| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Accelerates Autoimmune Type 1 Diabetes in Nonobese Diabetic Mice and Breaks the Tolerance to Cells in Nondiabetes-Prone Mice1
* Laboratory of Immunobiology for Research and Diagnostic Applications, Transfusion Center and Tissue Bank, Germans Trias i Pujol University Hospital, Barcelona, Spain; and
Biochemistry Department, Faculty of Veterinary, Autonomous University of Barcelona, Barcelona, Spain
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
cells develop diabetes. To determine the role of IFN in diabetes, we studied transgenic mice expressing human IFN in the cells. Autoimmune features were found: MHC class I islet hyperexpression, T and B cells infiltrating the islets and transfer of the disease by lymphocytes. Moreover, the expression of 2-microglobulin, preproinsulin, and glucagon in the thymus was not altered by IFN, thus suggesting that the disease is caused by a local effect of IFN, strong enough to break the peripheral tolerance to cells. This is the first report of the generation of NOD (a model of spontaneous autoimmune diabetes) and nonobese-resistant (its homologous resistant) transgenic mice expressing a type I IFN in the islets: transgenic NOD and nonobese-resistant mice developed accelerated autoimmune diabetes with a high incidence of the disease. These results indicate that the antiviral cytokine IFN breaks peripheral tolerance to cells, influences the insulitis progression and contributes to autoimmunity in diabetes and nondiabetes- prone mice. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
cells as a consequence of the complex interaction between genetic and environmental factors (1). Genetic susceptibility to the development of the disease has been described. At the present, more than 20 putative diabetes predisposing genes have been identified in the mouse and human genome, but only MHC genes are related to the disease (2, 3, 4). Studies on identical twins showed a concordance rate of diabetes lower than 50%, suggesting an important role for the environmental factors, e.g., viral infections (5). Viruses might be involved in the pathogenesis of the disease through exposure to sequestered Ags released by damaged cells, by altering some mechanisms of peripheral tolerance, by molecular mimicry, or by a direct destruction of the insulin-producing cells (6, 7).
Type I IFNs (IFN
, , 
, 
, 
, and 
) provide the first line of defense against viral infection but their mechanisms of action are only partially understood. Among other effects, type I IFNs increase the expression of MHC class I molecules and induce the activity of NK and CTLs and the production of cytokines. For years, an abnormal expression of type I IFNs has been reported in the sera of patients affected with autoimmune diseases (8). In previous studies, other scientists and we have detected type I IFNs in the pancreata of patients with a recent onset of diabetes (9, 10, 11). The expression of type I IFNs, IFN and IFN, by the cells induces autoimmune diabetes in nondiabetes-prone transgenic mice (12, 13). These mice developed insulitis and hypoinsulinemic diabetes with a cumulative incidence of 
50%. Transgenic mice expressing human IFN under the control of rat insulin promoter (RIP-HuIFN) were created (14) to avoid the sterility observed in mice expressing high levels of mouse IFN (15). HuIFN is active in mouse cells although its efficacy is 1000 times lower than that of mouse IFN (14). The transgenic mice in a C57BL6/SJL background showed features of a prediabetic state, with peri-insulitis but without spontaneous diabetes. When these mice were backcrossed to the outbred albino CD-1 mice, they spontaneously developed diabetes with an incidence of 11%. Since type I IFN induction takes place in cells infected mainly by viruses, this model suggests that the initial cell damage may induce the production of IFN by cells, thus triggering inflammation and cell-mediated autoimmunity.
Because IFNs act as both antiviral factors and inflammatory mediators, the aim of this study was to characterize the immunology of diabetes in the RIP-HuIFN transgenic mice and to evaluate the role of IFN in animal models with different susceptibility to diabetes: a spontaneous model of diabetes, the NOD mice (16) and a diabetes-free, MHC-matched, homologous strain, the Nonobese-resistant mice (NOR) (17). NOR/LtJ is an insulitis-resistant and diabetes-free strain where limited regions of the NOD/LtJ genome have been replaced by genome from the C57BLKS/J strain. In this study, we provide evidence that diabetes in RIP-HuIFN transgenic mice is autoimmune. The expression of this cytokine in cells accelerates diabetes in NOD mice and breaks the tolerance to insulin-producing cells in NOR mice. This type I IFN is a functional mediator for the development of autoimmunity to cells.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Transgenic mice expressing HuIFN under the control of the rat insulin I promoter (18) in outbred albino CD-1 background (CD-1 RIP-HuIFN) were generated by backcrossing the original C57BL6/SJL RIP-HuIFN transgenic mice to CD-1 mice (14). NOD mice, NOR mice, and NOD-SCID mice, unable to produce mature T and B lymphocytes (19), were obtained from The Jackson Laboratory (Bar Harbor, ME). Mice were kept in our facility under specific pathogen-free conditions in a 12-h light-dark cycle with free access to a standard diet. The Guidelines for the Use and Care of Laboratory Animals of the Generalitat de Catalunya were followed and animal studies have been approved by the Hospital Germans Trias i Pujol Review Board. To generate inbred NOD RIP-HuIFN and NOR RIP-HuIFN, the CD-1 RIP-HuIFN outbred mice were backcrossed at least six times onto the NOD or NOR strains. The acquisition of the genetic background was controlled by the analysis of microsatellites. NOD RIP-HuIFN mice were backcrossed twice onto the NOD-SCID strain to generate NOD-SCID RIP-HuIFN. Nontransgenic littermates were used as controls for each strain of mice.
Genotyping
The presence of the transgene was determined by standard PCR using genomic DNA obtained from mice tails and specific primers to HuIFN: HuIFN sense (5'-TCACCAGGGGAAAACTC-3') and HuIFN antisense (5'-CAGTCACTTAAACAGCATCT-3'). The PCR products were resolved by 2% agarose gel and ethidium bromide staining. The genotyping for the microsatellite markers linked to Idd loci (Idd1–Idd15) was performed using genomic DNA from the fourth generation of mice from the aforementioned NOD and NOR backcrosses (20, 21), with a set of 11 primers specific for polymorphic loci between NOD/CD-1 and NOR/CD-1: D6Mit52, D3Nds36, D2Mit107, D4Mit202, D1Mit24, D5Mit69, D11Mit320, D7Mit20, D7Nds6, D14Nds3, and D3Nds6. The chosen markers were evenly spaced along the genome. PCR products were electrophoresed in 4% agarose gel and stained with ethidium bromide. Homozygous mice for Idd NOD or NOR alleles (including Idd1 for H-2g7) were selected for the following backcross.
Insulin and glycemia assays
Blood glucose levels were measured weekly in CD-1 RIP-HuIFN mice and nontransgenic littermates between weeks 4 and 20 using Glucocard strips (Menarini, Barcelona, Spain). Mice with either successive blood glucose levels higher than 200 mg/dl or with a measure higher than 360 mg/dl were considered diabetic (n = 10 per condition). Starting at day 21, all CD-1, NOD, and NOR transgenic mice, after N6 backcross onto NOD or NOR background, were monitored daily for glycosuria using Chroma 1 Glucose test strips (Menarini). Insulin levels were determined in both CD-1 RIP-HuIFN mice and nontransgenic littermates. Nonfasting insulinemia was determined in sera and pancreatic insulin content was measured in total pancreas homogenized in HCl/ethanol by RIA (Linco Research, St. Charles, MO). The i.p. glucose tolerance test was performed in fasting conditions in control (NOR and NOD-SCID) and transgenic (NOR RIP-HuIFN and NOD-SCID RIP-HuIFN) adult mice. First, a blood sample was obtained from the tail vein to measure the basal level of glucose. Mice were subsequently given an i.p. injection of 2 mg of glucose per g of body weight. Blood glucose levels were measured at 15, 30, 60, 120, and 210 min after the injection.
Assessment of HuIFN in the sera of transgenic mice
The amount of HuIFN in the sera of CD-1 RIP-HuIFN transgenic mice was determined by ELISA (Fujirebio, Tokyo, Japan). The ELISA test showed no cross-reactivity with murine IFN (range of detection, 2.5–200 IU/ml).
Effect of HuIFN on the pancreas and the thymus expression profile
Real-time RT-PCR was performed to quantify the specific mRNA for preproinsulins I and II, glucagon and 2-microglobulin (2m). Total RNA was extracted from pancreas, thymus, and liver from CD-1 RIP-HuIFN using TRIzol reagent (Invitrogen Life Technologies, Gaithersburg, MD) according to the manufacturer’s instructions. RNA was then DNase treated (DNA-free kit; Ambion, Houston, TX). After denaturation, 3 µg of total RNA was reverse transcribed in a 20-µl volume using a Promega Reverse Transcriptase kit (Promega, Southampton, U.K.) for 1 h at 42°C in a solution containing 20 U of RNAsin inhibitor (Promega), 4 µl of 5x reverse transcriptase buffer, 1 mM dNTPs, 5 µM oligo(dT), and 100 U of Moloney murine leukemia virus. Real-time PCR was performed in a Light Cycler (Roche Diagnostic, Mannheim, Germany), using the FastStart DNA Master SYBR Green kit and specific primers (Table I). The results were analyzed using the LDCA software supplied with the machine. A PCR with -actin-specific primers was performed as control. A standard curve was produced for -actin, preproinsulins I and II, glucagon, HuIFN, and 2m with serial dilutions of the corresponding quantified cDNAs using a densitometric method (Quantity One, Huntington Station, NY). The real-time PCR was performed with an annealing temperature of 65°C. The results were expressed as the ratio between the level of relevant gene expression and the level of -actin expression.
|
Adoptive transfer experiments were performed transferring splenocytes or pancreatic-infiltrating cells from NOD RIP-HuIFN and NOR RIP-HuIFN diabetic mice into adult NOD-SCID mice. NOD-SCID mice were transferred only with splenocytes from NOR nontransgenic littermates because they are insulitis free. As positive control, splenocytes from wild-type NOD diabetic mice were injected in NOD-SCID mice. As negative control, NOD-SCID mice were sham transferred with physiological serum. Splenocytes or pancreatic-infiltrating cells from recent-onset diabetic NOD RIP-HuIFN mice were transferred into adult nondiabetic NOD RIP-HuIFN mice to assess the acceleration of the disease. The different transfer experiments were performed in a minimum of three recipients per condition (n = 3–6). To obtain splenocytes, a mechanically disrupted spleen was incubated in a hypotonic solution to remove erythrocytes. After washing, the cells were counted and injected into the tail vein. Pancreata from recent-onset diabetic animals were digested with collagenase P (Worthington Biochemical, Lakewood, NJ) and total digest was mechanically disrupted. Islet infiltrating lymphocytes were obtained by discontinuous density gradient using a lymphocyte isolation solution (Rafer, Barcelona, Spain). Lymphocytes were recovered from the interphase, washed twice, and transferred as above. In both cases, the cells to transfer were examined for viability by trypan blue exclusion and only the preparations with a viability >80% were transferred. Adult NOD-SCID males were injected i.v. with 2 x 107 splenocytes or 106 islet infiltrating cells resuspended in 200 µl of sterile physiological serum. After the transfer, the mice were monitored for assessment of diabetes every 2 days for 15 wk. Pancreata from mice that developed diabetes after the transfer were extracted and frozen for immunohistological analysis.
Autoantibodies
Five-micrometer mouse pancreatic cryosections were air dried and incubated with the sera from control (CD-1, NOD, and NOR nontransgenic littermates) and nondiabetic transgenic animals (CD-1 RIP-HuIFN, NOD RIP-HuIFN, and NOR RIP-HuIFN). The secondary Ab was a FITC-labeled goat anti-mouse Ig (Southern Biotechnology Associates, Birmingham, AL). All sections were blindly evaluated for the presence of autoantibodies to islet cells by two independent observers under an UV microscope and an image analyzer (OpenLab 2.0; Improvision, Coventry, U.K.).
Flow cytometric analysis
Splenocytes from transgenic mice (CD-1 RIP-HuIFN, NOD RIP-HuIFN, and NOR RIP-HuIFN) and nontransgenic littermates were isolated by mechanical disruption, and erythrocytes were lysed with a hypotonic solution. Cell clumps were allowed to sediment and the supernatant was centrifuged at 548 x g for 5 min. Aliquots of 105 spleen cells were stained with specific Abs (Table I) conjugated to FITC and PE for 30 min to detect T CD4 and CD8 subsets, respectively. A PE-labeled CD19 Ab was used to stain B cells. The controls included unstained cells (autofluorescence control) and cells stained with an irrelevant isotype-matched control PE/FITC. Dead cells were excluded by propidium iodide staining. The analysis was conducted in a FACScan Cell Analyzer (BD Biosciences, San Jose, CA) and the data were analyzed using CellQuest software (BD Biosciences).
Insulitis development
To determine the degree of islet infiltration, pancreata from transgenic mice in different genetic backgrounds were snap frozen in an isopentane/cold acetone bath and stored at –70°C. Five-micrometer cryostat sections were obtained at five nonoverlapping levels. The sections were stained with H&E. Groups of mice (n = 6) were analyzed at different ages (6, 9, and 12 wk), assessing 40–100 islets per animal. Insulitis was scored on a 0–4 scale as described elsewhere (22).
Immunohistological analysis
Consecutive pancreatic cryostat sections (5 µm) from different mice were air dried as described above. To block nonspecific binding, 2% FCS was added to the PBS used to dilute the Abs. The sections were sequentially incubated with 1) Ab to specific marker (Table I), 2) FITC-labeled goat anti-rat IgGs (Southern Biotechnology Associates) or FITC-labeled rabbit anti-sheep FITC (Zymed Laboratories, San Francisco, CA), 3) guinea pig anti-insulin, and 4) tetramethylrhodamine isothiocyanate- labeled goat anti-guinea pig Ab (Biogenesis, Eschwege, Germany). The preparations were assessed with a fluorescence UV microscope and an image analyzer (OpenLab 2.0; Improvision).
Statistical analysis
Statistical analyses to compare independent groups were performed using the t test when groups passed normality and showed equal variance tests. When these tests failed, Mann-Whitney U test was performed. Differences were considered significant when a value of p < 0.05 was reached.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
cells in transgenic mice
To discard a reduced capacity to produce insulin by the cells of transgenic mice, the glycemia, the insulinemia, and the pancreatic insulin content were determined in healthy CD-1 transgenic mice. No significant differences in glycemia were observed between CD-1 RIP-HuIFN mice and nontransgenic littermates at weeks 4–20 (Fig. 1a), except in animals progressing toward diabetes. The blood glucose concentration was 116.4 ± 6.4 mg/dl (mean ± SEM) in transgenic animals (133.2 ± 2.1, male and 96.7 ± 16, female, respectively) and 106.7 ± 8.7 mg/dl in the control group (120.6 ± 2.4, male and 94.4 ± 1.9, female, respectively). Two weeks before the clinical onset of the disease, diabetic-prone transgenic animals showed progressively increasing blood glucose levels. To rule out the possibility that cell destruction was caused by the deleterious effects of HuIFN, we backcrossed NOD RIP-HuIFN mice onto the NOD-SCID strain twice. No significant differences were observed in glycemia between NOD-SCID RIP-HuIFN mice (119.4 ± 2.4 mg/dl, mean ± SEM) and nontransgenic littermates (123.9 ± 3.7 mg/dl). In addition, to assess whether the transgene could influence cell metabolism, we tested the response of transgenic (NOD-SCID and NOR) and control animals to an acute increase of blood glucose levels (Fig. 1b). Glycemia of transgenic and control animals reached the highest value 15 min after glucose administration and decreased to basal values at 120 min in all groups. No significant differences were found between transgenic and control mice at any time, demonstrating that insulin production and secretion was unaffected by the transgene.
|
(156.9 ± 13.7 µg of insulin/g of pancreas, mean ± SEM) when compared with that of the controls (162.2 ± 23.8 µg of insulin/g of pancreas) except in recent-onset diabetic subjects (p < 0.05; Fig. 1c). Insulinemia was not altered in CD-1 transgenic animals (1.9 ± 0.3 ng/ml) compared with controls (1.3 ± 0.2 ng/ml) and, as expected, diabetic mice showed a significant decrease in the insulin levels in serum (0.3 ± 0.04 ng/ml, p < 0.05; Fig. 1d). In addition, no significant differences were observed in the levels of preproinsulin or glucagon RNA when compared to transgenic mice and nontransgenic littermates (see below). Apoptotic cells (TUNEL) were not detected in the noninfiltrated islets of healthy mice (data not shown). All of these data suggest that the development of diabetes is not a consequence of cell dysfunction caused by the effect of the transgene or its product.
HuIFN affects the islets but not the thymic environment
HuIFN was detected in the sera from CD-1 RIP-HuIFN transgenic animals. The amount of this cytokine in sera was 49.2 ± 12.7 IU/ml in transgenic animals, statistically different from 5.6 ± 1.6 IU/ml in nontransgenic littermates (p < 0.01). These data indicate a systemic distribution of HuIFN.
To assess the effects of circulating HuIFN on the molecular pattern expression in the thymus (primary lymphoid organ) and in the liver (control tissue), we determined the mRNA levels for preproinsulins I and II, glucagon, and 2m in CD-1 RIP-HuIFN transgenic mice. In addition to the presence of protein (HuIFN) in the sera, we confirmed the transcription of HuIFN mRNA in the pancreas and the thymus of transgenic mice (Table II). HuIFN did not alter the level of expression of preproinsulin I, which was only detected in the pancreas but not in the thymus or in the liver. However, preproinsulin II was detected in the pancreas, thymus, and liver but the differences between transgenic animals and controls were not significant, thus confirming that the transgene did not affect the transcription of insulin. Glucagon was detected in the pancreas and the thymus but not in the liver, and no significant differences were found when comparing transgenic and control animals. The amount of 2m mRNA was higher in the pancreata of transgenic mice than in the control group (p < 0.05), showing an enhanced transcription of MHC class I, due to the effect of HuIFN (these data were confirmed by immunofluorescence staining of the islets with Abs to MHC class I). By contrast, no significant differences were found in the levels of 2m in the thymus of transgenic and control mice. These data confirmed that the islets of Langerhans of transgenic animals have a normal expression profile of insulin I, insulin II, and glucagon and an increased MHC class I hyperexpression caused by the local release of HuIFN. Moreover, the expression of 2m, preproinsulin, and glucagon in the thymus of transgenic mice was unaffected by the expression of the transgene. These results suggests that, in terms of T cell development, no differences exist between transgenic and wild-type mice.
|
-transgenic mice
The incidence of the disease varies according to the genetic background (Fig. 2). In CD-1 RIP-HuIFN mice, 22% of males and 0% of females developed diabetes. The onset of the disease starts after 7 wk of age (Fig. 2a). The analysis of the progeny resulting from F6 backcross with NOR strain onward showed not only a spontaneous development of diabetes but also a significant acceleration of the onset of the disease and an increase of the incidence: Fifty-seven percent of male and 22% of female NOR RIP-HuIFN mice became diabetic. The onset of diabetes started at 3 wk of age (Fig. 2b). Mice resulting from F6 backcross with NOD strain (NOD RIP-HuIFN) onward showed a diabetes incidence of 53% in male and 51% in female mice, starting diabetes at 3 wk of age. The incidence of diabetes in NOD nontransgenic littermates was identical to that in wild-type NOD: 27% in males and 66% in females (Fig. 2c). As expected, neither CD-1 nor NOR nontransgenic mice developed diabetes during the study. Moreover, NOD-SCID RIP-HuIFN mice did not become diabetic.
|
mice that started diabetes at 7 wk of age. These data, along with the variation in the incidence of the disease in the three groups of transgenic mice, suggest that the genetic background influences the start point and the percentage of diabetes in mice expressing this transgene. The development of diabetes is lymphocyte dependent: evidence of autoimmunity
NOD-SCID RIP-HuIFN mice showed no signs of diabetes nor insulitis, thus indicating that the destruction of islet cells in transgenic mice is not caused by the presence of HuIFN per se and that it is essential they share immunocompetent T and B cells.
All NOD-SCID mice (four of four) transferred with splenocytes from diabetic nontransgenic NOD mice (positive control), become diabetic after 4–9 wk, as expected (23). All NOD-SCID mice (five of five) transferred with splenocytes or islets infiltrating lymphocytes from diabetic NOD RIP-HuIFN mice developed type 1 diabetes after 4–15 wk after transfer. NOR RIP-HuIFN mice were also able to transfer diabetes in most (three of four) treated NOD-SCID mice at 8–9 wk, starting 4 wk later than littermates that received cells from diabetic NOD RIP-HuIFN. NOD-SCID mice transferred with splenocytes from NOR nontransgenic littermates were insulitis free 3 wk after the injection and at the end of the follow up (15 wk) none of the animals (zero of six) developed diabetes. All NOD RIP-HuIFN mice (three of three) transferred with autologous splenocytes or pancreatic-infiltrating lymphocytes from diabetic mice developed diabetes in 3–6 days. As expected, sham-transplanted mice did not develop diabetes nor insulitis. Histological examination of pancreata from transferred diabetic mice showed a high degree of insulitis, confirming that development of the disease is lymphocyte dependent.
Islet cell Abs (ICA) are present in transgenic mice
ICA were detected in mice from all of the study groups, including diabetic and healthy animals: CD-1 RIP-HuIFN, NOD RIP-HuIFN, NOR RIP-HuIFN, and also their nontransgenic littermates. The expression of the transgene did not result in a disappearance of or significant increase of ICA. The ICA-positive staining pattern was weak. Occasionally, ICA were associated with antinuclear autoantibodies in endocrine and exocrine tissues. However, the autoantibodies to islet Ags are not a prerequisite for the subsequent emergence of the disease (24).
Peripheral lymphocyte subsets in transgenic mice
It has been described that type I IFNs have an important effect on the proliferation of different cell types, including lymphocyte subsets (25). To assess the effect of circulating HuIFN in lymphocyte subpopulations, we determined the percentages of splenic lymphocyte subsets in transgenic mice. Mice expressing the transgene in different genetic backgrounds (CD-1, NOR, and NOD) had normal percentages (26) of splenic B and T (CD4 and CD8) lymphocyte subsets (Fig. 3).
|
in islet cells increases the intensity of insulitis in different genetic backgrounds
The sequential histological examination of the pancreata of nondiabetic RIP-HuIFN transgenic mice reveals that the intensity of the insulitis depends on the genetic background. CD-1 RIP-HuIFN mice showed normal islets or just weak insulitis at 3 wk of age. Ten percent of the islets of 4-wk-old CD-1 RIP-HuIFN mice were weakly infiltrated with mononuclear cells around or inside the islets (peri-insulitis or mild insulitis); this pattern was maintained in adult mice in which 15% of the islets were weakly infiltrated (Table III). The insulitis score (Fig. 4a) was very low due to the lack of moderately and severely infiltrated islets and to the absence of infiltrated islets in some subjects. Only at 12 wk of age the intensity of the insulitis was significantly higher than that of controls (p < 0.05). As expected, the inflammation was more marked and the insulitis score significantly increased in transgenic mice developing diabetes (Fig. 4c).
|
|
mice showed severe infiltration compared with wild-type NOR mice and nontransgenic littermates (p < 0.01). In contrast to the pancreata from CD-1 RIP-HuIFN nondiabetic mice, in which most islets were not infiltrated (Table III), NOR transgenic mice showed mononuclear cells infiltrating most islets at 3 wk of age. Adult mice showed 86% of the islets infiltrated (43% moderately or severely). The insulitis score (Fig. 4b) was significantly higher (p < 0.01) in NOR RIP-HuIFN than in CD-1 transgenic mice. Both diabetic NOD and NOR RIP-HuIFN mice showed an insulitis score higher than transgenic CD-1 diabetic subjects (p < 0.05) and similar between them (Fig. 4c). A weak insulitis is enough to cause the disease in the CD-1 transgenic colony, whereas NOD and NOR transgenic strains require massive infiltration.
HuIFN in the islets causes MHC class I hyperexpression and recruits lymphomononuclear cells
As expected, HuIFN staining was only positive in cells of transgenic animals. The islets from CD-1 RIP-HuIFN healthy mice were predominantly noninfiltrated, peri-infiltrated, or poorly infiltrated. The phenotypic characterization of insulitis (Fig. 5b) showed that most mononuclear islet-infiltrating cells were CD4 and CD8 T lymphocytes (ratio 3:1); B cells were only occasionally detected. A few macrophages were also detected surrounding the endocrine islet cells. Dendritic cells were only observed inside the noninfiltrated islets. MHC class I molecules were hyperexpressed in the endocrine cells of all islets, correlating well with molecular data; no MHC class II hyperexpression was observed in islet cells. Diabetic mice showed normal, peri-infiltrated, and mildly infiltrated islets and a stronger infiltration (peri and poorly) than that observed in nondiabetic mice was detected (Fig. 5a). The composition of the insulitis in diabetic mice was almost the same as that in healthy animals except for a remarkable increase of B cell clusters in the periphery of the islets.
|
nondiabetic mice were moderately or severely infiltrated by CD18-positive cells (leukocytes). As expected, MHC class I, but not MHC class II, was highly hyperexpressed in endocrine islet cells from all of the islets. The insulitis mainly consisted of T and B cells (Fig. 5a). We observed B cells, T cells, more CD4 than CD8 cells (ratio 4:1), and few NK cells. B and CD4 cells, but not CD8 cells or NK cells, were distributed in clusters around the islets. These data were confirmed by flow cytometry of pancreatic-infiltrating mononuclear cells (data not shown). Only a few macrophages were observed in the periphery or inside the islets; no dendritic cells were found in the infiltrated islets. The islets from NOR RIP-HuIFN diabetic mice (Fig. 5b) were almost devoid of cells and showed a higher intensity of insulitis than healthy mice, but no differences in the composition of the infiltrate were observed. The islets from NOD RIP-HuIFN diabetic mice were almost devoid of cells and showed a high intensity of insulitis, but no differences in the infiltrate composition were observed when compared with NOR RIP-HuIFN diabetic mice.
Nontransgenic littermates did not show infiltrating leukocytes or MHC class I hyperexpression (Fig. 5c). Histological examination of pancreata from NOD-SCID RIP-HuIFN revealed that MHC class I was hyperexpressed in the islets due to the effect of HuIFN but, as expected, immunofluorescence staining revealed the lack of insulitis because SCID mice are unable to produce mature T and B cells (Fig. 5c).
Transgenic mice in the three analyzed genetic backgrounds showed a HuIFN staining pattern that correlates well with the MHC class I hyperexpression in the islets and it is independent of the presence of infiltrating cells.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
, , and (12, 13, 14) in cells developed type 1 diabetes, with T and B cells infiltrating the islets. These data suggest a local proinflammatory role for these type I IFNs and their ability to induce T1D.
In this study, we provide evidence that the local expression of HuIFN is involved in the development of autoimmune diabetes: 1) the islets from transgenic mice hyperexpress MHC class I molecules, 2) the islets are infiltrated by T and B lymphocytes, 3) the transfer of lymphocytes from a diabetic transgenic animal causes the disease in NOD-SCID recipient mice and accelerates the onset of diabetes in prediabetic recipient mice, and 4) transgenic NOD-SCID RIP-HuIFN mice do not develop diabetes. Moreover, systemic HuIFN does not alter the expression of analyzed pancreatic hormones or 2m in the thymus, thus supporting the idea that the disease could be caused by a local effect of IFN strong enough to break the peripheral tolerance to cells rather than by a change in central tolerance. Preliminary results show that HuIFN could modulate the expression of adhesion molecules, cytokines, and chemokines in the islets (data not shown). Therefore, we conclude that in this transgenic model, the disease is lymphocyte dependent since transgenic animals lacking functional T and B cells cannot develop diabetes or insulitis, although they increase the expression of MHC class I in the islets.
To study the role of IFN in diabetes, we also introduced the RIP-HuIFN transgene in different genetic backgrounds with different susceptibility to T1D. We demonstrated that the genetic background not only determines the onset and the severity of the lesions related to the disease but also that these genetic factors cannot protect individuals from T1D by themselves when an inducing factor triggers the autoimmune process.
It has been previously described that a small percentage of N3 generation obtained from the backcross of C57BL6/SJL RIP-HuIFN to CD-1 mice spontaneously developed T1D (14). In this study, we showed an increase of the incidence of diabetes in homozygous mice kept under specific pathogen-free conditions. The research in the field of T1D has developed many transgenic mice in nondiabetes-prone strains that were later backcrossed to diabetes-prone NOD mice. This is the first report of the generation of NOD and NOR transgenic mice expressing a type I IFN in the islets. NOD mice expressing HuIFN in cells showed an acceleration of the disease and, unexpectedly, NOR RIP-HuIFN mice developed early diabetes with a high incidence. The severity of the insulitis varied according to the genetic background and was correlated with the incidence of diabetes. The three models shared a MHC class I hyperexpression in the islet cells, but whereasdiabetic NOD and NOR transgenic mice islets were strongly infiltrated, diabetic CD-1 RIP-HuIFN mice were only mildly infiltrated. These results indicated that the genetic background was not strong enough to protect mice from diabetes when an injury triggered an inflammatory response in the islets. In humans, the lack of concordance of diabetes in identical twins indicates that nongenetic factors determine the development of the disease. These factors might be environmental agents but also random factors such as the generation of the immunological repertoire are subject to complex regulation.
The influence of HuIFN was always stronger in males than in females. It has been described that male mice have more difficulty maintaining glucose homeostasis under conditions of impaired cell function and that male islets are more susceptible to an autoimmune damage than are female islets but male islet-specific lymphocytes are less aggressive (31). Results similar to ours have been recently described in nondiabetes-prone transgenic mice expressing IFN in cells (13), where a higher incidence of the disease in males than in females was observed (G. Vassileva, personal communication). Assuming a viral etiology for T1D in humans, females seem to be less susceptible than males to the environmental infectious influences (32).
To date, 14 different viruses have been reported to be associated with the development of autoimmune diabetes in both humans and animal models (5). The most convincing data include an increased frequency of diabetes in patients with enterovirus infection (33), congenital rubella syndrome, measles (34), and the detection of CMV in the lymphocytes of diabetic patients (35). Viral infections result in an enhancement of the MHC class I expression in a number of cell types which normally do not express these molecules (36). This enhancement may occur by a direct interaction between the viral component and the MHC class I gene or indirectly by virus-induced soluble factors, IFN-
, IFN-, and TNF-- produced by the infected cells (37). Thus, these viruses induced an enhanced display of self-peptides in MHC class I, which led to the temporary activation of autoreactive T cells and autoimmunity (38).
Nevertheless, it is very difficult to connect viruses and T1D: it is well known that development of the disease in humans and animal models is reduced in the presence of high amounts of pathogens. Recently it has been demonstrated that viral infections could influence the progression of insulitis beneficially at the preclinical stage if produced at the correct location, time, and intensity, preventing autoimmune diabetes (39), correlating well with some protocols that prevent diabetes in NOD mice by administration of antiviral cytokines in mice (40, 41). The timing of the infection, the viral strain, the antigenic load, and other unknown factors must be crucial in the protection or predisposition to the disease.
The results presented here, along with previous and future studies, will provide a greater understanding on how the production of type I IFNs in the pancreas affects the recruitment of infiltrating cells and accelerates the onset of diabetes. The genetic background by itself does not protect mice from diabetes when damages, stresses, or dangers affect cells with intensity strong enough to trigger an inflammatory response. Our data indicate that the antiviral cytokine IFN expressed in insulin-producing cells triggers the break of peripheral tolerance to cells, leading to insulitis and autoimmunity in diabetes and nondiabetes prone mice.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
1 This work was supported by grants from the Fondo de Investigaciones Sanitarias (Projects 99/1066 and 02/0107 to M.V.-P) and Fundació La Marató de TV3 (Project 99/1810 to M.V.-P.), and Ministry of Science and Technology (Project SAF 2000-050), CICYT to R.P.-B. A.A. was supported by a fellowship from the Instituto de Salud Carlos III (BEFI 01/9065) and J.V. and M.V.-P. are associate researchers supported by Fondo de Investigacion Sanitaria, Spanish Ministry of Health. M.-C.P. was supported by a Comissio Interdepartamental de Recerca I Innovacio Tecnologica Predoctoral Fellowship (2001FI0002) Generalitat de Catalunya, Barcelona, Spain. 
2 Address correspondence and reprint requests to Dr. Marta Vives-Pi, Laboratory of Immunobiology for Research and Diagnostic Applications, Transfusion Centre and Tissue Bank, Germans Trias i Pujol University Hospital, 08916 Badalona, Barcelona, Spain. E-mail address: vivespi{at}ns.hugtip.scs.es
3 Abbreviations used in this paper: T1D, type 1 diabetes; RIP, rat insulin promoter; HuIFN, human IFN; NOR, nonobese resistant; 2m, 2-microglobulin; ICA, islet cell Ab.
Received for publication March 15, 2004. Accepted for publication September 24, 2004.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
cells activates autoreactive T cells in virus-mediated autoimmune diabetes. J. Clin. Invest. 109:79.[Medline]
-cell damage and immune infiltration. J. Autoimmun. 16:211.[Medline]
interferon in insulin secreting cells in type 1 diabetes mellitus. Lancet 2:1423.[Medline]
usage, and cytokine profile. J. Immunol. 153:1360.[Abstract]
in transgenic mice. Science 260:1942. may be involved in the onset of diabetes mellitus. J. Biol. Chem. 20:12332.
under the control of metallothionein enhancer-promoter. EMBO J. 7:3757.[Medline]
, IFN- and ligands. S. K. Durum, and T. Hirano, and J. Vilcek, and N. A. Nicol, eds. In Cytokine Reference Vol. 1:627. Academic, Cornwall, U.K..
receptor signaling is dispensable in the development of autoimmune type 1 diabetes in NOD mice. Diabetes 49:2007.. Intern. Med. 39:146.[Medline]
therapy for chronic hepatitis C. Gut 48:378. mediates enhanced HLA-class I expression in measles virus-infected cells. J. Immunol. 151:6248.[Abstract]
This article has been cited by other articles:
|
|
 |
|
  P. A. Lang, L. Cervantes-Barragan, A. Verschoor, A. A. Navarini, M. Recher, M. Pellegrini, L. Flatz, A. Bergthaler, K. Honda, B. Ludewig, et al. Hematopoietic cell-derived interferon controls viral replication and virus-induced disease Blood, January 29, 2009; 113(5): 1045 - 1052. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 X. Huang, D. J. Moore, R. J. Ketchum, C. S. Nunemaker, B. Kovatchev, A. L. McCall, and K. L. Brayman Resolving the Conundrum of Islet Transplantation by Linking Metabolic Dysregulation, Inflammation, and Immune Regulation Endocr. Rev., August 1, 2008; 29(5): 603 - 630. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Wang, L. Chen, E. Ahmed, L. Ma, D. Yin, P. Zhou, J. Shen, H. Xu, C.-R. Wang, M.-L. Alegre, et al. Prevention of Allograft Tolerance by Bacterial Infection with Listeria monocytogenes J. Immunol., May 1, 2008; 180(9): 5991 - 5999. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Casellas, A. Salavert, J. Agudo, E. Ayuso, V. Jimenez, M. Moya, S. Munoz, S. Franckhauser, and F. Bosch Expression of IGF-I in Pancreatic Islets Prevents Lymphocytic Infiltration and Protects Mice From Type 1 Diabetes Diabetes, December 1, 2006; 55(12): 3246 - 3255. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Rasschaert, L. Ladriere, M. Urbain, Z. Dogusan, B. Katabua, S. Sato, S. Akira, C. Gysemans, C. Mathieu, and D. L. Eizirik Toll-like Receptor 3 and STAT-1 Contribute to Double-stranded RNA+ Interferon-{gamma}-induced Apoptosis in Primary Pancreatic {beta}-Cells J. Biol. Chem., October 7, 2005; 280(40): 33984 - 33991. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
