A Salen-Manganese Catalytic Free Radical Scavenger Inhibits Type 1 Diabetes and Islet Allograft Rejection

The salen-manganese compound EUK-8 was a gift from Eukarion (Bedford, MA). EUK-8 was dissolved in autoclaved water at 30 mg/ml, filtered, and stored at room temperature for up to 2 months. For in vivo studies, EUK-8 was mixed with sterile PBS to the indicated concentration (5–100 mg/kg). For in vitro studies, EUK-8 was initially dissolved as a 20× stock in PBS and then diluted to a final concentration of 80 μmol/l in media for the NIT cell viability assays.

3-Morpholinosydnonimine-N-ethylcarbamide (SIN-1) was purchased from Cayman Chemical (Ann Arbor, MI). SIN-1 is an NO donor that rapidly releases superoxide anion and hydrogen peroxide concomitantly with NO when dissolved in aqueous solutions. This primarily leads to peroxynitrite formation but can also produce some nitrate and nitrite as well (33–35). DETA-NONOate [(Z)-1-[2-(2-aminoethyl)-N-(2-ammonioethyl)amino]diazen-1-ium-1,2-diolate (Alexis, San Diego, CA)] is an NO donor with a half-life of NO release that is comparable with that of SIN-1 (36,37), but it does not produce peroxynitrite. SIN-1 or DETA-NONOate was dissolved directly in culture medium (Click’s media containing 10% fetal bovine serum, 2 mmol/l l-glutamine, and 2 mmol/l penicillin-streptomycin) immediately before treatment of cells for the in vitro assays.

Female NOD mice (Taconic Farms) and BALB/c mice (The Jackson Laboratories) were bred and housed under specific pathogen-free conditions. Food and water were provided ad libitum, and the animals were housed under 12-h light and dark cycles. The UCLA Animal Research Committee approved all animal care and experimental procedures.

NOD mice were screened for the onset of hyperglycemia by monitoring urine glucose levels by Tes-tape (Lilly). NOD mice were tested every other day after reaching 12 weeks in age or after receiving an infusion of diabetogenic T-cells. After detection of abnormal glucose in the urine, blood glucose concentration was monitored daily by measurement of glucose with a Glucometer Elite (Bayer, Pittsburgh, PA). Two consecutive blood glucose levels of >13 mmol/l were considered as type 1 diabetes onset. NOD mice that received islet allografts were monitored daily for blood glucose levels. Diabetes incidence or recurrence data were analyzed by SAS statistical software (SAS, Cary, NC) using the Kaplan-Meier method to generate survival curves, followed by log-rank test comparisons to generate the P value.

Female NOD mice (6–8 weeks old) were irradiated (500 rad) and 24 h later received 10 million splenic mononuclear cells intravenously from newly diabetic female NOD mice. The recipients were randomly separated into experimental and control groups (n = 10 per group). One group was treated daily with 5 mg/kg EUK-8 in 0.1 ml PBS i.p., whereas the other received PBS alone. The EUK-8 dose was based on previous results from EAE studies (29). Diabetes incidence was monitored as described above.

To test the ability of EUK-8 to prevent spontaneous TID in NOD mice, female NOD mice were treated from 6 to 35 weeks of age with 100 mg/kg of EUK-8 given in 0.5 ml PBS i.p. every other day. Age-matched control female NOD mice received 0.5 ml PBS every other day. Their blood glucose levels were monitored for up to 52 weeks in age as described above.

Islets were isolated from pancreata of 6- to 8-week-old female BALB/c mice by digestion with collagenase P (Sigma) at 37°C in Dulbecco’s modified Eagle’s media (10% FCS, 2 mmol/l l-glutamine, and 2 mmol/l penicillin-streptomycin) as previously described (38). After digestion, islets were rinsed three times in chilled media (Click’s media with 10% FCS, 2 mmol/l l-glutamine, and 2 mmol/l penicillin-streptomycin) and maintained in chilled media during isolation by hand picking using a dissecting scope. Approximately 500 islets were implanted under the kidney capsule of newly diabetic female NOD mice. After islet transplantation, recipient mice were treated daily with PBS or 100 mg/kg EUK-8 in intraperitoneal PBS, and their blood glucose was monitored daily.

NIT-1 cells (ATCC) were plated at 2–5 × 10⁵ cells/well in 4 ml incubation media (Click’s media with 10% FCS, 2 mmol/l l-glutamine, and 2 mmol/l penicillin-streptomycin). Cultures were exposed to dose ranges of SIN-1 or DETA-NONOate. Some of the cultures were pretreated with EUK-8 (80 μmol/l) for 2 h before addition of SIN-1 or DETA-NONOate. Control cultures were incubated in media alone. Twelve hours after SIN-1 or DETA-NONOate addition, the NIT cells were harvested and stained with trypan blue, and their viability was determined (n = 4–6 wells/group). The experiments were repeated twice. The results are expressed as the average percentage of viable cells in each treatment group. Statistical analyses were performed using pairwise t tests.

A number of new salen-manganese compounds have been generated, of which EUK-8 is a promising therapeutic antioxidant based on its strong SOD, peroxidase, and catalase activities (27,31,39) (Fig. 1). To begin to evaluate whether EUK-8 may have therapeutic potential to protect β-cells from autoreactivity, we first tested its ability to inhibit the adoptive transfer of type 1 diabetes. The 6- to 8-week-old irradiated female NOD mice were transfused with 10 million splenic mononuclear cells from newly diabetic NOD mice. Recipient mice were treated daily with EUK-8 or PBS alone. All of the mice that were treated with PBS alone became diabetic within 32 days (Fig. 2). In contrast, only 10% of the EUK-8–treated mice were diabetic at this time. Over the 50-day observation period, 30% of the EUK-8–treated mice became diabetic, a significant protective effect compared with the PBS-treated group (survival analysis, P < 0.0001). Thus, EUK-8 treatment can significantly inhibit a highly diabetogenic immune cell population. Accordingly, inactivation of ROS, such as superoxide and hydroxyl radical, as well as prevention of nitrogen oxide products, are important components of therapeutic strategies aimed at protecting β-cells from immune responses.

We next examined the ability of EUK-8 treatment to inhibit the progression of spontaneous disease in NOD mice. We began treating female NOD mice at 6 weeks in age, when T-cell autoreactivity and insulitis are already well established. The mice received EUK-8 or PBS alone, every other day, from 6 to 35 weeks in age, and their blood glucose was monitored for up to 1 year in age. All of the mice that received EUK-8 thrived throughout the lengthy treatment period and displayed no loss in body weight. In the PBS treatment group, 66% of the mice became diabetic by 25 weeks in age, and 80% were diabetic within the 1-year observation period (Fig. 3). In contrast, none of the EUK-8–treated mice became hyperglycemic, even up to 1 year in age (P < 0.0001). Thus, even though EUK-8 treatment was discontinued at 35 weeks in age, the entire group of mice remained euglycemic until 52 weeks in age, at which point the study was terminated.

We next examined whether EUK-8 administration could promote the survival of islet allografts in newly diabetic NOD mice. A total of 500 islets isolated from BALB/c mice were implanted under the kidney capsule of newly diabetic female NOD mice. The allograft recipients were treated daily with EUK-8 or PBS. Islet implantation restored euglycemia in all recipients. However, hyperglycemia recurred in all of the control PBS treated mice within 7–10 days after alloislet transplantation, as is expected in immunocompetent mice (Fig. 4). In contrast, all of the EUK-8–treated allograft recipients remained euglycemic until 10 days after transplantation. Thereafter, hyperglycemia recurred, but at a slower rate, with three of six mice remaining euglycemic at 20 days after transplantation and one of six of the islet allograft recipients remaining disease free at 50 days after transplantation (P < 0.0011 by survival analysis). These data suggest that ROS, such as superoxide and hydroxyl radical, and RNS, such as peroxynitrite and nitrite, contribute to the islet graft rejection process.

The ability of EUK-8 to prevent spontaneous type 1 diabetes and inhibit islet allograft rejection suggests that the production of ROS and RNS contributes to β-cell death in these models. Conceivably, the reactivity of NO, superoxide, hydroxyl radical, and/or NO/superoxide-derived products may contribute to β-cell cytotoxicity. To further explore the mechanisms of β-cell injury and the protective mechanism(s) of EUK-8, we used an in vitro system in which we could modulate the putative free radicals that were present.

The insulinoma cell line NIT-1 was exposed to SIN-1, an NO donor that rapidly releases superoxide anion concomitantly with NO in aqueous solutions, producing peroxynitrite (predominantly) as well as nitrite and nitrate (33–35). We also added EUK-8 (80 μmol/l) to some of the cultures to inactivate superoxide and hydrogen peroxide and thereby prevent, or catalyze the breakdown of, cytotoxic NO products. Twelve hours after adding SIN-1 (0, 1, or 3 mmol/l), we determined cell viability by trypan blue exclusion. Cells grown in media alone or that were treated with EUK-8 had similar viability (82 and 76% viability, respectively; Fig. 5). Addition of SIN-1 to 1 mmol/l reduced cell viability to 63%. However, inclusion of EUK-8 with SIN-1 protected NIT cells, such that NIT viability was similar to that of NIT cells cultured in media alone (78% viability). When SIN-1 was added to 3 mmol/l, cell viability in cultures without EUK-8 was 47%, whereas in cultures with EUK-8, 64% of the cells remained viable (P < 0.0001). These results indicate that NIT cells are sensitive to ROS and/or RNS generated by SIN-1 and support the hypothesis that production of NO, superoxide, and their derived products may lead to β-cell death.

To further evaluate the contribution of NO versus superoxide to cytotoxicity in our model, we cultured NIT cells with DETA-NONOate, an agent that only generates NO (at a rate similar to that of SIN-1). We observed that increasing dosages of DETA-NONOate (0.01–0.1 mmol/l) induced progressively greater NIT cell death, but that addition of EUK-8 (80 μmol/l) effectively inhibited the cytotoxicity of DETA-NONOate (Fig. 6). For example, in cultures exposed to 1 mmol/l DETA-NONOate, only 12% of cultured NIT cells were viable. However, inclusion of EUK-8 with 1 mmol/l DETA-NONOate led to 67% cell viability. Therefore, in the absence of superoxide, NO production alone can lead to cytotoxicity, which can be prevented by a salen-manganese compound.

The formation of ROS and RNS has been implicated in the pathoetiology of type 1 diabetes. Accordingly, limitation of deleterious ROS and RNS products may provide new therapeutic strategies to mitigate β-cell destruction. Previous studies showed that adenovirus-mediated expression of manganese SOD in islets could protect them from diabetogenic T-cells after implantation in syngeneic mice (22). Recently, a metalloporphyrin-based SOD mimic was shown to inhibit the adoptive transfer of autoimmune diabetes by a diabetogenic T-cell clone (BDC2.5) (23). However, these studies did not examine the compound’s ability to prevent spontaneous type 1 diabetes or prolong islet graft survival. Here, we used a low–molecular weight synthetic salen-manganese compound that has SOD, peroxidase, and catalase activity and tested its ability to inhibit both spontaneous type 1 diabetes and islet allograft rejection in diabetic NOD mice. The enzymatic activities of this compound may make it well suited to protect islets from immune-mediated damage, because it inactivates superoxide and hydrogen peroxide and prevents the formation of a highly cytotoxic hydroxyl radical. This compound also catalyzes the breakdown of numerous NO-derived products, including cytotoxic nitrogen oxides.

Our initial studies demonstrated that EUK-8 treatment could inhibit the adoptive transfer of disease by diabetogenic spleen cells from newly diabetic mice. This finding supports the hypothesis that at late stages of the disease process, effector cells producing ROS and/or RNS play a major role in β-cell destruction. We went on to show that when given to pre-diabetic NOD mice with established autoimmunity, EUK-8 completely inhibited the spontaneous development of type 1 diabetes for up to 52 weeks in age, even though treatment was discontinued after the mice reached 35 weeks in age. Histological analysis of pancreata taken from 52-week-old EUK-8–treated NOD mice showed most islets to be free of infiltrates, although some had a slight insulitis (data not shown). Because EUK-8 treatment was initiated at 6 weeks in age, some insulitis is expected. It is unclear whether the slight insulitis at 52 weeks of age (17 weeks after ceasing EUK-8 treatment) represents a long-term benign insulitis or the beginning of a potentially pathogenic response.

We also found that salen-manganese treatment significantly delayed the recurrence of diabetes in newly diabetic NOD mice that received an islet allograft. The observation that salen-manganese treatment completely inhibited spontaneous disease progression in pre-diabetic NOD mice with established β-cell autoimmunity and delayed recurrent diabetes in islet allograft recipients suggests that ROS and RNS contribute to the pathogenesis of spontaneous type 1 diabetes and allograft rejection. Because EUK-8 has very low toxicity in vivo (27,31,32), it will be of interest to further investigate the potential of such salen-manganese treatments to inhibit type 1 diabetes and prolong islet allograft survival.

Nitrogen oxide formation is considered to contribute to cytotoxicity during type 1 diabetes pathogenesis because of NO and/or superoxide release. To model this process in vitro, we exposed insulinoma cells to two different NO donors (SIN-1 and DETA-NONOate), which have equivalent rates of NO release but vary in the proportions of their NO products. SIN-1 releases superoxide anion concomitantly with NO in aqueous solutions (40) and provides two potential pathways of cytotoxicity. The first is a coordinated release of superoxide with NO, leading to the formation of peroxynitrite predominantly, as well as some other nitrogen oxides. The second involves superoxide releasing iron from ferritin (which is known to be present at high levels in islets [12,13]), and concomitant dismutation of superoxide to generate hydrogen peroxide. The combined presence of hydrogen peroxide and iron in the ferrous state can then yield cytotoxic hydroxyl radicals by the Fenton reaction (14,15,41) (Fig. 1). Using cultured NIT cells, we found that SIN-1 mediated NIT cell death in a dose-dependent manner and that a micromolar concentration of EUK-8 could inhibit this cytotoxicity. DETA-NONOate, which only releases NO with no superoxide, had a higher cytotoxicity profile than that of SIN-1, whereas the median lethal dose for SIN-1 was ∼3 mmol/l, and the median lethal dose for DETA-NONOate was only ∼0.03 mmol/l. The cytotoxicity of DETA-NONOate was also inhibited by EUK-8. Collectively, these data indicate that NO production with or without the presence of superoxide is cytotoxic, either through nitrogen oxide formation or via the generation of hydroxyl radical.

The ability of EUK-8 to inhibit SIN-1–and DETA-NONOate–mediated cytotoxicity helps clarify previous conflicting evidence regarding the contribution of the joint products of superoxide and NO, namely peroxynitrite, to islet destruction. Previous studies have pointed to peroxynitrite as a primary mediator of β-cell cytotoxicity based on its ability to form hydroxyl radical, the formation of nitrotyrosine in NOD mouse islets, and the ability of pharmacological agents that block peroxynitrite formation to partially inhibit disease in NOD mice (5,8,42–45). Peroxynitrite can also give rise to the cytotoxic nitrogen dioxide. However, other lines of evidence suggest that peroxynitrite may not be particularly cytotoxic to β-cells. Peroxynitrite can readily degrade spontaneously to nitrate (which is innocuous), and, accordingly, its formation may not yield substantial cytotoxicity (46). Moreover, studies of peroxynitrite cytotoxicity have used supraphysiological concentrations in vitro, far exceeding the levels of peroxynitrite that are generated locally by inflammatory responses in vivo (6,47). Finally, nitrotyrosine may be formed by other nitrogen oxides besides peroxynitrite, such as nitrogen dioxide (48,49). In our studies, both DETA-NONOate and SIN-1 produce NO at similar rates (36,37), but SIN-1 also generates superoxide, leading to peroxynitrite formation. Our finding that DETA-NONOate was more cytotoxic than SIN-1 suggests that NO and its derived products (such as nitrogen dioxide, dinitrogen trioxide, and nitrite) are more cytotoxic than the release of superoxide and NO, leading to their peroxynitrite derivative (as well as its NO-derived products). Indeed, the formation of peroxynitrite may provide a detoxification mechanism, because its rapid spontaneous degradation to the innocuous nitrate limits the production of potentially more cytotoxic nitrogen oxides from NO (Fig. 1). This latter category of nitrogen oxides (e.g., nitrogen dioxide, dinitrogen trioxide, and nitrite) cannot spontaneously degrade to nitrate and thus are available to nitrate tyrosine residues in proteins (50).

Salen-manganese compounds (such as EUK-8) are able to convert nitrogen oxides (such as nitrogen dioxide, dinitrogen trioxide, and nitrite) to nitrate during NO “detoxification,” presumably by its peroxidase activity (28). Our finding that nitrogen oxides can be a major contributor to NIT cell toxicity may explain the ability of EUK-8 to inhibit the cytotoxicity of NO donors in vitro, as well as its therapeutic efficacy in preventing spontaneous type 1 diabetes and prolonging allograft survival in vivo.

In summary, we have shown that treatment with a salen-manganese antioxidant, even after the onset of autoimmunity, completely prevented spontaneous type 1 diabetes and delayed islet allograft rejection in newly diabetic NOD mice. Our study also broadens our understanding of the cytotoxicity of reaction products resulting from NO and/or superoxide formation in β-cell destruction. Specifically, the superoxide product hydroxyl radical, as well as its specific reaction products with NO, appears to contribute to cytotoxicity in our models. The SOD, peroxidase, and catalase activity of EUK-8 to prevent both hydroxyl radical formation and inactivate nitrogen oxides may account for its therapeutic efficacy. Salen-manganese compounds have been shown to inhibit EAE, skin allograft rejection, and neuronal damage in neuropathological conditions (27,29,30). This new class of antioxidants may provide a potent and safe pharmacological approach to inhibit spontaneous type 1 diabetes and aid in prolonging survival of islet allografts.

This work was supported by a grant from the Juvenile Diabetes Research Foundation International.

We thank Drs. John Corbett, Alexander Rabinovitch, Georgette Buga, Mike Collins, Hoa Dang, and Susan Doctrow for their comments and suggestions on the manuscript. We also thank Eukarion for generously providing EUK-8.