Target Cell Expression of Suppressor of Cytokine Signaling-1 Prevents Diabetes in the NOD Mouse

The generation of SOCS-1-transgencic (Tg) NOD mice and screening for Tg progeny has previously been described (7). SOCS-1-Tg and NODscid mice were bred and maintained in a specific pathogen-free environment at The Scripps Research Institute. All experiments were conducted in accordance with institutional guidelines for animal care and use.

Venous blood glucose concentrations were measured under nonfasting conditions using Glucometer Elite strips (Bayer, Pittsburgh, PA). Mice were considered diabetic after two consecutive measurements of blood glucose >13.8 mmol/l (250 mg/dl). Onset of diabetes was dated from the first of the sequential measurements.

Harvested organs were fixed in formalin and embedded in paraffin. Tissue sections 5 μm thick were stained with hematoxylin-eosin (H-E) or with a primary antibody against insulin, glucagon (DAKO, Carpinteria, CA), or SOCS-1 (J192; Immuno-Biological Laboratories, Fujioka, Japan) (7). Slides were counterstained in Mayer’s hematoxylin.

Pancreatic sections (2–3 levels/organ) stained with H-E were ranked for insulitis using the following grades: 0, normal islet morphology with no periinsulitis or insulitis; 1, periinsulitis; 2, insulitis; and 3, islet remnant (8).

Pancreatic islets were isolated and cultured as previously described (8). Protein extracts from islets exposed to IFN-γ (1,000 units/ml; Pharmingen, San Diego, CA) or vehicle were separated by SDS-PAGE and transferred to nitrocellulose membranes (9). Membranes were incubated with primary antibodies for murine phospho-STAT-1 and murine STAT-1 (Upstate Biotechnology, Lake Placid, NY), and signal detection was accomplished as previously described (7).

For T-cell tolerance studies, splenocytes were harvested from 14-week-old non-Tg or SOCS-1-Tg mice and injected intravenously (0.8–1.0 × 10⁷ or 1.8–2.0 × 10⁷ cells/mouse) in NODscid recipients. To determine β-cell susceptibility to activated effector cells, splenocytes from overtly diabetic NOD mice were injected in nonlethally irradiated (700 rad) SOCS-1-Tg or non-Tg mice within 2 h of irradiation. For transfers using cells infiltrating pancreata of 18-week-old prediabetic SOCS-1-Tg and non-Tg mice, pancreata were digested with collagenase followed by the isolation of islets and lymphocytes using Lympholyte-M (Cedarlane, Hornby, Canada). Single-cell suspensions were prepared using trypsin/EDTA (Sigma). Lymphocytes were pooled, counted, and injected into NODscid recipients (1.3–1.5. × 10⁶/mouse).

Accumulated incidence of diabetes was determined by χ² test, and insulitis was determined by unpaired Student’s t test. The Kaplan-Meier life-table analysis was used to test the difference in rates of disease development after cell transfers.

Two lines of SOCS-1-Tg NOD mice (lines A and B) were generated, as previously described (7). To test the effect of the SOCS-1 gene product on the development of type 1 diabetes, female mice were monitored for the development of hyperglycemia. As already reported (7), the SOCS-1-Tg mice developed diabetes with similar kinetics as non-Tg littermates (Fig. 1A and B); SOCS-1-Tg mice of both lines reached age 14–16 weeks before the first mice developed disease, paralleling the onset of diabetes in non-Tg mice. Interestingly, SOCS-1-Tg mice had a lower cumulated incidence of diabetes as compared with non-Tg mice. The first line, line A, demonstrated a reduction in diabetes incidence that was not significantly different from that of non-Tg littermates (Fig. 1A). However, the second Tg line, line B, showed a robust protection from disease (Fig. 1B). At age 32 weeks, only 11% of the SOCS-1-Tg mice were diabetic, whereas 75% of their non-Tg littermates had succumbed to diabetes (P < 0.001, χ² test).

Because the most prominent disease protection had been observed in mice of line B, the following studies on the mechanisms for protection were performed using mice from this line only. First, we performed functional studies to confirm the effect of the SOCS-1 gene product. Exposure of islets from non-Tg mice to IFN-γ resulted in the phosphorylation of STAT-1 (n = 2) (Fig. 2A). Conversely, this treatment failed to trigger STAT-1 phosphorylation in SOCS-1–expressing islet cells (n = 2) (Fig. 2A), verifying that JAK/STAT signaling was blocked by SOCS-1 in islet cells.

Histological evaluation revealed that SOCS-1 expression did not have any obvious effect on β-cell development, as the pancreatic islets of 8-week-old SOCS-1-Tg mice had a normal, round structure with a distribution of endocrine cells similar to that of non-Tg mice (Fig. 2B and C; data not shown). Moreover, their nonfasting blood glucose levels were similar to those of non-Tg mice (blood glucose values in mmol/l at age 8 weeks, mean ± SE: non-Tg, 6.6 ± 0.3, n = 17; SOCS-1-Tg, 6.8 ± 0.3, n = 19) and normoglycemia was retained even at age 32 weeks (7.7 ± 0.4 mmol/l, n = 24; mice with diabetes were excluded). In addition, the nonfasting blood glucose levels of 32-week-old SOCS-1-Tg mice backcrossed to a NODscid background (n = 5) lacking T- and B-cells and thereby islet-infiltrating lymphocytes did not differ from the blood glucose levels of non-Tg NODscid littermates (n = 7) (blood glucose levels in mmol/l at age 31–32 weeks, mean ± SEM: non-Tg, 6.6 ± 0.5, n = 7; SOCS-1-Tg, 6.5 ± 0.5, n = 5). These findings indicated that overexpression of SOCS-1 did not alter β-cell function leading to overt changes in glucose homeostasis.

We next determined whether the low rate of type 1 diabetes in the SOCS-1-Tg mice reflected reduced autoreactivity within the T-cell compartment. To this end, splenocytes from pre-diabetic SOCS-1-Tg and non-Tg littermate mice were isolated. We observed no differences in total spleen cell numbers or in the percentages of CD4+, CD8+, and B220+ cell populations between SOCS-1-Tg (n = 10) and non-Tg mice (n = 7) (data not shown). Moreover, when these cells were transferred to NODscid mice, they transferred disease with similar incidence and kinetics (Fig. 3A; data not shown). This suggested that protection was not conferred by overt T-cell tolerance.

To determine whether SOCS-1 could affect T-cell recruitment to the pancreas, the degree of insulitis was quantitated in SOCS-1-Tg and non-Tg mice. It was found that SOCS-1-Tg mice developed insulitis, demonstrating that β-cell expression of SOCS-1 did not prevent the early recruitment of lymphocytes to the pancreas (Table 1). Interestingly, although the degree of insulitis became more severe with time in non-Tg mice (Table 1), it remained constant between ages 14 and 20 weeks in SOCS-1-Tg mice (Table 1).

We asked whether SOCS-1–expressing β-cells were less susceptible to T-cell−mediated killing than their non-Tg counterparts. Splenocytes from overtly diabetic NOD donors were injected into nonlethally irradiated SOCS-1-Tg and non-Tg mice (Fig. 3B). Interestingly, in SOCS-1-Tg recipients, the onset of disease was markedly delayed compared with in non-Tg recipients (P < 0.01) (Fig. 3B). When 100% of the non-Tg mice had developed chronic hyperglycemia, only 50% of the SOCS-1-Tg mice were diabetic, indicating that SOCS-1-Tg β-cells were less sensitive to destruction by activated self-reactive immune cells.

To test whether the expression of SOCS-1 affects the diabetogenicity of lymphocytes locally within the pancreas, we compared the pathogenicity of pancreatic infiltrating populations (10). Lymphocytes infiltrating the pancreas of 18-week-old, nondiabetic SOCS-1-Tg and non-Tg littermates were isolated and transferred to NODscid recipients. Interestingly, cells isolated from SOCS-1-Tg mice triggered diabetes with a clearly reduced efficiency compared with cells isolated from non-Tg mice (P < 0.05) (Fig. 3C), thereby demonstrating a lower efficacy of these cells in causing β-cell destruction.

Our study showed that NOD mice harboring β-cells expressing SOCS-1 have a dramatically lowered incidence of diabetes. This finding provides an important insight into the pathogenesis of type 1 diabetes as it suggests that cytokines critically contribute to type 1 diabetes by acting locally on the target β-cell. Moreover, our study showed that SOCS-1 expression by β-cells is an efficient preventive treatment for type 1 diabetes in the NOD mouse.

Our adoptive transfer studies suggested that SOCS-1–expressing β-cells were less vulnerable to the destructive actions of activated T-cells. A likely explanation for this is that by blocking JAK/STAT signaling, SOCS-1 directly protects β-cells from the detrimental actions of inflammatory cytokines produced by these T-cells. This notion is consistent with the finding that overexpression of SOCS-1 in β-cell lines (5,6) and primary β-cells (7) blocks responses to several cytokines implicated in the pathogenesis of type 1 diabetes, such as IFN-γ. Another possibility is that SOCS-1 reduces the β-cell’s risk to being targeted by self-reactive CD8+ T-cells as SOCS-1 prevented IFN-γ signaling, and IFNs are potent inducers of major histocompatibility complex I expression (3,11). Recent studies have indicated that β-cells may contribute to their own demise by secreting chemokines after exposure to proinflammatory cytokines such as IFN-γ (12–14). These chemokines may be instrumental in attracting self-reactive T-cells to the pancreas of the NOD mouse (13). Although we did not observe a robust difference in insulitis between non-Tg and SOCS-1-Tg mice (Table 1), the possibility that SOCS-1 affords disease protection by preventing β-cell expression of such chemoattractants cannot be excluded. In fact, all the above-mentioned mechanisms could act simultaneously to prevent β-cell destruction in SOCS-1-Tg mice. Finally, disease protection may not be entirely explained by a block in β-cell responses to cytokine signaling through the JAK/STAT pathway, as studies have indicated that SOCS-1 may regulate additional signaling pathways besides the JAK/STAT pathway (4,15).

From our studies it is clear that SOCS-1 expression did not block the early recruitment of lymphocytes to the pancreas. However, we found that lymphocytes isolated from the pancreas of prediabetic SOCS-1-Tg mice had a reduced capacity to cause diabetes when transferred to NODscid mice. Self-reactive T-cells proliferate and acquire their capacity to home to the pancreas after an initial encounter with islet antigen in the pancreatic lymph node (16,17). However, a small percentage of cells undergoing such proliferation will acquire effector functions, and an expansion of the effector pool occurs only after restimulation with antigen (18,19). During the pre-diabetic stage in the NOD mouse, it is likely that macrophages, dendritic cells, and B-cells infiltrating the pancreas present antigens to recruited T-cells, thereby providing the restimulation necessary for the generation and maintenance of the effector T-cell pool. It is possible that a lower degree of cytokine-induced β-cell damage leads to reduced antigen release. This would lead to fewer antigens available for presentation to T-cells homing to the pancreas of SOCS-1-Tg mice. Consequently, SOCS-1-Tg mice would have reduced numbers of effector T-cells in their pancreata as compared with non-Tg mice. This would explain why cells reaching the target organ of SOCS-1-Tg were less aggressive than those recruited to the pancreas of non-Tg littermates. Importantly, this could further contribute to the diabetes protection observed in SOCS-1-Tg mice.

It is unclear whether SOCS-1 plays a role in the pathogenesis of type 1 diabetes; at the low level that SOCS-1 is naturally expressed by β-cells, it does not lower responsiveness to IFN-γ, but may affect TNF-α stimulation (20). Regardless of this, we have shown that forced expression of SOCS-1 in β-cells results in ameliorated β-cell destruction and prevention of diabetes development in the NOD mouse. This finding suggests that target cell responses to cytokines play an important role in the pathogenesis of type 1 diabetes and reinforces the idea that modulation of target cell actions can alter susceptibility to disease (1,2,7,13,17,21). Importantly, the SOCS-1-Tg mice displayed normal nonfasting blood glucose levels throughout their life. Hence, expression of SOCS-1 may be a way to suppress cytokine-induced β-cell damage without causing overt alterations in β-cell function. One important consideration will be that SOCS-1–expressing β-cells fail to mount an antiviral defense response after exposure to IFNs (7) and thereby succumb to diabetes after infection with the coxsackievirus (7). Therefore, the next step will be to delineate which specific cytokine(s) needs to be blocked to prevent β-cell destruction, but still allow for survival from exposure to pathogens.

This work was supported by National Institutes of Health Grant AI-42231 (N.S.), an advanced postdoctoral fellowship from the Juvenile Diabetes Research Foundation (M.F.-T.), the Swedish Research Council (M.F.-T.), the Swedish Diabetes Association Research Foundation (M.F.-T.), and the Swedish Foundation for Strategic Research (M.F.-T.).

The authors would like to thank L. Tucker, K. Van Gunst, and A. Maday for excellent technical assistance, as well as Drs. N. Hill and V. Judkowski and other members of the Sarvetnick lab for discussions and suggestions.