Endocrinology and Metabolism


We have generated transgenic mice that express green fluorescent protein (GFP) under the control of the mouse insulin I gene promoter (MIP). The MIP-GFP mice develop normally and are indistinguishable from control animals with respect to glucose tolerance and pancreatic insulin content. Histological studies showed that the MIP-GFP mice had normal islet architecture with coexpression of insulin and GFP in the β-cells of all islets. We observed GFP expression in islets from embryonic day E13.5 through adulthood. Studies of β-cell function revealed no difference in glucose-induced intracellular calcium mobilization between islets from transgenic and control animals. We prepared single-cell suspensions from both isolated islets and whole pancreas from MIP-GFP-transgenic mice and sorted the β-cells by fluorescence-activated cell sorting based on their green fluorescence. These studies showed that 2.4 ± 0.2% (n = 6) of the cells in the pancreas of newborn (P1) and 0.9 ± 0.1% (n = 5) of 8-wk-old mice were β-cells. The MIP-GFP-transgenic mouse may be a useful tool for studying β-cell biology in normal and diabetic animals.

  • insulin
  • diabetes

the insulin-producingβ-cell of the pancreas plays a central role in the pathophysiology of diabetes mellitus, with anatomic and functional loss of these cells leading to type 1 and type 2 diabetes mellitus, respectively (2). The identification and characterization of embryonic or adult stem cells that give rise to the β-cell could lead to cellular-based therapies for treating both forms of diabetes (1, 18, 22). Previous studies have shown that green fluorescent protein (GFP) from the jellyfish Aequorea victoria and its yellow and cyan derivatives could be utilized as reporter genes to label specific cell types including pancreatic β-cells by expressing GFP under the control of a tissue-specific promoter (5, 6, 8-10, 15, 23, 28). One advantage of these proteins is that they can be detected in living cells because they fluoresce brightly upon exposure to ultraviolet light or blue light without the addition of coactivators or exogenous substrate. In addition, pure populations of fluorescent-tagged cells can be isolated using a fluorescence-activated cell sorter (FACS) (10, 15,23). Rat and human β-cells treated with recombinant adenovirus expressing green fluorescent protein (GFP) under the control of the rat insulin I promoter appear to function normally, suggesting that expression of GFP may be well tolerated by these cells (9,23). Thus we reasoned that, if we could generate a mouse model in which we had genetically tagged the pancreatic β-cells with GFP, we would have a potentially valuable research tool for studying β-cell biology, including the identification of progenitor cells. Here, we describe a line of transgenic mice in which the pancreatic β-cells are genetically tagged with GFP.


Generation of mouse insulin I gene promoter-GFP-transgenic mice.

The mouse insulin I gene promoter (MIP)-GFP-transgenic construct was assembled using an 8.5-kb fragment of the MIP that includes a region from −8.5 to +12 bp (relative to the transcriptional start site), the coding region of enhanced GFP (EGFP) (0.76 kb; Clontech, Palo Alto, CA), and a 2.1-kb fragment of the human growth hormone (hGH) cassette gene for high-level expression (25, 26). The 11.2-kb MIP-EGFP-hGH fragment was isolated from the vector by digestion of the plasmid construct with SfiI and HindIII and agarose gel electrophoresis. The fragment was further purified using an Elutip-D column (Schleicher & Schuell, Keene, NH). The purified transgene DNA was microinjected into the pronuclei of CD-1 mice by the Transgenic Mouse/ES Core Facility of the University of Chicago Diabetes Research and Training Center (DRTC). Tail DNA from potential founder mice was screened for the presence of the transgene by PCR using forward and reverse primers 5′-GAAGACAATAGCAGGCATGCTG-3′ and 5′-ACTGGGCTTACATGGCGATACTC-3′, respectively.

All of the procedures involving mice were approved by the University of Chicago Institutional Animal Care and Use Committee.

Glucose tolerance testing.

Intraperitoneal glucose tolerance tests (IPGTTs) were performed after a 4-h fast. Blood was sampled from the tail vein before and 30, 60, 90, and 120 min after intraperitoneal injection of 2 mg/g body wt of dextrose. Glucose levels were measured using a Precision Q.I.D. Glucometer (MediSense, Waltham, MA).

Insulin assays.

Pancreatic insulin content was measured after acid ethanol extraction of the whole pancreas, as described previously (24). Insulin concentration was measured by a double-antibody radioimmunoassay using a rat insulin standard in the Radioimmunoassay Core Laboratory of the University of Chicago DRTC. The intra-assay coefficient of variation for this assay is 7%. All samples were assayed in duplicate.

Isolation of pancreatic islets of Langerhans.

Pancreatic islets were isolated using a modification of the procedure originally described by Lacy and Kostianovsky (21). Briefly, the pancreas was inflated with a solution containing 0.3 mg/ml collagenase (Type XI; Sigma, St. Louis, MO) in Hanks' balanced salt solution, injected via the pancreatic duct. The inflated pancreas was removed, incubated at 37°C for 10 min, and shaken vigorously to disrupt the tissue. After differential centrifugation through a Ficoll gradient to separate islets from acinar tissue, the islets were washed and then hand picked. They were plated on 12- or 35-mm coverslips to facilitate adherence for subsequent measurement of intracellular calcium or confocal microscopic visualization. The islets were cultured in RPMI 1640 supplemented with 10% (vol/vol) fetal bovine serum (FBS), 100 U/ml penicillin, and 100 μg/ml streptomycin in a humidified incubator at 37°C in 95% air and 5% CO2.

Preparation of single-cell suspensions from islets and pancreas.

Isolated islets were incubated in a solution of 0.05% trypsin-EDTA (GIBCO, Grand Island, NY) at 37°C for 3 min. The digestion was stopped by adding RPMI 1640 with 10% (vol/vol) FBS. The pancreata from newborn (P1) mice were removed and digested in a solution containing 0.3 mg/ml collagenase and then in 0.05% trypsin-EDTA. The resulting single cells were washed with PBS, resuspended in cold PBS with 10% (vol/vol) FBS, and filtered using 70-μm mesh. Cells were stained with trypan blue to check viability, and preparations showing >95% live cells were analyzed. The pancreatic cell suspension was diluted to 2 × 106 cells/ml with the PBS-FBS solution and then fixed in 4% paraformaldehyde.

Pancreatic islet morphology.

The pancreas was removed, embedded in optimum cutting temperature compound (Tissue-Tek O.T.C., Sakura Finetek, Torrance, CA) and frozen in isopentane at −70°C. Serial sections were cut at 6 μm in thickness and fixed in 4% paraformaldehyde. GFP fluorescence is well retained under these conditions (24). The sections were stained with a polyclonal guinea pig anti-porcine insulin antibody to identify β-cells and with polyclonal rabbit anti-human glucagon, somatostatin, and pancreatic polypeptide antibodies to identify α-, δ-, and PP cells, respectively (DAKO, Carpinteria, CA). Sections were also stained with a monoclonal anti-mouse GFP antibody (Clontech) to detect GFP expression and with a polyclonal goat anti-hGH antibody (DAKO) to test for expression of GH from the transgene construct. The primary antibodies were detected using Texas red-conjugated anti-guinea pig/rabbit/goat IgG (H+L) and biotin-streptavidin-conjugated anti-mouse IgG (H+L) followed by Texas red-conjugated streptavidin (Jackson ImmunoResearch Laboratory, West Grove, PA). Microscopic images were taken with a Nikon Eclipse E800 microscope with PCM-2000 (Nikon, New York, NY) and an Olympus SZX-RFL3 microscope (Olympus, Melville, NY).

Intracellular calcium measurements.

The GFP expression interferes with fura 2 signals, and as a consequence, the calcium indicator fura 2 cannot be used to monitor changes in intracellular calcium in GFP-expressing β-cells. The short-wavelength excitation of the EGFP excitation spectrum extends down to ∼350 nm, which is in the range of fura 2. Thus excitation of fura 2 will excite EGFP and contaminate the common emission spectrum. Fura red is a calcium indicator with a large stokes shift, thereby minimizing the contamination of the calcium signal by GFP (20,30). Fura red has been successfully used in combination with GFP expression in other studies (11, 12); however, there are reports of some signal cross talk when used with EGFP (3). Another calcium indicator, X-rhod-1, which has excitation/emission maxima of ∼580/602 nm, may also give a good separation of the signals from GFP and the calcium indicator (3); however, care must be taken, since X-rhod-1 has a certain selectivity for mitochondria (13). We used fura red in the studies described here. The isolated islets were loaded with 5 μmol/l fura red-AM (acetoxymethyl ester; Molecular Probes, Eugene, OR), and changes in intracellular calcium were monitored using a Fluoview scanning laser confocal lens with an inverted IX70 microscope (Olympus). Tiempo real-time acquisition software (Olympus) was used to collect and plot the data. The Cy5 filter set (700/75-nm bandpass filter set) was used to detect fura red signals and to eliminate the contamination with GFP signals. Solutions were perfused in a temperature-controlled chamber by use of a TC-344 Dual Heater Controller (Warner Instrument, Hamden, CT). All the measurements were performed at 34–35°C.

Flow cytometric analysis.

Flow cytometric analysis of GFP-labeled β-cells was carried out using a FACScan flow cytometer with Cell Quest software (Becton Dickinson, Franklin Lakes, NJ). GFP-expressing cells were detected using the FL1 channel (absorption spectra 530/30 nm). We have sorted both fixed and nonfixed cells. However, the studies described here were carried out using cells fixed as described above.

Statistical analysis.

Results are expressed as means ± SE. We compared groups by use of ANOVA (StatView software; SAS Institute, Cary, NC). Differences were considered to be significant at P < 0.05.


Expression of GFP in β-cells.

The MIP-GFP construct was injected into the fertilized eggs of CD-1 mice. We obtained three founders: one female (6504) and two males (6502 and 6508). GFP was expressed in the islets of all three founders. The founders 6502 and 6508 were estimated to have one copy of the transgene by quantitative real-time PCR. The founder 6504 was estimated to have five copies, and two of her male F1 pups, 6729 and 6719, each had one copy of the transgene. Progeny from transgenic lines 6502 and 6719 showed delayed recovery on IPGTT and nonuniform expression of GFP within their islets and were not studied further. The studies described here were carried out on line 6729 [Tg(MIP-GFP)6729Hara]. The transgene has been maintained on the CD-1 background, and mice have been housed under specific pathogen-free conditions with free access to food and water.

The pancreatic islets from the MIP-GFP transgenic animals were green when exposed to ultraviolet light (Fig.1 A). The overall islet architecture of the MIP-GFP mice was normal. The patterns of GFP and insulin expression within the islet were identical, and GFP expression was observed only in β-cells (Fig. 1 C). There was no evidence of expression of GFP in non-β-cells of the islet or in the exocrine pancreas (Fig. 1 D). A stacked confocal microscopic image of an isolated islet suggested that the steady-state levels of GFP were similar in all β-cells (Fig. 1 E).

Fig. 1.

Pancreatic β-cell specific expression of green fluorescent protein (GFP). A: confocal microscopic image of a pancreatic section shows GFP expression (green) in the central core of the islet. Scale bar = 50 μm. B: the same section was stained with an anti-insulin antibody (red). C: dual-color image showed overlapping expression of GFP and insulin in the β-cells of the islet (yellow). D: a pancreatic section was stained with a cocktail of antibodies against 3 non-β-cell hormones, i.e., glucagon, somatostatin, and pancreatic polypeptide (red). There was no overlapping expression of GFP with other endocrine hormones. Scale bar = 50 μm. E: isolated pancreatic islet from a mouse insulin I gene promoter (MIP)-GFP-transgenic mouse. A stacked confocal microscopic image is shown. The dark region within each cell is the nucleus. Scale bar = 10 μm.

We also tested for expression of GFP in the pancreas and other tissues at 6 wk of age by Western blotting. We detected GFP expression only in pancreas and not in any of the other tissues examined: brain, fat, muscle, small intestine, spleen, heart, kidney, liver, uterus, and testis (data not shown). In addition, we also examined sections from each of these tissues for the presence of isolated GFP-expressing cells and found none. The transgene construct includes the hGH gene at its 3′ end, since previous studies have shown that the presence of heterologous introns provided by the hGH gene resulted in enhanced expression of the transgene (25). Because there is no internal ribosomal entry site upstream of the GH gene, there should be no GH expression, and as expected, we found no evidence of GH in islets from MIP-GFP mice by immunohistochemical staining with sections from pituitary as a positive control (data not shown). Expression of the MIP-GFP transgene appears to be restricted to β-cells. The MIP fragment that we used appears to be more tissue specific in its expression than shorter fragments (which are more typically used to make insulin promoter-driven fusion genes), since no expression was observed in brain with the use of this construct, whereas expression in the brain is often observed with the shorter promoter construct (26).

Physiological characterization of MIP-GFP mice.

The MIP-GFP mice developed normally. At 6 wk of age, there were no significant differences in body weight, fasting blood glucose, and pancreatic insulin content between transgenic and nontransgenic CD-1 male mice (data not shown). IPGTT at 6 wk showed no statistically significant difference in the response between transgenic and nontransgenic male animals (Fig. 2). We followed body weight and performed IPGTTs on the transgenic mice up to 40 wk and body weight and nonfasting blood glucose levels up to 60 wk, and none of the animals has shown any evidence of abnormal weight or hypo- or hyperglycemia (data not shown.)

Fig. 2.

Intraperitoneal glucose tolerance testing in 6-wk-old male MIP-GFP-transgenic and age- and sex-matched wild-type (WT) mice. There was no significant difference in blood glucose levels at each time point (P > 0.05).

Glucose responsiveness of pancreatic islets from the MIP-GFP mice.

We examined the effects of glucose on intracellular calcium mobilization in the islets of the MIP-GFP mice as a measure of β-cell function. We used fura red (16) to monitor intracellular calcium levels to allow measurement of signals from the calcium indicator in the presence of GFP. The islets from both MIP-GFP and nontransgenic mice exhibited a robust mobilization of calcium in response to 10 mM glucose, with no apparent difference between the islets from transgenic and nontransgenic animals (Fig.3). Similar responses were observed in response to 20 mM glucose and 50 mM KCl, occasionally accompanied by oscillations in calcium levels (data not shown).

Fig. 3.

Pancreatic islets from the MIP-GFP mice retain normal glucose responsiveness. A: representative trace of fura red-AM intensity showing the intracellular calcium increase in response to 10 mM glucose in cells within an isolated islet from a WT mouse. Note that fluorescence of the fura red decreases once the indicator binds calcium. The duration of administration is indicated by the horizontal bar above the traces. Fura red-AM fluorescence was measured by drawing a region of interest around a group of the calcium indicator-loaded cells and plotted vs. time. Fluorescence intensity is expressed in arbitrary units. B: GFP-expressing β-cells show a similar response to 10 mM glucose. Inset: merged confocal images of an isolated islet from an MIP-GFP mouse showing GFP-expressing β-cells (green) and peripheral cells that were loaded with fura red-AM (5 μM; red), which appear yellow (green + red), demonstrating colocalization of indicator and GFP in β-cells. Scale bar = 20 μm.

FACS analysis of GFP-labeled β-cells.

The islets isolated from MIP-GFP mice (8 wk old) were dissociated into single cells. The GFP-labeled β-cells could be readily separated from the non-β-cells cells by FACS (Fig.4 shows a representative trace).

Fig. 4.

Flow cytometric analysis of a single-cell suspension prepared from islets from MIP-GFP-transgenic (right) and WT mice (left). FL2-H, medium wavelength channel on the FACS. Fluorescence intensity is expressed in arbitrary units on a logarithmic scale. The range of negative signals (autofluorescence) was determined using WT β-cells as a control (left). In the cell suspension prepared from isolated pancreatic islets from MIP-GFP-transgenic mice (8 wk old), 90% of the cells fluoresce green (right).

Measurement of pancreatic β-cell number.

A single-cell suspension was prepared from whole pancreas of a MIP-GFP-transgenic mouse at P1 (Fig.5 A), and the percentage of pancreatic β-cells was measured by FACS. Flow cytometric analysis revealed that 2.6% of the pancreatic cells expressed GFP (Fig.5 B). Further analyses on five additional transgenic littermates gave similar results and indicate that 2.4 ± 0.2% (n = 6) of the cells in the P1 pancreas were GFP-expressing β-cells (each trace not shown). We also prepared single-cell suspensions from whole pancreas of older mice and sorted the cells on the basis of GFP expression. We found that 0.9 ± 0.1% (n = 5) of the cells in pancreas of 8-wk-old animals were GFP positive (data not shown).

Fig. 5.

Flow cytometric analysis of a single-cell suspension prepared from whole pancreas of a P1 MIP-GFP-transgenic mouse. A: bright-field image of a pancreatic cell suspension from an MIP-GFP-transgenic mouse (left). Arrow indicates GFP-expressing β-cells that fluoresce green in the fluorescent image (right). The dissociated cells were concentrated by a brief centrifugation before viewing. Scale bar = 50 μm. B: representative flow cytometric analysis of a single-cell suspension prepared from a WT mouse (left) and an MIP-GFP mouse (right). GFP is expressed by 2.6% of the pancreatic cells from the MIP-GFP mouse at P1.

Expression of MIP-GFP transgene during development.

The pancreas begins to form at the 26-somite stage (E9.5) of gestation in the mouse (14). Although insulin mRNA can be detected at the 20-somite stage (14), significant expression of insulin begins only at E13.5 (17). We observed GFP-labeled cells in the pancreas at E13.5 (we did not examine earlier stages) and continuing throughout life (Fig. 6). At E13.5, there is a scattered mass of GFP-expressing cells in the pancreas adjacent to the duodenum. GFP-labeled islets are evident at E18.5 and in the adult (60 wk), where they can be readily seen distributed throughout the pancreas.

Fig. 6.

GFP expression in the pancreas from MIP-GFP transgenic mice at different developmental stages. A: dark-field image of an embryo at embryonic day E13.5. Scale bar = 500 μm. B: dark-field image of the gut region including duodenum and pancreas from the embryo in A. An enlarged image of the boxed area is shown in C. Scale bar = 500 μm. C: fluorescent image of the pancreas (boxed area inB) showing GFP-labeled green β-cells. Scale bar = 100 μm. D: dark-field image of the gut from an embryo at E18.5. Scale bar = 500 μm. E: fluorescent image of the gut in D. Note the mass of GFP-expressing cells scattered throughout the pancreas. F: dark-field image of gut from a mouse at birth (P0). Scale bar = 500 μm.G: fluorescent image of the gut in F. H: dark-field image of the pancreas, including spleen, from a 60-wk-old male mouse. Scale bar = 5 mm. I: fluorescent image of the gut in H. Note that the islets are readily evident as clusters of green cells dispersed throughout the pancreas.


We have generated transgenic mice in which pancreatic β-cells are genetically tagged with GFP. The phenotypic characterization of the MIP-GFP transgenic mice suggests that the presence of cytoplasmic GFP does not impair β-cell development or function, at least in the line of mice described in this report. We can envisage a number of uses of these mice for studying β-cells in situ and in isolation. In this regard, β-cell function can be studied in the MIP-GFP mice or in intercrosses with other genetically engineered and mutant mice. We believe that the MIP-GFP mice will be useful for isolating pancreatic β-cells at various stages of development from embryonic to adult. The purified β-cells can be used for molecular biological studies such as monitoring the changing pattern of gene expression during development with the use of microarrays (27) or for biophysical studies of their functional properties (19). The isolation of a pure population of β-cells is arduous. Van de Winkel et al. (29) have described a procedure that works well with rat islets, beginning with purified islets. The islets are disassociated into single cells, from which a highly enriched β-cell population can be obtained by FACS on the basis of the high intrinsic autofluorescence of β-cells. Meyer et al. (23) have described another procedure, in which dispersed islet cells (in this case human) were infected with a recombinant adenovirus expressing GFP driven by the rat insulin I promoter. The GFP is expressed only in β-cells, and these cells can be sorted by FACS to obtain a pure (>95%) population of β-cells. This procedure is generally applicable to isolating β-cells from many different species, including mice, and at various stages of development. However, special precautions must be taken when working with adenovirus, even the attenuated vectors that are commonly used in molecular biology. In contrast to these procedures, purified β-cells can be readily isolated from MIP-GFP mice, beginning with either islets or pancreas. The ability to flow sort disassociated pancreatic cells to isolate them will facilitate studies of embryonic β-cells when islet isolation is very difficult. A second use of the MIP-GFP-transgenic mice is in studies where real-time instant identification of β-cells is required, such as in electrophysiological studies, as the β-cells can be distinguished from non-β-cells on the basis of their green fluorescence. Another use is in studies of β-cell development. β-cell progenitors and/or stem cells have been described in ductal tissue (4), adult marrow (18), and embryonic stem cells (1,22), and the MIP-GFP mice may be useful in identifying the progenitor/stem cells in these tissues and cells. The preliminary data that we have presented here indicate that we can use FACS to quantify the number of β-cells in the pancreas, and this may be a useful method for following the changes in β-cell number that occur in different physiological states such as pregnancy and diabetes. We also believe that we may be able to use FACS analysis to monitor changes in β-cell size. Finally, the GFP fluorescence is sufficiently intense, especially in the adult, to be detected within a thick specimen, including a whole pancreas (Fig. 6). Thus it may be possible to carry out a three-dimensional reconstruction of the distribution of islets within the entire pancreas of the MIP-GFP mice.

In summary, we have generated a line of mice in which the β-cells are genetically tagged with GFP. We believe that the MIP-GFP mice will be a useful tool for studying β-cells and islets in normal and diabetic states.


This research was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants DK-20595, DK-44840, and DK-61245. M. Hara is a Naomi Berrie Fellow, and G. I. Bell is an Investigator of the Howard Hughes Medical Institute.


  • Address for reprint requests and other correspondence: M. Hara, Howard Hughes Medical Institute Research Laboratories, The Univ. of Chicago, 5841 South Maryland Ave., MC1028, Chicago, IL 60637 (E-mail: mhara{at}midway.uchicago.edu).

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

  • September 17, 2002;10.1152/ajpendo.00321.2002


View Abstract