The Islet Estrogen Receptor-α Is Induced by Hyperglycemia and Protects Against Oxidative Stress-Induced Insulin-Deficient Diabetes

The female steroid, 17β-estradiol (E2), is important for pancreatic β-cell function and acts via at least three estrogen receptors (ER), ERα, ERβ, and the G-protein coupled ER (GPER). Using a pancreas-specific ERα knockout mouse generated using the Cre-lox-P system and a Pdx1-Cre transgenic line (PERαKO−/−), we previously reported that islet ERα suppresses islet glucolipotoxicity and prevents β-cell dysfunction induced by high fat feeding. We also showed that E2 acts via ERα to prevent β-cell apoptosis in vivo. However, the contribution of the islet ERα to β-cell survival in vivo, without the contribution of ERα in other tissues is still unclear. Using the PERαKO−/− mouse, we show that ERα mRNA expression is only decreased by 20% in the arcuate nucleus of the hypothalamus, without a parallel decrease in the VMH, making it a reliable model of pancreas-specific ERα elimination. Following exposure to alloxan-induced oxidative stress in vivo, female and male PERαKO−/− mice exhibited a predisposition to β-cell destruction and insulin deficient diabetes. In male PERαKO−/− mice, exposure to E2 partially prevented alloxan-induced β-cell destruction and diabetes. ERα mRNA expression was induced by hyperglycemia in vivo in islets from young mice as well as in cultured rat islets. The induction of ERα mRNA by hyperglycemia was retained in insulin receptor-deficient β-cells, demonstrating independence from direct insulin regulation. These findings suggest that induction of ERα expression acts to naturally protect β-cells against oxidative injury.


Introduction
The female steroid, 17b-estradiol (E2), is important for pancreatic b-cell function in mammals [1][2][3][4].E2 acts through at least three estrogen receptor(ER)s in b-cells, ERa ER b and the Gprotein coupled ER (GPER).These ERs are expressed in rodent and human b -cells in both sexes, where they exhibit a predominant extranuclear localization [3,5].The islet ERa is important for enhancing insulin biosynthesis in vivo via an extranuclear ERa-dependent mechanism that amplifies the effect of glucose in stimulating the insulin gene promoter [6,7].The islet ERa also suppresses excess de novo lipogenesis, which prevents glucolipotoxic b-cell failure in rodent models of type 2 diabetes (T2D) [8].E2 also acts as a survival hormone that prevents b-cell apoptosis in vivo in both sexes at physiological concentrations.This protection is lost in mice globally deficient in ERa [9].In cultured mouse and human islets, E2 protection is mediated mainly via ERa and GPER, and it protects from diabetes-associated injury resulting from oxidative stress and pro-inflammatory cytokines [5,[9][10][11].Further, during pancreatic islet transplantation, use of an ERa-selective agonist enhances human islet graft survival, thus protecting islet functional mass [12].Overall, global expression of ERa is necessary for islet survival in mice, and pharmacological activation of ERa protects islet survival in culture and following in vivo treatment.Nonetheless, the direct and singular impact of ERa in islet b-cells on islet survival in vivo -without contribution from the effects of ERa action in other tissues-has not been addressed.In this study we used the PERaKO 2/2 mouse to examine the role of islet ERa in islet survival from alloxan induced-oxidative stress in vivo.

Generation of mutant mice and animal care
Pancreas specific ERa knockout mice were generated using the Cre-lox-P system and a Pdx1-Cre transgenic line (PERaKO 2/2 ) as previously described [7].Pdx1-Cre mice were bred onto the creinducible Rosa26-LacZ line at the University of Michigan.Animal had free access to food and water.They were kept on a 12-h light/ dark cycle.All animal experiments were approved by Northwestern University or University of Michigan Institutional Animal Care and Use Committee.

Induction of experimental diabetes and tissue collection
Diabetes was induced in 10-12 week-old female and male mice by a single intraperitoneal (IP) injection of 150 mg/kg of alloxan (ALX) (2,4,5,6-Tetraoxypyrimidine) (Sigma-Aldrich) freshly prepared in sterile cold saline (0.9%).Mice ERalox +/+ were used as control for PERaKO 2/2 mice.Blood glucose was measured every 48 h after ALX injection using One Touch Ultra Glucose Monitor (Lifescan).At day 11 after ALX injection, mice were killed and blood and pancreata were collected.

Pancreas insulin concentration
Tails of the pancreata were collected, weighed, and homogenized in acid/ethanol.Then, pancreas homogenates were centrifuged, and supernatants were used to measure pancreas insulin concentration by radioimmunoassay (Linco) as described [7].

Plasma insulin concentrations
Plasma insulin concentrations were measured by ELISA (Millipore).

Pancreas immunohistochemistry
Deparaffinized pancreatic sections (5 mm) were blocked for 30 min with blocking solution (20% Fetal Bovine Serum + 2% Roche Blocking Reagent).Sections were incubated overnight with primary antibodies and 1-2 h with secondary antibodies at room temperature with the following primary antibodies: guinea pig anti-human insulin (1:1000; Linco Research), rabbit anti-glucagon (1:1000, Linco Research), rat anti-mouse CD31 (1:400; BD Biosciences).Secondary antibodies FITC-conjugated donkey anti-guinea pig, CY3-conjugated donkey anti-rabbit, AMCAconjugated donkey anti-guinea pig, and CY3-conjugated goat anti-rat (Jackson ImmunoResearch Laboratories) were used at concentrations recommended by the manufacturer.The nuclei were stained with DAPI (Invitrogen, Molecular Probes).Images were obtained with either Nikon Eclipse E400 microscope or Tissue Genostics Tissue/Cell High Throughput Imaging and Analysis System at Northwestern University Cell Imaging Facility.

Brain immunohistochemistry
Perfusion and immunohistochemistry were performed as previously described [13].Briefly, mice were anesthetized with a lethal dose of intraperitoneal pentobarbital (150 mg/kg) and transcardially perfused with sterile PBS and then either 4% paraformaldehyde or 10% formalin.Brains were removed, postfixed overnight and dehydrated in a 30% sucrose solution.

Cell counts and statistic
Stained sections were imaged using Leica microscope using 10X and 20X air objectives and processed using Adobe Photoshop CSII (Adobe Systems, San Jose, CA).Photoshop was only used to overlay matched images in different RGB channels such that duallabeled cells would become apparent and could be quantified as described [13].
Calculation of pancreatic b-cell mass b-cell area was measured in insulin-stained 5 mm thick pancreatic sections.Three to four sections per tissue were randomly chosen for morphometric analysis.Insulin positive area was determined by using ImageJ 1.37v program.To calculate bcell mass (mg), insulin positive area was divided by pancreas area and then multiplied by pancreas weight.

Calculation of vessel density in islets
Blood vessel density was calculated by dividing the mouse-CD31-positive area by the insulin-positive islet area by using ImageJ 1.37v program.

Islets isolation
At the end of infusion, islets were isolated by pancreas collagenase digestion as described [14].

Rat islets culture
Wistar rat islets were pre-cultured for a week in serum-free RPMI medium supplemented with 5 g/L BSA (37uC, humidified atmosphere containing 5% CO2).Islets were further cultured for 18 h or 1 week in the same medium containing 5, 10 or 30 mM glucose (medium was renewed every other day) [15].

Mouse model of moderate hyperglycemia
To study ERa expression under mild hyperglycemia conditions, a 4 days glucose infusion in mice was performed as is described in [16].Briefly, C57bl/6J male mice of 8-12 weeks-old and 20-25 g received a 4 days infusion of saline or 50% glucose.After this, mice were anesthetized and islets isolated by digestion with 1.7 ml/cc Collagenase P (Sigma) [9].

bIRKO cell culture
We used insulin-secreting cell lines established from groups of bIRKO, and Lox control mice as was described previously [17,18].Cells were maintained at 37uC and 5% CO2 in Dulbecco's modified Eagle's medium (DMEM) containing 25 mM glucose, 10% fetal bovine serum, and penicillin and streptomycin.Experiments were performed using 80-90% confluent cells.Lox and bIRKO cells were seeded in 6-well plates and incubated for 24 h to recover.Cells were first washed with PBS before incubating in 16.7 or 33 mM glucose in DMEM containing 10% serum and penicillin and streptomycin for 3 days.

Q-PCR
Total RNA was extracted using RNeasyH Micro kit (Qiagen, Valencia, CA) for islets or RNeasyH kit (Qiagen, Valencia, CA) for cells according to the manufacturer's protocols.cDNA was prepared from 1 mg of total RNA using the High Capacity cDNA Reverse Transcription Kit (Invitrogen) with random hexamer primers, according to the manufacturer's instructions Real-time PCR amplification of ERa and Tbp (TATA-box binding protein) cDNAs was carried out i on a CFX96 using iQ-SYBR green supermix (Bio-Rad, Hercules, CA).Results were normalized to TBP expression and expressed as arbitrary units.Primer sequences are the following: Rat islets: 59ACCCTTCACCAATGACTCC-TATG-39 and 59-TCAGCATTTCTGGCACGAAGT-39for TBP and -59AATTCTGACAATCGACGCCAG39 and 59-GTGCT-TCAACATTCTCCCTCCTC-39 for Era.Mouse islets and cells: 59-ACCCTTCACCAATGACTCCTATG-39 and 59-ATGATG-ACTGCAGCAAATCGC-39 for TBP and 59-GCTTCTCTT-GGCCTGTACTT-39 and 59-CTCTCCCAGTTTCCACATC-TT-39 for ERa

Statistical analysis
Data are presented as mean 6 SEM unless otherwise stated.Data were analyzed by Student's t test.A value of p,0.05 was considered statistically significant.

Recombination of ERa in hypothalamic neurons of PERaKO 2/2 mice
To investigate the role of pancreatic ERa on b-cell biology in vivo, we used PERaKO 2/2 mice in which ERa was inactivated in all pancreatic lineages using a Pdx1-Cre transgenic mouse [7,8].Because Pdx1-Cre transgenic mice were reported to promote recombination in nutrient sensing hypothalamic neurons [19], we first sought to determine whether recombination of ERa occurs in the hypothalamus of PERaKO 2/2 mice.Accordingly, using a transgenic Pdx1-Cre/LacZ mouse [19], we observed that Pdx1 is co-expressed with ERa in ,26% of neurons of the ventromedial hypothalamus (VMH), ,17% of neurons in the preoptic area (POA), and ,15% of neurons of the arcuate nucleus (ARC) (Fig. 1A-B).Female PERaKO 2/2 mice exhibited a 20% decrease in the number of ERa positive cells in the ARC, without a parallel decrease in the VMH (Fig. 1D-F).They also exhibited decreased fertility (data not shown), suggesting that ERaexpression was also decreased in the POA.

No alteration in islet vascularization in absence of ERa
E2 stimulates angiogenesis and promotes endothelial cell recovery after injury [13][14][15][16][17].We previously observed that estrogens improve islet revascularization during islet transplantation [12].Thus, prior to exploring islet predisposition to oxidative stress, we sought to determine whether islet vascularization was altered in female PERaKO 2/2 mice.Because loss of ERa in bcells or in endothelial cells can alter endothelial cell function via paracrine or endocrine mechanisms, respectively, we studied vascular density in PERaKO 2/2 and mice globally deficient in ERa (ERaKO 2/2 ).When we quantified the endothelial cell area in pancreas section using the mouse endothelial cell marker CD31, we observed no difference in islet vascularization among ERaKO 2/2 , PERaKO 2/2 and female control mice (Fig. 2).The absence of islet vascular defects demonstrated that ERa was not essential for islet angiogenesis in mice.

The absence of islet ERa predisposes to oxidative stressinduced diabetes in mice
We induced oxidative stress in b-cells in vivo using a single highdose injection of alloxan (ALX; 150 mg/kg of body weight), which augments the generation of reactive oxygen species (ROS) in pancreatic islets [20].We initially observed that female C57BL/6 mice were protected from ALX-induced diabetes (Fig. 3).Next, we induced oxidative stress in b-cells of PERaKO 2/2 female mice.In basal conditions (time = 0, prior to ALX injection), control and PERaKO 2/2 female mice displayed similar blood glucose (Fig. 4A-B) and insulin concentrations (data no shown).They also exhibited normal islet architecture, with insulin-producing bcells in a central location and glucagon-producing a-cells at the periphery (Fig. 4D).PERaKO 2/2 female mice showed a trend toward decreased pancreatic insulin concentration, an observation that was consistent with the known effect of ERa in stimulating insulin synthesis [6,7].Following exposure to ALX, control female mice showed relative protection compared to PERaKO 2/2 female mice.Control female mice displayed only a minor increase in blood glucose despite hypoinsulinemia and an 87% decrease in b-cell mass and pancreatic insulin concentration (Fig. 4C-F).This finding was consistent with the fact that only 20% of b-cells are needed to maintain euglycemia [12].In contrast, relative to controls, exposure of PERaKO 2/2 female mice to ALX, induced marked hyperglycemia and insulin deficiency that resulted from a more severe b-cell destruction (97%) and decrease in pancreatic insulin concentrations (Fig. 4A-F).Thus, PERaKO 2/2 female mice exhibited a predisposition to alloxan-induced b-cell destruction.Note that we did not observe differences in a-cell density between alloxan-injected control and PERaKO 2/2 female mice.
Regarding males, control and PERaKO 2/2 mice were normoglycemic and normoinsulinemic in basal conditions (Fig. 5A-C), and displayed normal islet architecture (Fig. 5D).After ALX exposure, both control and PERaKO 2/2 male mice developed hyperglycemia and insulin deficiency and exhibited decreased b-cell mass and pancreatic insulin concentrations.However, the reduction in all of these parameters was more dramatic in PERaKO 2/2 than in control mice (Fig. 5A-F).In addition, after E2 administration, we observed partial protection from alloxan-induced b-cell destruction and insulin deficiency in both controls and PERaKO 2/2 male mice (Fig. 5A-F).Thus, as observed in females, male PERaKO 2/2 mice exhibited a predisposition to alloxan-induced b-cell destruction (although to a lesser extent), but estrogen still provided some protection from alloxan in the absence of islet ERa.Note that unlike in the case of Fig. 3, experiments of ALX injections described in Fig. 4 and 5 were performed independently in male and female mice.Therefore, males and female mice described in in Fig. 4 and 5 are not comparable with regard to the female protection from diabetes observed in Fig. 2.

Altered islet ERa expression during hyperglycemia and hyperinsulinemia
Having determined that islet ERa is important to oxidative stress protection in vivo, we next sought to determine whether ERa mRNA expression was altered in islets during hyperglycemiainduced oxidative stress in vivo.We used two established rodent models of glucotoxicity and glucolipotoxicity.We first studied ERa mRNA expression in islets from non-diabetic Wistar rats that received a 72 h glucose and intralipid co-infusion to mimic glucolipotoxicity (mean glucose 15 mM) [14].Under these conditions, hyperglycemia was associated with increased ERa mRNA expression in 2 month-old rat islets (Fig. 6A).However, hyperglycemia did not increase ERa mRNA in islets from 6 month-old rats.We next studied ERa mRNA expression in a mouse model of mild hyperglycemia that was achieved by a 4-day glucose infusion (mean glucose 7 mM) [16].In this model, we observed no increase in islet ERa mRNA (Fig. 6B).To ascertain whether ERa mRNA induction under severe hyperglycemic conditions resulted from a direct glucose effect on islets, we further studied ERa mRNA expression in Wistar rat islets cultured one week in hyperglycemic conditions [15].ERa expression was increased when glucose was raised from 5 mM to 10 mM, but there was no further increase at 30 mM (Fig. 6C).Therefore, moderate to severe hyperglycemia [15] is associated with increased ERa mRNA expression in vitro and in vivo in rats.

ERa expression in insulin resistant b-cells
We hypothesized that the increased ERa mRNA expression in islets exposed to hyperglycemia could be due to the stimulatory effect of high glucose or to the impact of elevated insulin on the IR in the islets.To address this question, we quantified ERa expression in b-cells isolated from normal and b-cell IR knockout (bIRKO) mice [18].These islets were cultured in hyperglycemic conditions to increase insulin secretion.Consistent with the effect of glucose described above (Fig. 5A), ERa mRNA expression was increased in both control (lox/lox) and bIRKO b-cells when glucose was increased from 16.7 mM to 33 mM (Fig. 6D).However, at both glucose concentrations, ERa mRNA expression was higher in bIRKO compared to lox/lox b-cells, demonstrating that insulin action in b-cells inhibits ERa mRNA expression.

Discussion
Having established that ERa is not essential to islet angiogenesis in mice, we focused on the role of ERa in protecting islets from glucotoxicity and oxidative stress in vivo and report that both male and female mice lacking ERa selectively in the pancreas are more susceptible to alloxan-induced b-cell destruction, insulin deficiency, and hyperglycemia.Although these experiments demonstrated a mild decrease in ERa mRNA expression in hypothalamus of PERaKO 2/2 mice, the absence of alteration in energy homeostasis [7,8] and the pancreas-specific phenotype observed in this model both suggest that the PERaKO 2/2 phenotype results exclusively from pancreatic elimination of ERa.Since alloxan induces oxidative stress, these findings demonstrate that normal islet ERa expression is required to protect b-cells from oxidative stress-induced apoptosis in vivo in both sexes.The harmful effect of ERa deletion is more pronounced in female mice, presumably as a result of higher E2 serum concentrations that are required to activate the islet ERa in this gender.Nonetheless, the negative effect of ERa deletion on islet cells is also observed in males, demonstrating that islet protection by ERa is sex independent.We previously reported that mice of both sexes globally lacking ERa (aERKO 2/2 ) were predisposed to streptozotocininduced b-cell apoptosis and insulin-deficient diabetes [9].However, the beneficial actions of estrogen on glucose homeostasis results from the combined actions of ERa in different tissues [2].Thus aERKO 2/2 mice globally lacking estrogen action in skeletal muscle, adipose tissue, and the brain become obese and insulin resistant as well as mildly hyperglycemic.This could produce additional b-cell stress that would synergize streptozotocin toxicity to alter b-cell survival.The current study demonstrates that loss of ERa selectively in islets -while ERa is normally expressed in other tissues -is sufficient to induce b-cell destruction in the presence of another b-cell stress and without any influence of altered body weight [2,7].Given the mild decrease in ERa expression observed in PERaKO 2/2 hypothalamic ARC, this abnormality is unlikely to play a role in the PERaKO 2/2 phenotype.
We previously reported that ERa gene dosage plays a role in the islet protection from streptozotocin injury because heterozygous aERKO 2/2 mice of both sexes were predisposed to streptozotocin-induced diabetes [9].Thus, increased ERa expression could function to protect b-cells against oxidative stress.To evaluate this hypothesis, we used established models of glucolipotoxicity and moderate hyperglycemia.We observed that in both cultured rat islets and in mouse islets in vivo, moderate to severe hyperglycemia increased ERa mRNA.In cultured mouse islets and MIN6 cells, short term exposure to high glucose is also associated with an increase in ERa mRNA expression [21].Overexpression of ERa prevents apoptosis in PC12 neuronal cells, [22] and in the SK-N-MC human neuroblastoma cell line [23].In contrast, in a model of moderate hyperglycemia, ERa mRNA is not upregulated in islets.
During hyperglycemia, is ERa induced by glucose itself or by insulin?In IR-deficient b-cells cultured in high glucose, ERa mRNA was increased compared to normal cells.This demonstrates that compared to hyperglycemia, insulin action in b-cells is unlikely to play a direct role in inducing ERa mRNA in b-cells.Thus, the induction of ERa expression by hyperglycemia could function as a b-cell protection against oxidative injury when hyperglycemia reaches a threshold beyond which oxidative injury occurs.Further, glucolipotoxicity, upregulates ERa mRNA in young rats, but this feature is lost in older animals.Since ERa improves b-cell survival [3,5,9], the loss of ERa induction in old islets may alter their resistance to diabetic injuries, as we observe in the PERaKO 2/2 mouse.This weakness may further increase bcell susceptibility to oxidative injuries such as glucotoxicity, setting the stage for b-cell failure in old age.
ERa protection from oxidative stress could involve a combination of rapid antiapoptotic actions that are independent of nuclear events and that potentially lead to alteration in protein phosphorylation [3,24].Alternatively, it could involve a more classical genomic mechanism that induces an anti-inflammatory cascade via expression of the liver receptor homolog [25].
In conclusion, ERa mRNA expression is induced in islets from young mice by exposure to hyperglycemia and oxidative stress, and mice of both sexes that selectively lack ERa in the islets are susceptible to both oxidative stress in b-cells and insulin-deficient diabetes.