Interleukin-21 Is Critically Required in Autoimmune and Allogeneic Responses to Islet Tissue in Murine Models Open Access

NOD Ltj, C57BL/6, and BALB/c mice were obtained from Animal Resources Centre, Perth. NOD/Scid mice were obtained from Walter and Eliza Hall Institute of Medical Research, Melbourne. The Il21^−/− mice were created through a National Institutes of Health initiative with Lexicon and Deltagen, on the 129 background and backcrossed to C57BL/6. These mice were backcrossed to NOD N10 and selected for known NOD Idd loci (vs. B6 regions) by PCR of genomic DNA (speed congenics), as previously described (32). Il21r^−/− mice (33) were kindly provided by M. Smyth (Melbourne) at C57BL/6 N6 and backcrossed to N7 for experimental use; both Il21r^−/− and Il21^−/− NOD mice underwent an additional genome scan of +600 NOD microsatellite markers. Major histocompatibility complex (MHC) class II–deficient mice (stock 3584, Jackson Laboratories) were kindly provided by R. Brink (Sydney). Animals were housed under specific pathogen-free conditions and handled in accordance with the Garvan Institute of Medical Research and St. Vincent Hospital’s Animal Experimentation and Ethics Committee. Blood glucose levels were determined using Accu-chek Advantage blood glucose strips (Roche).

IL-21R/Fc chimera was generated in-house, as previously described (34). In brief, the DNA encoding the predicted extracellular domain (aa 1–235) of mouse IL-21R with a GSGS linker were amplified by PCR, and linked to mIgG2a Fc. The Fc domain contains four amino acid mutations (L285E, E368A, K370A, and K372A) to minimize Fc binding and complement fixation (34). The resulting construct was subcloned into pEE12.4 (Lonza), a mammalian glutamine synthase expression vector, transfected into Chinese hamster ovary-K1SV cells, and grown in the presence of 25 μmol/L methionine sulphoximine. Following protein A purification, purity of IL-21R/Fc was checked by SDS/PAGE (Coomassie blue staining and silver stain, Lonza) (Supplementary Fig. 1) and tested for presence of endotoxin (Lonza). In vitro BAF-3 proliferation assay measuring biological activity of IL-21R/Fc construct was conducted as previously described (35) (Supplementary Fig. 1).

Complete treatment of NOD mice consisted of a total of 20 μg IL-21R/Fc given every other day for 11 days. Flow cytometric analyses were conducted at time points relative to this treatment regimen. Treatment of NOD transplant recipients consisted of a total of 80 μg IL-21R/Fc given on day −1, day 0, and every other day until day 12. A monoclonal Ig control was provided with a nonspecific antibody (affinity purified, produced in house). Mice received 60 μg polyclonal IL-21 antibody (goat IgG; R&D Systems) as an alternative IL-21 neutralizing reagent. The frequency of diabetes was compared between treatment groups with the Kaplan-Meier log-rank test using Prism software (GraphPad).

Paraffin-embedded sections of pancreas and engrafted kidney were conventionally stained with hematoxylin and eosin or insulin by immunocytochemical techniques. For insulitis scoring, serial sections of pancreata were matched and stained for insulin (DakoCytomation) and glucagon (Chemicon). For islet mass scoring, sections of pancreata were stained for insulin. Images were processed using the Leica acquisition and analysis software ImageJ (Freeware; National Institutes of Health, Bethesda, MD).

We stained lymphocytes extracted from the pancreas and pancreatic lymph nodes with conjugated antibodies to Annexin V, B220, CD4, CD8, CD45 (BD Biosciences), and CD11b, CD11c, CD44, CD69, Foxp3, and IL-17 (eBioscience). For flow-cytometric analysis of lymphocytes isolated from the pancreas, mice were killed by intraperitoneal injection of ketamine and perfused with ice-cold PBS for 10 min. Pancreas samples were then subjected to lymphocyte isolation as described (36). For IL-21 analysis, after ex vivo staining of surface markers, cells were fixed and permeated (BD Biosciences), followed by intracellular staining with biotin labeled polyclonal IL-21 Ab (R&D Systems). For IL-17 analysis, cells were stimulated with CD3 (2.5 μg/mL) and CD28 mAb (2.5 μg/mL) in culture medium containing monensin (GolgiStop 1 μL/mL; BD Biosciences) at 37°C for 6 h. Intracellular staining was performed according to the manufacturer’s instructions, for IL-17 (BD Biosciences) and Foxp3 (eBioscience). Data were collected on a Canto flow cytometer (BD Biosciences), and analyzed using FlowJo software (Tree Star).

Total RNA was harvested from tissues or cells using TRIzol (Invitrogen). Pancreas samples were perfused through the bile duct with RNAlater (Ambion). Tail vein bleeds were collected into RNAlater. Preparation of complementary DNA and analysis of IL-21 gene expression was performed with Taqman gene expression assays (Applied Biosystems) as described previously, with values normalized to the level of expression of the housekeeping gene GAPDH (37).

Autoimmune diabetic NOD mice received islets prepared from NOD/Scid mice (for syngeneic transplants), and chemically induced diabetic mice (C57BL/6 and Il21r^−/− mice, H-2^b), injected with 180 mg/kg streptozotocin (Sigma-Aldrich), received islets prepared from BALB/c (H-2^d) (for allogeneic transplants). Islets were prepared as described previously (38).

We used an F test to compare variances, and when variances were significantly different t tests were performed with Welch correction (without assuming equal variances). P values between datasets were determined by two-tailed Student t test assuming equal variance, with log-rank test used in Figs. 1B, D, and G, 4A, 5A, and 6A. Data are reported as the mean ± SEM, along with the calculated P values.

We have shown previously that NOD mice express higher levels of IL-21 compared with diabetes-resistant strains (37,39). NOD mice made genetically deficient for IL-21R lack insulitis and do not develop diabetes (10–12). We used an IL-21R/Fc chimeric protein to therapeutically neutralize IL-21 at different periods of time in the disease process. IL-21 was neutralized from 7 to 9 weeks of age in one group, and IL-21 was neutralized from 14 to 16 weeks of age in another group. We also compared groups treated with IL-21R/Fc with NOD mice made genetically deficient in IL-21 that do not develop insulitis (Fig. 1A) or diabetes during the 40 week observation period (Fig. 1B).

Histological analysis of pancreata 5 weeks after treatment of 7-week-old mice revealed islets with mild peri-insulitis, but very few islets exhibited insulitis compared with control mice (Fig. 1C). However, this early short-term treatment was reversible and ultimately had little impact on diabetes incidence (Fig. 1D). The 11 day neutralization of IL-21 in mice treated on the brink of clinical diabetes (14-weeks-of-age) also reduced insulitis in the pancreatic islets compared with control mice (Fig. 1E). Because NOD islets at the start of treatment exhibited a considerably greater degree of insulitis than those at the end of treatment, our results indicate that insulitis was reversed whether we treated with the IL-21R/Fc chimera or an IL-21 neutralizing polyclonal antibody (Fig. 1F). Furthermore, in contrast to the 7-week-old group, the effect of the 11 day treatment was not reversible in mice treated at a late preclinical stage, delaying the onset and reducing the incidence of diabetes to 40%, compared with 90% of control mice by 40 weeks of age (Fig. 1G). These findings indicate that IL-21 was required for the transition from peri-insulitis to insulitis.

To further examine the reduction of inflammation caused by neutralization of IL-21, we investigated the effect of IL-21 neutralization on the islet infiltrate from the start of treatment (day 0) and up to 12-weeks after treatment (day 90). One of the early effects of treatment (day 5) was to reduce the percentages and total numbers of lymphocytes in the islet infiltrate comprising CD4+ T-cells, CD8+ T-cells, and B220+ B-cells (Fig. 2A and B, and Supplementary Fig. 2). Analyses of the long-term effects of IL-21 neutralization showed the sustained reduction in CD8+ T-cells and B-cells, but not CD4+ T-cells, 90 days after treatment (Fig. 2A and B, and Supplementary Fig. 2). The percentage of macrophages and dendritic cells were proportionally increased, however, absolute numbers of both subsets remained largely unchanged (Fig. 2C and D, and Supplementary Fig. 2).

Analyses of cell surface molecules that distinguish activated T-cells from naïve T-cells indicated that IL-21 neutralization preferentially reduced the number of activated lymphocytes in the islet infiltrate (Table 1). The reduction of CD44^hi CD8+ T-cells preceded that of CD4+ T-cells and B-cells (data not shown) and was also sustained until 90 days after treatment (Fig. 2E and F, and Supplementary Fig. 2).

IL-21 neutralization resulted in a profound reduction in IL-21 expression (Fig. 2G) and IL-21–producing cells (Fig. 2H and I) in the islets. The population of IL-21+ cells had significantly declined after treatment and remained decreased in mice that had not developed diabetes by 40 weeks of age (Fig. 2I). By contrast, mice that did go on to develop diabetes were distinguished from those who remained protected from disease by an elevated fraction of IL-21–producing CD4+ T-cells in the islet infiltrate (Fig. 2I). Measuring IL-21 transcript in secondary lymphoid organs, blood, and the pancreas over the course of disease in the NOD mouse demonstrated that an increase in IL-21 in the blood and pancreas corresponded with the age of heightened destruction of the islets and coincided with the onset of diabetes in our colony (Supplementary Fig. 3). Moreover, the analyses of IL-21 mRNA in blood demonstrated that high IL-21 expression was predictive of the subsequent development of clinical diabetes (Fig. 2J).

The generation and survival of several T_H subsets, including T follicular helper cells (40), T_H17 cells and T_H2 cells (20,21), are influenced by IL-21. The analyses of pancreata from NOD mice also revealed a small fraction of T_H17 cells that were significantly diminished during IL-21R/Fc treatment (Fig. 3A and B). However, a greater than 50% reduction in the fraction of T_H17 cells early in the treatment regimen was consistent with an overall loss of activated CD4+ T-cells rather than a specific effect on T_H17 cells. By contrast, there was no specific decline or retention of Foxp3+ regulatory T-cells in either the pancreatic lymph nodes or pancreas (Fig. 3C and D).

These data show the merit of neutralizing IL-21 at a late preclinical stage. To be of relevance to human patients with type 1 diabetes, however, it would be more beneficial to reverse diabetes once it is clinically revealed, yet treatment of newly diabetic NOD mice (15–20 weeks of age) with IL-21R/Fc failed to reverse diabetes (data not shown).

Given the slow regeneration of the depleted β-cell mass (41), we assessed whether IL-21 neutralization combined with pancreatic islet transplantation could be a solution. Newly diabetic NOD mice (with two consecutive blood glucose readings above 18 mmol/L) were treated with IL-21R/Fc (10 mg/IV) on day −1, day 0, and every other day until day 12. Pancreatic islets from MHC-matched, lymphocyte deficient NOD/Scid mice were grafted under the kidney capsule of NOD mice (day 0). In mice treated with IL-21R/Fc, destruction of the graft and subsequent diabetes was delayed and reduced (Fig. 4A). The majority of treated mice maintained a stable normal range of blood glucose while undergoing treatment (Fig. 4B) with three of seven mice returning to hyperglycemia, compared with the entire untreated group with a mean survival time of 16.1 days.

To determine whether the graft was still necessary in long-term (100 days) diabetes-free IL-21R/Fc–treated NOD mice, islet grafted kidneys were removed in these mice and blood glucose was measured. No considerable modulation of blood glucose level was observed after the nephrectomy, and mice remained diabetes free for another 100 days (Fig. 4C). This finding raised the possibility that IL-21R/Fc treatment had affected the residual endogenous islets, either by increasing islet mass or restoring function. Assessment of islet mass from histological sections of pancreata immunostained for insulin indicated that IL-21R/Fc–treated NOD mice had equivalent islet mass to that of 15–20-week-old diabetic NOD mice, which were both significantly reduced compared with prediabetic NOD mice (Fig. 4D). Taken together these data indicated that NOD pancreatic islets had regained functionality through the combined effects of IL-21 neutralization and transplantation of syngeneic islets and were capable of maintaining normoglycemia without the islet graft.

Histological analysis of islet grafts from IL-21R/Fc–treated long-term survivors revealed healthy insulin producing grafts free from mononuclear cell infiltration (Fig. 4E). Furthermore, the pancreas revealed islet mass with mild peri-insulitis (Fig. 4F). These findings showed that IL-21R/Fc treatment was capable of inducing tolerance. Furthermore, these data demonstrate that combining islet transplantation (as a source of insulin) with specific immunomodulation of IL-21 could reverse diabetes.

These exciting findings have potential application for human type 1 diabetic patients, however this circumstance would have the additional challenge of allograft rejection. Given this challenge, we assessed the role of IL-21 in an islet allograft immune response. A fully mismatched H-2^d BALB/c islet allograft was transplanted into H-2^b wild type C57BL/6 (WT) and H-2^b Il21r^−/− streptozotocin-induced diabetic recipients.

All C57BL/6 (H-2^b) mice rejected the (H-2^d) islet allograft with a mean survival time of day 16.6 (n = 8). In comparison, 66% of Il21r^−/− (H-2^b) mice failed to acutely reject their H-2^d islet allografts (Fig. 5A). As indicated by the individual blood glucose readings (Fig. 5B), the mice that survived long-term consistently maintained steady normoglycemic levels. Nephrectomies were performed on 100 day surviving Il21r^−/− mice, resulting in loss of normoglycemia, confirming that the islet allograft was functional and responsible for controlling blood glucose in these mice.

In order to address whether lacking IL-21 responsiveness and an inability to reject the islet allograft was due to a problem in immune priming capacity, we rechallenged the nephrectomized mice with another H-2^d BALB/c islet allograft, on the contralateral kidney. These secondary islet allografts remained unchallenged, surviving a further 100 days (Fig. 5C). Histological analyses of the long-term surviving islet allograft in Il21r^−/− mice revealed strong insulin production with very mild mononuclear cell infiltrate peripheral to the graft (Fig. 5D). Taken together, these results indicated that responsiveness to IL-21 was crucial for successful rejection of islet allografts.

Given the important role IL-21 has in CD8+ T-cell survival and effector function (16,42), and the well documented critical role CD8+ T-cells have in the rejection of pancreatic islet allografts (43), we determined whether IL-21 responsiveness in CD8+ T-cells alone was sufficient to induce allograft rejection. The number and activation status of CD8+ T-cells are not impaired in resting unchallenged Il21r^−/− mice (14,33). However, in the absence of IL-21 responsiveness, CD8+ T-cells become “exhausted” in chronic viral infections, exhibiting “impaired self-maintenance” (17–19).

To address whether IL-21 responsiveness in CD8+ T-cells was sufficient to provoke islet allograft rejection, we adoptively transferred IL-21R–sufficient CD8+ T-cells into Il21r^−/− recipient mice and repeated the H-2^d BALB/c islet allograft. We used MHC class II deficient mice as the source of CD8+ T-cells, thus eliminating the possibility of contaminating CD4+ T-cells. As is shown in Fig. 6, restoring IL-21 signaling to CD8+ T-cells alone resulted in rapid rejection of the allograft.

Cytokines can play both destructive and immunomodulatory roles in graft rejection. Roles for cytokines such as IFNγ, IL-1β, and tumor necrosis factor-α in allograft rejection have previously been described (44), but this study is the first to demonstrate a role for IL-21 in islet allograft rejection. IL-21 was important for islet targeted autoimmune inflammation in diabetes, and both autoimmune and allograft responses against islet grafts. Pharmacological neutralization of IL-21 at a late preclinical stage inhibited insulitis and delayed diabetes. Moreover, IL-21 neutralization when combined with syngeneic islet transplantation reversed diabetes. It has previously been shown that diabetic NOD mice retain some islet mass (45,46). Removal of the syngeneic islet graft from long-term (100 days) diabetes-free IL-21R/Fc–treated NOD mice revealed that the residual endogenous pancreatic islets (from previously clinically diabetic NOD mice) had regained functionality and were capable of supporting the insulin requirements without the islet graft.

Short-term neutralization of IL-21 reset the clock of autoimmune infiltration in mice on the brink of clinical disease indicating that the treatment both inhibits the ongoing accumulation of infiltrating immune cells as well as affecting resident lymphocytes, which emphasizes the dynamic nature of the islet lesion. Modulation of IL-21 for the brief time period suggests that the autoimmune inflammation is continually reliant on IL-21. The numbers of IL-21–producing cells in the pancreatic lesion were increased at the preclinical stage and were reduced by neutralization of IL-21. An autocrine role for IL-21 on the survival of IL-21–producing CD4+ T-cells has been demonstrated previously and might offer an explanation as to why the relatively short neutralization treatment regimen could perpetuate a sustained effect at a preclinical stage.

In an effort to explore the therapeutic effect of IL-21 neutralization on islet inflammation, we analyzed the endogenous pancreatic islet infiltrate and observed an immediate reduction in the number of activated lymphocytes. In particular, a sustained reduction in activated/memory-phenotype CD8+ T-cells was observed. Indeed, there was an approximately threefold decline in CD8+ T-cells 90 days following short-term treatment, which indicates that this population was less capable of recovering from treatment than CD4+ T-cells or B-cells. CD8+ T-cells have a well-supported role in both the early stage of disease, and in the final effector stage of diabetes (47,48). The broad effect of IL-21 neutralization on lymphocytes in the islet lesion probably reflects that IL-21 is produced by activated CD4+ T-helper cells and, as well as acting in an autocrine fashion, provides soluble help by facilitating the differentiation and survival of B cells and CD8+ cytotoxic T lymphocytes (33,49,50). Furthermore, our findings suggest that measurement of IL-21 could be a predictive marker for diabetes.

The influence of IL-21 on B-cells and T-cells indicates possible roles in both cellular and humoral rejection. However, IL-21 responsiveness by CD8+ T-cells alone could restore islet allograft destruction. This finding provides insight into the important clinical problem of organ allograft rejection, suggesting a potential role for IL-21 neutralization combined with pancreatic islet grafts for patients with type 1 diabetes. Interestingly, elevated IL-21 and IL-21R mRNA have been shown in biopsies from cardiac allograft recipients, with the highest mRNA expression levels found in rejection specimens (51). These reports support our finding that biological activities of the IL-21 pathway contribute to the graft rejection process.

It remains to be shown in our study how IL-21 promotes destruction of islet allografts by CD8+ T-cells. It is likely that IL-21–mediated rescue from ‘exhaustion’ of the graft specific CD8+ T-cells occurs, by a similar mechanism to that observed in chronic viral infections (17–19). Our finding of a role for IL-21 responsiveness in allograft rejection is in line with previous studies, which found that loss of IL-21 signaling attenuated the pathogenesis of CD4+ T-cell mediated graft versus host disease (GVHD) (23–25). Although these studies vary in the detail of their findings, both groups report a reduction in IFNγ-expressing T-cells (24,25). Even though the majority of Il21r^−/− mice accepted islet allografts long-term, these studies indicate a degree of redundancy in the IL-21–IL-21R network, an outcome that can be further explored with combined therapies.

Studies collectively demonstrate that a variety of immunomodulatory reagents can protect the NOD mouse from diabetes. However, there are few examples of reagents that can reverse insulitis (52,53) and only one other reagent, namely anti-CD3 monoclonal antibody (54), that has been shown to prevent the progress of type 1 diabetes once blood glucose levels have begun to rise, leading to clinical trials for type 1 diabetes (55). In contrast to anti-CD3 treatment, neutralization of IL-21 did not reverse disease in overtly diabetic NOD mice (52). However, in accordance with anti-CD3 treatment, syngeneic islets could be transplanted without recurrence of disease or continuous administration of IL-21R/Fc. These influential studies place the blockade of IL-21:IL-21R interactions in an elite group of immunomodulatory agents in type 1 diabetes. The ability of short-term blockade of a single cytokine to have such a profound effect on islet inflammation is unprecedented and raises possibilities for intervention in type 1 diabetes at a late preclinical stage and for the protection of both allogeneic islet graft tissue and transplanted islet tissue during recurrent autoimmunity.

C.K. is supported by grants from the National Health and Medical Research Council of Australia and the Juvenile Diabetes Research Foundation.

No potential conflicts of interest relevant to this article were reported.

H.M.M. researched data and wrote the manuscript. S.W., A.V., and C.M.Y.L. researched data. D.C. reviewed and edited the manuscript. K.E.W., J.S., and S.G. contributed to discussion. C.K. planned experiments and wrote the manuscript.

The authors thank the Australian Cancer Research Fund (ACRF) Unit for the Molecular Genetic of Cancer for use of the ABI Prism 7900 HT Sequence Detection System and the staff at the Biological Testing Facility (Garvan Institute of Medical Research, Sydney) for help with animal breeding and care.