Expression of IGF-I in Pancreatic Islets Prevents Lymphocytic Infiltration and Protects Mice From Type 1 Diabetes

Transgenic CD1 mice (N5 generation) overexpressing IGF-I in β-cells (20) and transgenic CD1 mice (N10 generation) expressing human IFN-β in β-cells (12) were used. Double-transgenic mice were obtained by cross-breeding these two lines. All experiments were performed in male mice. When stated, nontransgenic and transgenic mice were given, on 5 consecutive days, an intraperitoneal injection of STZ (15 or 20 mg/kg body wt), dissolved in 0.1 mol/l citrate buffer (pH 4.5), immediately before administration. Mice were fed ad libitum with a standard diet (Panlab, Barcelona, Spain) and kept under a 12-h light-dark cycle (lights on at 8:00 a.m.). Animal care and experimental procedures were approved by the ethics committee in animal and human experimentation of the Universitat Autònoma de Barcelona.

For immunohistochemical detection of IFN-β, IGF-I, insulin, glucagon, somatostatin, pancreatic polypeptide, and Fas proteins, pancreata were fixed for 12–24 h in formalin, embedded in paraffin, and sectioned. Sections were then incubated overnight at 4°C with sheep anti–human IFN-β antibody (Biosource International, Camarillo, CA) diluted at 1:500, with rabbit anti–rat IGF-I antibody (GroPep, North Adelaide, Australia) diluted at 1:50, with guinea pig anti–porcine insulin antibody (Sigma Chemical, St Louis, MO) at 1:100 dilution, with rabbit anti–human glucagon antibody (ICN Biomedicals, Irvine, CA) at 1:4,500 dilution, with rabbit anti–human somatostatin antibody (Serotec, Oxford, U.K.) at 1:750 dilution, with rabbit anti–human pancreatic polypeptide antibody (ICN Biomedicals) at 1:2,000 dilution, or with goat anti–mouse Fas antibody (R&D Systems, Abigdon, U.K.) at 1:10 dilution. As secondary antibodies, rabbit anti–guinea pig IgG, coupled to peroxidase (Dako, Glostrup, Denmark), or biotinylated goat anti-rabbit antibody (Pierce, Rockford, IL) and ABC complex (Vector Laboratories, Burlingame, CA) were used. The substrate chromogen was 3′3′-diaminobenzidine. Sections were counterstained in Mayer’s hematoxylin. The localization of IFN-β, IGF-I, and insulin in islets of transgenic mice was determined using fluorescein isothiocyanate–labeled goat anti-rabbit (Molecular Probes, Eugene, OR) or Texas Red isothiocyanate (TRITC)-labeled goat anti–guinea pig (Molecular Probes) as secondary antibodies. To study the coexpression of Fas and insulin in islets, biotinylated donkey anti-goat (sc-2040; Santa Cruz Biotechnology, Santa Cruz, CA) and TSA Fluorescence System (Perkin Elmer, Boston, MA) or TRITC-labeled goat anti–guinea pig (Molecular Probes) were used as secondary antibodies. For immunohistochemical detection of CD4 and CD8, pancreata were fixed for 4 h in 2% paraformaldehyde at 4°C, incubated overnight in 30% sucrose at 4°C, and then embedded in OCT. The cryostat sections (5 μm) were incubated overnight at 4°C with biotinylated rat anti–mouse CD4 antibody (Pharmingen International) diluted at 1:300 or with biotinylated rat anti–mouse CD8 antibody (Pharmingen International) diluted at 1:300.

The incidence and severity of insulitis was analyzed in four paraffin sections per pancreas, separated by 150 μm, and stained with hematoxylin. The degree of mononuclear cell infiltration (insulitis score) was ranked as follows: noninfiltrated, peri-insulitis (mononuclear cells surrounding islets and ducts but no infiltration of the islet architecture), moderate insulitis (mononuclear cells infiltrating <50% of the islet architecture), and severe insulitis (>50% of the islet tissue infiltrated by lymphocytes and/or loss of islet architecture).

Pancreata were obtained from control and transgenic mice, and immunohistochemical detection of insulin was performed in three (2–3 μm) sections, separated by 200 μm. The area (in micrometers squared) of each islet and the area of each section were determined using a Nikon Eclipse E800 microscope (Nikon, Tokio, Japan) connected to a video camera with a color monitor and an image analyzer (analySIS 3.0; Soft Imaging System, Lakewood, CO). The percentage of β-cell area in the pancreas was calculated by dividing the area of all insulin-positive cells in one section by the total area of this section and multiplying the result by 100. The β-cell mass was calculated by multiplying the pancreas weight by the percent of β-cell area.

Real-time PCR was performed to quantify the specific mRNA for β₂-microglobulin. Pancreatic islets were isolated from 3-month-old mice by intraductal infusion of collagenase solution (1 mg/ml; Collagenase P; Roche Molecular Biochemicals, Mannheim, Germany) and digestion in Hank’s solution for 10 min at 37°C. Total RNA was extracted from islets using TriPure Isolation Reagent (Roche Molecular Biochemicals) according to the manufacturer’s instructions. After denaturation, 1 μg total RNA was reverse transcribed for 1 h at 37°C using the Ommiscript Reverse Transcriptase kit (d-40724; Qiagen, Hilden, Germany). Real-time PCR was performed in a SmartCycler II (Cepheid, Sunnyvale, CA) using QuantiTect SYBR Green kit (Qiagen) and β₂-microglobulin specific primers: forward CCGGAGAATGGGAAGC and reverse GTAGACGGTCTTGGGC (Proligo Primers and Probes, Boulder, CO). The β₂-microglobulin cDNA levels were normalized to 36B4 cDNA using the following primers: forward GGCCCTGCACTCTCGCTTT and reverse TGCCAGGACGCGCTTGT. The results were analyzed using the mathematical model of Pfaffl (25).

Apoptotic cells were identified in deparaffinized pancreatic sections using transferase-mediated dUTP nick-end labeling (TUNEL) (in situ cell death detection kit; Roche Molecular Biochemicals). Non–β-cells were stained with an antibody cocktail (anti-glucagon, anti-somatostatin, and anti–pancreatic polypeptide). Nuclei were counterstained with Hoechst (no. 332588; Sigma Chemical). We considered the β-cell apoptotic when it was positive for TUNEL and negative for an antibody cocktail staining. On these sections, the mantle of non–β-cells showed red cytosolic staining, while the TUNEL-positive cells had green nuclei. At least 1,000 islet β-cell nuclei per pancreas were counted, and five animals per group were examined.

Blood glucose levels were measured with a Glucometer Elite analyzer (Bayer, Leverkusen Germany). Serum insulin concentrations were determined by radioimmunoassay (CIS Biointernational, Gif-Sur-Yvette, France). To determine pancreatic insulin content, whole pancreata were removed from the mice, weighed, and homogenized in 20 vol/vol of cold acidic ethanol (75% ethanol and 1.5% concentrated HCl) followed by 48 h of agitation at 4°C. Afterward, insulin was quantified in the supernatants of the samples diluted in phosphate buffer by radioimmunoassay (CIS Biointernational). For glucose tolerance test, awake mice fasted overnight (16 h) were given an intraperitoneal injection of glucose (1 g/kg body wt). At the times indicated, blood samples were obtained from the tail vein of the same animals and glucose concentration measured.

All values are expressed as the means ± SE. Differences between groups were compared by Student’s t test. A P value <0.05 was considered statistically significant.

To examine the role of IGF-I in counteracting type 1 autoimmune diabetes, an animal model with established recurrent infiltration of islets should be used. To this end, we took advantage of transgenic mice developed in our laboratory that express human IFN-β in β-cells, which show islet infiltration and increased susceptibility to develop diabetes (12,13). The initial experiments focused on establishing the suitability of IFN-β–expressing transgenic mice CD1 genetic background as a model of autoimmune diabetes. These mice expressed IFN-β specifically in β-cells (Fig. 1A) and showed circulating levels of human IFN-β (75 ± 10 IU/ml). Although ∼20% of transgenic mice developed overt diabetes at 2 months of age, the rest remained normoglycemic (135 ± 4 vs. 140 ± 5 mg/dl, control vs. transgenic, respectively) and normoinsulinemic (3 ± 0.5 vs. 2.8 ± 0.4 ng/ml). However, ∼45% of islets from 3-month-old normoglycemic transgenic mice showed lymphocytic infiltration (25% with peri-insulitis [<25% infiltration], 15% with moderate [<50% infiltration], and 5% with severe [>50% infiltration] insulitis) (Fig. 1B). In type 1 diabetes, autoreactive CD4⁺ and CD8⁺ T lymphocytes are activated in pancreatic lymph nodes and then recruited to pancreatic islets. Immunohistochemical analysis performed in the pancreas of 3-month-old transgenic mice clearly showed that cells surrounding the islets were positive for CD4 and CD8 T-cell markers (Fig. 1C). In contrast, neither CD8⁺ nor CD4⁺ T-cells were observed in the control pancreas (Fig. 1C). In spite of islet infiltration and β-cell destruction, these transgenic mice were normoinsulinemic, indicating that the remaining β-cells compensate for insulin production. Similarly, in humans, islet infiltration takes place months, or even years, before the clinical onset of the disease (3).

Because of the presence of insulitis, normoglycemic transgenic mice may exhibit high susceptibility to develop overt diabetes when treated with very-low doses of STZ, which does not cause diabetes in controls. To this end, 2-month-old control and transgenic mice received injections of either 20 or 15 mg/kg body wt STZ for 5 consecutive days. In contrast to STZ-treated control mice, which remained normoglycemic, all transgenic mice progressively developed hyperglycemia after treatment with either 15 or 20 mg/kg STZ (Fig. 2A and B). However, the increase in blood glucose levels was higher after treatment with five doses of 20 mg/kg. These animals reached hyperglycemic values of >600 mg/dl, the upper limit of measurement, at ∼10 weeks after STZ treatment (Fig. 2B). The degree of lymphocytic infiltration of islets was determined 2 and 4 months after treatment with 20 mg/kg STZ (Fig. 2C). Transgenic islets showed a progressive increase in insulitis, which was more severe at 4 months, when all islets were highly infiltrated (Fig. 2C). Thus, ∼60% of islets presented severe insulitis, 35% presented moderate, and only 5% exhibited peri-insulitis (Fig. 2C). In contrast, only ∼15% of islets from control mice showed mild peri-insulitis (Fig. 2C). These results indicate that the expression of IFN-β in β-cells increased islet sensitivity to STZ, which exacerbated infiltration and led to type 1 diabetes. Moreover, 4 months after STZ treatment, transgenic mice presented decreased serum insulin levels (∼90%), while controls remained normoinsulinemic (Fig. 3A). These results were consistent with the reduction in pancreatic insulin content and β-cell mass in STZ-treated transgenic mice (Fig. 3B–D). Whereas control mice treated with 20 mg/kg STZ presented normal levels of β-cell mass, transgenic mice showed a severe reduction (∼94%) (Fig. 3C and D). Thus, these results suggest that transgenic mice expressing IFN-β in β-cells treated with very-low doses of STZ may be used as a model of type 1 diabetes with recurrent islet infiltration.

To determine whether IGF-I expression in β-cells protected from recurrent lymphocytic infiltration, we examined double-transgenic mice overexpressing both IFN-β and IGF-I in β-cells. Islets from these mice showed colocalization of IFN-β and IGF-I in insulin-expressing cells (Fig. 4A). In contrast to islets from 3-month-old IFN-β transgenic mice that showed lymphocytic infiltration, islets from IFN-β+IGF-I–expressing mice did not show insulitis, indicating that IGF-I expression protected β-cells from immune response (Fig. 4B). Furthermore, islets from IFN-β transgenic mice showed about a ninefold increase in β₂-microglobulin expression, which indicated an upregulation of major histocompatibility complex (MHC) class I antigen by IFN-β (Fig. 4C). In contrast, the presence of IGF-I in IFN-β–expressing islets decreased β₂-microglobulin expression ∼50%. These findings suggest that IGF-I expression may protect double-transgenic mice in part by decreasing MHC class I antigen–induced T-cell immunity against islets.

IGF-I production from β-cells also protected double-transgenic mice from development of hyperglycemia after treatment with multiple very-low doses of STZ (5 × 20 mg/kg) (Fig. 5A). Thus, 100% of IFN-β transgenic mice developed diabetes by 10 weeks after STZ treatment (glucose >500 mg/dl), while all IFN-β+IGF-I double-transgenic mice remained normoglycemic (142 ± 4 mg/dl), similarly to STZ-treated control (143 ± 5 mg/dl) and IGF-I (139 ± 4 mg/dl) transgenic mice. When lymphocytic infiltration of islets was determined 4 months after STZ treatment, a strong reduction of insulitis was observed in IFN-β+IGF-I double-transgenic mice (Fig. 5B). These mice only showed ∼10% of islets with peri-insulitis, while ∼90% of islets from IFN-β mice presented infiltration (Fig. 5C). This was also consistent with normal insulinemia (Fig. 5D), pancreatic insulin content (Fig. 5E), and β-cell mass (Fig. 6A) in IFN-β+IGF-I–expressing mice 4 months after STZ treatment. Furthermore, in contrast to STZ-treated IFN-β–expressing mice, STZ-treated IFN-β+IGF-I double-transgenic mice showed normal glucose tolerance, indicating that they responded to an intraperitoneal glucose load similarly to STZ-treated control and IGF-I transgenic mice (Fig. 6B). These findings indicate that IGF-I protected β-cells from lymphocytic infiltration and double-transgenic mice from development of diabetes. Non–STZ-treated IGF-I+IFN-β double-transgenic mice showed normal islet cell distribution. However, when these mice were treated with STZ, their islets presented altered distribution of α- and β-cells, since α-cells no longer formed a mantle around the β-cell core (Fig. 6C). In contrast, STZ-treated control and IGF-I transgenic mice showed islets with normal distribution of insulin-expressing cells in the core and glucagon-expressing cells in the periphery (Fig. 6C). These results suggest that a few β-cells were probably slowly destroyed and regenerated in double transgenic mice leading to disorganization of islet cell distribution.

To study if IGF-I expression protected β-cells from STZ- and IFN-β–induced apoptosis in double-transgenic mice, we performed immunohistochemical analysis to detect non-β-cells (with an anti-glucagon, anti-somatostatin, and anti-pancreatic polypeptide antibody cocktail) and TUNEL to identify apoptotic cells. One month after STZ treatment, a great increase in TUNEL-positive β-cells was detected in IFN-β transgenic islets (Fig. 7A and B). In contrast, STZ-treated IGF-I+IFN-β–expressing islets showed very-low levels of apoptotic cells, similar to STZ-treated control and IGF-I–expressing islets (Fig. 7B). Furthermore, in IFN-β transgenic mice, most β-cells showed high Fas immunostaining 1 month after STZ treatment (Fig. 7C). In contrast, Fas expression was neither observed in islets from STZ-treated control and IGF-I transgenic mice nor in islets from STZ-treated IGF-I+IFN-β double-transgenic mice (Fig. 7C), indicating that IGF-I protected β-cells from Fas-induced apoptosis. Therefore, these results indicate that IGF-I expression clearly protected β-cells from lymphocytic infiltration and apoptosis.

In this study, we show that IGF-I expression in β-cells counteracts recurrent lymphocytic infiltration of islets. We have established that transgenic mice expressing IFN-β in β-cells, in CD1 genetic background, are an appropriate model of autoimmune diabetes used to examine the role of IGF-I. In such transgenic mice, expression of IFN-β results in a pre-diabetic state, with functional alterations in islets and insulitis (12,13). IFN-β transgenic mice in C57Bl6/SJL genetic background show no lymphocytic infiltration of islets and decreased susceptibility to diabetes onset compared with CD1 mice (12), indicating that diabetes susceptibility is related in part to the mouse strain. Here, we found that treatment with very-low doses of STZ, which do not affect control animals, triggered the diabetes process in CD1 IFN-β transgenic mice. These mice developed marked mononuclear infiltration in islets with the presence of CD4⁺ and CD8⁺ T-cells, which progressed from peri-insulitis to severe insulitis, similar to type 1 diabetes in humans (26,27). This led to decreased β-cell mass and overt diabetes, with chronic hyperglycemia. Furthermore, hyperexpression of Fas and increased β-cell apoptosis were detected in islets of IFN-β transgenic mice. Fas and FasL expression have also been found in pancreatic byopsies from recent-onset type 1 diabetic patients (28), and apoptosis is the main mechanism by which the autoimmune system induces β-cell death in type 1 diabetes (29). Thus, IFN-β transgenic mice share many similarities to autoimmune diabetes in humans and may be considered a useful model of this disease.

The NOD mouse is probably the most widely used animal model of autoimmune diabetes. However, NOD mice present several limitations. The incidence of spontaneous diabetes is 60–80% in females and 20–30% in males (30,31). Furthermore, the incidence of disease is highest when mice are maintained in a germ-free environment but dramatically decreases when NOD mice are kept in conventional housing facilities (31–34). In addition, colony management and animal reproduction is difficult because of diabetes, and daily insulin administration is necessary for survival. Although ∼20% of 2-month-old IFN-β transgenic mice developed overt diabetes, the rest remained normoglycemic, showed normal reproduction and lifespan, and both male and female (data not shown) developed diabetes after treatment with very-low doses of STZ (15 mg/kg body wt). Thus, the use of these transgenic mice offers advantages, since studies in diabetic male or female mice can be easily programmed, depending on the experimental needs. Furthermore, when IFN-β transgenic mice are backcrossed with NOD mice, development of diabetes is clearly accelerated, and the expression of IFN-β breaks the tolerance to insulin-producing cells in NOR mice (13). Therefore, these features indicate that IFN-β transgenic mice may be an important tool for dissecting tolerance mechanisms and assaying new treatments for this disease.

We have previously shown (20) that expression of IGF-I in β-cells leads to recovery from type 1 diabetes by a mechanism involving β-cell replication and neogenesis. The present study also demonstrates that IGF-I counteracted/prevented lymphocytic infiltration of islets in double-transgenic mice expressing both IGF-I and IFN-β in β-cells. The expression of IFN-β led to hyperexpression of the β₂-microglobulin in islets, which indicated enhanced transcription of MHC class I. However, when IGF-I was coexpressed in β-cells of IFN-β transgenic mice, β₂-microglobulin expression was markedly decreased, which was parallel to a dramatic reduction of islet lymphocytic infiltration. This suggests that the high level of intracellular IGF-I suppressed expression of MHC class I resulting in a level of antigen presentation below the threshold necessary to induce immune recognition. In this regard, β₂-microglobulin–deficient NOD mice, which lack class I expression, not only fail to develop diabetes but are completely devoid of insulitis (35,36). However, transgenic mice that express the costimulatory molecule CD80 (mB7-1) in pancreatic β-cells display an extreme sensitivity to low doses of STZ, leading to complete β-cell destruction associated with islet-infiltrating CD4⁺ and CD8⁺ T-cells (37). Similar effects of IGF-I to downregulate MHC-I has also been shown in other cell types, such as thyroid and glioma cells (38–41). In the latter, a correlation between increased IGF-I expression and reduction in the expression of MHC class I and B7 antigens, as well as apoptosis, is observed (39–41).

IGF-I can block the apoptotic pathways in different cell types. Here, we found that, in contrast to IFN-β transgenic mice, apoptosis of β-cells was not detected in islets from double–IGF-I+IFN-β transgenic mice, suggesting that IGF-I production protected β-cells from death. This is also consistent with the fact that in vitro treatment of islets with IGF-I protects from cytokine-mediated β-cell apoptosis (21). Furthermore, we found that IGF-I protected islets from STZ-treated double-transgenic mice from Fas-induced apoptosis. Similarly, adenovirus-mediated expression of IGF-I in cultured human islets prevents IL-1β–induced Fas-mediated apoptosis (24). In addition, decreased IGF-I–mediated signaling in β-cells of mice with both reduced expression of IGF-I receptor and lack of IRS-2 led to increased β-cell apoptosis, reduced β-cell mass, and hyperglycemia (42). Furthermore, consistent with decreased β-cell death, double–IGF-I+IFN-β transgenic mice did not develop hyperglycemia when treated with very-low doses of STZ and showed serum and pancreatic insulin levels similar to those of healthy control mice. In addition, neither IGF-I nor IGF-I+IFN-β transgenic mice developed pancreatic tumors, and they had a normal lifespan.

In summary, this study demonstrates that expression of IGF-I, specifically in β-cells, protected mice from lymphocytic infiltration and apoptosis in vivo. Since IGF-I also induces β-cell replication and neogenesis (20), this growth factor may be considered a good candidate for gene therapy approaches to treat type 1 diabetes. In this regard, we and others have recently shown that mouse endocrine pancreas (43,44) and diabetic canine pancreas (45) can efficiently be manipulated to express genes of interest in vivo, by either systemic or local administration of viral vectors. Thus, progress in diabetes therapy focused in endocrine pancreas regeneration by transferring key “therapeutic genes” may be envisaged in the future.

A.S. and E.A. were recipients of predoctoral fellowships from Direcció General de Recerca, Generalitat de Catalunya and J.A. and V.J. from Ministerio de Educación, Cultura y Deporte (Spain). This work was supported by grants from Plan Nacional I+D+I (SAF2002-20389) from Instituto de Salud Carlos III (Red Grupos Diabetes Mellitus G03/212 and Red Centros Metabolismo y Nutrición C03/08) and from La Marató de TV3 Foundation (992710), Spain, and the European Community (BetaCellTherapy, FP6-2004-512145).

We thank M. Watford for helpful discussions and C.H. Ros for technical assistance.