bile duct (20). GK 5' ... GGUGAUGCUGaUC AAAGUGGGAGAG... 3' Islet homogenates from 1- to 3-month-old mice were incu- GKRZ 3'...CCACUACGACCA UUUCACCCUCUC... 5' bated with glucose at various concentrations in the presence A C ofATP, glucose-6-phosphate dehydrogenase, and NAD, and A G the formation of NADH was monitored fluorimetrically as A described (5). G U Pancreas Perfusion. Anesthetized 4- to 7-month-old mice G were cannulated through the aorta and portal vein and A perfused with oxygenated Krebs-Ringer buffer containing C-G glucose in an increasing concentration gradient. Samples A-U were collected and assayed for glucose by a glucose analyzer G-C and for insulin by RIA.