SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

Objective

Beta-site amyloid precursor protein cleaving enzyme (BACE1) is highly expressed in pancreatic β-cells. The BACE1 gene is located in a region associated with a high diabetes risk in PIMA Indians.

Design and Methods

INS-1E cells were used to study the impact of siRNA-mediated BACE1 knockdown and glucose metabolism was characterized in Bace1–/– mice. BACE1 gene was sequenced in DNA samples from 48 subjects and 13 representative single nucleotide polymorphisms (SNPs) were then genotyped for association studies in 1,527 Caucasians.

Results

Reduction of Bace1 expression results in a significant decrease in insulin mRNA expression in INS-1E cells. Bace1–/– mice display significantly lower body weight, lower plasma insulin concentrations, but normal glucose tolerance and insulin sensitivity. In a case-control study including 538 healthy controls and 989 patients with type 2 diabetes (T2D), one SNP (rs535860) was significantly associated with T2D (P < 3.5 × 10−5, adjusted for age, sex, and BMI).

Conclusions

Reduced Bace1 expression causes impaired insulin expression in pancreatic β-cells of Bace1–/– mice, suggesting that BACE1 plays a role in the regulation of insulin biogenesis. The functionally relevant rs535860 SNP may decrease BACE1 expression by creating a new miR-661 binding site and could therefore contribute to T2D development.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

Beta-site APP-cleaving enzyme 1 (BACE1) is an enzyme responsible for initiating β-amyloid generation, which is the major constituent of amyloid plaques in Alzheimer's disease (AD) brain and is likely to play a central role in the pathogenesis of this neurodegenerative disorder [1].

Molecular cloning of the β-secretase has been reported in parallel by five independent groups, naming the enzyme either BACE [2], β-secretase [3], Asp2 [4, 5], or memapsin 2 [6]. BACE1 is expressed in the brain, in the exocrine pancreas and pancreatic β-cells [2, 5, 7-10].

BACE1 colocalizes with insulin in the human pancreas and pharmacological inhibition of BACE results in decreased insulin protein content [7]. However, this effect has been attributed to inhibition of BACE2, a homologue of BACE1 that belongs to the same family of transmembrane aspartic proteases [2]. BACE1 deficient mice (Bace1–/–) were initially described without any major deficiencies in development, physiology and behavior [11-13]. Subsequent studies have demonstrated that disruption of the BACE1 gene is associated with mild impairments in spatial learning and memory [14] as well as with a more anxious phenotype [15]. Recently, Meakin et al. reported that mice lacking Bace1 are lean, protected against diet-induced obesity and display higher degree of peripheral insulin sensitivity compared to controls [16]. Bace1–/– mice are protected against higher expression and activity of BACE1 in skeletal muscle and liver in response to high fat diet (HFD), which is associated with improved glucose tolerance [16]. Moreover, in humans the BACE1 gene is located on chromosome 11q23, which has been identified as a region associated with a high risk for diabetes in PIMA Indians [17]. Taken together these data suggest that BACE1 may play a role in the pathophysiology of type 2 diabetes (T2D). We therefore tested the hypothesis that genetic variants in the BACE1 gene are related to T2D by evaluating the association between representative variants in the BACE1 gene and diabetes related traits in a case control study of 1,527 individuals. To further dissect the contribution of BACE1 on pathogenic factors in the development of T2D, we studied glucose metabolism and insulin sensitivity in Bace1–/– mice as well as siRNA based reduction of Bace1 in INS-1E cells in vitro.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

Subjects

Nine hundred and eighty-nine unrelated patients with T2D and 749 healthy subjects were recruited at the University Hospital in Leipzig, Germany. The ethics committee at the Medical Faculty of the University of Leipzig specifically approved this study (#031-2006) and all subjects gave written informed consent before taking part in the study.

The healthy subjects had a mean age 47.2 ± 14.6 years and mean body mass index (BMI) 27.4 ± 5.3 kg m−2. Patients with T2D had a mean age 64.4 ± 10.4 years and mean BMI 29.9 ± 5.4 kg m−2. In addition, oral glucose tolerance test (OGTT) and fasting plasma insulin measurements were performed in all nondiabetic subjects as described elsewhere [18]. Two hundred eleven out of 749 subjects had impaired glucose tolerance. Because impaired glucose tolerance can be considered a prediabetic state, only the remaining 538 subjects with normal glucose tolerance were included as healthy controls in the T2D case-control study. In a subgroup of 403 nondiabetic subjects, body fat content was measured by dual X-ray absorptiometry (DEXA). Insulin sensitivity was assessed with the euglycemic-hyperinsulinemic clamp method as previously described [19]. All studies were approved by the ethics committee of the University of Leipzig and all subjects gave written informed consent before taking part in the study.

Cell culture and BACE 1 siRNA

The clonal β-cell line INS-1E, derived from parental INS-1 cells [20] were provided by Jürgen Klammt (University of Leipzig). INS-1E cells were isolated from INS1 cells based on both their insulin content and their secretory response to glucose and cultured as described [20]. INS-1E cells were cultured in a humidified atmosphere containing 5% CO2 in complete medium composed of RPMI 1640 with previously described supplements and a glucose concentration of ∼12 mM [20]. Before the glucose stimulated insulin secretion (expression) experiments, cells were maintained for 2 h in glucose-free culture medium. The cells were then washed twice and preincubated for 30 min at 37°C in glucose-free Krebs-Ringer bicarbonate HEPES buffer. siRNA specifically targeting Bace1 was synthesized using target sequence 5′-AAGCTTTGTGGAGAT GGTGGA-3′ (Qiagen Hilden, Germany). About 500,000 INS-1E cells per well were plated in 12-well dishes and transfected (HiPerFect Transfection Reagent, Qiagen Hilden, Germany) 48 h prior to further analysis with 5.4 nM siRNA or the pcDNA3 plasmid (AF190725).

Stimulation of insulin expression

Nearly 48 h after transfection, cells were washed with KRBH buffer and incubated with KRBH for 30 min at 37°C. Intracellular insulin was extracted by incubation in 0.18M HCl in 70% ethanol for 4 h at 4°C and quantified by ELISA (DRG Instruments, Marburg, Germany). Plasma insulin concentrations were measured with an ultrasensitive ELISA (Crystal Chem, Downers Grove, IL). Experiments were performed in triplicates.

RNA isolation and RT-PCR

Total RNA from INS-1E cells and mouse pancreatic islets (n = 7 per genotype) was isolated using RNeasyMini Kit (Qiagen, Hilden, Germany). RT-PCR was performed with the LightCycler FastStart DNA Master SYBR Green I Kit (Roche, Mannheim, Germany). Insulin and Bace1 gene expression was calculated relative to the expression of both hypoxanthine-guanine phosphoribosyltransferase (Hprt) and 18S rRNA using the delta Ct (2−ΔΔCt) method [21]. The following primers were used: Bace1: 5′-TTTGTGGAGATGGT GGACAA-3′ (forward) and 5′-CCTGGGTGTAGGGCACATA-3′ (reverse); insulin 5′-GGGGAACGT GGTTTCTTCT-3′ (forward) and 5′-AGTGGTGGACTCAGTTGCAG-3′ (reverse); Hprt 5′-GCAGACTTTGCTTTCCTTGG-3′ (forward) and 5′-TCCACTTTCGCT GATGACAC-3′ (reverse). For expression analyses of pancreatic-duodenal homeobox-1 (Pdx1), NeuroD1, Pax6, glucose transporter 2 (Glut2), Glucagon like peptide 1 receptor (Glp1r), islet amyloid polypeptide (Iapp), glucagon (Gcg) in islets of Bace1–/– and control mice, RT-PCR was performed with QuantiTect Primer Assays (Qiagen, Hilden, Germany) for Pdx1 (ID:QT00102235), NeuroD1 (ID:QT00251265), Pax6 (ID:QT01052786), Glut2 (ID: QT00103537), Glp1r (ID: QT00130767), Iapp (ID: QT00101052), glucagon (ID: QT00124033). Relative gene expression (using Hprt and 18S rRNA) was calculated using standard curve method.

Animal studies

All animal experiments were performed according to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996). The local ethics committee (Regierungspräsidium Leipzig) of the state of Saxony, Germany specifically approved this study.

In vivo characterization of glucose metabolism in Bace1–/– mice

Bace1–/– mice were generated as previously described (22). We measured fasted and fed glucose plasma concentrations and performed ip glucose (GTT) and insulin tolerance tests (ITT) in 12-weeks-old male (n = 6) and female Bace1–/– (n = 5) and C57BL/6 control (males n = 6, female, n = 6) mice. ipGTT and ipITT have been performed as previously described [23].

Isolation of islets from Bace1–/– and control mice

Islets were isolated by collagenase digestion of the pancreas and studied fresh. Islets were used to compare insulin response in 1.5, 5.5, 11, or 22 mM glucose after culture for 18 h in RPMI 1640 medium (Invitrogen, Merelbeke, Belgium) kept at 37°C in a 95% air-5% CO2 atmosphere. Effluent fractions from islet preparations were saved for insulin ELISA-assay (Crystal Chem) using mouse insulin as standard. At the end of the experiments, islets were recovered, and insulin content was determined after extraction.

miRNA-661 expression in human pancreatic islets

Human islets were isolated from one healthy organ donor and cultured in CMRL-1066 medium as described previously [24]. Islet purity was >95% as judged by dithizone staining. After a culture period of 24 h, total RNA was isolated with the miRNeasy kit (QIAGEN, Hilden, Germany). RT was done with the miScript RT kit according to the manufacturer's protocols (QIAGEN, Hilden, Germany) in a 10-µl reaction using 1 µg RNA. This kit was selected because it enables the RT of miRNAs, mRNAs, and other noncoding RNAs into cDNA in a single step. RT-qPCR was performed on a Roche (Mannheim, Germany) LightCycler 480 in a 10-µl reaction using miScript primer assays (QIAGEN) for Homo sapiens-miR-661, and the reference RNU6B and the miScript SYBR Green PCR kit according to the manufacturer's instructions. RNA of breast cancer cell line (MCF-10A) cells served as positive and no template as negative control. High miR-661 expression has been previously demonstrated in MCF-10A cells [25].

Reporter gene assay

We measured BACE1 expression in the presence of miR-661 in human embryonic kidney HEK 293 cells [#CRL-1573, American Type Culture Collection (ATCC), Manassas, VA] which either carried the rs535860 SNP (A allele) or the wild type (T) allele using a luciferase reporter gene assay. Site directed mutagenesis (QuikChange Kit, Stratagene) was performed to create the SNP rs535860 (Primer: 5′-CCTGTGGTACCCAGGCAGAGAAGAGBACE-3′; 5′-CTCTTCTCTGCCTGGGTACCACAGG-3′). The sequences of the construct were verified by direct sequencing. HEK 293 cells were incubated for 24 h and then transfected with 200 ng plasmid pGL4.10 either containing wildtype 3′UTR or the SNP rs535860 of BACE1. The sequence was cloned into pGl4.1 (luc 2) vector (Promega) between luciferase gene and SV40 late PolyA by restriction with Xbal. Orientation of the insert was controlled by sequencing. Site directed mutagenesis (QuikChange Site-Directed Mutagenesis Kit Stratagene) was performed to create the SNP rs535860. Cells were additionally transfected with 5nM mi-RNA-661 or 5nM siRNA as a negative control. After transfection, cells were incubated for another 24 h. Luciferase quantification was performed using the Dual-Glo Luciferase Assay (Promega) in an ELISA reader (Tecan).

Genetic analyses

Sequencing of the BACE1 gene was performed in DNA from 48 non-related subjects (27 with and 21 without diabetes) using Big Dye Terminator (Applied Biosystems, Foster City, CA) on an automated DNA capillary sequencer (ABI PRISM 3100 Avant; Applied Biosystems, Foster City, CA). Twenty-one genetic known variants were found. There were no new SNPs identified. When combined with the HapMap tagging SNPs, 13 polymorphisms were representative for their linkage disequilibrium blocks (rs473210, rs11601511, rs638405, rs676134, rs551662, rs525493, rs522843, rs593245, rs687740, rs535860, rs490460, rs45506400, rs28989504) and covered 100% of common genetic variation within the BACE1 locus. HapMap tagging SNPs were selected from the HapMap Phase II using the Tagger software according to the following selection criteria: minor allele frequency (MAF) > 0.05 and r2 > 0.8. SNP genotyping was done using the TaqMan SNP Genotyping assay (Applied Biosystems, Foster City, CA). To assess genotyping reproducibility, a random ∼5% selection of the sample was re-genotyped in all SNPs; all genotypes matched initial designated genotypes.

Statistical analyses

Data are given as mean ± SD unless otherwise stated. Data sets were analyzed for statistical significance using a two-tailed unpaired Student's t test or differences were assessed by one-way analysis of variance using the Statistical Package for Social Science (SPSS, Chicago, IL), version 14.0. Differences in genotype frequencies between patients with diabetes and healthy controls were compared using logistic regression analyses. P values <0.05 were considered significant. For association studies, prior to statistical analysis, non-normally distributed parameters were ln-transformed to approximate a normal distribution. Differences in genotype frequencies between patients with diabetes and healthy controls were compared using logistic regression analyses. Multivariate linear relationships were assessed by generalized linear regression models. All analyses were done under the additive model and the presented P values are adjusted for age, sex (and BMI for glucose traits). Two-sided P values <0.05 were considered to provide nominal evidence for association and are presented without correction for multiple testing. The analysis of associations with quantitative traits was restricted to nondiabetics to avoid diabetes status or treatment masking potential effects of the variants on these phenotypic traits.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

BACE1 knockdown causes reduced insulin expression in INS-1E cells

To study the effects of reduced Bace1 mRNA expression in INS-1E cells, we used a siRNA approach leading to a ∼60% reduction of Bace1 mRNA expression compare to controls (Figure 1A). Reduced Bace1 gene expression resulted in significantly decreased ∼60% insulin mRNA expression in INS-1E cells (Figure 1B). Reduction of Bace1 expression caused a significant 30% decreased insulin protein content (Figure 2). Insulin protein content of glucose stimulated cells was decreased after Bace1 knockdown from 50.2 ± 4.8 to 38.9 ± 3.9 nmol l−1 (P < 0.05).

image

Figure 1. Reduction of BACE1 expression by siRNA reduces insulin mRNA expression in INS-1E cells. siRNA mediated knockdown of BACE1 in INS-1E cells resulted in significant reduction in (A) BACE1 and (B) insulin mRNA expression. Nearly 48 h after transfection INS-1E cells were treated with KRBH buffer. *P < 0.05.

Download figure to PowerPoint

image

Figure 2. Knockdown of BACE1 by siRNA reduces intracellular insulin protein content in INS-1E cells. siRNA mediated knockdown of BACE1 in INS-1E cells resulted in significant reduction in insulin protein content. Nearly 48 h after transfection INS-1E cells were treated with either KRBH buffer. Insulin content was measured using a specific insulin-ELISA after HCl/alcohol extraction. *P < 0.05.

Download figure to PowerPoint

Bace1–/– mice have normal glucose tolerance and insulin sensitivity

We further sought to investigate the consequences of Bace1 gene disruption on whole body glucose homeostasis in vivo. Bace1–/– mice are significantly leaner than controls (Table 1). Mean fasting glucose concentrations were not significantly different between Bace1–/– and control mice (P = 0.09), whereas Bace1–/– mice display significantly lower fed plasma glucose concentrations (Table 1). We find significantly lower fasting and fed plasma insulin concentrations in Bace1–/– mice compared to controls (Table 1). However, after adjusting for body weight, differences in fed plasma glucose and insulin plasma concentrations between Bace1–/– and control mice lost significance. No significant differences between Bace1–/– mice and controls were found at any time point of the ipGTT (Figure 3A). Although no significant differences in glucose concentrations were found at any time point of the ipGTT, there was a significant increase in glucose AUC in Bace1–/– mice as compared to controls (Bace1–/– mice: 464 ± 41mg/dl/120 min; controls: 350 ± 59 mg/dl/120 min, P < 0.05). This difference in the AUC GTT glucose became even more significant after adjusting for differences in body weight (P < 0.01). Insulin tolerance tests revealed a trend for better insulin sensitivity in Bace1–/– compared to control mice (Figure 3B) (P = 0.15). However, this trend was completely lost after adjusting for body weight. We further sought to determine whether lower body weight in Bace1–/– mice is due to reduced fat mass. Surprisingly, we found significantly increased relative epigonadal and subcutaneous adipose tissue mass in Bace1–/– compared to control mice (Figure 3C). This further suggests that lower body weight of Bace1–/– mice is due to lower lean body mass, which could contribute to differences glucose metabolism.

image

Figure 3. Bace1–/– mice have normal glucose tolerance and insulin sensitivity. (A) Glucose tolerance tests performed on 16-h fasted 12-week-old male and female wild type (control; n = 12) and Bace1–/– mice (n = 11). (B) Insulin tolerance tests on random fed 12-week-old male and female control and Bace1–/– mice. Results are expressed as mean + SEM. There was no difference between the genders both in the GTT and in the ITT. (C) Relative subcutaneous (SC) and epigonadal (EPI) adipose tissue, but not brown adipose tissue (BAT) masses are significantly elevated in Bace1–/– (N = 10) compared to wildtype mice (N = 6) at an age of 12 weeks. *P < 0.05, **P < 0.01.

Download figure to PowerPoint

Table 1. Body weight and metabolic parameters in 12 weeks old male and female Bace1-/- (n = 11) and C57BL/6 (control) mice (n = 12)
 ControlBace1-/-
  1. a

    P < 0.05.

Body weight (g)28.6 ± 4.9a22.4 ± 2.3
Blood glucose, fasted (mmol/l)5.0 ± 0.94.7 ± 0.3
Blood glucose, fed (mmol/l)7.1 ± 0.6a6.4 ± 0.4
Serum insulin, fasted (ng ml−1)0.94 ± 0.4a0.46 ± 0.1
Serum insulin, fed (ng ml−1)2.3 ± 1.1a1.1 ± 0.4

Reduced insulin content of isolated islets from Bace1–/– mice

Isolated islets from Bace1–/– mice have a significantly lower insulin protein content both at 0 and 22 mM glucose concentrations in the medium (Figure 4A). In addition, insulin mRNA expression is significantly lower in isolated islets from Bace1–/– (0.2 ± 0.1 AU) compared to control mice (0.79 ± 0.2 AU, P < 0.05). In contrast to reduced insulin expression, islets of Bace1–/– mice display significantly higher expression of the β-cell genes Pdx and NeuroD, whereas expression of Pax6, Glut2, Glp1r, Iapp, and glucagon are indistinguishable between the genotypes (Figure 4B). Insulin secretion of Bace1–/– islets was not affected upon glucose concentrations between 1.5 and 11mM (data not shown).

image

Figure 4. Reduced BACE1 expression causes impaired insulin secretion upon high glucose concentrations. (A) Insulin content of isolated islets (ng/islet) from BACE1–/– and C57BL/6 control mice at different glucose concentrations (0 mM versus 22 mM). (B) Relative mRNA levels of pancreatic-duodenal homeobox-1 (Pdx1), NeuroD1, Pax6, glucose transporter 2 (Glut2), Glucagon like peptide 1 receptor (Glp1r), islet amyloid polypeptide (Iapp), glucagon (Gcg) in islets of Bace1–/– (N = 6) knockout mice compared to wild type mice (N = 6) at an age of 12 weeks. *P < 0.05, **P < 0.01, ***P < 0.001.

Download figure to PowerPoint

SNP rs535860 in the BACE1gene is associated with T2D

In a case-control study including 538 healthy controls with normal glucose tolerance and 989 patients with T2D, one SNP (rs535860) in the BACE1 gene is significantly associated with T2D (P < 3.5 × 10−5, adjusted for age, sex, BMI) (Supporting Information Table 1). Carriers of the minor allele A were at higher risk of developing diabetes [OR = 2.10 (1.47; 3.00)].

BACE1 gene variants and association with quantitative phenotypes

In subjects without T2D, the diabetes risk allele A of the rs535860 was associated with higher fasting and 2-hr plasma glucose (P < 0.05; Supporting Information Table 2). Furthermore, rs11601511 was nominally associated with fasting plasma glucose and HbA1c levels, rs525493 with fasting plasma glucose and rs473210 with glucose infusion rate during the steady state of an euglycemic-hyperinsulinemic clamp (P < 0.05; Supporting Information Table 2).

A-allele of rs535860 causes decreased BACE1 expression

The SNP rs535860 is within 3′UTR of the BACE1 gene and has been shown to create a new binding site for gene expression regulating miR-661 [26] (Figure 5A). We therefore compared the expression of a reporter gene under control of BACE1-3′UTR in HEK 293 cells either carrying the major T-allele or the minor A-allele of rs535860 in the presence or absence of miR-661 using a luciferase reporter gene assay. We found a significantly lower expression in cells with the A-allele compared to controls in the presence of miR-661, whereas expression of the plasmid containing the T-allele of rs535860 in 3′UTR is not altered by miR-661 (Figure 5B). Importantly, we detected expression of miR-661 in human pancreatic islets (Figure 5C).

image

Figure 5. The rs535860 maps within 3′UTR of the BACE1 gene creating a new binding site for miR-661. (A) miR-661 binding site within the 3′-UTR sequence of BACE1 gene (rs535860, variant allele). (B) Reporter gene activity under control of the BACE1-3′UTR: HEK 293 cells were transfected with luciferase reporter gene containing BACE1-3′UTR carrying the rs535860 SNP or the wild type allele in the presence or absence of miR-661. In cells, transfected with the SNP rs535860 containing plasmid, the luciferase expression was significantly reduced by miR-661. (C) Human miRNA-661 relative to RNU6B expression in a breast cancer cell line (MCF-10), pancreatic islets, and no template controls (NTC) (data from three independent experiments).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

We identified Bace1 as a potential novel candidate gene for impaired insulin expression in pancreatic β-cells. In addition, we found that one SNP (rs535860) in the human BACE1 gene is significantly associated with T2D (P < 3.5 × 10−5, adjusted for age, sex, and BMI). Moreover, in individuals without T2D including subjects with normal and impaired glucose tolerance, the diabetes risk allele A of the rs535860 was associated with higher fasting and 2-h plasma glucose and three additional SNPs are associated with either fasting plasma glucose (rs11601511 and rs525493), HbA1c values (rs11601511), and insulin sensitivity as determined by glucose infusion rate during the steady state of an euglycemic-hyperinsulinemic clamp (rs473210). These genetic associations between BACE1 gene variants and parameters of glucose homeostasis and insulin sensitivity support the hypothesis that BACE1 may play a role in the pathogenesis of type 2 diabetes. We therefore tried to dissect whether BACE1 maybe involved in the regulation of insulin expression and/or insulin sensitivity. In 3-months-old Bace1–/– mice, we did not find significant alterations of glucose tolerance or insulin sensitivity, although significantly lower fasted and fed insulin plasma concentrations may be suggestive for better insulin sensitivity of Bace1–/– mice. Interestingly, the AUC glucose of an ipGTT was significantly higher in Bace1–/– compared to control mice. This was even more pronounced after adjusting the AUC GTT glucose to body weight. Noteworthy, we confirm previous data [22] that Bace1–/– mice are significantly smaller and there was no body weight difference between males and females at an age of 12 weeks. The absence of gender differences and the lower body weight Bace1–/– mice may be due to a hyperactive behavior [22]. In addition, we found that relative fat mass is significantly higher in Bace1–/– compared to control mice, suggesting that lower body weight of Bace1–/– mice is due to lower lean body mass, which could contribute to differences glucose metabolism.

However, BACE1 has been recently shown to play a significant role in the protection against diet-induced obesity and insulin resistance [16]. At ages of 8 and 12 months, Bace1–/– mice were significantly more insulin sensitive and had improved glucose tolerance compared to controls [16]. Together with our results in younger mice, these data suggest that improved insulin sensitivity and glucose metabolism in relation to reduced Bace1 expression may develop with ageing. Because we did not study mice up to an age of 8 months or older, we can not directly confirm the effect of BACE1 reduction on insulin sensitivity. Therefore, the exact mechanisms how BACE1 may alter insulin sensitivity still need to be explored. Increased expression of uncoupling proteins in brown adipose tissue and skeletal muscle have been suggested to link reduced Bace1 expression to improvements in glucose homeostasis upon challenges including high fat diet and ageing [16].

The potential contribution of BACE1 on altered insulin expression has not been studied in Bace1–/– mice yet. Casas et al. [7] recently demonstrated that inhibition of BACE causes reduced insulin protein content in islets in vitro. However, this effect was due to reduced Bace2, but not Bace1 expression. In contrast to that, we found that reducing BACE1 expression in vitro and in Bace1–/– mice causes decreased insulin content of islets. Reduction of Bace1 expression in INS-1E cells by specific siRNA in vitro resulted in a significant decrease in insulin expression under different conditions of glucose or insulin secretagogue stimulation. INS-1E cells were used because they express Bace1 and exhibit typical characteristics of β-cells including release of insulin in response to glucose [20]. Moreover, we found decreased insulin expression in pancreatic islets from Bace1–/– compared to islets from control mice. We demonstrate that isolated islets from Bace1–/– mice express less insulin than controls under conditions of both glucose depletion and high glucose concentrations. To exclude a general down-regulation of β-cell genes in Bace1–/– mice, we analyzed expression developmentally regulated genes Pdx 1, NeuroD, and Pax6, but also critical genes for β-cell maturation and secretory function (Glut2, Glp1r, Iapp) in islets of these mice. Reduced insulin mRNA expression in islets of Bace1–/– mice maybe due to an effect of BACE1 on β-cell development, differences in the general composition of Bace1–/– compared to control islets, and the possibility that genes associated with β-cell maturation/function are being primarily affected. However, reduced Bace1 expression is associated with significantly higher expression of the developmental genes Pdx 1 and NeuroD, whereas Pax6 expression was not affected. In addition, expression of the β-cell maturation genes Glut2, Glp1r, and Iapp were not affected by reduced Bace1 expression further supporting the hypothesis that BACE1 plays a distinct role in regulation of insulin expression. We did not find evidence for the hypothesis that changes in islet composition may contribute to reduced insulin expression in islets of Bace1–/– mice, because expression of glucagon was indistinguishable between islets of Bace1–/– and control mice. Taken together, reduced expression of insulin despite higher or normal expression of other β-cell genes, which are important of the development and/or maturation of β-cells in response to reduced Bace1 suggest a specific effect of Bace1 on insulin independently of potential effects at different stages in β-cell development.

Bace1–/– mice display lower fasted and fed plasma insulin concentrations, further supporting our hypothesis that BACE1 plays a significant role in insulin expression. In addition, reduced body weight in Bace1–/– mice might be associated with impaired growth due to chronically reduced insulin levels. The mechanisms how BACE1 may affect insulin expression are less clear. Further studies are necessary to investigate whether BACE1 overexpression may increase β-amyloid in pancreatic islets thereby mediating reduced insulin expression.

Noteworthy, in carriers of the A-allele of rs535860, which is significantly and independently of age, sex and BMI associated with T2D, we did not find a significant association with fasting plasma insulin concentrations. However, since we did not systematically measure postprandial or post glucose challenge insulin plasma concentrations in this cohort, we can not exclude an effect of rs535860 on fed insulin concentrations. Therefore future studies should test the hypothesis that carriers of rs535860 may have reduced postprandial insulin response. Interestingly, rs535860 creates a new miR-661 binding site [26], contributing to reduced BACE1 expression in carriers of the risk allele at least in vitro. We found miR-661 to be highly expressed in human pancreatic islets from a healthy donor suggesting that the new miR-661 binding site in rs535860 carriers may contribute to reduced BACE1 expression in vivo. However, further studies are necessary to further discriminate which specific islet cell type(s) express miR-661.

However, there are some limitations of our genetic analyses. Given the relatively small sample size, we see the values for odds ratios with caution and are aware that they might be overestimated. In addition, the observed association does not seem to be a signal from recent GWAS for diabetes [27-29]. Neither does the gene and its polymorphism show up in the publically available GWAS type 2 diabetes datasets that can be readily accessed, including an imputed dataset from the WTCCC and the FUSION dataset in The database of Genotypes and Phenotypes (http://www.ncbi.nlm.nih.gov/sites/entrez?db=gap). It remains to be clarified whether this might be due to genetic heterogeneity between various study populations. In a prior linkage study in Pima Indians, the chromosomal region on 11q23 which encompasses BACE1 was linked to BMI and T2D [17]. Rs535860 has not been directly genotyped in Pima Indians but rs1056136, which is complete linkage disequilibrium with rs535860 in full heritage Pima Indians (Leslie Baier, personal communication) was genotyped in a recent GWAS [30] (P = 0.003 adjusted for age, sex, and family membership) in 412 nondiabetic Pima Indians. In our study of Caucasians, the diabetes risk allele A of rs535860 was associated with other quantitative traits relevant to type 2 diabetes such as fasting and 2-h plasma glucose in a cohort which not only included subjects with normal but also with impaired glucose tolerance. Therefore, although the association with diabetes has not been directly replicated, it has been indirectly supported by quantitative trait analyses in our study and Pima Indian data. In addition, our study provides functional evidence on the potential biological role of the diabetes-associated variant thereby further supporting the validity of the association findings. In conclusion, reduced BACE1 expression causes decreased insulin expression in pancreatic β-cells. Our data therefore suggest that BACE1 may in addition to previously reported adverse effects on insulin sensitivity [16] contribute to impaired insulin biogenesis in the development of type 2 diabetes.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

We are grateful to Paul Saftig who provided BACE1 KO mice for the study. We thank Michael Willenborg and Ingo Rustenbeck from the University Braunschweig, Germany for their help in establishing the technique of isolation of islets from mice at our laboratory. Human islets were distributed by the Integrated Islet Distribution Program (IIDP). INS-1E cells were kindly provided by Jürgen Klammt, Leipzig.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
oby20482-sup-0001-suppinfo.doc82KSupplementary Table 1
oby20482-sup-0002-suppinfo.doc126KSupplementary Table 2

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.