β-Cell-Specific Glucocorticoid Reactivation Attenuates Inflammatory β-Cell Destruction

Progression and severity of type 1 diabetes is dependent upon inflammatory induction of nitric oxide production and consequent pancreatic β-cell damage. Glucocorticoids (GCs) are highly effective anti-inflammatory agents but have been precluded in type 1 diabetes and in islet transplantation protocols because they exacerbated insulin resistance and suppressed β-cell insulin secretion at the high-doses employed clinically. In contrast, physiological-range elevation of GC action within β-cells ameliorated lipotoxic β-cell failure in transgenic mice overexpressing the intracellular enzyme 11β-hydroxysteroid dehydrogenase type 1 (MIP-HSD1tg/+ mice). Here, we tested the hypothesis that elevated β-cell 11beta-HSD1 protects against the β-cell destruction elicited by streptozotocin (STZ), a toxin that dose-dependently mimics aspects of inflammatory and autoimmune β-cell destruction. MIP-HSD1tg/+ mice exhibited an episodic protection from the severe hyperglycemia caused by a single high dose of STZ associated with higher and sustained β-cell survival, maintained β-cell replicative potential, higher plasma and islet insulin levels, reduced inflammatory macrophage infiltration and increased anti-inflammatory T regulatory cell content. MIP-HSD1tg/+ mice also completely resisted mild hyperglycemia and insulitis induced by multiple low-dose STZ administration. In vitro, MIP-HSD1tg/+ islets exhibited attenuated STZ-induced nitric oxide production, an effect reversed with a specific 11beta-HSD1 inhibitor. GC regeneration selectively within β-cells protects against inflammatory β-cell destruction, suggesting therapeutic targeting of 11beta-HSD1 may ameliorate processes that exacerbate type 1 diabetes and that hinder islet transplantation.

Progression and severity of type 1 diabetes is dependent upon inflammatory induction of nitric oxide production and consequent pancreatic β-cell damage. Glucocorticoids (GCs) are highly effective anti-inflammatory agents but have been precluded in type 1 diabetes and in islet transplantation protocols because they exacerbated insulin resistance and suppressed β-cell insulin secretion at the high-doses employed clinically. In contrast, physiological-range elevation of GC action within β-cells ameliorated lipotoxic β-cell failure in transgenic mice overexpressing the intracellular enzyme 11β-hydroxysteroid dehydrogenase type 1 (MIP-HSD1 tg/+ mice). Here, we tested the hypothesis that elevated β-cell 11beta-HSD1 protects against the β-cell destruction elicited by streptozotocin (STZ), a toxin that dose-dependently mimics aspects of inflammatory and autoimmune tg/ β-cell destruction. MIP-HSD1 + mice exhibited an episodic protection from the severe hyperglycemia caused by a single high dose of STZ associated with higher and sustained β-cell survival, maintained β-cell replicative potential, higher plasma and islet insulin levels, reduced inflammatory macrophage infiltration and increased anti-inflammatory T regulatory cell content. MIP-HSD1 tg/+ mice also completely resisted mild hyperglycemia and insulitis induced by multiple low-dose STZ administration. In vitro, MIP-HSD1 tg/+ islets exhibited attenuated STZ-induced nitric oxide production, an effect reversed with a specific 11beta-HSD1 inhibitor. GC regeneration selectively within β-cells protects against inflammatory β-cell destruction, suggesting therapeutic targeting of 11beta-HSD1 may ameliorate processes that exacerbate type 1 diabetes and that hinder islet transplantation.

INTRODUCTION
Type 1 diabetes is a chronic disease characterized by inflammatory β-cell destruction secondary to an initial autoimmune targeting of the islets (1). Inflammatory macrophages are key to the development and maintenance of islet damage (2). Pro-inflammatory cytokines derived from macrophages and damaged β-cells further suppress β-cell function in part through induction of nitric oxide production (3,4). As type 1 diabetes progresses, pro-inflammatory cytokines inhibit β-cell regeneration, stimulate peripheral insulin resistance and maintain insulitis (1).
Glucocorticoids (GCs) are used clinically due to their potent anti-inflammatory and immunosuppressive effects (5) but were excluded as a treatment for type 1 diabetes and in transplant protocols (6) because they promoted peripheral insulin resistance and suppressed β-cell function at the high-doses employed (7)(8)(9). However, the prevailing dogma that GC action on β-cells is purely deleterious has been increasingly challenged (10)(11)(12)(13)(14). Exposure of normal mouse islets to GCs can improve aspects of secretory function through suppression of inflammatory signaling (10). Moreover, pre-treatment of islets with GCs (11) or localized exposure of transplanted islets to GCs contained within their surrounding implant matrix (12) has shown improved efficacy and graft survival. Crucially, transgenic mice with modest β-cell-specific elevation of the intracellular GC regenerating enzyme 11β-hydroxysteroid dehydrogenase (HSD11b1; 11beta-HSD1; MIP-HSD1 mice) exhibited protection from lipotoxic βcell failure in vivo as a result of increased islet number, arising from a post-developmental effect, and function, due to enhanced secretory capacity and cell survival signaling (14).
The beneficial effects of β-cell-specific 11beta-HSD1 elevation (14) were manifest in a chronic high-fat feeding obesity model. Although obesity is associated with a low-grade inflammation of the islets (15), the protective mechanisms found in MIP-HSD1 islets were not obviously anti-inflammatory (14). Therefore, the impact of intra-β-cell GC regeneration on the processes of cellular damage occurring in profoundly inflammatory contexts relevant to type 1 diabetes remains unknown. To address this we tested the hypothesis that elevated β-cell 11beta-HSD1 protects against the profound β-cell destruction or inflammatory insulitis driven by distinct doses of the β-cell toxin streptozotocin (STZ).

ANIMALS
All experiments conformed to local ethical guidelines of the University of Edinburgh and the UK Home Office Animals (Scientific Procedures) Act (1986). Male MIP-HSD1 tg/+ and C57BLKS/J (KsJ) littermate control mice (14) were housed in standard conditions on a 12 h light/dark cycle and fed standard rodent chow (Special Diet Services, Edinburgh, UK). Age matched 10-12-week-old male mice were used for all the experiments.

STREPTOZOTOCIN TREATMENTS
Mice were injected intraperitoneally with a single bolus of STZ (180 mg/kg/body weight) or for five consecutive days with 40 mg/kg/body weight STZ dissolved in 10 mmol/l sodium citrate (pH 4.5) or vehicle. Blood glucose was measured (OneTouch Ultra, Johnson and Johnson, Bucks, UK) from a tail venesection. Mice were sacrificed at 3 and 10 days (single dose) or 15 days (multiple dose) after injection. Insulin was measured by ELISA (Crystal Chem, Downers Grove, IL, USA).

ISLET ISOLATION AND PREPARATION
Pancreata were digested with collagenase XI (Sigma Aldrich) and islets were hand-picked under a stereomicroscope in Hank's Balanced Salt Solution, 10% FBS (Lonza, Berkshire, UK). Batches of 80 islets were incubated in RPMI-1640 (Gibco, Life Technologies, Paisley, UK), 10% FBS, 6.1 mmol/l d-glucose, 2 nmol/l 11-dehydrocorticosterone with or without 10 mmol/l STZ diluted in sodium citrate 10 mmol/l and with or without L-NAME (Sigma) 5 mmol/l for 72 h on 8 µm inserts (Millipore). Pictures of the islets were taken using a Zeiss microscope and media were collected for measurement of nitric oxide.

NITRIC OXIDE (NO) PRODUCTION
Total NO in the media was assayed as nitrite, the stable breakdown product of NO, using a Sievers chemiluminescence analyzer (Analytix, Sunderland, UK). Islets were homogenized in lysis buffer as described in (14) and protein content evaluated by Biorad assay (BioRad Laboratories, Hercules, CA, USA).

STATISTICS
Data are expressed as mean ± SEM and were analyzed using one-way ANOVA (Newman-Keuls post hoc test).

MIP-HSD1 tg/+ MICE RESIST HIGH-DOSE STZ-INDUCED HYPERGLYCEMIA
We began by administering a high-dose of STZ (180 mg/kg body weight) known to completely ablate β-cell function (16). Highdose STZ caused marked and comparable hyperglycemia by 2 days in MIP-HSD1 tg/+ and control non-transgenic littermates (KsJ), indicating comparable ablation of β-cell function. However, after day 3, during the inflammatory-response phase of islet destruction, MIP-HSD1 tg/+ mice began to exhibit episodic phases of significantly less severe hyperglycemia than KsJ mice (Figures 1A,B) suggestive of partial recovery of function in existing β-cells and/or spontaneous regeneration of new β-cells. Plasma insulin levels were significantly higher in MIP-HSD1 tg/+ than in KsJ mice at day 3 and 10 ( Figure 1C), consistent with their residual islet insulin staining and further supporting an islet-specific β-cell recovery ( Figure 1D). Notably, circulating corticosterone levels were markedly elevated by high STZ, but to a similar degree in KsJ and MIP-HSD1 tg/+ mice (nmol/l: KsJ vehicle: 131 ± 28, KsJ high STZ: 535 ± 155, MIP-HSD1 tg/+ vehicle: 138 ± 45 MIP-HSD1 tg/+ high STZ: 683 ± 125, no significant effect of genotype), supporting a role for local β-cell GC regeneration as the underlying driver of genotype-specific effects.

MIP-HSD1 tg/+ MICE MAINTAIN HIGHER β-CELL MASS AND REPLICATIVE CAPACITY AFTER STZ
The remarkable resilience of MIP-HSD1 tg/+ β-cells against highdose STZ could be due to higher β-cell survival or increased spontaneous β-cell regeneration. Of note, there was low and comparable β-cell proliferation (~2% of total islet cell number double-positive for the proliferation marker Ki67 and the β-cell marker PDX1) in vehicle-treated KsJ and MIP-HSD1 tg/+ mice (Figure 2A, left lane and quantification Figure 2B). Islet PDX1positive β-cell number was severely reduced by day 3 and continued to fall by day 10 to undetectable levels in KsJ mice treated with STZ. Replicating Ki67 single-positive cells in islets of STZtreated KsJ mice are likely infiltrating immune cells. Ki67/PDX1 double-positive cells were undetectable in KsJ islets indicating a complete loss of β-cell replicative capacity (Figure 2A, upper row, middle, and right lane and quantification, Figure 2B). Despite a marked reduction in PDX1-positive cells, a substantial β-cell number remained in STZ-treated MIP-HSD1 tg/+ islets, although the comparable hyperglycemia between genotypes from day 2 to 3 indicates these β-cells underwent a period of initial secretory dysfunction (Figure 1). MIP-HSD1 tg/+ mice also maintained their Ki67/PDX1 double-positive cell number in their remaining islets (Figure 2A bottom row and quantification Figure 2B) suggesting that their β-cell replicative capacity is maintained after STZ. This could not be accounted for by β-cell neogenesis from potential progenitor cell types, as assessed with SOX9 Frontiers in Endocrinology | Genomic Endocrinology and NEUROG3 (17,18) immunostaining that did not differ by genotype 3 or 10 days after STZ (Figures S1 and S2 in Supplementary Material). Apoptosis was not detected in islets of either genotype by a number of methods suggesting that this process does not account for the differences between the genotypes under these experimental conditions ( Figure S3 in Supplementary Material).

MIP-HSD1 tg/+ ISLETS HAVE REDUCED MACROPHAGE AND INCREASED T REGULATORY CELL INFILTRATION AFTER STZ
Inflammatory macrophage infiltration is an early event in autoimmune (2) and STZ-induced islet damage (19). MIP-HSD1 tg/+ islets had fewer (~40%) infiltrated macrophages, as assessed by the macrophage marker Mac-2, compared to KsJ controls at 3 and 10 days after the single high-dose STZ injection (Figure 3A and quantification, Figure 3B).

MIP-HSD1 tg/+ MICE RESIST MULTIPLE LOW-DOSE STZ-INDUCED HYPERGLYCEMIA
To test responses to a more subtle diabetic insult that recapitulates some of the inflammatory and autoimmune aspects of type 1 diabetes (21, 22), a low-dose of STZ (40 mg/kg/BW) was administered for five consecutive days. KsJ mice showed significant hyperglycemia from day 4 that reached a modestly diabetic plateau by day 12 (Figure 4A), had reduced islet insulin staining ( Figure 4B) and increased islet macrophage numbers, whereas MIP-HSD1 tg/+ mice maintained normal glycemia and pancreatic morphology and displayed abrogated macrophage infiltration ( Figure 4C, and quantification, Figure 4D).

DISCUSSION
Modest elevation of 11beta-HSD1 activity in β-cells (14) has conferred improved β-cell survival and a sustained capacity for spontaneous β-cell regeneration in the context of severe inflammatory β-cell destruction. Although the protection against hyperglycemia is modest across the short-time course of the present studies, the continued survival and replenishment of functional β-cells after high-dose STZ points to a remarkable and unexpectedly effective protective role for local GC regeneration. Hyperglycemia induced by multiple low-dose STZ, a regimen that invokes some of the autoimmune and inflammatory processes (21,22,24) to that found in type 1 diabetes, was completely prevented in MIP-HSD1 tg/+ mice. The protective effect of elevated β-cell 11beta-HSD1 involves both increased β-cell survival and maintained β-cell replicative potential. Improved MIP-HSD1 tg/+ islet survival is mediated by GC-mediated suppression of pro-inflammatory NO production (3,4), an effect of GCs also found in the vasculature (25). Increased expression of heat shock and other cellular stress-resistance pathways in MIP-HSD1 tg/+ islets (14) is consistent with their augmented survival, and may counteract triggering of more aggressive inflammatory responses after β-cell insult.
Under normal physiological conditions, β-cell mass is maintained through slow rates of renewal and turnover (26). Hyperglycemia can prompt islet mass compensation predominantly through hypertrophy of existing β-cells (27), β-cell proliferation (26), and neogenesis from progenitors, at least in pancreatic injury models (18). However, we found no evidence for altered ductal (SOX9) or islet endocrine (NEUROG3) progenitor cells as the basis of maintained β-cell mass in STZ-treated MIP-HSD1 tg/+ mice, in agreement with recent findings that βcell neogenesis does not come from ductal cell progenitors in the adult pancreas (17). Thus, altered GC regeneration selectively impacts upon β-cell replication in the STZ-injured adult pancreas. Moreover, MIP-HSD1 tg/+ islets exhibit a hyper-functionality associated with increased Cdkn1a (P21) expression suggesting accelerated functional maturation of MIP-HSD1 β-cells (14), consistent with the role of GCs in terminal differentiation. Newly generated MIP-HSD1 β-cells likely achieve functionality more rapidly than normal β-cells. In support of this notion, overexpression of the pro-differentiation factor P21 in β-cells promotes resistance to high-dose STZ by increasing progenitor differentiation (28). Notably, apoptosis was undetectable in STZ-treated KsJ mouse islets and pancreas, despite a strong signal in our positive control spleen tissue ( Figure S3B in Supplementary Material). It cannot be ruled out that our time points missed a significant incidence of apoptosis, as noted by others (29) or that rapid turnover of apoptotic cells markers [within minutes; (30)] may have caused the signal to fall below our limit of detection. Experimental designs more suited to inducing β-cell apoptosis will be www.frontiersin.org FIGURE 5 | Isolated MIP-HSD1 tg/+ islets show attenuated induction of nitric oxide production and resistance to disintegration in culture after STZ challenge. Islets from 12-week-old mice were isolated and incubated in vitro with vehicle or STZ (10 mmol/l) with or without the nitric oxide synthase inhibitor L-NAME (5 mmol/l) in the presence of 2 nmol/l 11-dehydrocorticosterone 11beta-HSD1 substrate for 72 h. (A) Analysis of nitric oxide (NO) production in media after 72 h incubation with vehicle (−), STZ or STZ + L-NAME from islets from KsJ (white bars) or MIP-HSD1 tg/+ mice (black bars). Values represent mean ± SEM. Differences were analyzed by one-way ANOVA, Newman-Keuls test, ***P < 0.001 for STZ vs. Veh; † † † P < 0.001 for MIP-HSD1 tg/+ STZ vs. KsJ STZ, § § P < 0.01 STZ + L-NAME vs. STZ (n = 6). (B) NO production in media of isolated islets from MIP-HSD1 tg/+ mice treated with STZ with or without 11beta-HSD1 inhibitor UE2316 (1 µmol/l) for 72 h. Values represent mean ± SEM, differences analyzed by one-way ANOVA, Newman-Keuls test, ***P < 0.001 for STZ vs. Veh; **P < 0.01 for STZ + UE2316 vs. STZ (n = 6). needed to determine the contribution of altered β-cell GCs to this process.
Macrophage infiltration in diabetic islets is an early event (1-4). Depletion (19) or inactivation (2) of macrophages prevents the progression of type 1 diabetes, highlighting the importance of the inflammatory mechanism per se in islet destruction. Reduction of macrophage infiltration in MIP-HSD1 tg/+ islets is consistent with a reduced inflammatory insult that will work, at least in part, through longer-term curtailment of cytokine-mediated NO generation long after the STZ has been metabolized. GCs suppress a number of distinct pro-inflammatory signaling pathways that will also contribute to overall improvement in islet function (5). The higher Treg cell influx into MIP-HSD1 tg/+ pancreas is consistent with increased islet GC levels influencing infiltrating macrophages to signal for resolution of inflammation (31). Increased T regulatory cell numbers also suggests that a greater suppression of pathogenic pro-inflammatory CD8 + Th1 cells and, intriguingly, may indicate enhanced suppression of self-antigens (32) and autoimmune responses over a longer time frame, which can only be addressed in future work in models such as the NOD mouse strain and in transplant models.
Of clinical significance, new technological advances in biohybrid β-cell transplant methods have achieved improvement of graft β-cell survival using slow-release of local corticosteroids, reportedly as a consequence of local immunosuppression (12). Our previous (14) and current data suggest that the secretory function and survival of encapsulated or transplanted human β-cells, which appear to express 11beta-HSD1 (33) could be potentially augmented through pre-treatment or local release of low-dose cortisone, thus, avoiding the deleterious effects of high dose or systemic GC administration that has hitherto precluded their use in this context.