Release of Transgenic Human Insulin from Gastric G Cells: A Novel Approach for the Amelioration of Diabetes
http://www.100md.com
内分泌学杂志 2005年第6期
Department of Medicine, Division of Digestive Diseases, David Geffen School of Medicine at University of California, and CURE: Digestive Diseases Research Center, University of California, Los Angeles, California 90095
Address all correspondence and requests for reprints to: Elena Zhukova, 900 Veteran Avenue, Warren Hall, Room 14-109, Department of Medicine, University of California School of Medicine, Los Angeles, California 90095-1786. E-mail: ezhukova@mednet.ucla.edu.
Abstract
We explored the hypothesis that meal-regulated release of insulin from gastric G cells can be used for gene therapy for diabetes. We generated transgenic mice in which the coding sequence of human insulin has been knocked into the mouse gastrin gene. Insulin was localized specifically to antral G cells of G-InsKi mice by double immunofluorescence staining using antibodies against insulin and gastrin. Insulin extracted from antral stomach of G-InsKi mice decreased blood glucose upon injection into streptozotocin-diabetic mice. Intragastric administration of peptone, a known potent luminal stimulant of gastrin secretion, induced an increase in circulating levels of transgenic human insulin from 10.7 ± 2 to 23.3 ± 4 pM in G-InsKi mice. Although G cell-produced insulin decreased blood glucose in G-InsKi mice, it did not cause toxic hypoglycemia. Proton pump inhibitors, pharmacological agents that increase gastrin output, caused a further increase in the circulating levels of gastric insulin (41.5 ± 2 pM). G cell-produced insulin was released into circulation in response to the same meal-associated stimuli that control release of gastrin. The most striking aspect of the results presented here is that in the presence of the G-InsKi allele, Ins2Akita/+ mice exhibited a marked prolongation of life span. These results imply that G cell-derived transgenic insulin is beneficial in the amelioration of diabetes. We suggest that an efficient G cells-based insulin gene therapy can relieve diabetic patients from daily insulin injections and protect them from complications of insulin insufficiency while avoiding episodes of toxic hypoglycemia.
Introduction
THE PREVALENCE OF diabetes mellitus has reached epidemic proportions in many parts of the world (1). This chronic group of disorders is characterized by no insulin secretory capacity (type I) or insulin resistance and defective endogenous insulin secretion (type II). The primary focus for the treatment of type I diabetes is to replace insulin secretion by the administration of exogenous insulin, and treatment of advanced type II often requires daily injections of insulin. Improved glycemic control significantly decreases long-term diabetes-associated microvascular complications (2) but is often difficult to achieve. Consequently, novel therapeutic approaches to diabetes are being explored, and several laboratories have focused on the possibility of using gene therapy to achieve optimal glycemic control. Despite considerable effort to engineer insulin production in a variety of tissues and cells, including muscle (3), exocrine pancreas (4), liver (5, 6, 7, 8, 9, 10, 11, 12, 13, 14), pituitary (15, 16, 17), and intestinal K cells (18), a well-characterized system that provides optimal glycemic control has not been generated yet.
The systems favored so far use glucose-regulated insulin secretion (5, 9, 10, 11, 12, 13, 14, 18). Each of these systems is characterized by certain advantages over the alternatives, albeit simultaneously tempered by particular limitations (reviewed in Ref. 19). Nevertheless, glucose-regulated systems have an enormous potential for the treatment of type I diabetes.
Type II diabetes, which represents about 90% of cases, poses different challenges. Our understanding of the complex molecular mechanisms leading to type II diabetes is increasing and evolving; consequently, the cellular targets selected for engineering insulin production must evolve as well. In type II diabetes, the glucose-sensing mechanism itself can be the target of alteration (reviewed in Refs. 20, 21, 22). Consequently, it would be advantageous to engineer physiologically regulated insulin secretion by cells that will not be affected by changes in glucose sensing.
Regulated exocytosis is a crucial aspect of instant and timely release of insulin from ?-cells and should be an essential feature of a cell selected as a target for the production of extrapancreatic insulin. Therefore, in searching for a candidate surrogate ?-cell, an important property would be the ability to respond to food intake with physiologically regulated exocytosis. Ideally, insulin release should mimic the normal pattern of insulin secretion, namely a basal level during day and night with brief increases coinciding with ingestion of meals. Consequently, in this paper, we explored a new putative target for diabetes gene therapy and focused on reengineering gastrin-secreting G cells into secretagogue-responsive insulin-secreting cells.
G cells, endocrine cells located in the antral part of the stomach, release the peptide hormone gastrin in response to protein-rich meals. Although G cells are insensitive to changing glucose levels, the secretion of gastrin from these cells is tightly linked to food intake and occurs within minutes (23, 24, 25, 26, 27, 28). These cells synthesize preprogastrin and process it with the same endopeptidases that are implicated in proinsulin processing, namely PC1/PC3, PC2, and carboxypeptidase H (29, 30, 31, 32). Similarly to pancreatic ?-cells that store and secrete insulin, G cells store gastrin in secretory vesicles and release it upon stimulation into the portal circulation. Importantly, G cells are open to the gastric lumen, through which gene delivery agents can directly access them.
Therefore, we hypothesized that transgenic human insulin selectively produced by G cells under the regulation of the gastrin gene promoter will be released into circulation in response to the same meal-associated stimuli that control release of gastrin. In this study, we generated a novel mouse model (G-InsKi) in which human insulin is expressed in the gastric G cells under the regulation of the mouse gastrin gene promoter. G cell-produced insulin is functional and is released into circulation by intragastric administration of peptone, a potent stimulant of gastrin release. The most striking aspect of the results presented here is that in the presence of the G-InsKi allele, Ins2Akita/+ mice exhibited a marked prolongation of life span. These results imply that G cell-derived transgenic insulin is beneficial in the amelioration of diabetes.
Materials and Methods
Assembly of the targeting vector pG-InsKi
The pG-InsKi vector was assembled using the mouse gastrin gene (33) (a gift by Dr. Timothy C. Wang, University of Massachusetts Medical Center, Worcester, MA), insulin cDNA (a gift from Dr. Graeme I. Bell, University of Chicago, Chicago, IL), and a loxP-pgk-Neo cassette from the plasmid p1338 (kindly provided by Dr. Timothy Ley, Washington University, St. Louis, MO). Specifically, the 346-bp fragment of the human insulin cDNA (position +12 to +358) containing the coding sequence but not the polyadenylation signal was amplified from oligonucleotide primers InsFW (ACTGAATTCATCACTGTCCTTCTGCCATGG, sense) and InsRV (ACTTCTAGACGTCTAGTTGCAGTAGTTCTC, antisense) (Fig. 1A). This fragment was subcloned into the EcoRI and XbaI sites of the pIRESneo vector (BD Biosciences, Palo Alto, CA), which was used as an intermediate carrier and a donor of bovine GH polyadenylation signal (Fig. 1B). The EcoRI-XhoI fragment, containing the insulin coding sequence and the bovine growth hormone polyadenylation signal, was subcloned into the BamHI site of the plasmid loxP-pgk-Neo (version p1338) upstream from the loxP-pgk-Neo cassette (Fig. 1, C and D). The resulting 2-kb insulin/loxP-pgk-neo cassette was excised from the p1338 vector at NsiI sites and inserted into the exon 2 of the mouse gastrin gene at the HindIII site (Fig 1, D and E). The resulting targeting vector G-InsKi is shown in Fig. 1F. It has approximately 3 kb homology to the targeted mouse gastrin gene, both 5' and 3' from the point of knock-in insertion. The correct assembly of the G-InsKi construct was confirmed by DNA sequencing.
FIG. 1. Generation of the targeting construct. A, The 346-bp fragment of the human insulin cDNA (position +12–+358) containing the coding sequence but not the polyadenylation signal was amplified using oligonucleotide primers InsFW and InsRV. B, The amplified fragment was subcloned into the EcoRI and XbaI sites of the pIRESneo vector, which was used as an intermediate carrier and a donor of bovine growth hormone polyadenylation signal (hatched box). C and D, The EcoRI-XhoI fragment, containing the insulin coding sequence and the bovine growth hormone polyadenylation signal, was subcloned into the BamHI site of the plasmid loxP-pgk-Neo upstream from the loxP-pgk-Neo cassette. D and E, The resulting-2 kb INS/loxP-pgk-Neo cassette was excised from the vector at NsiI sites and inserted into the exon 2 of the mouse gastrin gene at the HindIII site. F, The targeting of vector G-InsKi.
Gene targeting and blastocyst injection
The G-InsKi targeting construct was linearized with SacII, electroporated into germline-competent LW-1ES cells (pure inbred 129/SVJae), and selected with G418. DNA microinjection and G418 selection were performed at the University of California (Los Angeles) Embryonic Stem Cell Facility. DNA was isolated from 288 surviving clones. ES clones that have undergone homologous recombination were identified by Southern blot analysis using a 200-bp NheI-BglII fragment as a 5'-external probe. Seven positive clones were identified and expanded. Two of those, clones 18 and 27, were microinjected into blastocysts from CB6F1 (a hybrid of C57Bl/6 and BALB/c) donors at the University of California (Irvine) Transgenic Mouse Facility. High-percentage male chimeras were obtained and mated with C57Bl/6 females. Germline transmission was identified by PCR analysis of genomic DNA obtained from ear snips. It should be pointed out that the Gas-Ins transgenic mouse described in a previous study from our group is an independent animal model that was generated by pronuclear injection (34). The transgene used in that study contained the genomic sequence of human insulin DNA. The aberrant splicing of insulin introns in G cells resulted in the production of a truncated form of the human insulin RNA and, consequently, led to the production of truncated insulin molecules (34). Consequently, we used the cDNA of human insulin for the generation of knock-in mice G-InsKi described in the present study.
Animals
All mice were housed in specific pathogen-free barrier facilities, maintained on a 12-h light, 12-h dark cycle, and fed a standard autoclavable rodent diet (PMI Feeds, Inc., Richmond, IN). G-InsKi mice were maintained on C57Bl/6 background as heterozygotes to avoid creation of the gastrin-null phenotype. Akita diabetic mice, strain C57Bl/6-Ins2Akita, were purchased form the Jackson Laboratory (Bar Harbor, ME) and maintained as heterozygotes. The studies were approved by UCLA’s Office for Protection of Research Subjects.
Genotype analysis
Genotyping was performed by PCR analysis of genomic DNA obtained from ear snips or tail biopsies. Genomic DNA was purified as described by Miller and Polesky (35). PCRs included 200 ng DNA and 2.5 U AmpliTaq Polymerase (PerkinElmer Life and Analytical Sciences, Boston, MA). PCR conditions were: 94 C, 1 min, followed by 35 cycles of 94 C, 30 sec; 55 C, 30 sec; and 72 C, 30 sec. The G-InsKi allele was detected using primers Ins 25, CTGTCCTTCTGCCATGG; and Ins 33, GTTGCAGTAGTTCTCCAGCTG. The identity of the PCR product was confirmed by digestion with the restriction endonuclease SexAI, which was added directly to the aliquot of the PCR mix. DNA samples negative for the presence of the G-InskKi transgene were analyzed for the presence of the ?-globin by PCR analysis from primers 5'CCAATGTGCTCACACAGGATAGAGAGGGCAGG 3' and 5' CCTTGAGGCTGTCCAAGTGATTCAGGCCATCG 3', which produce a 494-bp fragment. PCR identification of Ins2Akita allele was performed as described elsewhere (36).
Immunocytochemistry
Mice were deeply anesthetized with ip sodium pentobarbital (50 mg/kg Nembutal; Abbott Laboratories, Chicago, IL) (50 mg/kg ip) and then perfused intracardially with 4% paraformaldehyde in 0.1 M phosphate buffer (PB) (pH 7.4). The stomach and pancreas were dissected, postfixed with the same fixative for 1 h at room temperature, and then stored at 4 C in 25% sucrose in PB overnight. Tissue sections were then cut with a cryostat at 10–12 μm thickness, mounted on slides, and stored at –20 C until processing for immunohistochemistry. Cryostat sections were processed by the immunofluorescence method for single and double labeling as previously described (37). For single labeling, cryostat sections of the stomach or pancreas tissues were washed in PB (3 x 15 min), incubated in 10% normal donkey serum for 1 h at room temperature to minimize the background, incubated in either guinea pig antiporcine insulin antibody GPDI (kindly provided by Dr. K. Polonsky, University of Chicago) at 1:1250 or guinea pig antihuman insulin antibody (Linco Research, Inc., St. Charles, MO) (1:200) in PB containing 0.5% Triton X-100 overnight at 4 C, washed in PB (3 x 30 min), then incubated in affinity-purified donkey anti-guinea pig IgG conjugated with ALEXA 488 (1:1000) (Molecular Probes, Eugene, OR), washed again for 45 min, and coverslipped. Immunoblocking experiments were performed with human insulin peptide at 10–5 μM (Sigma, St. Louis, MO). For double-label immunofluorescence, cryostat sections of the stomach were incubated in a mixture of guinea pig antihuman insulin or GPDI antibody and rabbit gastrin antiserum (gastrin 1802 from CURE Antibody Core; 1:500) overnight, followed by a 2-h incubation with a mixture of donkey antirabbit IgG-ALEXA 488 (1:1000) (Molecular Probes) and donkey anti-guinea pig IgG conjugated with rhodamine Red X (Jackson Immunoresearch Laboratories, West Grove, PA; 1:300). Tissues were examined with a Zeiss Axioplan 2 research microscope for fluorescence with an Axiocam color digital camera fluorescence microscope (Carl Zeiss Inc., Thornwood, NY), equipped with fluorescein isothiocyanate and Red-X cubes, or a Zeiss 410 laser scanning confocal microscope equipped with a krypton/argon laser and attached to a Zeiss Axiovert 100 microscope with a 100x Plan Apo 1.4 numerical aperture objective (Carl Zeiss Inc.).
Analytical procedures
Blood for analyses was collected by saphenous vein or retroorbital bleeding (38, 39). Blood glucose concentrations were measured using the Accu-Chek Compact blood glucose monitoring system (Roche Diagnostic Corp., Indianapolis, IN). Human insulin in serum and in antral tissue extracts was determined using a human insulin ELISA kit, which does not recognize mouse insulin (Linco Research, Inc.). This kit has no cross-reactivity to intact human proinsulin and des proinsulin(31, 32); however, there can be cross-reactivity with des proinsulin(64, 65). Mouse insulin in serum was determined using the Mouse Insulin ELISA kit (Mercodia AB, Uppsala, Sweden), which also cross-reacts with human insulin (120%). Values of mouse serum insulin in transgenic mice were adjusted by subtracting respective values of human insulin. Mouse gastrin in serum and in antral tissue extracts was measured by human gastrin I (G17) immunoassay (R&D Systems, Minneapolis, MN), which has 70.7% cross-reactivity with rat gastrin I and 0.8% cross-reactivity with Big Gastrin (G34-I).
Extraction of insulin from stomachs
Gastric antrums were collected from mice, immediately frozen in liquid nitrogen, and stored at –70 C until used. Insulin was extracted by the modified procedure described by Baggio et al. (40). Specifically, the antrums were homogenized with PowerGen homogenizer (Model 700) on ice in 5 volumes of extraction buffer [1 N HCl, 5% (vol/vol) formic acid, 1% (vol/vol) trifluoroacetic acid, and 1% (wt/v) NaCl]. Insoluble material was removed by centrifugation at 12,000 x g at 4 C for 15 min. Supernatant (1 ml) was loaded onto a Sep Pak C18 cartridge. The Sep Pak cartridges were washed with 2 ml methanol and 2 ml water followed by 2 ml extraction buffer. After loading the samples, the cartridges were washed with 2 ml extraction buffer before eluting the samples with 1 ml 80% (vol/vol) isopropanol containing 0.1% (vol/vol) trifluoroacetic acid. Human insulin in tissue extracts was measured using the specific ELISA kit as described above. Antrums from 19 transgenic mice were pooled in six groups, and each of the groups was processed independently.
Western blot analysis
For immunoblotting, extracts of gastric antrum containing at least 50 ng human insulin as measured by ELISA were loaded onto a 10% NuPAGE minigel in a reducing sample buffer containing 50 mM dithiothreitol and subjected to electrophoresis for 35 min at 200 V using 2-(N-morpholino)ethanesulfonic acid running buffer (Invitrogen, Carlsbad, CA). Proteins were transferred to Immobilon-PSQ membranes (Millipore, Billerica, MA) in Tris-glycine transfer buffer at 35 V for 1 h. Membranes were blocked with 5% (wt/vol) nonfat skim milk in PBS (pH 7.2) for at least 2 h at room temperature and then incubated at room temperature for 3 h with guinea pig antipig insulin whole antiserum (I8510, Sigma-Aldrich) diluted 1:100 in 3% (wt/vol) nonfat skim milk in PBS. This antiserum has 100% cross-reactivity with human insulin. Immunoreactive bands were visualized using horseradish peroxidase-conjugated rabbit anti-guinea pig IgG (A5545, Sigma-Aldrich) and enhanced immunofluorescence methods. Specificity of the staining was confirmed by preabsorbtion with 10 μg/ml Humulin (Eli Lilly, Indianapolis, IN) for 1 h at 37 C.
In vivo experimental procedures
Male mice, 6–8 wk old, were used for experiments. Peptone solution in water (Becton Dickinson, Sparks, MD) or omeprazole suspension in hydroxypropyl methylcellulose (AstraZeneca, Wilmington, DE) were instilled into anesthetized mice by intragastric gavage at the volume of 1% of body weight. For ip injections, human insulin in stomach extracts was quantified by specific ELISA as described above, vacuum dried, made up in 3 μM HCl, and injected ip at a dose of 75 mU/g body weight into streptozotocin (STZ)-diabetic mice. Control groups of STZ-diabetic mice were injected with antral extract from nontransgenic mice normalized by the weight of the wet tissue used for the extraction either with or without Humulin.
Six-week-old male mice, either wild-type C57Bl/6 or GinsKi, were made diabetic by ip injections of STZ at a dose of 100 mg/kg of body weight per day, on 2 d, 1 d apart. The total amount of STZ injected on 2 separate d was 200 mg/kg body weight. Mice were fasted overnight before the injections. At d 14 after the injection, mice developed diabetes. Development of diabetes was assessed by measuring fasting glucose levels and mouse insulin in serum.
Statistical analysis
All results are presented as the means ± SEM. Statistical significance was determined by Student’s t test. P < 0.05 was considered significant. Welch’s correction was used to assess the variances wherever necessary. Survival analysis was performed using GraphPad Prism version 3.02 for Windows (GraphPad Software, San Diego, CA).
Results and Discussion
Generation of G-InsKi mice
The G-InsKi transgenic mice in which human insulin is produced in G cells under the regulation of the gastrin gene promoter were generated by knocking-in the coding sequence of the human insulin gene into the exon 2 of the mouse gastrin gene (Fig. 1). The knock-in approach was selected over the transgenic overexpression by pronuclei injection to ensure that all components of the gastrin gene-specific transcriptional regulation were preserved.
Initially, the G-InsKi mice were identified by Southern blot analysis using a 200-bp NheI-BglII fragment as a 5'-external probe (Fig. 2A). The substitution of the mouse gastrin gene with the G-InsKi allele increased the NheI restriction fragment size from 7 to 9 kb (Fig. 2B). Subsequently, genotyping was performed by PCR analysis using primers Ins 25 and Ins 33, which amplify a 340-bp fragment of human insulin cDNA (Fig. 2, A and C). The Ins 25 primer corresponds to the 5'-untranslated region of the human insulin (c)DNA and has no homology with the mouse insulin genes I and II. As a result, this set of primers distinguishes between human and mouse insulin nucleotide sequence. The identity of the 340-bp PCR product was confirmed by digestion with the restriction endonuclease SexAI, which generates diagnostic fragments of 230 and 110 bp, characteristic of this fragment of human insulin cDNA (Fig. 2C).
FIG. 2. Gene targeting and genotyping. A, Partial map of the G-InsKi knock-in allele. The gastrin gene is shown in black. Gastrin exons are shown as filled boxes. Gastrin exon 2 is interrupted by the insulin cDNA linked to the bovine growth hormone polyadenylation signal (INS) and a phosphoglycerol kinase-neomycin resistance gene cassette flanked by two 34-bp loxP sequences (Neo), both shown as white boxes. The location of the DNA fragment used as probe is indicated. The position of primers used for the PCR genotyping is shown. B, Southern blot analysis. Genomic DNA was digested with NheI and then probed with a RNA probe corresponding to the NheI/BglII fragment of the 5'-untranslated region of the mouse gastrin gene (200 bp). The location of the probe is indicated in A. Expected alleles and sizes are shown. G-InsKi, Transgenic allele; WT, wild-type gastrin allele. C, Genotyping by PCR. Genomic DNA was amplified from primers Ins25 and Ins33, shown in A. Note, that Ins25 has no homology with the mouse insulin genes I and II. Authenticity of the 340-bp PCR fragment was validated by the digest with SexAI that produced 230- and 110-bp fragments. ML, DNA mass ladder; WT, negative control, amplification of the wild-type genomic DNA; WT + PL, positive control, amplification of the wild-type genomic DNA spiked with 0.06 pg targeting vector G-InsKi; TG, – and +, Transgenic genomic DNA amplified and incubated without or with SexAI.
The G-InsKi mice were maintained as heterozygotes to avoid creation of the gastrin-null phenotype, which is associated with abnormalities of gastric function, as well as changes in architecture and differentiation of the gastric mucosa (41, 42). Mice that carry a single normal allele of the gastrin gene have physiological levels of circulating gastrin and normal gastric function (42). G-InsKi mice had a normal reproduction life and were healthy.
Localization and antral content of gastric insulin
In accordance with the tissue distribution of the gastrin-producing G cells, in G-InsKi mice, human insulin was expressed by cells located at the bottom of gastric glands in the antrum, as revealed by immunocytochemistry (Fig. 3A). The expression of human insulin in gastric G cells was specific. A few positively stained cells were localized in the duodenum, a minor site of gastrin expression. No insulin-expressing cells were found in oxyntic mucosa, a region of stomach devoid of G cells, or in jejunum, ileum, colon, and liver. Furthermore, insulin-expressing cells in the antrum were identified as G cells by double immunofluorescence staining with antibodies against insulin and gastrin (Fig. 3, B–D). Importantly, every cell positive for gastrin staining was also positive for insulin, indicating the high targeting efficiency of the G-InsKi transgene. Notably, cells positive for somatostatin staining (D cells) did not express any detectable immunoreactive insulin (data not shown). Insulin expressing cells were not found in the stomachs of nontransgenic mice (data not shown).
FIG. 3. Tissue and cell specificity of human insulin production in G-InsKi mice. A, Immunofluorescence staining of the antral tissue section for insulin, low-power view. Mus, Muscle wall. Arrows point to G cells. B–D, Double immunofluorescence staining of the tissue section of antral stomach for insulin (B, red) and gastrin (C, green) shown in high-power view. Colocalization of the signal is shown in D (yellow).
Human insulin content measured in the antral extract by human insulin-specific immunoassay was 207 ± 37.0 pmol/g wet weight (mean ± SEM, n = 6) in G-InsKi mice, whereas it was undetectable in controls (n = 3) (Fig. 4A). This value is comparable with the content of gastrin in antrums of heterozygous gastrin knockout mice (339 ± 94 pmol/g) (42). Pancreatic content of mouse insulin measured for comparison was 3590 ± 640 pmol/g wet weight (mean ± SEM, n = 6). However, these data are not corrected for the number of cells that produce insulin in each tissue.
FIG. 4. Antral content, molecular forms, and biological activity of G cell-produced human insulin. A, Detection and quantification of human insulin in the extracts of antral stomachs by human insulin-specific immunoassay. Values are means ± SEM, n = 19. Note that this kit has no cross-reactivity to intact human proinsulin and des proinsulin(31 32 ); however, there can be cross-reactivity with des proinsulin(64 65 ). B, Western blot analysis of human insulin expression in the extract of antral stomach from G-InsKi mice and nontransgenic littermates. Arrowheads point at the product corresponding to the size of human insulin chains A and B (3.5 and 2.5 kDa, respectively), as well as at the expected location of intermediately processed forms (5.5–6.5 kDa) and intact proinsulin (9 kDa). TG, Transgenic; WT, wild-type. C, G cell-produced human insulin in antral extracts has hypoglycemic activity. Wild-type C57Bl/6 mice were made diabetic by injection of STZ. Extracts of antral stomach from transgenic G-InsKi and nontransgenic littermates were injected into STZ-diabetic mice. Transgenic extracts provided a dose of 75 mU human insulin/kg body weight. Amount of nontransgenic extract was normalized by the weight of the wet tissue used for the extraction. The third group of mice was injected with nontransgenic extract supplemented with Humulin (Eli Lilly) to provide a dose of 75 mU/kg body weight. Blood glucose levels at 20, 40, and 60 min after the injection are shown. White bars, Mice injected with the nontransgenic extract (n = 3). Gray bars, Mice injected with transgenic extract (n = 5). Hatched bars, Mice injected with nontransgenic extract supplemented with Humulin (n = 3). Results are means ± SEM. Number of animals in each group is shown above.
G cells express prohormone convertases PC2 and PC1/PC3 as well as carboxypeptidase H (29, 30, 31, 32), suggesting that proinsulin could be completely processed in G cells. Western blot analysis, illustrated in Fig. 4B, showed that human insulin from antral extracts of G-InsKi mice migrated as a broad band with an apparent molecular mass of approximately 2.5–3.5 kDa, which corresponds to the size of fully processed insulin, namely chains A and B. Preincubation of the antibody with human insulin completely blocked the appearance of this immunoreactive band (data not shown). No immunoreactive bands corresponding to the size of proinsulin (9 kDa) or its intermediately processed forms (5.5–6.5 kDa) were observed in extracts from G-InsKi mice. We did not detect any immunoreactive band when antral extracts from nontransgenic littermates were analyzed.
Circulating gastric insulin in G-InsKi mice
Circulating levels of human insulin in G-InsKi mice were measured with the same human insulin-specific immunoassay, which has no cross-reactivity with mouse insulin, human proinsulin, or des proinsulin (31, 32). The levels of human insulin in samples collected from individual mice were 10.8 ± 0.7 pM in fasting state and 20.7 ± 3 pM in postabsorbtive state (values are means ± SEM; n = 12 and 19, respectively). Fasting blood glucose levels were similar in transgenic and nontransgenic mice, 89 ± 3.4 and 88 ± 2.5 mg/dl, respectively (means ± SEM, n = 24 and 39). Levels of circulating mouse insulin in G-InsKi mice and wild-type controls were 115 ± 13 and 113 ± 16 pM in fasting state and 192 ± 29 and 274 ± 55 pM in postabsorbtive state, respectively (means ± SEM, from at least eight animals in each group). Although the reduction in postabsorbtive levels of pancreatic mouse insulin in G-InsKi mice is in line with the expected down-regulation due to the presence of transgenic human insulin, it did not reach statistical significance in our experiments.
Biological activity of gastric insulin
The biofunctionality of G cell-produced insulin was assessed by its ability to lower blood glucose upon injection into diabetic mice. STZ-diabetic mice were fasted overnight and injected via ip route with the antral extracts of either transgenic G-InKi mice or nontransgenic littermates. As shown in Fig. 4C, antral extract from the nontransgenic mice did not decrease blood glucose levels (white bars), whereas injection of the G-InsKi extract caused a dramatic decrease in blood glucose levels at 20, 40, and 60 min after the injection (gray bars). Interestingly, the efficiency with which G cell-produced insulin regulates glucose levels is virtually identical with the efficiency of an equivalent dose of human insulin added to the control extracts (hatched bars, Fig. 4C). Collectively, results presented in Fig. 4 indicate that insulin is fully processed in gastric G cells.
Regulation of gastric insulin release
Having established that G cell-produced insulin is functional, we next examined whether the release of human insulin by G cells is regulated by the same meal-associated stimuli that control gastrin release. G cells are known to secrete gastrin in response to meal-associated stimuli that act through luminal and neural pathways. The main luminal stimulants of gastrin secretion are small peptides and, in particular, free aromatic amino acids (24, 25, 43). Recent studies of the Ca2+/amino acid-sensing receptor and bitter taste receptors and their localization to the gastric mucosa, and specifically to G cells, provided an insight into the initial molecular recognition event(s) by which G cells apparently sense chemical composition of luminal contents (44, 45, 46, 47). Interestingly, the Ca2+/amino acid-sensing receptor was also localized to ?-cells (48); intriguingly, ?-cells also respond to bitter tastants (49).
To assess whether luminal stimulation causes insulin release from G cells, we used peptone, an enzymatic digest of protein, which is a potent luminal stimulant of gastrin release. As shown in Fig. 5A, gastric instillation of 8% peptone solution induced a 2.2-fold increase in circulating levels of human insulin 40 min after treatment, demonstrating that release of human insulin is physiologically regulated by components of meals. Instillation of vehicle into the control group of animals had no effect on circulating insulin levels (data not shown). As expected, gastric instillation of 8% peptone solution also induced a marked increase in the level of serum gastrin, shown in Fig. 5B.
FIG. 5. Regulation of gastric insulin release. A, Liquid peptone meal induces release of G cell-produced insulin in G-InsKi mice. Mice were fasted overnight, and 8% peptone was instilled by intragastric gavage. Circulating human insulin was measured before (white bar) and 40 min after (black bar) the treatment. The same experiment was repeated with G-InsKi mice made diabetic by the injection of STZ (hatched bar, only levels after the peptone treatment are shown). Note that the human insulin-specific ELISA used in these studies does not cross-react with mouse insulin. Results are means ± SEM, n = 5. B, Liquid peptone meal induces release of gastrin in G-InsKi mice. Mice were fasted overnight, and 8% peptone was instilled by intragastric gavage. Circulating gastrin was measured before (white bar) and 40 min after (black bar) the treatment. The same experiment was repeated with G-InsKi mice made diabetic by the injection of STZ. Only levels after the peptone treatment are shown (hatched bar). Note that peptone did not induce the release of human insulin or mouse gastrin in STZ-treated mice. Results are means ± SEM, n = 7. C, PPIs increase circulating levels of G cell-produced insulin in G-InsKi mice. Mice in postabsorptive state received either omeprazole or vehicle by intragastric gavage once daily for 4 consecutive d at a dose of 100 mg/kg body weight. After the completion of omeprazole treatment, 8% peptone was instilled into the lumen. Circulating human insulin was measured before (data not shown) and after the treatment with vehicle (white bar), omeprazole (black bar), and peptone (gray bar). Results are means ± SEM, and each group included at least five animals. Vh, Vehicle; OZ, omeprazole; Pep, peptone.
Furthermore, we examined whether pharmacological agents that increase gastrin output also release G cell-produced insulin into the circulation. Proton pump inhibitors (PPIs) are commonly used to reduce gastric acid secretion by directly inhibiting H+,K+-ATPase in parietal cells (50, 51). Physiological consequences of PPI action include increase in the number of gastric G cells and an increase in levels of circulating gastrin (52, 53). To determine whether circulating levels of gastric insulin are regulated by PPIs, omeprazole was instilled into G-InsKi mice at 100 mg/kg body weight by intragastric gavage once daily for 4 consecutive d. Treatment with omeprazole induced a 2.3-fold increase in the basal circulating levels of human insulin in G-InsKi mice. The administration of peptone after the completion of omeprazole treatment further increased circulating human insulin (41.5 ± 2 pM). There was no change in human insulin levels in mice that were treated with the vehicle only (Fig. 5C). Collectively, these experiments demonstrate the release of G cell-produced insulin in response to protein hydrolyzates and to suppression of gastric acid production.
Regulation of blood glucose by meal-induced release of gastric insulin
We next explored whether meal-induced release of gastric insulin into portal circulation regulates blood glucose. Consequently, G-InsKi mice and nontransgenic controls were given a peptone meal containing 25 g/liter glucose. Blood glucose was measured before and 40 min after intragastric gavage. In wild-type mice, blood glucose levels were elevated 40 min after gavage, whereas in G-InsKi mice blood glucose levels were approximately 30% lower than in the wild-type controls (Fig. 6A). These results suggest that the administration of the peptone-glucose meal induced the release of both gastric and pancreatic insulin in G-InsKi mice. In agreement with this interpretation, treatment with glucose in the absence of peptone (i.e. in the absence of G cell stimulation) caused equal glycemia in G-InsKi mice and nontransgenic controls (Fig. 6B).
FIG. 6. Regulation of blood glucose in G-InsKi mice. A, Blood glucose is regulated by meal-induced release of gastric insulin. G-InsKi mice were fasted overnight, and liquid meal containing 8% peptone and 25 g/liter glucose was instilled by intragastric gavage. Blood glucose was measured before and 40 min after the treatment. Results are means ± SEM from at least eight animals in each group. B, Glucose meal in the absence of peptone causes equal glycemia in wild-type and G-InsKi mice. Mice were fasted overnight, and a glucose solution (25 g/liter) was instilled by intragastric gavage. Blood glucose was measured before and 40 min after the treatment. Results are means ± SEM from at least eight animals in each group. C, Left panel, Regulation of blood glucose by mixed peptone-glucose meal as a function of time. Mice were fasted overnight, and liquid meal containing 8% peptone and 25 g/liter glucose was instilled by intragastric gavage. Blood glucose was measured before and at 20, 40, 60, 120, and 180 min after the treatment. Results are means ± SEM, from four to eight animals per group. *, P < 0.04 (20 min) and P < 0.02 (40 min). Right panel, Levels of circulating human insulin released by peptone-glucose meal as a function of time. Liquid meal containing 8% peptone and 25 g/liter glucose was instilled by intragastric gavage. Human insulin was measured before and at 20, 40, and 60 min after the treatment. Results are means ± SEM from at least four animals per group.
To substantiate these findings, we next examined blood glucose levels of wild-type and G-InsKi mice at various times. As shown in Fig. 6C (left panel), blood glucose levels were markedly elevated at 20 and 40 min after the administration of the peptone-glucose meal in wild-type mice. In contrast, the blood glucose levels of G-InsKi mice did not show a significant change at either 20 or 40 min after the administration of the peptone-glucose meal. The unchanged blood glucose levels in G-InsKi mice corresponded to the release of human insulin (Fig. 6C, right panel). At 60 min, when human insulin declined to basal level, blood glucose levels in two groups of mice were no longer statistically different. Importantly, at 2 and 3 h after the administration of the meal, blood glucose levels were at the basal level in both groups of mice. Taken together, these results suggest that meal-stimulated release of G cell-produced insulin prevented the increase in blood glucose seen in the control mice. Importantly, the release of gastric insulin did not cause toxic hypoglycemia.
Gastric insulin reduces hyperglycemia and prolongs life span in Akita diabetic mice
Next, we explored whether gastric insulin could ameliorate diabetes in G-InsKi mice made diabetic by injection of STZ. We found that in STZ-diabetic G-InsKi mice, peptone stimulation failed to induce any increase in serum levels of either human insulin or mouse gastrin, suggesting that STZ interferes with G cell function (Fig. 5, A and B, hatched boxes). Indeed, treatment with STZ caused a 60% decrease in antral content of gastric insulin in G-InsKi mice (84 ± 19 pmol/g wet weight, n = 8) compared with untreated G-InsKi mice (207 ± 37.0 pmol/g wet weight, n = 6, see above). Accordingly, there was no difference in fasting blood glucose levels between STZ-diabetic G-InsKi and WT (347 ± 33 and 326 ± 22 mg/dl, respectively). Therefore, it appears that STZ-induced diabetes is not an appropriate model for evaluating the role of G cell-produced insulin in the development of diabetes in G-InsKi mice. It is known that somatostatin exerts inhibitory action on gastrin release from G cells (54). Because STZ induces an increase in pancreatic and antral somatostatin (55, 56, 57, 58, 59), it is likely that increased somatostatin could explain the lack of response of gastric (human) insulin and mouse gastrin in this system (Fig. 5, A and B, hatched bars). Furthermore, an effect of STZ treatment on the number of G cells cannot be ruled out either (60).
These results and considerations prompted us to test whether gastric insulin can ameliorate diabetes in a different model system. As an alternative approach, we used a genetic model of ?-cell loss, Akita diabetic mice (Ins2Akita), and assessed the effect of gastric insulin on the development of diabetes in these mice. In Ins2Akita mice, diabetes results from the altered function of pancreatic ?-cells. The Ins2Akita mice carry an autosomal dominant mutation Ins2C96Y in one of the alleles of the gene for insulin 2 (36, 61). This mutation produces gradual accumulation of malfolded proinsulin-2, which is retained in the pancreatic ?-cell endoplasmic reticulum and causes progressive loss of ?-cells due to proteotoxicity (36, 62). Consequently, islets from Ins2Akita mice are depleted of ?-cells, and those remaining release very little mature insulin, causing juvenile-onset hyperglycemia and insulinopenia in the absence of obesity (36, 61). Although untreated homozygotes rarely survive beyond 12 wk of age, heterozygous Ins2Akita mice are viable and fertile. The diabetic phenotype is more severe and progressive in males than in females, and the mean life span of diabetic male mice on the C57Bl/NJcl background is significantly shorter than that of nondiabetic males in another colony of the same strain (305 vs. 690 d).
Consequently, G-InsKi mice were crossed to Ins2Akita mice that are maintained on the same C57Bl/6 background as G-InsKi mice, eliminating the effect of the undefined genetic background on diabetogenic potential of single-gene mutations (63). Mice heterozygous for the presence of the Ins2Akita allele (Ins2Akita/+), G-Inski allele (G-InsKi+/–), or both (Ins2Akita/+/G-InsKi+/–) were produced and identified as shown in Fig. 7A. Specifically, the presence of the Ins2Akita allele was identified by PCR amplification of the 280-bp fragment of the mouse insulin gene 2 and analyzed by digestion with restriction endonuclease Fnu4HI, as described elsewhere (36) and shown in Fig. 7A (upper and middle panels). The digest of wild-type alleles produced two fragments, 140 bp each, whereas mutated alleles did not have the Fnu4HI site. In the example shown in Fig. 7A (middle and bottom panels), mice 2 and 5 were Ins2Akita/+/G-InsKi+/–, mouse 6 was Ins2Akita/+, and mice 1, 3, 4, and 7 were G-InsKi+/– mice.
FIG. 7. Gastric insulin prolongs life span in Akita diabetic mice. A, Molecular identification of Ins2Akita/+/G-InsKi+/– mice. Upper and middle panels, detection of the Ins2Akita allele by PCR amplification of the 280-bp fragment of the mouse insulin gene 2 followed by restriction digest with Fnu4HI. Bottom panel, PCR identification of the G-InsKi allele with primers Ins25 and Ins33. In this example, mice 2 and 5 are Ins2Akita/+/G-InsKi+/–, mouse 6 is Ins2Akita/+, and mice 1, 3, 4, and 7 are G-InsKi+/– mice. B, Survival analysis of Ins2Akita/+ mice in the absence or presence of G-InsKi allele. Solid line, Ins2Akita/+ mice (n = 15, seven censored subjects, eight deaths); dotted line, Ins2Akita/+/G-InsKi+/– mice (n = 11, eleven censored subjects, no deaths), the ticks portray censored time points. Survival curves were created using the method of Kaplan and Meier. A logrank test was used to determine statistical significance between two curves (P < 0.002).
We found that the presence of gastric insulin caused statistically significant reduction in fasting hyperglycemia from 327 ± 36 to 243 ± 23 mg/dl in Ins2Akita/+/G-InsKi+/– compared with Ins2Akita/+ (n = 10 and 16, respectively; P < 0.05). Subsequently, we confirmed that the Ins2Akita/+ allele did not have any effect on circulating fasting levels of gastric insulin in Ins2Akita/+/G-InsKi+/– (11.2 ± 1 pM, n = 5) compared with G-InsKi mice (10.8 ± 0.7, n = 12, see above).
The salient feature of the result shown in Fig. 7 is the striking difference in the survival in Ins2Akita mice in the presence or absence of G-Inski allele, i.e. gastric insulin. The median survival of mice in the Ins2Akita/+ group was 10.3 months, which is in excellent agreement with the mean life span of diabetic male mice on the C57Bl/NJcl background. In striking contrast, mice carrying the G-InsKi allele were all alive at the end of the study that lasted over 13 months (Fig. 7B). The prolongation of life span in the Ins2Akita/+/G-InsKi+/– mice compared with Ins2Akita/+ mice was highly statistically significant (P < 0.002), as judged by the Logrank test.
Concluding remarks
Diabetes mellitus is a debilitating disease affecting more than 150 million people worldwide. Gene therapy is one of the attractive novel approaches to its treatment that are being explored by many laboratories. A fundamental question in this approach is the identification of an ideal target for the conversion into an insulin-producing cell. Ostensibly, glucose-sensing cells that have the same glucose feedback as original ?-cells would be desirable targets in gene therapy for type I diabetes. However, in type II diabetes, the glucose-sensing mechanisms itself appears to be the possible target of alteration (20), suggesting that release of insulin from other glucose-sensing cells engineered to produce this hormone could be impaired as well. Consequently, meal-regulated secretion of insulin by cells functioning outside of the glucose-sensing loop may be advantageous or complementary in the treatment of non-insulin-dependent diabetes.
In this study, we describe an animal model of meal-stimulated secretion of insulin from enteroendocrine cells. G-InsKi mice produce and process insulin in antral G cells. G cell-produced insulin retains full biological activity as evidenced by the ability of transgenic gastric extract to correct glucose levels in STZ-diabetic mice. Importantly, gastric insulin is released into circulation in response to meal-associated stimuli and regulates meal-induced increase in blood glucose levels. Notably, levels of gastric insulin return to basal at 60 min after the meal and do not cause toxic hypoglycemia even 2–3 h later. The circulating levels of gastric insulin in G-InsKi mice are further increased after 4 d of treatment with PPIs, pharmacological agents widely used to inhibit gastric acid production. In this case, peptone meal releases levels of transgenic insulin that are over 20% of the postprandial circulating pancreatic insulin in the G-InsKi mice.
The effect of gastric insulin on the development of diabetes was assessed in Ins2Akita mice, in which ?-cell loss is caused by endoplasmic reticulum stress induced by impaired folding of mutated proinsulin-2 (36, 61, 62). We generated Ins2Akita/+ mice that carry the G-InsKi allele (Ins2Akita/+/G-InsKi+/–). The most striking aspect of the findings presented here is that in the presence of the G-InsKi allele, Ins2Akita/+ mice exhibited a marked prolongation of life span. These results imply that G cell-derived transgenic insulin is beneficial in the amelioration of diabetes.
In summary, the results presented here demonstrate the feasibility of meal-regulated physiological secretion of insulin by G cells. Moreover, our increasing knowledge of G cell-specific receptors, including receptors for neurotransmitters, ions, amino acids, and tastants (46, 54, 64, 65), suggests that meals may be adjusted to the needs of individual patients by adding defined nutrients that either promote or diminish hormonal secretion from G cells. It is conceivable that in future clinical applications, the therapeutic gastric insulin transgene could be delivered to patients noninvasively into the lumen by endoscopy, e.g. in the form of a viral vector carrying an insulin expression cassette under the regulation of the G cell-specific gastrin promoter. Consequently, it is plausible that an efficient G cell-based insulin gene therapy could protect diabetic patients from complications of insulin insufficiency and defend vulnerable ?-cells from exhaustion by increased secretory demand caused by insulin resistance (66, 67).
Acknowledgments
The authors are grateful to Drs. S. Guha, A. Lunn, and I. Kurland for comments on the manuscript.
References
King H, Rewers M 1993 Global estimates for prevalence of diabetes mellitus and impaired glucose tolerance in adults. WHO Ad Hoc Diabetes Reporting Group. Diabetes Care 16:157–177
The Diabetes Control and Complications Trial Research Group 1993 The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med 329:977–986
Abai AM, Hobart PM, Barnhart KM 1999 Insulin delivery with plasmid DNA. Hum Gene Ther 10:2637–2649
Goldfine ID, German MS, Tseng HC, Wang J, Bolaffi JL, Chen JW, Olson DC, Rothman SS 1997 The endocrine secretion of human insulin and growth hormone by exocrine glands of the gastrointestinal tract. Nat Biotechnol 15:1378–1382
Valera A, Fillat C, Costa C, Sabater J, Visa J, Pujol A, Bosch F 1994 Regulated expression of human insulin in the liver of transgenic mice corrects diabetic alterations. FASEB J 8:440–447
Kolodka TM, Finegold M, Moss L, Woo SL 1995 Gene therapy for diabetes mellitus in rats by hepatic expression of insulin. Proc Natl Acad Sci USA 92:3293–3297
Muzzin P, Eisensmith RC, Copeland KC, Woo SL 1997 Hepatic insulin gene expression as treatment for type 1 diabetes mellitus in rats. Mol Endocrinol 11:833–837
Lu D, Tamemoto H, Shibata H, Saito I, Takeuchi T 1998 Regulatable production of insulin from primary-cultured hepatocytes: insulin production is up-regulated by glucagon and cAMP and down-regulated by insulin. Gene Ther 5:888–895
Chen R, Meseck M, McEvoy RC, Woo SL 2000 Glucose-stimulated and self-limiting insulin production by glucose 6-phosphatase promoter driven insulin expression in hepatoma cells. Gene Ther 7:1802–1809
Ferber S, Halkin A, Cohen H, Ber I, Einav Y, Goldberg I, Barshack I, Seijffers R, Kopolovic J, Kaiser N, Karasik A 2000 Pancreatic and duodenal homeobox gene 1 induces expression of insulin genes in liver and ameliorates streptozotocin-induced hyperglycemia. Nat Med 6:568–571
Thule PM, Liu J, Phillips LS 2000 Glucose regulated production of human insulin in rat hepatocytes. Gene Ther 7:205–214
Thule PM LJ 2000 Regulated hepatic insulin gene therapy of STZ-diabetic rats. Gene Ther 7:1744–1752
Chen R, Meseck ML, Woo SL 2001 Auto-regulated hepatic insulin gene expression in type 1 diabetic rats. Mol Ther 3:584–590
Kojima H, Fujimiya M, Matsumura K, Younan P, Imaeda H, Maeda M, Chan L 2003 NeuroD-betacellulin gene therapy induces islet neogenesis in the liver and reverses diabetes in mice. Nat Med 9:596–603
Lipes MA, Cooper EM, Skelly R, Rhodes CJ, Boschetti E, Weir GC, Davalli AM 1996 Insulin-secreting non-islet cells are resistant to autoimmune destruction. Proc Natl Acad Sci USA 93:8595–8600
Faradji RN, Havari E, Chen Q, Gray J, Tornheim K, Corkey BE, Mulligan RC, Lipes MA 2001 Glucose-induced toxicity in insulin-producing pituitary cells that coexpress GLUT2 and glucokinase. Implications for metabolic engineering. J Biol Chem 276:36695–36702
Wu L, Nicholson W, Wu CY, Xu M, McGaha A, Shiota M, Powers AC 2003 Engineering physiologically regulated insulin secretion in non-beta cells by expressing glucagon-like peptide 1 receptor. Gene Ther 10:1712–1720
Cheung AT, Dayanandan B, Lewis JT, Korbutt GS, Rajotte RV, Bryer-Ash M, Boylan MO, Wolfe MM, Kieffer TJ 2000 Glucose-dependent insulin release from genetically engineered K cells. Science 290:1959–1962
Chan L, Fujimiya M, Kojima H 2003 In vivo gene therapy for diabetes mellitus. Trends Mol Med 9:430–435
Kahn SE 2003 The relative contributions of insulin resistance and ?-cell dysfunction to the pathophysiology of type 2 diabetes. Diabetologia 46:3–19
Chan CB, Saleh MC, Koshkin V, Wheeler MB 2004 Uncoupling protein 2 and islet function. Diabetes 53:S136–S142
Bell GI, Pilkis SJ, Weber IT, Polonsky KS 1996 Glucokinase mutations, insulin secretion, and diabetes mellitus. Annu Rev Physiol 58:171–186
Taylor IL, Singer M, Kauffman Jr GL 1982 Time-dependent effects of vagotomy on pancreatic polypeptide release. Dig Dis Sci 27:491–494
McArthur KE, Isenberg JI, Hogan DL, Dreier SJ 1983 Intravenous infusion of L-isomers of phenylalanine and tryptophan stimulate gastric acid secretion at physiologic plasma concentrations in normal subjects and after parietal cell vagotomy. J Clin Invest 71:1254–1262
Eysselein VE, Kovacs TO, Kleibeuker JH, Maxwell V, Reedy T, Walsh JH 1992 Regulation of gastric acid secretion by gastrin in duodenal ulcer patients and healthy subjects. Gastroenterology 102:1142–1148
Varner AA, Modlin IM, Walsh JH 1981 High potency of bombesin for stimulation of human gastrin release and gastric acid secretion. Regul Pept 1:289–296
Hirschowitz BI, Molina E 1983 Relation of gastric acid and pepsin secretion to serum gastrin levels in dogs given bombesin and gastrin-17. Am J Physiol 244:G546–G551
Yang H, Wong H, Wu V, Walsh JH, Tache Y 1990 Somatostatin monoclonal antibody immunoneutralization increases gastrin and gastric acid secretion in urethane-anesthetized rats. Gastroenterology 99:659–665
Stepan V, Sugano K, Yamada T, Park J, Dickinson CJ 2002 Gastrin biosynthesis in canine G cells. Am J Physiol Gastrointest Liver Physiol 282:G766–G775
Sawada M, Finniss S, Dickinson CJ 2000 Diminished prohormone convertase 3 expression (PC1/PC3) inhibits progastrin post-translational processing. Regul Pept 89:19–28
Rehfeld JF, Lindberg I, Friis-Hansen L 2002 Progastrin processing differs in 7B2 and PC2 knockout animals: a role for 7B2 independent of action on PC2. FEBS Lett 510:89–93
Macro JA, Bate GW, Varro A, Vaillant C, Seidah NG, Dimaline R, Dockray GJ 1997 Regulation by gastric acid of the processing of progastrin-derived peptides in rat antral mucosa. J Physiol 502:409–419
Koh TJ, Wang TC 1995 Molecular cloning and sequencing of the murine gastrin gene. Biochem Biophys Res Commun 216:34–41
Zhukova E, Afshar A, Ko J, Popper P, Pham T, Sternini C, Walsh JH 2001 Expression of the human insulin gene in the gastric G cells of transgenic mice. Transgenic Res 10:329–341
Miller SA DD, Polesky HF 1988 A simple salting out procedure for extracting DNA from human nucleated cells. Nucleic Acids Res 16:1215
Wang J, Takeuchi T, Tanaka S, Kubo SK, Kayo T, Lu D, Takata K, Koizumi A, Izumi T 1999 A mutation in the insulin 2 gene induces diabetes with severe pancreatic beta-cell dysfunction in the Mody mouse. J Clin Invest 103:27–37
Pham T, Guerrini S, Wong H, Reeve Jr J, Sternini C 2002 Distribution of galanin receptor 1 immunoreactivity in the rat stomach and small intestine. J Comp Neurol 450:292–302
Campbell DH,et al. 1963 Puncture of the ophthalmic venous plexus. In: Methods in immunology: a laboratory text for instruction and research. New York: W. A. Benjamin
Hem A, Smith AJ, Solberg P 1998 Saphenous vein puncture for blood sampling of the mouse, rat, hamster, gerbil, guinea pig, ferret and mink. Lab Anim 32:364–368
Baggio L, Kieffer TJ, Drucker DJ 2000 Glucagon-like peptide-1, but not glucose-dependent insulinotropic peptide, regulates fasting glycemia and nonenteral glucose clearance in mice. Endocrinology 141:3703–3709
Koh TJ, Goldenring JR, Ito S, Mashimo H, Kopin AS, Varro A, Dockray GJ, Wang TC 1997 Gastrin deficiency results in altered gastric differentiation and decreased colonic proliferation in mice. Gastroenterology 113:1015–1025
Friis-Hansen L, Sundler F, Li Y, Gillespie PJ, Saunders TL, Greenson JK, Owyang C, Rehfeld JF, Samuelson LC 1998 Impaired gastric acid secretion in gastrin-deficient mice. Am J Physiol 274:G561–G568
Taylor IL, Byrne WJ, Christie DL, Ament ME, Walsh JH 1982 Effect of individual l-amino acids on gastric acid secretion and serum gastrin and pancreatic polypeptide release in humans. Gastroenterology 83:273–278
Conigrave AD, Quinn SJ, Brown EM 2000 L-amino acid sensing by the extracellular Ca2+-sensing receptor. Proc Natl Acad Sci USA 97:4814–4819
Kobilka B 2000 Allosteric activation of the CaR by L-amino acids. Proc Natl Acad Sci USA 97:4419–4420
Young SH, Rozengurt E 2002 Amino acids and Ca2+ stimulate different patterns of Ca2+ oscillations through the Ca2+-sensing receptor. Am J Physiol Cell Physiol 282:C1414–C1422
Buchan AMJ, Squires PE, Ring M, Meloche RM 2001 Mechanism of action of the calcium-sensing receptor in human antral gastrin cells. Gastroenterology 120:1128–1139
Squires PE, Harris TE, Persaud SJ, Curtis SB, Buchan AM, Jones PM 2000 The extracellular calcium-sensing receptor on human beta-cells negatively modulates insulin secretion. Diabetes 49:409–417
Straub SG, Mulvaney-Musa J, Yajima H, Weiland GA, Sharp GW 2003 Stimulation of insulin secretion by denatonium, one of the most bitter-tasting substances known. Diabetes 52:356–364
Wallmark B, Larsson H, Humble L 1985 The relationship between gastric acid secretion and gastric H+,K+-ATPase activity. J Biol Chem 260:13681–13684
Lorentzon P, Eklundh B, Brandstrom A, Wallmark B 1985 The mechanism for inhibition of gastric (H+ + K+)-ATPase by omeprazole. Biochim Biophys Acta 817:25–32
Sundler F, Hakanson R, Carlsson E, Larsson H, Mattsson H 1986 Hypergastrinemia after blockade of acid secretion in the rat: trophic effects. Digestion 35(Suppl 1):56–69
Pawlikowski M, Karbownik M, Lewinski A, Pisarek H, Wajs E, Szkudlinski M 1992 Effects of omeprazole on the number of immunoreactive gastrin- and somatostatin-cells in the rat gastric mucosa. Histol Histopathol 7:153–156
Sachs G, Zeng N, Prinz C 1997 Physiology of isolated gastric endocrine cells. Annu Rev Physiol 59:243–256
Patel YC, Cameron DP, Bankier A, Malaisse-Lagae F, Ravazzola M, Studer P, Orci L 1978 Changes in somatostatin concentration in pancreas and other tissues of streptozotocin diabetic rats. Endocrinology 103:917–923
Chiba T, Kadowaki S, Taminato T, Abe H, Chihara K, Matsukura S, Goto Y, Seino Y, Fujita T 1981 Concentration and secretion of gastric somatostatin in streptozotocin-diabetic rats. Diabetes 30:188–191
Chiba T, Taminato T, Kadowaki S, Chihara K, Matsukura S, Nozawa M, Seino Y, Fujita T 1981 Reversal of increased gastric somatostatin in streptozotocin-diabetic rats by whole pancreas transplantation. Diabetes 30:724–727
Michels JE, Bauer GE, Johnson D, Dixit PK 1986 Morphometric analysis of the endocrine cell composition of rat pancreas following treatment with streptozotocin and nicotinamide. Exp Mol Pathol 44:247–258
Aughsteen AA 2000 An ultrastructural study on the effect of streptozotocin on the islets of Langerhans in mice. J Electron Microsc (Tokyo) 49:681–690
Nwokolo CU, Debnam ES, Booth JD, Sim R, Sankey EA, Dhillon AP, Pounder RE 1992 Neuroendocrine changes in rat stomach during experimental diabetes mellitus. Dig Dis Sci 37:751–756
Yoshioka M, Kayo T, Ikeda T, Koizumi A 1997 A novel locus, Mody4, distal to D7Mit189 on chromosome 7 determines early-onset NIDDM in nonobese C57BL/6 (Akita) mutant mice. Diabetes 46:887–894
Ron D 2002 Proteotoxicity in the endoplasmic reticulum: lessons from the Akita diabetic mouse. J Clin Invest 109:443–445
Leiter EH 2002 Mice with targeted gene disruptions or gene insertions for diabetes research: problems, pitfalls, and potential solutions. Diabetologia 45:296–308
Rozengurt E, Walsh JH 2001 Gastrin, CCK, signaling and cancer. Annu Rev Physiol 63:49–76
Wu SV, Rozengurt N, Yang M, Young SH, Sinnett-Smith J, Rozengurt E 2002 Expression of bitter taste receptors of the T2R family in the gastrointestinal tract and enteroendocrine STC-1 cells. Proc Natl Acad Sci USA 99:2392–2397
DeFronzo RA, Ferrannini E 1991 Insulin resistance. A multifaceted syndrome responsible for NIDDM, obesity, hypertension, dyslipidemia, and atherosclerotic cardiovascular disease. Diabetes Care 14:173–194
Kruszynska YT, Olefsky JM 1996 Cellular and molecular mechanisms of non-insulin-dependent diabetes mellitus. J Investig Med 44:413–428(Yu-Chun Lu, Catia Sternin)
Address all correspondence and requests for reprints to: Elena Zhukova, 900 Veteran Avenue, Warren Hall, Room 14-109, Department of Medicine, University of California School of Medicine, Los Angeles, California 90095-1786. E-mail: ezhukova@mednet.ucla.edu.
Abstract
We explored the hypothesis that meal-regulated release of insulin from gastric G cells can be used for gene therapy for diabetes. We generated transgenic mice in which the coding sequence of human insulin has been knocked into the mouse gastrin gene. Insulin was localized specifically to antral G cells of G-InsKi mice by double immunofluorescence staining using antibodies against insulin and gastrin. Insulin extracted from antral stomach of G-InsKi mice decreased blood glucose upon injection into streptozotocin-diabetic mice. Intragastric administration of peptone, a known potent luminal stimulant of gastrin secretion, induced an increase in circulating levels of transgenic human insulin from 10.7 ± 2 to 23.3 ± 4 pM in G-InsKi mice. Although G cell-produced insulin decreased blood glucose in G-InsKi mice, it did not cause toxic hypoglycemia. Proton pump inhibitors, pharmacological agents that increase gastrin output, caused a further increase in the circulating levels of gastric insulin (41.5 ± 2 pM). G cell-produced insulin was released into circulation in response to the same meal-associated stimuli that control release of gastrin. The most striking aspect of the results presented here is that in the presence of the G-InsKi allele, Ins2Akita/+ mice exhibited a marked prolongation of life span. These results imply that G cell-derived transgenic insulin is beneficial in the amelioration of diabetes. We suggest that an efficient G cells-based insulin gene therapy can relieve diabetic patients from daily insulin injections and protect them from complications of insulin insufficiency while avoiding episodes of toxic hypoglycemia.
Introduction
THE PREVALENCE OF diabetes mellitus has reached epidemic proportions in many parts of the world (1). This chronic group of disorders is characterized by no insulin secretory capacity (type I) or insulin resistance and defective endogenous insulin secretion (type II). The primary focus for the treatment of type I diabetes is to replace insulin secretion by the administration of exogenous insulin, and treatment of advanced type II often requires daily injections of insulin. Improved glycemic control significantly decreases long-term diabetes-associated microvascular complications (2) but is often difficult to achieve. Consequently, novel therapeutic approaches to diabetes are being explored, and several laboratories have focused on the possibility of using gene therapy to achieve optimal glycemic control. Despite considerable effort to engineer insulin production in a variety of tissues and cells, including muscle (3), exocrine pancreas (4), liver (5, 6, 7, 8, 9, 10, 11, 12, 13, 14), pituitary (15, 16, 17), and intestinal K cells (18), a well-characterized system that provides optimal glycemic control has not been generated yet.
The systems favored so far use glucose-regulated insulin secretion (5, 9, 10, 11, 12, 13, 14, 18). Each of these systems is characterized by certain advantages over the alternatives, albeit simultaneously tempered by particular limitations (reviewed in Ref. 19). Nevertheless, glucose-regulated systems have an enormous potential for the treatment of type I diabetes.
Type II diabetes, which represents about 90% of cases, poses different challenges. Our understanding of the complex molecular mechanisms leading to type II diabetes is increasing and evolving; consequently, the cellular targets selected for engineering insulin production must evolve as well. In type II diabetes, the glucose-sensing mechanism itself can be the target of alteration (reviewed in Refs. 20, 21, 22). Consequently, it would be advantageous to engineer physiologically regulated insulin secretion by cells that will not be affected by changes in glucose sensing.
Regulated exocytosis is a crucial aspect of instant and timely release of insulin from ?-cells and should be an essential feature of a cell selected as a target for the production of extrapancreatic insulin. Therefore, in searching for a candidate surrogate ?-cell, an important property would be the ability to respond to food intake with physiologically regulated exocytosis. Ideally, insulin release should mimic the normal pattern of insulin secretion, namely a basal level during day and night with brief increases coinciding with ingestion of meals. Consequently, in this paper, we explored a new putative target for diabetes gene therapy and focused on reengineering gastrin-secreting G cells into secretagogue-responsive insulin-secreting cells.
G cells, endocrine cells located in the antral part of the stomach, release the peptide hormone gastrin in response to protein-rich meals. Although G cells are insensitive to changing glucose levels, the secretion of gastrin from these cells is tightly linked to food intake and occurs within minutes (23, 24, 25, 26, 27, 28). These cells synthesize preprogastrin and process it with the same endopeptidases that are implicated in proinsulin processing, namely PC1/PC3, PC2, and carboxypeptidase H (29, 30, 31, 32). Similarly to pancreatic ?-cells that store and secrete insulin, G cells store gastrin in secretory vesicles and release it upon stimulation into the portal circulation. Importantly, G cells are open to the gastric lumen, through which gene delivery agents can directly access them.
Therefore, we hypothesized that transgenic human insulin selectively produced by G cells under the regulation of the gastrin gene promoter will be released into circulation in response to the same meal-associated stimuli that control release of gastrin. In this study, we generated a novel mouse model (G-InsKi) in which human insulin is expressed in the gastric G cells under the regulation of the mouse gastrin gene promoter. G cell-produced insulin is functional and is released into circulation by intragastric administration of peptone, a potent stimulant of gastrin release. The most striking aspect of the results presented here is that in the presence of the G-InsKi allele, Ins2Akita/+ mice exhibited a marked prolongation of life span. These results imply that G cell-derived transgenic insulin is beneficial in the amelioration of diabetes.
Materials and Methods
Assembly of the targeting vector pG-InsKi
The pG-InsKi vector was assembled using the mouse gastrin gene (33) (a gift by Dr. Timothy C. Wang, University of Massachusetts Medical Center, Worcester, MA), insulin cDNA (a gift from Dr. Graeme I. Bell, University of Chicago, Chicago, IL), and a loxP-pgk-Neo cassette from the plasmid p1338 (kindly provided by Dr. Timothy Ley, Washington University, St. Louis, MO). Specifically, the 346-bp fragment of the human insulin cDNA (position +12 to +358) containing the coding sequence but not the polyadenylation signal was amplified from oligonucleotide primers InsFW (ACTGAATTCATCACTGTCCTTCTGCCATGG, sense) and InsRV (ACTTCTAGACGTCTAGTTGCAGTAGTTCTC, antisense) (Fig. 1A). This fragment was subcloned into the EcoRI and XbaI sites of the pIRESneo vector (BD Biosciences, Palo Alto, CA), which was used as an intermediate carrier and a donor of bovine GH polyadenylation signal (Fig. 1B). The EcoRI-XhoI fragment, containing the insulin coding sequence and the bovine growth hormone polyadenylation signal, was subcloned into the BamHI site of the plasmid loxP-pgk-Neo (version p1338) upstream from the loxP-pgk-Neo cassette (Fig. 1, C and D). The resulting 2-kb insulin/loxP-pgk-neo cassette was excised from the p1338 vector at NsiI sites and inserted into the exon 2 of the mouse gastrin gene at the HindIII site (Fig 1, D and E). The resulting targeting vector G-InsKi is shown in Fig. 1F. It has approximately 3 kb homology to the targeted mouse gastrin gene, both 5' and 3' from the point of knock-in insertion. The correct assembly of the G-InsKi construct was confirmed by DNA sequencing.
FIG. 1. Generation of the targeting construct. A, The 346-bp fragment of the human insulin cDNA (position +12–+358) containing the coding sequence but not the polyadenylation signal was amplified using oligonucleotide primers InsFW and InsRV. B, The amplified fragment was subcloned into the EcoRI and XbaI sites of the pIRESneo vector, which was used as an intermediate carrier and a donor of bovine growth hormone polyadenylation signal (hatched box). C and D, The EcoRI-XhoI fragment, containing the insulin coding sequence and the bovine growth hormone polyadenylation signal, was subcloned into the BamHI site of the plasmid loxP-pgk-Neo upstream from the loxP-pgk-Neo cassette. D and E, The resulting-2 kb INS/loxP-pgk-Neo cassette was excised from the vector at NsiI sites and inserted into the exon 2 of the mouse gastrin gene at the HindIII site. F, The targeting of vector G-InsKi.
Gene targeting and blastocyst injection
The G-InsKi targeting construct was linearized with SacII, electroporated into germline-competent LW-1ES cells (pure inbred 129/SVJae), and selected with G418. DNA microinjection and G418 selection were performed at the University of California (Los Angeles) Embryonic Stem Cell Facility. DNA was isolated from 288 surviving clones. ES clones that have undergone homologous recombination were identified by Southern blot analysis using a 200-bp NheI-BglII fragment as a 5'-external probe. Seven positive clones were identified and expanded. Two of those, clones 18 and 27, were microinjected into blastocysts from CB6F1 (a hybrid of C57Bl/6 and BALB/c) donors at the University of California (Irvine) Transgenic Mouse Facility. High-percentage male chimeras were obtained and mated with C57Bl/6 females. Germline transmission was identified by PCR analysis of genomic DNA obtained from ear snips. It should be pointed out that the Gas-Ins transgenic mouse described in a previous study from our group is an independent animal model that was generated by pronuclear injection (34). The transgene used in that study contained the genomic sequence of human insulin DNA. The aberrant splicing of insulin introns in G cells resulted in the production of a truncated form of the human insulin RNA and, consequently, led to the production of truncated insulin molecules (34). Consequently, we used the cDNA of human insulin for the generation of knock-in mice G-InsKi described in the present study.
Animals
All mice were housed in specific pathogen-free barrier facilities, maintained on a 12-h light, 12-h dark cycle, and fed a standard autoclavable rodent diet (PMI Feeds, Inc., Richmond, IN). G-InsKi mice were maintained on C57Bl/6 background as heterozygotes to avoid creation of the gastrin-null phenotype. Akita diabetic mice, strain C57Bl/6-Ins2Akita, were purchased form the Jackson Laboratory (Bar Harbor, ME) and maintained as heterozygotes. The studies were approved by UCLA’s Office for Protection of Research Subjects.
Genotype analysis
Genotyping was performed by PCR analysis of genomic DNA obtained from ear snips or tail biopsies. Genomic DNA was purified as described by Miller and Polesky (35). PCRs included 200 ng DNA and 2.5 U AmpliTaq Polymerase (PerkinElmer Life and Analytical Sciences, Boston, MA). PCR conditions were: 94 C, 1 min, followed by 35 cycles of 94 C, 30 sec; 55 C, 30 sec; and 72 C, 30 sec. The G-InsKi allele was detected using primers Ins 25, CTGTCCTTCTGCCATGG; and Ins 33, GTTGCAGTAGTTCTCCAGCTG. The identity of the PCR product was confirmed by digestion with the restriction endonuclease SexAI, which was added directly to the aliquot of the PCR mix. DNA samples negative for the presence of the G-InskKi transgene were analyzed for the presence of the ?-globin by PCR analysis from primers 5'CCAATGTGCTCACACAGGATAGAGAGGGCAGG 3' and 5' CCTTGAGGCTGTCCAAGTGATTCAGGCCATCG 3', which produce a 494-bp fragment. PCR identification of Ins2Akita allele was performed as described elsewhere (36).
Immunocytochemistry
Mice were deeply anesthetized with ip sodium pentobarbital (50 mg/kg Nembutal; Abbott Laboratories, Chicago, IL) (50 mg/kg ip) and then perfused intracardially with 4% paraformaldehyde in 0.1 M phosphate buffer (PB) (pH 7.4). The stomach and pancreas were dissected, postfixed with the same fixative for 1 h at room temperature, and then stored at 4 C in 25% sucrose in PB overnight. Tissue sections were then cut with a cryostat at 10–12 μm thickness, mounted on slides, and stored at –20 C until processing for immunohistochemistry. Cryostat sections were processed by the immunofluorescence method for single and double labeling as previously described (37). For single labeling, cryostat sections of the stomach or pancreas tissues were washed in PB (3 x 15 min), incubated in 10% normal donkey serum for 1 h at room temperature to minimize the background, incubated in either guinea pig antiporcine insulin antibody GPDI (kindly provided by Dr. K. Polonsky, University of Chicago) at 1:1250 or guinea pig antihuman insulin antibody (Linco Research, Inc., St. Charles, MO) (1:200) in PB containing 0.5% Triton X-100 overnight at 4 C, washed in PB (3 x 30 min), then incubated in affinity-purified donkey anti-guinea pig IgG conjugated with ALEXA 488 (1:1000) (Molecular Probes, Eugene, OR), washed again for 45 min, and coverslipped. Immunoblocking experiments were performed with human insulin peptide at 10–5 μM (Sigma, St. Louis, MO). For double-label immunofluorescence, cryostat sections of the stomach were incubated in a mixture of guinea pig antihuman insulin or GPDI antibody and rabbit gastrin antiserum (gastrin 1802 from CURE Antibody Core; 1:500) overnight, followed by a 2-h incubation with a mixture of donkey antirabbit IgG-ALEXA 488 (1:1000) (Molecular Probes) and donkey anti-guinea pig IgG conjugated with rhodamine Red X (Jackson Immunoresearch Laboratories, West Grove, PA; 1:300). Tissues were examined with a Zeiss Axioplan 2 research microscope for fluorescence with an Axiocam color digital camera fluorescence microscope (Carl Zeiss Inc., Thornwood, NY), equipped with fluorescein isothiocyanate and Red-X cubes, or a Zeiss 410 laser scanning confocal microscope equipped with a krypton/argon laser and attached to a Zeiss Axiovert 100 microscope with a 100x Plan Apo 1.4 numerical aperture objective (Carl Zeiss Inc.).
Analytical procedures
Blood for analyses was collected by saphenous vein or retroorbital bleeding (38, 39). Blood glucose concentrations were measured using the Accu-Chek Compact blood glucose monitoring system (Roche Diagnostic Corp., Indianapolis, IN). Human insulin in serum and in antral tissue extracts was determined using a human insulin ELISA kit, which does not recognize mouse insulin (Linco Research, Inc.). This kit has no cross-reactivity to intact human proinsulin and des proinsulin(31, 32); however, there can be cross-reactivity with des proinsulin(64, 65). Mouse insulin in serum was determined using the Mouse Insulin ELISA kit (Mercodia AB, Uppsala, Sweden), which also cross-reacts with human insulin (120%). Values of mouse serum insulin in transgenic mice were adjusted by subtracting respective values of human insulin. Mouse gastrin in serum and in antral tissue extracts was measured by human gastrin I (G17) immunoassay (R&D Systems, Minneapolis, MN), which has 70.7% cross-reactivity with rat gastrin I and 0.8% cross-reactivity with Big Gastrin (G34-I).
Extraction of insulin from stomachs
Gastric antrums were collected from mice, immediately frozen in liquid nitrogen, and stored at –70 C until used. Insulin was extracted by the modified procedure described by Baggio et al. (40). Specifically, the antrums were homogenized with PowerGen homogenizer (Model 700) on ice in 5 volumes of extraction buffer [1 N HCl, 5% (vol/vol) formic acid, 1% (vol/vol) trifluoroacetic acid, and 1% (wt/v) NaCl]. Insoluble material was removed by centrifugation at 12,000 x g at 4 C for 15 min. Supernatant (1 ml) was loaded onto a Sep Pak C18 cartridge. The Sep Pak cartridges were washed with 2 ml methanol and 2 ml water followed by 2 ml extraction buffer. After loading the samples, the cartridges were washed with 2 ml extraction buffer before eluting the samples with 1 ml 80% (vol/vol) isopropanol containing 0.1% (vol/vol) trifluoroacetic acid. Human insulin in tissue extracts was measured using the specific ELISA kit as described above. Antrums from 19 transgenic mice were pooled in six groups, and each of the groups was processed independently.
Western blot analysis
For immunoblotting, extracts of gastric antrum containing at least 50 ng human insulin as measured by ELISA were loaded onto a 10% NuPAGE minigel in a reducing sample buffer containing 50 mM dithiothreitol and subjected to electrophoresis for 35 min at 200 V using 2-(N-morpholino)ethanesulfonic acid running buffer (Invitrogen, Carlsbad, CA). Proteins were transferred to Immobilon-PSQ membranes (Millipore, Billerica, MA) in Tris-glycine transfer buffer at 35 V for 1 h. Membranes were blocked with 5% (wt/vol) nonfat skim milk in PBS (pH 7.2) for at least 2 h at room temperature and then incubated at room temperature for 3 h with guinea pig antipig insulin whole antiserum (I8510, Sigma-Aldrich) diluted 1:100 in 3% (wt/vol) nonfat skim milk in PBS. This antiserum has 100% cross-reactivity with human insulin. Immunoreactive bands were visualized using horseradish peroxidase-conjugated rabbit anti-guinea pig IgG (A5545, Sigma-Aldrich) and enhanced immunofluorescence methods. Specificity of the staining was confirmed by preabsorbtion with 10 μg/ml Humulin (Eli Lilly, Indianapolis, IN) for 1 h at 37 C.
In vivo experimental procedures
Male mice, 6–8 wk old, were used for experiments. Peptone solution in water (Becton Dickinson, Sparks, MD) or omeprazole suspension in hydroxypropyl methylcellulose (AstraZeneca, Wilmington, DE) were instilled into anesthetized mice by intragastric gavage at the volume of 1% of body weight. For ip injections, human insulin in stomach extracts was quantified by specific ELISA as described above, vacuum dried, made up in 3 μM HCl, and injected ip at a dose of 75 mU/g body weight into streptozotocin (STZ)-diabetic mice. Control groups of STZ-diabetic mice were injected with antral extract from nontransgenic mice normalized by the weight of the wet tissue used for the extraction either with or without Humulin.
Six-week-old male mice, either wild-type C57Bl/6 or GinsKi, were made diabetic by ip injections of STZ at a dose of 100 mg/kg of body weight per day, on 2 d, 1 d apart. The total amount of STZ injected on 2 separate d was 200 mg/kg body weight. Mice were fasted overnight before the injections. At d 14 after the injection, mice developed diabetes. Development of diabetes was assessed by measuring fasting glucose levels and mouse insulin in serum.
Statistical analysis
All results are presented as the means ± SEM. Statistical significance was determined by Student’s t test. P < 0.05 was considered significant. Welch’s correction was used to assess the variances wherever necessary. Survival analysis was performed using GraphPad Prism version 3.02 for Windows (GraphPad Software, San Diego, CA).
Results and Discussion
Generation of G-InsKi mice
The G-InsKi transgenic mice in which human insulin is produced in G cells under the regulation of the gastrin gene promoter were generated by knocking-in the coding sequence of the human insulin gene into the exon 2 of the mouse gastrin gene (Fig. 1). The knock-in approach was selected over the transgenic overexpression by pronuclei injection to ensure that all components of the gastrin gene-specific transcriptional regulation were preserved.
Initially, the G-InsKi mice were identified by Southern blot analysis using a 200-bp NheI-BglII fragment as a 5'-external probe (Fig. 2A). The substitution of the mouse gastrin gene with the G-InsKi allele increased the NheI restriction fragment size from 7 to 9 kb (Fig. 2B). Subsequently, genotyping was performed by PCR analysis using primers Ins 25 and Ins 33, which amplify a 340-bp fragment of human insulin cDNA (Fig. 2, A and C). The Ins 25 primer corresponds to the 5'-untranslated region of the human insulin (c)DNA and has no homology with the mouse insulin genes I and II. As a result, this set of primers distinguishes between human and mouse insulin nucleotide sequence. The identity of the 340-bp PCR product was confirmed by digestion with the restriction endonuclease SexAI, which generates diagnostic fragments of 230 and 110 bp, characteristic of this fragment of human insulin cDNA (Fig. 2C).
FIG. 2. Gene targeting and genotyping. A, Partial map of the G-InsKi knock-in allele. The gastrin gene is shown in black. Gastrin exons are shown as filled boxes. Gastrin exon 2 is interrupted by the insulin cDNA linked to the bovine growth hormone polyadenylation signal (INS) and a phosphoglycerol kinase-neomycin resistance gene cassette flanked by two 34-bp loxP sequences (Neo), both shown as white boxes. The location of the DNA fragment used as probe is indicated. The position of primers used for the PCR genotyping is shown. B, Southern blot analysis. Genomic DNA was digested with NheI and then probed with a RNA probe corresponding to the NheI/BglII fragment of the 5'-untranslated region of the mouse gastrin gene (200 bp). The location of the probe is indicated in A. Expected alleles and sizes are shown. G-InsKi, Transgenic allele; WT, wild-type gastrin allele. C, Genotyping by PCR. Genomic DNA was amplified from primers Ins25 and Ins33, shown in A. Note, that Ins25 has no homology with the mouse insulin genes I and II. Authenticity of the 340-bp PCR fragment was validated by the digest with SexAI that produced 230- and 110-bp fragments. ML, DNA mass ladder; WT, negative control, amplification of the wild-type genomic DNA; WT + PL, positive control, amplification of the wild-type genomic DNA spiked with 0.06 pg targeting vector G-InsKi; TG, – and +, Transgenic genomic DNA amplified and incubated without or with SexAI.
The G-InsKi mice were maintained as heterozygotes to avoid creation of the gastrin-null phenotype, which is associated with abnormalities of gastric function, as well as changes in architecture and differentiation of the gastric mucosa (41, 42). Mice that carry a single normal allele of the gastrin gene have physiological levels of circulating gastrin and normal gastric function (42). G-InsKi mice had a normal reproduction life and were healthy.
Localization and antral content of gastric insulin
In accordance with the tissue distribution of the gastrin-producing G cells, in G-InsKi mice, human insulin was expressed by cells located at the bottom of gastric glands in the antrum, as revealed by immunocytochemistry (Fig. 3A). The expression of human insulin in gastric G cells was specific. A few positively stained cells were localized in the duodenum, a minor site of gastrin expression. No insulin-expressing cells were found in oxyntic mucosa, a region of stomach devoid of G cells, or in jejunum, ileum, colon, and liver. Furthermore, insulin-expressing cells in the antrum were identified as G cells by double immunofluorescence staining with antibodies against insulin and gastrin (Fig. 3, B–D). Importantly, every cell positive for gastrin staining was also positive for insulin, indicating the high targeting efficiency of the G-InsKi transgene. Notably, cells positive for somatostatin staining (D cells) did not express any detectable immunoreactive insulin (data not shown). Insulin expressing cells were not found in the stomachs of nontransgenic mice (data not shown).
FIG. 3. Tissue and cell specificity of human insulin production in G-InsKi mice. A, Immunofluorescence staining of the antral tissue section for insulin, low-power view. Mus, Muscle wall. Arrows point to G cells. B–D, Double immunofluorescence staining of the tissue section of antral stomach for insulin (B, red) and gastrin (C, green) shown in high-power view. Colocalization of the signal is shown in D (yellow).
Human insulin content measured in the antral extract by human insulin-specific immunoassay was 207 ± 37.0 pmol/g wet weight (mean ± SEM, n = 6) in G-InsKi mice, whereas it was undetectable in controls (n = 3) (Fig. 4A). This value is comparable with the content of gastrin in antrums of heterozygous gastrin knockout mice (339 ± 94 pmol/g) (42). Pancreatic content of mouse insulin measured for comparison was 3590 ± 640 pmol/g wet weight (mean ± SEM, n = 6). However, these data are not corrected for the number of cells that produce insulin in each tissue.
FIG. 4. Antral content, molecular forms, and biological activity of G cell-produced human insulin. A, Detection and quantification of human insulin in the extracts of antral stomachs by human insulin-specific immunoassay. Values are means ± SEM, n = 19. Note that this kit has no cross-reactivity to intact human proinsulin and des proinsulin(31 32 ); however, there can be cross-reactivity with des proinsulin(64 65 ). B, Western blot analysis of human insulin expression in the extract of antral stomach from G-InsKi mice and nontransgenic littermates. Arrowheads point at the product corresponding to the size of human insulin chains A and B (3.5 and 2.5 kDa, respectively), as well as at the expected location of intermediately processed forms (5.5–6.5 kDa) and intact proinsulin (9 kDa). TG, Transgenic; WT, wild-type. C, G cell-produced human insulin in antral extracts has hypoglycemic activity. Wild-type C57Bl/6 mice were made diabetic by injection of STZ. Extracts of antral stomach from transgenic G-InsKi and nontransgenic littermates were injected into STZ-diabetic mice. Transgenic extracts provided a dose of 75 mU human insulin/kg body weight. Amount of nontransgenic extract was normalized by the weight of the wet tissue used for the extraction. The third group of mice was injected with nontransgenic extract supplemented with Humulin (Eli Lilly) to provide a dose of 75 mU/kg body weight. Blood glucose levels at 20, 40, and 60 min after the injection are shown. White bars, Mice injected with the nontransgenic extract (n = 3). Gray bars, Mice injected with transgenic extract (n = 5). Hatched bars, Mice injected with nontransgenic extract supplemented with Humulin (n = 3). Results are means ± SEM. Number of animals in each group is shown above.
G cells express prohormone convertases PC2 and PC1/PC3 as well as carboxypeptidase H (29, 30, 31, 32), suggesting that proinsulin could be completely processed in G cells. Western blot analysis, illustrated in Fig. 4B, showed that human insulin from antral extracts of G-InsKi mice migrated as a broad band with an apparent molecular mass of approximately 2.5–3.5 kDa, which corresponds to the size of fully processed insulin, namely chains A and B. Preincubation of the antibody with human insulin completely blocked the appearance of this immunoreactive band (data not shown). No immunoreactive bands corresponding to the size of proinsulin (9 kDa) or its intermediately processed forms (5.5–6.5 kDa) were observed in extracts from G-InsKi mice. We did not detect any immunoreactive band when antral extracts from nontransgenic littermates were analyzed.
Circulating gastric insulin in G-InsKi mice
Circulating levels of human insulin in G-InsKi mice were measured with the same human insulin-specific immunoassay, which has no cross-reactivity with mouse insulin, human proinsulin, or des proinsulin (31, 32). The levels of human insulin in samples collected from individual mice were 10.8 ± 0.7 pM in fasting state and 20.7 ± 3 pM in postabsorbtive state (values are means ± SEM; n = 12 and 19, respectively). Fasting blood glucose levels were similar in transgenic and nontransgenic mice, 89 ± 3.4 and 88 ± 2.5 mg/dl, respectively (means ± SEM, n = 24 and 39). Levels of circulating mouse insulin in G-InsKi mice and wild-type controls were 115 ± 13 and 113 ± 16 pM in fasting state and 192 ± 29 and 274 ± 55 pM in postabsorbtive state, respectively (means ± SEM, from at least eight animals in each group). Although the reduction in postabsorbtive levels of pancreatic mouse insulin in G-InsKi mice is in line with the expected down-regulation due to the presence of transgenic human insulin, it did not reach statistical significance in our experiments.
Biological activity of gastric insulin
The biofunctionality of G cell-produced insulin was assessed by its ability to lower blood glucose upon injection into diabetic mice. STZ-diabetic mice were fasted overnight and injected via ip route with the antral extracts of either transgenic G-InKi mice or nontransgenic littermates. As shown in Fig. 4C, antral extract from the nontransgenic mice did not decrease blood glucose levels (white bars), whereas injection of the G-InsKi extract caused a dramatic decrease in blood glucose levels at 20, 40, and 60 min after the injection (gray bars). Interestingly, the efficiency with which G cell-produced insulin regulates glucose levels is virtually identical with the efficiency of an equivalent dose of human insulin added to the control extracts (hatched bars, Fig. 4C). Collectively, results presented in Fig. 4 indicate that insulin is fully processed in gastric G cells.
Regulation of gastric insulin release
Having established that G cell-produced insulin is functional, we next examined whether the release of human insulin by G cells is regulated by the same meal-associated stimuli that control gastrin release. G cells are known to secrete gastrin in response to meal-associated stimuli that act through luminal and neural pathways. The main luminal stimulants of gastrin secretion are small peptides and, in particular, free aromatic amino acids (24, 25, 43). Recent studies of the Ca2+/amino acid-sensing receptor and bitter taste receptors and their localization to the gastric mucosa, and specifically to G cells, provided an insight into the initial molecular recognition event(s) by which G cells apparently sense chemical composition of luminal contents (44, 45, 46, 47). Interestingly, the Ca2+/amino acid-sensing receptor was also localized to ?-cells (48); intriguingly, ?-cells also respond to bitter tastants (49).
To assess whether luminal stimulation causes insulin release from G cells, we used peptone, an enzymatic digest of protein, which is a potent luminal stimulant of gastrin release. As shown in Fig. 5A, gastric instillation of 8% peptone solution induced a 2.2-fold increase in circulating levels of human insulin 40 min after treatment, demonstrating that release of human insulin is physiologically regulated by components of meals. Instillation of vehicle into the control group of animals had no effect on circulating insulin levels (data not shown). As expected, gastric instillation of 8% peptone solution also induced a marked increase in the level of serum gastrin, shown in Fig. 5B.
FIG. 5. Regulation of gastric insulin release. A, Liquid peptone meal induces release of G cell-produced insulin in G-InsKi mice. Mice were fasted overnight, and 8% peptone was instilled by intragastric gavage. Circulating human insulin was measured before (white bar) and 40 min after (black bar) the treatment. The same experiment was repeated with G-InsKi mice made diabetic by the injection of STZ (hatched bar, only levels after the peptone treatment are shown). Note that the human insulin-specific ELISA used in these studies does not cross-react with mouse insulin. Results are means ± SEM, n = 5. B, Liquid peptone meal induces release of gastrin in G-InsKi mice. Mice were fasted overnight, and 8% peptone was instilled by intragastric gavage. Circulating gastrin was measured before (white bar) and 40 min after (black bar) the treatment. The same experiment was repeated with G-InsKi mice made diabetic by the injection of STZ. Only levels after the peptone treatment are shown (hatched bar). Note that peptone did not induce the release of human insulin or mouse gastrin in STZ-treated mice. Results are means ± SEM, n = 7. C, PPIs increase circulating levels of G cell-produced insulin in G-InsKi mice. Mice in postabsorptive state received either omeprazole or vehicle by intragastric gavage once daily for 4 consecutive d at a dose of 100 mg/kg body weight. After the completion of omeprazole treatment, 8% peptone was instilled into the lumen. Circulating human insulin was measured before (data not shown) and after the treatment with vehicle (white bar), omeprazole (black bar), and peptone (gray bar). Results are means ± SEM, and each group included at least five animals. Vh, Vehicle; OZ, omeprazole; Pep, peptone.
Furthermore, we examined whether pharmacological agents that increase gastrin output also release G cell-produced insulin into the circulation. Proton pump inhibitors (PPIs) are commonly used to reduce gastric acid secretion by directly inhibiting H+,K+-ATPase in parietal cells (50, 51). Physiological consequences of PPI action include increase in the number of gastric G cells and an increase in levels of circulating gastrin (52, 53). To determine whether circulating levels of gastric insulin are regulated by PPIs, omeprazole was instilled into G-InsKi mice at 100 mg/kg body weight by intragastric gavage once daily for 4 consecutive d. Treatment with omeprazole induced a 2.3-fold increase in the basal circulating levels of human insulin in G-InsKi mice. The administration of peptone after the completion of omeprazole treatment further increased circulating human insulin (41.5 ± 2 pM). There was no change in human insulin levels in mice that were treated with the vehicle only (Fig. 5C). Collectively, these experiments demonstrate the release of G cell-produced insulin in response to protein hydrolyzates and to suppression of gastric acid production.
Regulation of blood glucose by meal-induced release of gastric insulin
We next explored whether meal-induced release of gastric insulin into portal circulation regulates blood glucose. Consequently, G-InsKi mice and nontransgenic controls were given a peptone meal containing 25 g/liter glucose. Blood glucose was measured before and 40 min after intragastric gavage. In wild-type mice, blood glucose levels were elevated 40 min after gavage, whereas in G-InsKi mice blood glucose levels were approximately 30% lower than in the wild-type controls (Fig. 6A). These results suggest that the administration of the peptone-glucose meal induced the release of both gastric and pancreatic insulin in G-InsKi mice. In agreement with this interpretation, treatment with glucose in the absence of peptone (i.e. in the absence of G cell stimulation) caused equal glycemia in G-InsKi mice and nontransgenic controls (Fig. 6B).
FIG. 6. Regulation of blood glucose in G-InsKi mice. A, Blood glucose is regulated by meal-induced release of gastric insulin. G-InsKi mice were fasted overnight, and liquid meal containing 8% peptone and 25 g/liter glucose was instilled by intragastric gavage. Blood glucose was measured before and 40 min after the treatment. Results are means ± SEM from at least eight animals in each group. B, Glucose meal in the absence of peptone causes equal glycemia in wild-type and G-InsKi mice. Mice were fasted overnight, and a glucose solution (25 g/liter) was instilled by intragastric gavage. Blood glucose was measured before and 40 min after the treatment. Results are means ± SEM from at least eight animals in each group. C, Left panel, Regulation of blood glucose by mixed peptone-glucose meal as a function of time. Mice were fasted overnight, and liquid meal containing 8% peptone and 25 g/liter glucose was instilled by intragastric gavage. Blood glucose was measured before and at 20, 40, 60, 120, and 180 min after the treatment. Results are means ± SEM, from four to eight animals per group. *, P < 0.04 (20 min) and P < 0.02 (40 min). Right panel, Levels of circulating human insulin released by peptone-glucose meal as a function of time. Liquid meal containing 8% peptone and 25 g/liter glucose was instilled by intragastric gavage. Human insulin was measured before and at 20, 40, and 60 min after the treatment. Results are means ± SEM from at least four animals per group.
To substantiate these findings, we next examined blood glucose levels of wild-type and G-InsKi mice at various times. As shown in Fig. 6C (left panel), blood glucose levels were markedly elevated at 20 and 40 min after the administration of the peptone-glucose meal in wild-type mice. In contrast, the blood glucose levels of G-InsKi mice did not show a significant change at either 20 or 40 min after the administration of the peptone-glucose meal. The unchanged blood glucose levels in G-InsKi mice corresponded to the release of human insulin (Fig. 6C, right panel). At 60 min, when human insulin declined to basal level, blood glucose levels in two groups of mice were no longer statistically different. Importantly, at 2 and 3 h after the administration of the meal, blood glucose levels were at the basal level in both groups of mice. Taken together, these results suggest that meal-stimulated release of G cell-produced insulin prevented the increase in blood glucose seen in the control mice. Importantly, the release of gastric insulin did not cause toxic hypoglycemia.
Gastric insulin reduces hyperglycemia and prolongs life span in Akita diabetic mice
Next, we explored whether gastric insulin could ameliorate diabetes in G-InsKi mice made diabetic by injection of STZ. We found that in STZ-diabetic G-InsKi mice, peptone stimulation failed to induce any increase in serum levels of either human insulin or mouse gastrin, suggesting that STZ interferes with G cell function (Fig. 5, A and B, hatched boxes). Indeed, treatment with STZ caused a 60% decrease in antral content of gastric insulin in G-InsKi mice (84 ± 19 pmol/g wet weight, n = 8) compared with untreated G-InsKi mice (207 ± 37.0 pmol/g wet weight, n = 6, see above). Accordingly, there was no difference in fasting blood glucose levels between STZ-diabetic G-InsKi and WT (347 ± 33 and 326 ± 22 mg/dl, respectively). Therefore, it appears that STZ-induced diabetes is not an appropriate model for evaluating the role of G cell-produced insulin in the development of diabetes in G-InsKi mice. It is known that somatostatin exerts inhibitory action on gastrin release from G cells (54). Because STZ induces an increase in pancreatic and antral somatostatin (55, 56, 57, 58, 59), it is likely that increased somatostatin could explain the lack of response of gastric (human) insulin and mouse gastrin in this system (Fig. 5, A and B, hatched bars). Furthermore, an effect of STZ treatment on the number of G cells cannot be ruled out either (60).
These results and considerations prompted us to test whether gastric insulin can ameliorate diabetes in a different model system. As an alternative approach, we used a genetic model of ?-cell loss, Akita diabetic mice (Ins2Akita), and assessed the effect of gastric insulin on the development of diabetes in these mice. In Ins2Akita mice, diabetes results from the altered function of pancreatic ?-cells. The Ins2Akita mice carry an autosomal dominant mutation Ins2C96Y in one of the alleles of the gene for insulin 2 (36, 61). This mutation produces gradual accumulation of malfolded proinsulin-2, which is retained in the pancreatic ?-cell endoplasmic reticulum and causes progressive loss of ?-cells due to proteotoxicity (36, 62). Consequently, islets from Ins2Akita mice are depleted of ?-cells, and those remaining release very little mature insulin, causing juvenile-onset hyperglycemia and insulinopenia in the absence of obesity (36, 61). Although untreated homozygotes rarely survive beyond 12 wk of age, heterozygous Ins2Akita mice are viable and fertile. The diabetic phenotype is more severe and progressive in males than in females, and the mean life span of diabetic male mice on the C57Bl/NJcl background is significantly shorter than that of nondiabetic males in another colony of the same strain (305 vs. 690 d).
Consequently, G-InsKi mice were crossed to Ins2Akita mice that are maintained on the same C57Bl/6 background as G-InsKi mice, eliminating the effect of the undefined genetic background on diabetogenic potential of single-gene mutations (63). Mice heterozygous for the presence of the Ins2Akita allele (Ins2Akita/+), G-Inski allele (G-InsKi+/–), or both (Ins2Akita/+/G-InsKi+/–) were produced and identified as shown in Fig. 7A. Specifically, the presence of the Ins2Akita allele was identified by PCR amplification of the 280-bp fragment of the mouse insulin gene 2 and analyzed by digestion with restriction endonuclease Fnu4HI, as described elsewhere (36) and shown in Fig. 7A (upper and middle panels). The digest of wild-type alleles produced two fragments, 140 bp each, whereas mutated alleles did not have the Fnu4HI site. In the example shown in Fig. 7A (middle and bottom panels), mice 2 and 5 were Ins2Akita/+/G-InsKi+/–, mouse 6 was Ins2Akita/+, and mice 1, 3, 4, and 7 were G-InsKi+/– mice.
FIG. 7. Gastric insulin prolongs life span in Akita diabetic mice. A, Molecular identification of Ins2Akita/+/G-InsKi+/– mice. Upper and middle panels, detection of the Ins2Akita allele by PCR amplification of the 280-bp fragment of the mouse insulin gene 2 followed by restriction digest with Fnu4HI. Bottom panel, PCR identification of the G-InsKi allele with primers Ins25 and Ins33. In this example, mice 2 and 5 are Ins2Akita/+/G-InsKi+/–, mouse 6 is Ins2Akita/+, and mice 1, 3, 4, and 7 are G-InsKi+/– mice. B, Survival analysis of Ins2Akita/+ mice in the absence or presence of G-InsKi allele. Solid line, Ins2Akita/+ mice (n = 15, seven censored subjects, eight deaths); dotted line, Ins2Akita/+/G-InsKi+/– mice (n = 11, eleven censored subjects, no deaths), the ticks portray censored time points. Survival curves were created using the method of Kaplan and Meier. A logrank test was used to determine statistical significance between two curves (P < 0.002).
We found that the presence of gastric insulin caused statistically significant reduction in fasting hyperglycemia from 327 ± 36 to 243 ± 23 mg/dl in Ins2Akita/+/G-InsKi+/– compared with Ins2Akita/+ (n = 10 and 16, respectively; P < 0.05). Subsequently, we confirmed that the Ins2Akita/+ allele did not have any effect on circulating fasting levels of gastric insulin in Ins2Akita/+/G-InsKi+/– (11.2 ± 1 pM, n = 5) compared with G-InsKi mice (10.8 ± 0.7, n = 12, see above).
The salient feature of the result shown in Fig. 7 is the striking difference in the survival in Ins2Akita mice in the presence or absence of G-Inski allele, i.e. gastric insulin. The median survival of mice in the Ins2Akita/+ group was 10.3 months, which is in excellent agreement with the mean life span of diabetic male mice on the C57Bl/NJcl background. In striking contrast, mice carrying the G-InsKi allele were all alive at the end of the study that lasted over 13 months (Fig. 7B). The prolongation of life span in the Ins2Akita/+/G-InsKi+/– mice compared with Ins2Akita/+ mice was highly statistically significant (P < 0.002), as judged by the Logrank test.
Concluding remarks
Diabetes mellitus is a debilitating disease affecting more than 150 million people worldwide. Gene therapy is one of the attractive novel approaches to its treatment that are being explored by many laboratories. A fundamental question in this approach is the identification of an ideal target for the conversion into an insulin-producing cell. Ostensibly, glucose-sensing cells that have the same glucose feedback as original ?-cells would be desirable targets in gene therapy for type I diabetes. However, in type II diabetes, the glucose-sensing mechanisms itself appears to be the possible target of alteration (20), suggesting that release of insulin from other glucose-sensing cells engineered to produce this hormone could be impaired as well. Consequently, meal-regulated secretion of insulin by cells functioning outside of the glucose-sensing loop may be advantageous or complementary in the treatment of non-insulin-dependent diabetes.
In this study, we describe an animal model of meal-stimulated secretion of insulin from enteroendocrine cells. G-InsKi mice produce and process insulin in antral G cells. G cell-produced insulin retains full biological activity as evidenced by the ability of transgenic gastric extract to correct glucose levels in STZ-diabetic mice. Importantly, gastric insulin is released into circulation in response to meal-associated stimuli and regulates meal-induced increase in blood glucose levels. Notably, levels of gastric insulin return to basal at 60 min after the meal and do not cause toxic hypoglycemia even 2–3 h later. The circulating levels of gastric insulin in G-InsKi mice are further increased after 4 d of treatment with PPIs, pharmacological agents widely used to inhibit gastric acid production. In this case, peptone meal releases levels of transgenic insulin that are over 20% of the postprandial circulating pancreatic insulin in the G-InsKi mice.
The effect of gastric insulin on the development of diabetes was assessed in Ins2Akita mice, in which ?-cell loss is caused by endoplasmic reticulum stress induced by impaired folding of mutated proinsulin-2 (36, 61, 62). We generated Ins2Akita/+ mice that carry the G-InsKi allele (Ins2Akita/+/G-InsKi+/–). The most striking aspect of the findings presented here is that in the presence of the G-InsKi allele, Ins2Akita/+ mice exhibited a marked prolongation of life span. These results imply that G cell-derived transgenic insulin is beneficial in the amelioration of diabetes.
In summary, the results presented here demonstrate the feasibility of meal-regulated physiological secretion of insulin by G cells. Moreover, our increasing knowledge of G cell-specific receptors, including receptors for neurotransmitters, ions, amino acids, and tastants (46, 54, 64, 65), suggests that meals may be adjusted to the needs of individual patients by adding defined nutrients that either promote or diminish hormonal secretion from G cells. It is conceivable that in future clinical applications, the therapeutic gastric insulin transgene could be delivered to patients noninvasively into the lumen by endoscopy, e.g. in the form of a viral vector carrying an insulin expression cassette under the regulation of the G cell-specific gastrin promoter. Consequently, it is plausible that an efficient G cell-based insulin gene therapy could protect diabetic patients from complications of insulin insufficiency and defend vulnerable ?-cells from exhaustion by increased secretory demand caused by insulin resistance (66, 67).
Acknowledgments
The authors are grateful to Drs. S. Guha, A. Lunn, and I. Kurland for comments on the manuscript.
References
King H, Rewers M 1993 Global estimates for prevalence of diabetes mellitus and impaired glucose tolerance in adults. WHO Ad Hoc Diabetes Reporting Group. Diabetes Care 16:157–177
The Diabetes Control and Complications Trial Research Group 1993 The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med 329:977–986
Abai AM, Hobart PM, Barnhart KM 1999 Insulin delivery with plasmid DNA. Hum Gene Ther 10:2637–2649
Goldfine ID, German MS, Tseng HC, Wang J, Bolaffi JL, Chen JW, Olson DC, Rothman SS 1997 The endocrine secretion of human insulin and growth hormone by exocrine glands of the gastrointestinal tract. Nat Biotechnol 15:1378–1382
Valera A, Fillat C, Costa C, Sabater J, Visa J, Pujol A, Bosch F 1994 Regulated expression of human insulin in the liver of transgenic mice corrects diabetic alterations. FASEB J 8:440–447
Kolodka TM, Finegold M, Moss L, Woo SL 1995 Gene therapy for diabetes mellitus in rats by hepatic expression of insulin. Proc Natl Acad Sci USA 92:3293–3297
Muzzin P, Eisensmith RC, Copeland KC, Woo SL 1997 Hepatic insulin gene expression as treatment for type 1 diabetes mellitus in rats. Mol Endocrinol 11:833–837
Lu D, Tamemoto H, Shibata H, Saito I, Takeuchi T 1998 Regulatable production of insulin from primary-cultured hepatocytes: insulin production is up-regulated by glucagon and cAMP and down-regulated by insulin. Gene Ther 5:888–895
Chen R, Meseck M, McEvoy RC, Woo SL 2000 Glucose-stimulated and self-limiting insulin production by glucose 6-phosphatase promoter driven insulin expression in hepatoma cells. Gene Ther 7:1802–1809
Ferber S, Halkin A, Cohen H, Ber I, Einav Y, Goldberg I, Barshack I, Seijffers R, Kopolovic J, Kaiser N, Karasik A 2000 Pancreatic and duodenal homeobox gene 1 induces expression of insulin genes in liver and ameliorates streptozotocin-induced hyperglycemia. Nat Med 6:568–571
Thule PM, Liu J, Phillips LS 2000 Glucose regulated production of human insulin in rat hepatocytes. Gene Ther 7:205–214
Thule PM LJ 2000 Regulated hepatic insulin gene therapy of STZ-diabetic rats. Gene Ther 7:1744–1752
Chen R, Meseck ML, Woo SL 2001 Auto-regulated hepatic insulin gene expression in type 1 diabetic rats. Mol Ther 3:584–590
Kojima H, Fujimiya M, Matsumura K, Younan P, Imaeda H, Maeda M, Chan L 2003 NeuroD-betacellulin gene therapy induces islet neogenesis in the liver and reverses diabetes in mice. Nat Med 9:596–603
Lipes MA, Cooper EM, Skelly R, Rhodes CJ, Boschetti E, Weir GC, Davalli AM 1996 Insulin-secreting non-islet cells are resistant to autoimmune destruction. Proc Natl Acad Sci USA 93:8595–8600
Faradji RN, Havari E, Chen Q, Gray J, Tornheim K, Corkey BE, Mulligan RC, Lipes MA 2001 Glucose-induced toxicity in insulin-producing pituitary cells that coexpress GLUT2 and glucokinase. Implications for metabolic engineering. J Biol Chem 276:36695–36702
Wu L, Nicholson W, Wu CY, Xu M, McGaha A, Shiota M, Powers AC 2003 Engineering physiologically regulated insulin secretion in non-beta cells by expressing glucagon-like peptide 1 receptor. Gene Ther 10:1712–1720
Cheung AT, Dayanandan B, Lewis JT, Korbutt GS, Rajotte RV, Bryer-Ash M, Boylan MO, Wolfe MM, Kieffer TJ 2000 Glucose-dependent insulin release from genetically engineered K cells. Science 290:1959–1962
Chan L, Fujimiya M, Kojima H 2003 In vivo gene therapy for diabetes mellitus. Trends Mol Med 9:430–435
Kahn SE 2003 The relative contributions of insulin resistance and ?-cell dysfunction to the pathophysiology of type 2 diabetes. Diabetologia 46:3–19
Chan CB, Saleh MC, Koshkin V, Wheeler MB 2004 Uncoupling protein 2 and islet function. Diabetes 53:S136–S142
Bell GI, Pilkis SJ, Weber IT, Polonsky KS 1996 Glucokinase mutations, insulin secretion, and diabetes mellitus. Annu Rev Physiol 58:171–186
Taylor IL, Singer M, Kauffman Jr GL 1982 Time-dependent effects of vagotomy on pancreatic polypeptide release. Dig Dis Sci 27:491–494
McArthur KE, Isenberg JI, Hogan DL, Dreier SJ 1983 Intravenous infusion of L-isomers of phenylalanine and tryptophan stimulate gastric acid secretion at physiologic plasma concentrations in normal subjects and after parietal cell vagotomy. J Clin Invest 71:1254–1262
Eysselein VE, Kovacs TO, Kleibeuker JH, Maxwell V, Reedy T, Walsh JH 1992 Regulation of gastric acid secretion by gastrin in duodenal ulcer patients and healthy subjects. Gastroenterology 102:1142–1148
Varner AA, Modlin IM, Walsh JH 1981 High potency of bombesin for stimulation of human gastrin release and gastric acid secretion. Regul Pept 1:289–296
Hirschowitz BI, Molina E 1983 Relation of gastric acid and pepsin secretion to serum gastrin levels in dogs given bombesin and gastrin-17. Am J Physiol 244:G546–G551
Yang H, Wong H, Wu V, Walsh JH, Tache Y 1990 Somatostatin monoclonal antibody immunoneutralization increases gastrin and gastric acid secretion in urethane-anesthetized rats. Gastroenterology 99:659–665
Stepan V, Sugano K, Yamada T, Park J, Dickinson CJ 2002 Gastrin biosynthesis in canine G cells. Am J Physiol Gastrointest Liver Physiol 282:G766–G775
Sawada M, Finniss S, Dickinson CJ 2000 Diminished prohormone convertase 3 expression (PC1/PC3) inhibits progastrin post-translational processing. Regul Pept 89:19–28
Rehfeld JF, Lindberg I, Friis-Hansen L 2002 Progastrin processing differs in 7B2 and PC2 knockout animals: a role for 7B2 independent of action on PC2. FEBS Lett 510:89–93
Macro JA, Bate GW, Varro A, Vaillant C, Seidah NG, Dimaline R, Dockray GJ 1997 Regulation by gastric acid of the processing of progastrin-derived peptides in rat antral mucosa. J Physiol 502:409–419
Koh TJ, Wang TC 1995 Molecular cloning and sequencing of the murine gastrin gene. Biochem Biophys Res Commun 216:34–41
Zhukova E, Afshar A, Ko J, Popper P, Pham T, Sternini C, Walsh JH 2001 Expression of the human insulin gene in the gastric G cells of transgenic mice. Transgenic Res 10:329–341
Miller SA DD, Polesky HF 1988 A simple salting out procedure for extracting DNA from human nucleated cells. Nucleic Acids Res 16:1215
Wang J, Takeuchi T, Tanaka S, Kubo SK, Kayo T, Lu D, Takata K, Koizumi A, Izumi T 1999 A mutation in the insulin 2 gene induces diabetes with severe pancreatic beta-cell dysfunction in the Mody mouse. J Clin Invest 103:27–37
Pham T, Guerrini S, Wong H, Reeve Jr J, Sternini C 2002 Distribution of galanin receptor 1 immunoreactivity in the rat stomach and small intestine. J Comp Neurol 450:292–302
Campbell DH,et al. 1963 Puncture of the ophthalmic venous plexus. In: Methods in immunology: a laboratory text for instruction and research. New York: W. A. Benjamin
Hem A, Smith AJ, Solberg P 1998 Saphenous vein puncture for blood sampling of the mouse, rat, hamster, gerbil, guinea pig, ferret and mink. Lab Anim 32:364–368
Baggio L, Kieffer TJ, Drucker DJ 2000 Glucagon-like peptide-1, but not glucose-dependent insulinotropic peptide, regulates fasting glycemia and nonenteral glucose clearance in mice. Endocrinology 141:3703–3709
Koh TJ, Goldenring JR, Ito S, Mashimo H, Kopin AS, Varro A, Dockray GJ, Wang TC 1997 Gastrin deficiency results in altered gastric differentiation and decreased colonic proliferation in mice. Gastroenterology 113:1015–1025
Friis-Hansen L, Sundler F, Li Y, Gillespie PJ, Saunders TL, Greenson JK, Owyang C, Rehfeld JF, Samuelson LC 1998 Impaired gastric acid secretion in gastrin-deficient mice. Am J Physiol 274:G561–G568
Taylor IL, Byrne WJ, Christie DL, Ament ME, Walsh JH 1982 Effect of individual l-amino acids on gastric acid secretion and serum gastrin and pancreatic polypeptide release in humans. Gastroenterology 83:273–278
Conigrave AD, Quinn SJ, Brown EM 2000 L-amino acid sensing by the extracellular Ca2+-sensing receptor. Proc Natl Acad Sci USA 97:4814–4819
Kobilka B 2000 Allosteric activation of the CaR by L-amino acids. Proc Natl Acad Sci USA 97:4419–4420
Young SH, Rozengurt E 2002 Amino acids and Ca2+ stimulate different patterns of Ca2+ oscillations through the Ca2+-sensing receptor. Am J Physiol Cell Physiol 282:C1414–C1422
Buchan AMJ, Squires PE, Ring M, Meloche RM 2001 Mechanism of action of the calcium-sensing receptor in human antral gastrin cells. Gastroenterology 120:1128–1139
Squires PE, Harris TE, Persaud SJ, Curtis SB, Buchan AM, Jones PM 2000 The extracellular calcium-sensing receptor on human beta-cells negatively modulates insulin secretion. Diabetes 49:409–417
Straub SG, Mulvaney-Musa J, Yajima H, Weiland GA, Sharp GW 2003 Stimulation of insulin secretion by denatonium, one of the most bitter-tasting substances known. Diabetes 52:356–364
Wallmark B, Larsson H, Humble L 1985 The relationship between gastric acid secretion and gastric H+,K+-ATPase activity. J Biol Chem 260:13681–13684
Lorentzon P, Eklundh B, Brandstrom A, Wallmark B 1985 The mechanism for inhibition of gastric (H+ + K+)-ATPase by omeprazole. Biochim Biophys Acta 817:25–32
Sundler F, Hakanson R, Carlsson E, Larsson H, Mattsson H 1986 Hypergastrinemia after blockade of acid secretion in the rat: trophic effects. Digestion 35(Suppl 1):56–69
Pawlikowski M, Karbownik M, Lewinski A, Pisarek H, Wajs E, Szkudlinski M 1992 Effects of omeprazole on the number of immunoreactive gastrin- and somatostatin-cells in the rat gastric mucosa. Histol Histopathol 7:153–156
Sachs G, Zeng N, Prinz C 1997 Physiology of isolated gastric endocrine cells. Annu Rev Physiol 59:243–256
Patel YC, Cameron DP, Bankier A, Malaisse-Lagae F, Ravazzola M, Studer P, Orci L 1978 Changes in somatostatin concentration in pancreas and other tissues of streptozotocin diabetic rats. Endocrinology 103:917–923
Chiba T, Kadowaki S, Taminato T, Abe H, Chihara K, Matsukura S, Goto Y, Seino Y, Fujita T 1981 Concentration and secretion of gastric somatostatin in streptozotocin-diabetic rats. Diabetes 30:188–191
Chiba T, Taminato T, Kadowaki S, Chihara K, Matsukura S, Nozawa M, Seino Y, Fujita T 1981 Reversal of increased gastric somatostatin in streptozotocin-diabetic rats by whole pancreas transplantation. Diabetes 30:724–727
Michels JE, Bauer GE, Johnson D, Dixit PK 1986 Morphometric analysis of the endocrine cell composition of rat pancreas following treatment with streptozotocin and nicotinamide. Exp Mol Pathol 44:247–258
Aughsteen AA 2000 An ultrastructural study on the effect of streptozotocin on the islets of Langerhans in mice. J Electron Microsc (Tokyo) 49:681–690
Nwokolo CU, Debnam ES, Booth JD, Sim R, Sankey EA, Dhillon AP, Pounder RE 1992 Neuroendocrine changes in rat stomach during experimental diabetes mellitus. Dig Dis Sci 37:751–756
Yoshioka M, Kayo T, Ikeda T, Koizumi A 1997 A novel locus, Mody4, distal to D7Mit189 on chromosome 7 determines early-onset NIDDM in nonobese C57BL/6 (Akita) mutant mice. Diabetes 46:887–894
Ron D 2002 Proteotoxicity in the endoplasmic reticulum: lessons from the Akita diabetic mouse. J Clin Invest 109:443–445
Leiter EH 2002 Mice with targeted gene disruptions or gene insertions for diabetes research: problems, pitfalls, and potential solutions. Diabetologia 45:296–308
Rozengurt E, Walsh JH 2001 Gastrin, CCK, signaling and cancer. Annu Rev Physiol 63:49–76
Wu SV, Rozengurt N, Yang M, Young SH, Sinnett-Smith J, Rozengurt E 2002 Expression of bitter taste receptors of the T2R family in the gastrointestinal tract and enteroendocrine STC-1 cells. Proc Natl Acad Sci USA 99:2392–2397
DeFronzo RA, Ferrannini E 1991 Insulin resistance. A multifaceted syndrome responsible for NIDDM, obesity, hypertension, dyslipidemia, and atherosclerotic cardiovascular disease. Diabetes Care 14:173–194
Kruszynska YT, Olefsky JM 1996 Cellular and molecular mechanisms of non-insulin-dependent diabetes mellitus. J Investig Med 44:413–428(Yu-Chun Lu, Catia Sternin)