A transgenic mouse line, FVB/N-Tg(RIP-luc) containing the firefly luciferase gene under the regulation of the rat insulin promoter II (RIP, 760 bp) that specifically and constitutively expresses firefly luciferase in the pancreatic islet β-cells (Caliper Life Sciences, Alameda, CA), was used for both experimental and control groups. All mice were housed in a barrier facility at Northwestern University and were used at ages 3-4 months. All animal procedures were approved by the Animal Care and Use Committee at Northwestern University.

FVB/N-Tg(RIP-luc) mice were made diabetic by a single intraperitoneal (i.p.) injection of the pancreatic β-cell toxin streptozotocin (STZ, 220 mg/kg body weight dissolved in saline, Sigma-Aldrich, St. Louis, MO). Serum glucose levels were measured daily using a glucometer (One Touch Basic, Lifescan Inc., Milpitas, CA). Mice that demonstrated serum glucose levels greater than 300 mg/dL on two consecutive days were considered diabetic. Control mice received an i.p. injection of saline solution.

In preparation for BLI, control and diabetic FVB/N-Tg(RIP-luc) mice were anesthetized with 2.5% isofluorane in air. Substrate luciferin potassium salt (Molecular Therapeutics Inc., Ann Arbor, MI) dissolved in PBS (40 mg/ml) was administered via a single i.p. injection at a dose of 200 mg/kg body weight. The mice were placed in the camera chamber, where a controlled flow of 2.0% isofluorane in air was administered through a nosecone via a gas anesthesia system designed to work in conjunction with the BLI system (IVIS® 200, Caliper Life Sciences). Mice were imaged for one-minute durations on both the dorsal and ventral sides at medium-resolution with a field of view of 20 cm at 10 minutes post luciferin injection. A gray-scale body image was collected and overlaid by a pseudo-color image representing the spatial distribution of detected photons. The BLI signal intensities emitted from the liver region of the mice were quantified using a fixed region of interest (ROI) of 2×1.5 cm in size from dorsal images and expressed as photons per second of light. Background images were collected daily and background subtractions were automatically calculated by the Living Image software. Ventral images were collected solely for visual purposes.

Autopsy mouse liver tissue samples were collected from either control normal glycemic mice or mice that had been treated with STZ and exposed to hyperglycemia for 8 days. The randomly collected liver samples (∼5 mm²) were placed in 400 µl of Passive Lysis Buffer (Promega Corp, Madison, WI) and were homogenized using a handheld tissue homogenizer (Pellet Pestle® Motor, Kontes). Tissue homogenates were centrifuged and the supernatant were used for luciferase activity and protein concentration assays. For the luciferase activity measurement, the In Vitro Luciferase Assay System (Promega) was used according to instructions provided by the manufacturer. Tissue homogenates (20 µl) and the luciferase substrate (100 µl) were mixed and the luminescence generated from the reaction was measured using a Sirius Luminomter (Berthold Detection Systems). Background luminescent readings were obtained and were subtracted from the luminescent data. Protein concentrations were determined using the BCA Protein Assay Kit (Pierce, Rockford, IL) according to the manufacturer's protocol and analyzed using a Coultar® UV Max microplate reader and the associated Softmax Pro version 3.1.2 software (Molecular Devices, Sunnyvale, CA). The luminescence of each of the sample lysates was calculated and expressed as relative light units (RLU) per second per microgram of protein.

Autopsy mouse pancreas and liver tissue samples were collected from Tg(RIP-luc) mice that were either control normoglycemic or that had been treated with STZ and exposed to hyperglycemia for 8 days. The samples were fixed in 10% buffered-formalin for 48 hours prior to paraffin embedding. Tissue sections of 3∼4 microns were cut and placed on charged slides, allowed to air dry before placing in a 60°C oven for at least 1 hour for adhesion. After deparaffinization and antigen retrieval, the tissue sections were incubated with primary rabbit polyclonal antibody to Firefly Luciferase (Abcam Inc, Cat#ab21176, 1∶500) for 30 min at room temperature, followed by incubation with a ready-to-use EnVision™+ Single Reagents (Dako, Cat#K4003) for 15 min at room temperature. Images of the graft sections were acquired using a Zeiss fluorescence axial microscope attached with a digital camera.

Total RNA was prepared from frozen liver samples using TRI Reagent (Molecular Research Center) with bromochloropropane according to the manufacturer's protocol. RNA quantity and quality was determined spectrophotometrically and 0.7 µg total RNA was used for reverse transcription in the presence of MMLV-Reverse Transcriptase and random hexanucleotides (Promega). The reverse-transcribed samples were denatured and amplified by PCR in the presence of GoTaq Flexi DNA Polymerase (Promega) in a PCR System 6700 (Applied Biosystems) for over 30 cycles (94°C, 45 sec; 58.5°C, 45 sec; 72°C, 45 sec) with a final extension time (72°C, 4 min). A sham reaction without the reverse transcriptase was performed in order to ensure that PCR amplification did not arise from genomic DNA contamination. PCR primers used for the detection of insulin-I were: forward 5′-ccatcagcaagcaggtca-3′; reverse 5′-ccacacaccaggtagagagc-3′; insulin-II: forward 5′-cctgctggccctgctctt-3′ and reverse 5′-ggctgggtagtggtgggtcta-3′. The PCR products were separated by electrophoresis in 1.8% agarose stained with ethidium bromide. All reaction templates were normalized using the same PCR conditions with GAPDH expression using primers: forward 5′-accacagtccatgccatcac-3′ and reverse 5′-tccaccaccctgttgctgta-3′.

The mean ± standard error (SE) of the BLI signal intensity and luminescence for both groups were calculated and the unpaired two-tailed Student's t-test was used for statistical analysis. A probability (p) value of <0.05 was considered to be statistically significant.

Mouse serum glucose levels and luciferase expression in the liver were evaluated daily for 8-10 days following treatment with saline solution (Control, n = 6) or STZ (n = 6). Figures 1A and 1B show the ventral and dorsal bioluminescence images of a representative control mouse taken on days 0, 4 and 8 after the start of the experiment. The blood glucose reading and the BLI signal intensity of the circular area (ROI, the liver region of interest) were recorded under each image. As expected, strong and consistent levels of BLI signals were detected from the pancreas region on all days tested as the serum blood glucose level stabled. The strongest signal was detected from the tail-portion of the pancreas while the head of the pancreas can be more clearly visualized from the dorsal side. The BLI signal intensity from the liver ROI was near the background reading (8.0∼9.0×10⁴ photons/second) at all times tested.

In comparison, Figures 1C and 1D represent the luminescence imaging of a STZ-treated FVB/N-Tg(RIP-luc) mouse in ventral and dorsal positions on days 0, 2, 4, 6, and 8 as serum blood glucose levels increased over a period of eight days. As demonstrated by Figure 1C, there was a dramatic decrease in pancreatic luminescent signal soon after the STZ treatment as the β-cell function was lost, and a concomitant spatial shift of the luminescent intensity to the area where liver is located as serum blood glucose levels rose above 300 mg/dl. The location where the luminescence was detected was very similar to that of luminescence emitted from non-transgenic FVB/NJ mice transplanted with the luciferase-tagged islets in the liver [13]. The shift of luminescent expression also occurred on the dorsal side of the mouse as shown in Figure 1D. To quantify the changes in bioluminescence intensity, a fixed circle (ROI) was placed over the area of the upper right abdomen on the dorsal side as shown in the figures. This specific area of the liver was targeted because of its distant proximity from the pancreas, thus minimizing pancreatic signal interference. As shown in Figure 1D, the BLI signal obtained from the liver ROI of a representative mouse on day 0 was 8.1×10⁴ photons/second, similar to that of the controls. The signal intensity increased 5.43 fold to 4.4×10⁵ photons/second over the course of eight days under hyperglycemia. Enhanced luminescence was detected by BLI in the liver region as early as 24 hours following treatment with STZ when the blood glucose level rose near or above 200 mg/dl (data not shown).

We also observed enhanced luminescence signal on the left side of the mouse (Figure 1D) with the progression of hyperglycemia. The location and the shape of the BLI images suggested that it was emitted from the spleen. The spleen is one of the organs that demonstrated positive insulin gene expression under diabetic conditions [5]. In addition, some of the STZ-treated mice showed an enhanced luminescence signal from the brain area as detected by BLI (data not shown). As discussed above, brain tissue is known to be an active site for insulin II gene.

Figure 2A summarizes the average serum glucose levels from both the saline- and the STZ-treated mice on days 2, 4, 6, and 8 following treatment. As shown, the average serum glucose levels on day 0 for the control and treatment groups were 126±11.7 and 123±15.4 mg/dl respectively (p = 0.81). While the saline-treated control mice maintained normoglycemia during the entire study period, the STZ-treated mice had a progressive increase in the blood glucose level each day thereafter (230±14, 295±14, 329±16, and 334±15 mg/dl, respectively), resulting in a significant difference from that of the control mice on days 2, 4, 6 and 8 (p = 0.002, 2.98×10⁻⁵, 4.49×10⁻⁶ and 1.25×10⁻⁶ respectively). Moreover, the average serum glucose level of the STZ-treated mice on day 0 before treatment was significantly different from that of days 2, 4, 6 and 8 (p<0.001 for all time points).

With the increase in the blood glucose level, there was a significant increase in luciferase expression in the liver of STZ-treated mice. Figure 2B represents the quantified average BLI signal intensity from the mouse liver region on days 0, 2, 4, 6, and 8 following treatment with saline or STZ. The BLI signal of the liver was quantified from the fixed ROI on the dorsal image as demonstrated in Figures 1B and 1D. The mean (± SE) BLI intensities on day 0 for the control and treated mice were 8.56±0.02×10⁴ and 11.10±0.04×10⁴ photons/second respectively. No statistical differences between the groups were found (p = 0.55). The signal intensity from the liver area of STZ-treated mice increased significantly each day thereafter (23.30±0.66×10⁴, 47.0±1.85×10⁴, 58.5±2.30×10⁴ and 62.60±2.19×10⁴ photons/second respectively), resulting in a statistical difference from the control mice on days 2, 4, 6, and 8 (p = 0.044, 0.046, 0.012, and 0.031 respectively). An average of 5.64 fold increase in the liver signal intensity was detected in the mice that were exposed to hyperglycemia for 8 days. The average BLI of the STZ-treated mice on day 0 before treatment was statistically different from days 2, 4, 6, and 8 (p = 0.03, 0.0003, 0.0001, and 0.0002 respectively), whereas no significant differences were observed in the control mice.

To confirm the in vivo BLI finding that luciferase expression under the insulin promoter was elevated in the liver under hyperglycemia, liver autopsy samples were obtained on day 8 and luciferase assays were performed on tissue lysates from the liver to quantify the luciferase expression in control and treated mice. A total of 14 random control samples and 35 random STZ-samples from 6 mice in each group were analyzed. Figure 3 represents the mean (± SE) luciferase activity expressed in RLU per gram of protein. As shown by this figure, the mean luciferase activity for the control liver was 0.30±0.03×10⁹ whereas the luciferase activity for the hyperglycemia-liver produced an average of 10.00±0.10×10⁹. There was a 34-fold induction in the luciferase activity in the liver after an eight-day exposure to hyperglycemia with a very large statistical difference (p = 6.76×10⁻¹⁰) compared to that of the controls. Luciferase activity in the brain tissue homogenate was also measured, quantified and shown to be 2-fold higher in the STZ-treated mice than that of the controls (data not shown).

For the identification of cells expressing luciferase, liver tissue samples from the Tg(RIP-luc) mice, both diabetic and non-treated, were subject to immunohistochemical staining with an anti-luciferse antibody. As shown in Figure 4, luciferase-positive cells were detected in the liver samples from the STZ-induced diabetic mice, but not in the liver samples of the non-treated control transgenic mice. These cells are likely hepatic oval cells since they are small in size (approximately 10 µm), with a large nucleus-to-cytoplasm ratio and an oval-shaped nucleus, and are located in the periportal region of the liver. No hepatocytes were detected positive for luciferase expression. In the pancreas samples of the Tg(RIP-luc) mice, luciferase-positive cells were only detected in islets (data not shown).

To determine if the transgene expression detected by BLI was associated with the endogenous insulin gene promoter activity, the expression of insulin-I and insulin-II were examined by semiquantitative RT-PCR analysis of RNA samples prepared from the liver samples of control and STZ-treated mice. As shown in Figure 5, the expression of both the endogenous insulin-I and insulin-II genes were elevated in the liver RNA samples from animals exposed to hyperglycemia for 8 days (samples 3, 4, and 5). No insulin transcripts were detected in the control liver samples (samples 1 and 2).