Reporter islets in the eye reveal the plasticity of the endocrine pancreas

Significance

An adaptive plasticity of the insulin-secreting pancreatic beta cells, located in the islets of Langerhans, is essential for maintaining proper blood glucose homeostasis under various functional demands. Understanding the molecular mechanisms underlying the regulation of beta cell mass and function is critically dependent on an in vivo monitoring system that can continuously report on the functional status of the islets. Because direct inspection of these islets is obstructed due to the fact that they are deeply embedded and scattered within the pancreas, we propose the use of “reporter islets” transplanted into the anterior chamber of the eye, serving as optically accessible indicators of islet plasticity in the pancreas and as a tool to individually assess effectiveness of specific treatment regimens.

Abstract

The islets of Langerhans constitute the endocrine part of the pancreas and are responsible for maintenance of blood glucose homeostasis. They are deeply embedded in the exocrine pancreas, limiting their accessibility for functional studies. Understanding regulation of function and survival and assessing the clinical outcomes of individual treatment strategies for diabetes requires a monitoring system that continuously reports on the endocrine pancreas. We describe the application of a natural body window that successfully reports on the properties of in situ pancreatic islets. As proof of principle, we transplanted “reporter islets” into the anterior chamber of the eye of leptin-deficient mice. These islets displayed obesity-induced growth and vascularization patterns that were reversed by leptin treatment. Hence, reporter islets serve as optically accessible indicators of islet function in the pancreas, and also reflect the efficacy of specific treatment regimens aimed at regulating islet plasticity in vivo.

Normal fluctuations in blood glucose concentration trigger an orchestrated release of hormones from cells in the islets of Langerhans, the endocrine part of the pancreas. Beta cells produce and secrete insulin, an essential hormone regulating glucose uptake. A decrease in functional beta cell mass leads to impaired glucose homeostasis and diabetes, a serious disease that has grown into a worldwide epidemic. An adequate regulation of beta cell mass is thus of paramount importance to the ability of the pancreas to adapt to various functional demands and ensure maintenance of a normal blood glucose concentration (1).

The major challenge in functional studies of the islets of Langerhans in health and disease is the fact that they are deeply embedded in the exocrine pancreas, which limits their accessibility. Thus, understanding the regulation of function and survival and assessing the clinical outcome of individual treatment strategies for diabetes requires a monitoring system that can continuously report on the status of the endocrine pancreas in the living organism. We have developed a technical platform for noninvasive, longitudinal, in vivo imaging at single-cell resolution (2, 3). In the present study, we applied this in vivo imaging technique to test our hypothesis that islets transplanted into the anterior chamber of the eye can report on the functional status of the endogenous endocrine pancreas, and that intervention with pathological processes in the islets of Langerhans can be monitored in the eye.

For this purpose we used the obese mutant mouse (ob/ob) as a model system. This mouse, first described in 1950 (4), displays impressive islet plasticity during its lifetime and has been extensively studied as a model for obesity and insulin resistance. At a very young age ob/ob mice are hyperinsulinemic, hyperglycemic, and have a higher-than-average body weight (5). In addition to a distinct obesity coupled to a strong appetite, these mice exhibit various impaired functions, including reduced metabolic rate, impaired thermogenesis, impaired immunity, and infertility (6).

In 1994, Friedman et al. (7) identified the ob gene encoding for the hormone leptin, produced mainly in adipose tissue. This gene is mutated in the ob/ob mouse, leaving this mouse incapable of expressing functional leptin. Under normal conditions, leptin has many different physiological roles, but one of its most remarkable functions is in appetite regulation. In response to nutrition, leptin is released from adipose tissue and activates leptin receptors in the hypothalamus, leading to suppression of appetite and, consequently, reduced food intake. In addition, leptin receptors are expressed in beta cells and are involved in an adipo-insular feedback loop inhibiting insulin expression and release after food intake (8, 9). The lack of leptin in the ob/ob mouse thus results in a rapid increase in body weight and, in an effort to compensate for the increased demand for insulin, beta cell hyperplasia (10). As a result the islet cell population is altered in the ob/ob mouse; the percentage of beta cells compared with other endocrine cells is particularly high, accounting for more than 90% of the total islet cells (11). The absence of a leptin-driven feedback loop in the ob/ob mouse contributes to the high insulin release from beta cells, resulting in apparent degranulation (10), and the development of insulin resistance.

Using the ob/ob mouse as a model system, we demonstrated the versatile potential of optically accessible “reporter islets” in the anterior chamber of the eye. These reporter islets can be used to successfully monitor beta cell plasticity in situ in the endocrine pancreas and also to facilitate follow-up of specific treatment regimens.

Results

The ob/ob Mouse Displays Abnormal Physiological Properties.

The ob/ob mouse can be morphologically differentiated from its control littermate already at 4 wk of age, and its increased body mass becomes increasingly noticeable with advancing age (Fig. 1 A and B). Physiological studies revealed elevated fasting blood glucose and insulin levels compared with control littermates (Fig. 1C). This finding is in agreement with earlier reports (5) and is indicative of excessive food intake and insulin resistance.

Fig. 1.

Increases in islet size in response to high insulin demand in the ob/ob mouse pancreas. (A) Representative morphological appearance of a 4-mo-old ob/ob mouse (Left) compared with a control littermate (Right), showing a strong obese phenotype. (B) Fasted body weight of ob/ob and control mice at different ages shows a rapid increase in body mass in the ob/ob mice. (C) Fasted blood glucose and plasma insulin levels in 3-mo-old ob/ob and control mice. (D and E) Image montages of 5-µm-thick sections of 8-mo-old ob/ob (D) and control littermate (E) pancreata, stained with H&E. Note the staining of islet sections in light gray. (Insets) Magnified views of typical islet dimensions and morphology. B and C show values for mixed males and females (n = 5 males + 4 females for ob/ob; n = 5 males + 5 males for control littermates). Values are average ± SEM. **P < 0.01; ***P < 0.001. (Scale bar: 1 mm; inset dimensions, 1 mm × 1 mm.)

Paraffin-embedded sections of ob/ob mouse and control mouse pancreata were stained with H&E to investigate differences in islet morphology. Islet dimensions were increased in the ob/ob mouse, providing a large potential insulin secretory capacity in an effort to compensate for increased food intake (Fig. 1 D and E). Although examination of pancreas sections provides information on the morphological status of islets at specific time points, dynamic changes cannot be appreciated.

Intraocular Islet Transplants Mirror the Adaptive Morphological Plasticity of in Situ Endogenous Islets.

To study the morphological plasticity of these islets over time, we transplanted a few reporter islets into the anterior chamber of the mouse eye, which can be optically accessed for longitudinal in vivo imaging. Pancreatic islets were isolated from donor mice at the age of 4 wk and transplanted into the anterior chamber of the eye of age-matched recipients. This in vivo environment contains a rich capillary network, as well as a nerve network, that connect to the engrafted islets to permit not only intercellular paracrine input, but also endocrine and nervous input (2, 12).

Transplanted islets rapidly engrafted onto the iris, and individual islet transplants could be identified repeatedly at various time points after transplantation (Fig. 2A). In addition to increased growth, these islets exhibited increased intra-islet vessel diameters in the ob/ob recipients (Fig. 2B), as previously documented from dissected pancreatic tissue in vitro (13, 14). The increase in intra-islet vessel diameter could be part of a compensatory mechanism designed to increase blood perfusion under hyperglycemic conditions (15). It also might be an approach to optimizing endocrine signaling, ensuring that each beta cell will have direct communication with blood flow under this strong proliferative condition. Individual islets were imaged in vivo at different time points in syngeneically transplanted mice by confocal microscopy, and the large and tortuous blood vessels in the ob/ob islet grafts were confirmed by fluorescence imaging (Fig. 2C). The reflective properties of hormone-containing vesicles permit the characterization of islet morphology by backscatter imaging (16), as well as insulin secretion status (Fig. S1). The ob/ob mouse islets displayed a seemingly degranulated and uneven pattern compared with control (Fig. 2C). The poor reflection of light is indicative of rapid insulin secretion after food intake, which is further intensified by disruption of the adipo-insular feedback loop under leptin-deficient conditions (9). Immunohistochemistry analysis of in situ islets in paraffin-embedded sections showed that islets in the pancreas have similar distinctive properties in ob/ob and control mice; CD31 staining showed large vessels in the ob/ob islets, and insulin staining was irregular (Fig. S2).

Fig. 2.

In vivo longitudinal imaging of islet growth. (A) Islets from 4-wk-old mice were transplanted into the anterior chamber of the eye of control and ob/ob mice at the age of 4 wk. Photography of transplanted eyes at different time points shows that individual islets can be identified and followed longitudinally (see yellow dashed circle). (B) Magnified views of islet grafts (marked by red frames in A) show the clearly visible large and tortuous blood vessels in the islet engrafted onto the iris of ob/ob recipient. (C) In vivo imaging of single islets 1 mo after transplantation by confocal microscopy shows morphological differences between islet grafts in control versus ob/ob mice. Vascularization is visualized by i.v. injection of FITC-labeled dextran prior to imaging. Note differences in backscatter intensity and vessel diameters. (D) In vivo imaging of islet grafts at different time points after transplantation by confocal microscopy. (E) Quantification of islet volumes by analysis of backscatter images reveals a significantly increased growth in ob/ob (solid lines) compared with control (dashed lines). Gray lines represent average islet volumes in single mice; black lines represent averaged values obtained per genotype (n = 3). (F) Immunohistochemistry analysis shows a strong proliferation of beta cells in both ob/ob transplanted eye and pancreas, as seen by insulin and Ki67 staining. (G) Average beta cell area was quantified from insulin and DAPI staining, demonstrating that beta cells from ob/ob mice (solid lines) were significantly larger than those from their control littermates (dashed lines). This hypertrophy was similar and independent of whether the islets were located in situ in the pancreas or in the transplanted eye (no significant differences). All images are representative. Confocal images are displayed as maximum intensity projections (MIPs) of optical Z-stacks. Immunohistochemistry experiments were performed using 4-mo-old mice (n ≥ 3 per group). Error bars represent SEM. *P < 0.05; **P < 0.01. (Scale bars: 100 µm.)

In vivo imaging of transplanted islets by confocal microscopy provides 3D morphological information that allows for precise quantification of islet volume at any given time point after transplantation (Movie S1). The quantification of average islet volumes over time revealed significant differences between ob/ob and control mice starting at 1 mo after transplantation, illustrating a dramatic islet growth in the ob/ob mouse (Fig. 2 D and E). Interestingly, this growth proved to be independent of the mouse donor genotype; we observed a similar growth after transplanting control mouse islets into ob/ob mice, demonstrating that in this particular case, (i) signaling factors originating from the recipient mouse dictate morphological changes of transplanted islets, and (ii) the transplanted islets reflect morphological behavior of the recipient’s in situ pancreatic islets (Fig. S3).

To examine whether this observed plasticity in the anterior chamber of the eye indeed reflects the plasticity occurring in situ in islets located in the pancreas, we compared paraffin-embedded sections of islet transplants and endogenous pancreatic islets by immunohistochemistry. We observed a strong proliferation of beta cells in the ob/ob mouse islets by staining for insulin and Ki67 (Fig. 2F). This proliferation rate was not significantly different when measured from in situ pancreatic islets or from islets engrafted in the anterior chamber of the eye (mean ± SEM, 0.82 ± 0.10%; n = 5 vs. 0.95 ± 0.15%; n = 4 mice, respectively). We also assessed morphological changes in beta cells by immunohistochemistry, which revealed a ∼1.35-fold increase in cell area in the ob/ob mice compared with control mice in both in situ islets located in the pancreas and islet grafts (Fig. 2G). The lack of leptin signaling is a likely contributor to this volume expansion. Indeed, using pancreas-specific leptin receptor KO mice, it has been shown that the disruption of leptin action in islets results in enhanced PI3K/AKT signaling, associated with increases in beta cell size and islet mass (17). Although this compensatory islet growth has been demonstrated in vitro, we are now able to verify this pattern in the living ob/ob mouse longitudinally through the analysis of single reporter islets. Importantly, our data are compatible with previous studies based on isolated islets (18) or on imaging of histological sections spanning the entire pancreas (19) showing a fourfold difference in mean islet volume between control and ob/ob mice at the age of 2 mo.

Leptin Treatment Reduces Body Weight and Corrects Blood Glucose Homeostasis in the ob/ob Mouse.

Having witnessed the behavior of individual islets in response to food overconsumption and insulin resistance, we next questioned whether this growth could be reduced or even reversed by treating the ob/ob mice with leptin. Mice were injected i.p. daily between 3 and 4 mo of age, and the physiological effects of this treatment were monitored regularly using different parameters. In concert with a visibly reduced appetite, we found decreases in body weight, blood glucose levels, and insulin concentrations during treatment (Fig. 3A). This beneficial effect of leptin was not permanent, however; body weight increased rapidly after the end of the treatment period, reaching levels similar to those in age-matched untreated ob/ob mice approximately 1 mo later (compare with Fig. 1B).

Fig. 3.

Physiological effects of leptin treatment on ob/ob mice. (A) The ob/ob mice received daily i.p. injections of leptin between 3 and 4 mo of age. Body weight and blood glucose and plasma insulin levels were monitored before, during, and after the treatment; the beginning and end of treatment are represented by vertical dashed lines. (B) The i.p. glucose tolerance tests show impaired glucose handling in the ob/ob mice compared with control littermates at the age of 4 mo (Top), which is normalized by leptin treatment but not by sham treatment (Middle). (Bottom) Bar graph showing area under the curve (AUC) values from the above traces, demonstrating the beneficial effect of leptin on glucose handling. Values are average ± SEM. **P < 0.01; ***P < 0.001.

The leptin treatment also had a beneficial impact on insulin sensitivity. At the end of the treatment period, i.p. glucose tolerance tests in ob/ob mice demonstrated glucose excursions similar to those measured in age-matched control mice (Fig. 3B).

Longitudinal in Vivo Imaging Reveals the Reversed Dysregulation of ob/ob Mouse Islets by Leptin.

The leptin treatment also exerted an influence on islet growth. Whereas sham-treated mice displayed a continuous adaptive growth of islet transplants, this growth was abolished and partially reversed under leptin treatment (Fig. 4 A and B), in agreement with earlier reports (20, 21). At the end of the treatment, the proliferation of beta cells was virtually abolished in both the engrafted islets and the in situ pancreatic islets from leptin-treated ob/ob mice (Fig. 4C). After leptin treatment, the average beta cell size was decreased and was not significantly different from that in age-matched untreated control mice, independent of whether the islets were transplanted or were present endogenously in the pancreas. The decrease in islet size could be explained in part by this decrease in individual beta cell size, as reported previously (22), as well as by morphological changes in intra-islet vasculature (Fig. 5). We measured a decrease in vessel diameter in islet grafts that paralleled the decrease in blood glucose levels. Interestingly, we observed angiogenesis in the islet grafts at an early stage after leptin treatment, a biological phenomenon known to enhance blood perfusion and to be positively influenced by leptin via a synergistic stimulation with the angiogenic factors FGF-2 and VEGF (23).

Fig. 4.

Leptin treatment reverses dysregulation of ob/ob islets. (A) Longitudinal in vivo imaging of islet grafts in ob/ob mice receiving leptin treatment (Upper) or sham treatment (Lower) between 3 and 4 mo of age. Vasculature was visualized by tail vein injection of dextran-FITC. (B) Islet volume analysis showing reversal of islet growth during leptin treatment (gray bars) compared with sham treatment (white bars). There was no difference in islet growth after the end of leptin or sham treatment. (C) Immunohistochemistry analysis of transplanted eye and pancreas samples from ob/ob mice at the end of the leptin treatment demonstrating the attenuated beta cell proliferation by immunostaining for insulin and Ki67 (compare with Fig. 2F). Confocal images are representative and shown as MIPs. Values are average ± SEM. ***P < 0.001. (Scale bars: 100 µm.)

Fig. 5.

Effect of leptin treatment on intra-islet vascularization. Islets were transplanted into the anterior chamber of the eye of 4-wk-old ob/ob mice. At the age of 3 mo, the mice received daily i.p. injections of leptin. (A) Leptin administration had a rapid effect on blood vessel diameters, as demonstrated by in vivo imaging of the same islet before treatment (Left) and after 1 wk of treatment (Right). The vasculature was visualized by tail vein injection of dextran-FITC. The same vessel segments could be identified and their diameters measured at different time points (red lines). Note that angiogenesis is evident after leptin administration (arrows). (B) Longitudinal analysis of individual vessel diameters showing vessel morphological changes before, during, and after leptin treatment. (Upper) Traces of 20 vessel segments from islet grafts in the ob/ob mouse over time. (Lower) Bar graph showing corresponding average monthly diameter increases. Confocal images are representative and shown as MIPs. Values are average ± SEM. ***P < 0.001; ****P < 0.0001. (Scale bars: 100 µm.)

Owing to the restored leptin-driven adipo-insular feedback loop and the decrease in blood glucose levels, the resulting regranulation of islet cells in vivo in transplanted islets was observed by their stronger reflective optical properties (Fig. 4A). This is in accordance with the increased immunostaining for insulin perceived in islets from leptin-treated ob/ob mice (20). The normalization of insulin release in vivo after leptin treatment in the ob/ob mouse has been linked to mechanisms both dependent on and independent of food intake (21).

Discussion

To test our concept of reporter islets revealing the status of the in situ endocrine pancreas, we made use of the significant remodeling capacity of the endocrine pancreas in the leptin-deficient ob/ob mouse. Reporter islets from the ob/ob mouse model were transplanted into the anterior chamber of the eye of a similar type of mouse for the investigation of islet cell mass regulation, representative of that occurring in the endocrine pancreas. The reporter islets displayed obesity-induced growth and vascularization patterns identical to those observed in situ in the pancreas. These obesity-induced growth and vascularization patterns were reversed by leptin treatment in both the anterior chamber of the eye and the in situ pancreas, providing evidence of a highly effective islet remodeling aiming at the expansion or reduction of the insulin secretory potential.

Importantly, using this technique, we were able to witness over time and in vivo the impressively fast response of single islets to functional demands, calibrating not only their individual overall beta cell mass but also at more subtle levels, reshaping vascular architecture and beta cell sizes. This remodeling capacity has so far been only indirectly suggested by cross-sectional studies or by the analysis of pancreatic islet populations at relatively low imaging and temporal resolutions in vitro. Thus, reporter islets can successfully reveal the molecular mechanisms regulating adaptive plasticity of the endocrine pancreas in vivo.

Various different experimental approaches allow for the quantification of beta cell mass and for the possible remodeling of islets. In vitro techniques include the analysis of histological sections of pancreas by point counting morphometry (19) and the direct measurements of islet dimensions in dissected pancreatic tissue (24, 25). Ex vivo imaging techniques involve the imaging of exteriorized pancreas by confocal microscopy (26) or optical coherence microscopy (27). Finally, several noninvasive in vivo imaging techniques aim to longitudinally quantify total beta cell mass, including magnetic resonance imaging (28), positron emission tomography (29), bioluminescence imaging (30), and combined multimodal imaging (31). All of these techniques deliver precious information either at the islet population level or at the islet and cellular level. However, they provide only indirect evidence of islet plasticity, without the capability to follow morphological changes in individual islets over time.

We have shown that a few reporter islets can reveal, in a representative way, the remodeling of the in situ pancreatic islet population, expanding or reducing their insulin-secretory potential. Thus, this technique enables us to “merge” the study of individual islets with the study of islet populations, and has the potential to replace multiple cross-sectional experiments with longitudinal studies. We thus propose in vivo imaging of reporter islets transplanted into the anterior chamber of the eye as a versatile tool to clarify molecular mechanisms as well as identify pharmacologic compounds in the regulation of beta cell function and survival. Importantly, a homologous use of reporter islets in humans, to both diagnose islet malfunction and monitor effects of specific individually based treatment regimens, may be viewed as a unique personalized medicine approach. Our concept of reporter islets can be developed further by implementation of multiple cellular biomarkers for function and proliferation with high-temporal resolution and also extended to other organs, with the goal of identifying and understanding in detail molecular interactions and adaptive mechanisms at the cell and organ levels, with implications for both physiology and pathology.

Materials and Methods

Mouse Model.

The ob/ob mice used in our experiments originated from Umeå, Sweden, and were inbred in the animal core facility at Karolinska Hospital. Discrimination from control lean littermates was achieved by phenotypic and genotypic analysis (32). All experiments were performed in accordance with the Karolinska Institutet’s guidelines for the care and use of animals in research and were approved by the institute’s Animal Ethics Committee.

Physiological Measurements and Leptin Treatment.

Body weight and blood glucose levels were measured in a minimum of seven mice. Glucose concentrations were obtained using the Accu-Chek Aviva monitoring system (Roche). Glucose tolerance tests were performed by i.p. injection of 2 mg of glucose per 1 g of body weight in mice that had been fasted for 16 h (n ≥ 5). Plasma insulin concentrations in a minimum of three mice per condition were measured using mouse insulin ELISA plates (Mercodia). Leptin treatment was performed by daily i.p. injections of 1.5 µg/g body weight of recombinant human leptin (Amylin Pharmaceuticals), and sham treatment was performed by i.p. injection of water instead of leptin.

Transplantation of Pancreatic Islets into the Anterior Chamber of the Eye.

Islets were isolated from female donor mice and transplanted into the anterior chamber of the eye of male recipients, using a technique described previously (2). During anesthesia using isoflurane (Baxter), ∼10–20 islets were transplanted into the anterior chamber of the mouse eye. Islets used for transplantation were selected to be of similar size for each experiment, independent of the mouse genotype. The mice were injected s.c. with Temgesic (Schering-Plough) for postoperative analgesia.

In Vivo Imaging of Intraocular Islet Grafts.

Islet grafts in mice were imaged in vivo at specific time points after transplantation as described previously (2), using an upright laser scanning confocal microscope based on a Leica TCS-SP2-AOBS with a long-distance water-dipping objective (Leica HXC-APO 10×/0.30 NA) and a custom-built stereotaxic head holder allowing positioning of the mouse eye containing the engrafted islets toward the objective. Viscotears (Novartis) was used as an immersion liquid between the eye and the objective, and isoflurane was used to anesthetize the mice during in vivo imaging. Imaging of islet morphology was done by laser illumination at 633 nm, and backscattered light was collected at the same wavelength.

For visualization of blood vessels, 100 μL of a solution containing 2.5 mg/mL of 500-kDa FITC-labeled dextran (Invitrogen) was injected into the tail vein, followed by fluorescence imaging using a 496-nm excitation wavelength. Scanning speed and laser intensities were adjusted to avoid cellular damage to the mouse eye or islet graft.

Tissue Sections and Immunohistochemistry.

Dissected tissues were rinsed briefly with PBS, fixed with formalin for 48 h at room temperature, dehydrated, and embedded in paraffin. Then 5-μm-thick sections mounted on precoated microscope slides were dewaxed by xylene and progressively rehydrated before processing. Insulin was stained using chicken anti-insulin (1:200 dilution; Abcam) followed by goat anti-chicken Alexa Fluor 488 (1:1,000 dilution; Invitrogen) antibodies; blood vessels were stained using rat anti-mouse CD31 (1:50 dilution; BD Pharmingen), followed by incubation of biotinylated goat anti-rat (5 µg/mL; Vector Laboratories) and amplification with HRP-streptavidin/Alexa Fluor 647-tyramide (Invitrogen). Proliferation was assessed using mouse anti-human Ki67 (1:50 dilution; Novocastra) together with M.O.M. biotinylated anti-mouse IgG (Vector Laboratories), and amplified with HRP-streptavidin/Alexa Fluor 647-tyramide. Nuclei were stained and slides were preserved by cover glass mounting using ProLong Gold Antifade Reagent with DAPI (Invitrogen). Slides were imaged with a BD Pathway 855 system (BD Biosciences). Insulin, CD31, Ki67, and DAPI stainings were imaged using an Olympus 20×/0.75 NA UApo/340 objective, and overview of H&E-stained pancreatic sections was obtained in montage capture mode with an Olympus 4×/0.16 NA UPlan SApo objective.

Image Processing and Analysis.

AutoQuant X2 software (Media Cybernetics) was used for blind deconvolution of all in vivo confocal images before image analysis. Islet volume was analyzed based on the backscatter signal channel, using MATLAB and the Image Processing Toolbox (MathWorks). The islet “equatorial volume” (i.e., the volume from the top of the islet down to the calculated equator) was used to represent the volume and calculate islet growth. A minimum of three islet grafts per animal were analyzed to determine islet growth, with a minimum of three mice per category. Individual vessel segments were identified at different imaging time points, and their diameters were measured using Volocity image analysis software (PerkinElmer). Image analysis protocols were established in Volocity for automated analysis of histological sections. The average beta cell section area was obtained by dividing the insulin-stained area by the total number of DAPI-stained nuclei enclosed in this area (minimum of 1,000 cells per tissue). The beta cell proliferation rate was calculated by counting Ki67-positive nuclei and dividing by the total number of DAPI-stained nuclei in insulin-positive cells (minimum of 10,000 cells per transplanted eye or pancreas). Volocity was used for image display, and Adobe Photoshop CS5 was used for image assembly.

Statistical Analysis.

All results are presented as average ± SEM. The Student t test was used to determine statistical significance, with a P value < 0.05 considered to indicate significance.