Deletion of STAT-1 Pancreatic Islets Protects Against Streptozotocin-Induced Diabetes and Early Graft Failure but Not Against Late Rejection

STAT-1 knockout (^−/−) mice (C57BL/6 background) were a gift of Dr. David Levy (New York University School of Medicine, New York, NY). Eight-week-old wild-type and BALB/c mice were produced from stocks purchased from Harlan Nederland (Horst, the Netherlands). NOD mice, inbred in our animal facility (Proefdierencentrum, Leuven, Belgium) since 1989, were used as diabetes-prone animals, and diabetes was defined and detected as previously described (11). STAT-1^−/− mice were kept in specific pathogen-free conditions. Experiments were conducted with the approval of the animal ethical committee of Katholieke Universiteit Leuven, Belgium.

Freshly isolated STAT-1^−/− or wild-type islets (n = 500) were isolated and transplanted under the kidney capsule of recipients, as previously described (11). Alloxan diabetes was induced and graft function evaluated, as previously described (11). Graft destruction was defined as the return to hyperglycemia (blood glucose >200 mg/dl on 2 consecutive days after initial normoglycemia). Recipients were killed the day of graft rejection or in a separate experiment for gene analysis in grafts 8 h posttransplantation.

Islet grafts, retrieved 8 h posttransplantation, were used for RNA extraction using SV Total RNA Isolation kit (Promega Benelux, Leiden, the Netherlands). cDNA was created and quantitative PCR analysis performed as described (7). Primer and probe sequences for the determination of mouse cDNAs for housekeeping genes β-actin, IL-1β, IL-15, IFN-γ, inducible nitric oxide synthase (iNOS), MCP-1 (monocyte chemoattractant protein-1), interferon-γ–inducible protein-10 (IP-10), and MIP-3α were as previously described (7).

Freshly dissolved STZ (50 mg/kg) (Sigma, St. Louis, MO) was injected intraperitoneally for 5 consecutive days in male wild-type mice that had received an islet transplantation with 500 STAT-1^−/− or wild-type islets (after being rendered diabetic by alloxan) and had been normoglycemic for at least 14 days. Graft function was followed for 40 days after the last STZ injection. At that time point, grafts were removed, embedded in paraffin, and used for hematoxylin/eosin and insulin staining using a guinea pig anti-insulin antibody (Dakocytomation, Glostrup, Denmark), as previously described (12). After 2 additional days, pancreata were removed for histology and insulin content determination (12).

As recommended by Ulrich Feige (Amgen, Thousand Oaks, CA), human recombinant IL-1ra (Kineret) was delivered by a subcutaneously implanted in vitro–primed osmotic pump, as previously described, at a dose of 100 mg · kg⁻¹ · day⁻¹ for 15 days, beginning 1 day pretransplantation (10). A subtherapeutic dose of 7.5 mg/kg CsA (Sandimmune; Novartis, Basel, Switzerland) was administered daily by gavage, beginning 1 day pretransplantation for the duration of normoglycemia.

NCSS 2000 (Kaysville, UT) software was used for statistical analysis. Data are expressed as means ± SEM. Log-rank test was performed for graft survival. χ² test was used for the incidence of multiple low-dose STZ-induced diabetes. Student's t test and ANOVA were used for multiple comparisons, whenever appropriate. Significance was defined at the 0.05 level.

STAT-1^−/− or wild-type islets were transplanted under the kidney capsule of alloxan-diabetic syngeneic C57BL/6 recipients. Normoglycemia was rapidly reached by STAT-1^−/− (n = 7) and wild-type (n = 7) islets (3.9 ± 2.9 vs. 4.3 ± 2.8 days, respectively; P = NS). After 2–3 weeks of normoglycemia, multiple low-dose STZ-induced diabetes was induced. At the start of the multiple low-dose STZ treatments, no differences in blood glucose were present between mice transplanted with either STAT-1^−/− or wild-type islets (132 ± 32 vs. 121 ± 20 mg/dl, respectively; P = NS). However, 10 days after the last injection with STZ, there were clear differences in blood glucose levels (159 ± 32 vs. 250 ± 45 mg/dl in STAT-1^−/− vs. wild-type islet recipients; P < 0.05). This difference was maintained until the end of the study, 40 days after the last STZ injection (182 ± 55 vs. 305 ± 80 mg/dl in STAT-1^−/− vs. wild-type islet recipients; P = 0.07). After 40 days, 100% of wild-type islet recipients became diabetic compared with 28% in STAT-1^−/− islet recipients (P < 0.05) (Fig. 1). Maintenance of normoglycemia in mice transplanted with STAT-1^−/− islets was not due to regeneration of recipients’ pancreata, as demonstrated by low insulin content in pancreata (data not shown). Moreover, clear insulin positivity was seen in removed grafts (supplemental Fig. 1 [found in an online appendix at http://dx.doi.org/10.2337/db07-0052]). In separate mice where multiple low-dose STZ-induced diabetes was not induced, STAT-1^−/− and wild-type islets maintained normoglycemia for >30 days (Fig. 1).

Absence of STAT-1 in islet cells did not prolong graft survival in chemically diabetic recipients compared with islets from wild-type mice (Table 1). Wild-type islet transplantations in spontaneously diabetic NOD mice, where autoimmune recurrence also occurs, failed to normalize glycemia in 2 of 12 mice, representing 17% of primary islet nonfunction (PNF). No PNF was observed in transplantations with STAT-1^−/− islets. However, islets of STAT-1^−/− and wild-type mice were rejected at the same time (Table 2). mRNA analysis of STAT-1^−/− and wild-type islet grafts 8 h after transplantation revealed analogous levels of inflammatory cytokines in both recipients, illustrating comparable immune responses to the grafts (data not shown). Figure 2 (BALB/c recipients) and supplemental Fig. 2 (NOD recipients) show that MCP-1, IL-15, IP-10, and iNOS genes were strongly upregulated in islet allografts of wild-type mice. Induction of IL-15, IP-10, and iNOS mRNA were markedly suppressed in STAT-1^−/− grafts (P < 0.05).

All chemically diabetic recipients treated with a combination of IL-1ra and CsA had prolonged graft survival (Table 1). However, mean islet graft survival was identical for IL-1ra–and CsA-treated mice transplanted with STAT-1^−/− and wild-type islets (Table 1). The combination of IL-1ra and CsA prevented the PNF observed in wild-type into NOD transplants, but no additional benefit from STAT-1 deletion was observed (Table 2).

Pancreatic islets are markedly susceptible to the cytotoxic effects of inflammatory cytokines (e.g., IFN-γ and IL-1β) secreted by infiltrating immune cells (2,13). Microarray studies demonstrate that β-cells under attack are not passive bystanders but actively participate in their destruction, up- and downregulating a complex pattern of genes (2). This altered gene expression profile is most likely responsible for activation of proapoptotic genes that eventually elicit specific destruction of the β-cells (5,14). The transcription factor STAT-1 is a key mediator of biological responses to IFN-γ, with IFN-γ activating STAT-1 via Janus kinase–mediated tyrosine phosphorylation, allowing dimerization, nuclear translocation, and binding of STAT to γ-activated sequences within target promoters (6). As a result, STAT-1 modulates the expression of primary response genes (among them other transcription factors, such as interferon regulatory factor-1), sustaining the transcriptional response and driving expression of secondary response genes (6,15–17).

Here, we used STAT-1^−/− mice to determine the contribution of this transcription factor to immune-mediated β-cell death in vivo (18). We previously showed that both islets and fluorescence-activated cell sorter–purified β-cells lacking STAT-1 are completely protected against IFN-γ plus IL-1β–or IFN-γ plus dsRNA–mediated β-cell death in vitro (7,19). Furthermore, STAT-1^−/− mice are partially resistant to development of diabetes after multiple low-dose STZ treatment (7). It was, however, impossible to discriminate in this systemic knockout model between effects of STAT-1 deletion on islet cells and immune cells. When multiple low-dose STZ-induced diabetes is induced in wild-type mice transplanted with STAT-1^−/− islets, only the effect of the STAT-1 deletion in the islets contributes to differences in diabetes. The higher percentage of diabetic animals at the end of the study and the higher glycemia at all time points studied after multiple low-dose STZ in mice transplanted with wild-type islets point toward a higher resistance of STAT-1^−/− islets to the immune consequences of multiple low-dose STZ-induced diabetes. We have previously shown that STAT-1^−/− and wild-type mice are equally sensitive to a single high dose of STZ (7), making it unlikely that the present results are due to specific resistance of STAT-1^−/− islets against the toxic effects of the drug. In the multiple low-dose STZ model, cytokine and chemokine expression by resident inflammatory cells, endocrine and endothelial cells, are proposed to be the detrimental trigger leading to islet immune infiltration causing development of hyperglycemia (20). In this regard, we have shown that STAT-1^−/− islets produce less chemokines compared with C57BL/6 islets when exposed to stress signals (7). Therefore, we hypothesize that our observations in multiple low-dose STZ model can be explained by differences in chemokine expression and recruitment of immune cells by those chemokines in the graft of STAT-1^−/− islets compared with wild-type islets. We have previously demonstrated that PNF is mediated by local inflammation (11), and the hypothesis that STAT-1 is a crucial mediator in the cytokine-induced component of β-cell destruction in vivo is further strengthened by the observation that STAT-1^−/− islets were fully protected against PNF after transplantation in spontaneously diabetic NOD mice. Importantly, iNOS expression, which is positively correlated with the prevalence of PNF in islet transplantation (11), was decreased 8 h posttransplantation in STAT-1^−/− islets grafted in spontaneously diabetic NOD mice. Taken together, the multiple low-dose STZ and the PNF data indicate that cytokine-mediated cell death is an important component of β-cell demise in vivo. Indeed, inflammatory cytokines probably contribute to β-cell death in type 1 diabetes (13). This form of β-cell destruction may be particularly relevant when viral triggers are involved, and transcription factors such as STAT-1 may be crucial messengers in this phenomenon (19).

The absence of prolongation of graft survival by blocking STAT-1 in β-cells indicates that the early beneficial effects of STAT-1 deletion are not sufficient to prevent late and more complex phenomena involved in islet graft destruction. Even additional blocking of IL-1β signaling by IL-1ra and partial blocking of T-cell activation by CsA could not prevent islet graft destruction. Other mediators such as Fas/Fas ligand and perforin/granzyme probably contribute to late islet graft loss (4,21,22). In fact, recent data involving islet-specific CD8+ T-cells from T-cell receptor transgenic NOD8.3 mice indicate that both Fas and perforin are implicated in β-cell destruction (23).

In conclusion, the present findings point to the very diverse mechanisms that contribute to β-cell destruction in vivo. Cytokine-mediated mechanisms clearly contribute to β-cell death, and interfering with cytokine signaling cascades may help create stronger β-cells. Our data suggest, however, that to prevent late β-cell loss, long-term integrated approaches targeting both β-cells and different players in the immune system will be necessary.

This study was funded, in part, by grants from the Flemish Research Foundation (Fonds Voor Wetenschappelijk Onderzoek) Vlaanderen G.0084.02, G.0233.04, and G.0552.06; the Belgium Program on Interuniversity Poles of Attraction initiated by the Belgian State (IAP P6/40); the Katholieke Universiteit Leuven (GOA 2004/10 and EF-05-011); the European Union Project SAVEBETA (contract no. 036903 in the Framework Programme 6 of the European Community); the Belgian Research Foundation (Fonds National de la Recherche Scientifique); the Communauté française de Belgique, Actions de Recherche Concertées (ARC); an aspirant doctoral scholarship (to H.C.); a postdoctoral fellowship (to C.G.); and a clinical fellowship (to C.M.).

We thank W. Cockx, J. Depovere, J. Laureys, and D. Valckx for their technical support.