Increased insulin demand promotes while pioglitazone prevents pancreatic beta cell apoptosis in Wfs1 knockout mice
- 1.5k Downloads
The WFS1 gene encodes an endoplasmic reticulum (ER) membrane-embedded protein called Wolfram syndrome 1 protein, homozygous mutations of which cause selective beta cell loss in humans. The function(s) of this protein and the mechanism by which the mutations of this gene cause beta cell death are still not fully understood. We hypothesised that increased insulin demand as a result of obesity/insulin resistance causes ER stress in pancreatic beta cells, thereby promoting beta cell death.
We studied the effect of breeding Wfs1 −/− mice on a C57BL/6J background with mild obesity and insulin resistance, by introducing the agouti lethal yellow mutation (A y /a). We also treated the mice with pioglitazone.
Wfs1 −/− mice bred on a C57BL/6J background rarely develop overt diabetes by 24 weeks of age, showing only mild beta cell loss. However, Wfs1 −/− A y /a mice developed selective beta cell loss and severe insulin-deficient diabetes as early as 8 weeks. This beta cell loss was due to apoptosis. In Wfs1 +/+ A y /a islets, levels of ER chaperone immunoglobulin-binding protein (BiP)/78 kDa glucose-regulated protein (GRP78) and phosphorylation of eukaryotic translation initiation factor 2, subunit α (eIF2α) apparently increased. Levels of both were further increased in Wfs1 −/− A y /a murine islets. Electron micrography revealed markedly dilated ERs in Wfs1 −/− A y /a murine beta cells. Interestingly, pioglitazone treatment protected beta cells from apoptosis and almost completely prevented diabetes development.
Wfs1-deficient beta cells are susceptible to ER stress. Increased insulin demand prompts apoptosis in such cells in vivo. Pioglitazone, remarkably, suppresses this process and prevents diabetes. As common WFS1 gene variants have recently been shown to confer a risk of type 2 diabetes, our findings may be relevant to the gradual but progressive loss of beta cells in type 2 diabetes.
KeywordsApoptosis Diabetes mellitus Endoplasmic reticulum Endoplasmic reticulum stress Insulin Pancreatic beta cell Pioglitazone Wolfram syndrome
Eukaryotic translation initiation factor 2,subunit α
78 kDa Glucose-regulated protein
94 kDa Glucose-regulated protein
Intraperitoneal glucose tolerance test
Inositol-requiring protein 1
Insulin tolerance test
Unfolded protein response
Wolfram syndrome 1
Many obese individuals with marked insulin resistance do not develop overt diabetes. However, in individuals destined to develop type 2 diabetes, pancreatic beta cells fail to produce enough insulin to meet systemic demand. This beta cell failure is caused by insufficient beta cell response to glucose and inadequate beta cell mass expansion . Recently, several reports have convincingly confirmed the contribution of beta cell mass reduction to type 2 diabetes [2, 3], at least in its advanced stages. Although the cause of this decrease is unknown, increased apoptosis may play an important role [1, 2]. In addition to genetic factors, processes involving glucotoxicity and/or lipotoxicity are likely to be contributory . Endoplasmic reticulum (ER) stress has also recently emerged as a candidate mechanism [5, 6, 7].
Wolfram syndrome is a rare recessively inherited genetic disorder, characterised by juvenile-onset diabetes mellitus and progressive optic atrophy . Several neuro-psychiatric illnesses may also be present . Beta cells are selectively lost from the pancreatic islets and this loss is genetically programmed . Our group and others previously identified the Wolfram syndrome gene, designating it WFS1  or wolframin  and showing that it is localised primarily in the ER [13, 14]. Homozygous Wfs1 knockout mice developed glucose intolerance. However, the diabetic phenotype of the Wfs1 −/− mouse was milder than that seen in Wolfram syndrome patients and was largely dependent on genetic background .
Despite intensive studies, the precise function of the Wolfram syndrome 1 (WFS1) protein is still largely unknown. It may regulate ER calcium homeostasis by serving as an ER cation channel [16, 17]. Recently the WFS1 protein was also suggested to be a putative chaperone for the Na/K ATPase b1 subunit in the ER . However, it is certain that its function is closely related to ER stress responses. ER stress induces Wfs1 expression and, in turn, loss of WFS1 protein exacerbates ER stress [15, 19, 20, 21]. Furthermore, Wfs1 −/− islet cells are susceptible to ER stress-induced apoptosis [15, 21, 22]. ER stress is induced under conditions such as overload of protein synthesis/processing, accumulation of structurally abnormal proteins, disturbance of post-translational modification and ER calcium homeostasis abnormalities. When ER stress develops, cells respond by unfolded protein response (UPR), facilitating protein folding via the induction of chaperone proteins, attenuation of translations, as well as degradation of misfolded proteins, a process called ER-associated degradation [23, 24, 25]. If the stress is severe, apoptosis is induced. Accumulating evidence suggests that a high level of ER stress or defective ER stress signalling causes beta cell death and that diabetes thereby develops [26, 27, 28, 29, 30].
We wished to test in vivo the hypothesis that beta cell loss in Wfs1 −/− mice is caused by an inability to cope with ER stress. The agouti yellow (A y /a) mouse is a genetic model of mild obesity/insulin resistance with compensatory beta cell hyperplasia . In this mouse, ectopically expressed agouti protein promotes food intake and weight gain by antagonising signalling at melanocortin-4 receptor in the hypothalamus. ER stress is likely to have been induced in these beta cells by increased insulin synthesis and secretion demands, and also by elevated serum NEFA . As Wfs1 −/− mice on the C57BL/6J background rarely develop overt diabetes by 24 weeks, we introduced the A y mutation into C57BL/6J Wfs1 −/− mice and assessed the consequences. Furthermore, we treated Wfs1 −/− mice with pioglitazone, which ameliorates insulin resistance.
All experimental protocols were approved by the Ethics of Animal Experimentation Committee at Yamaguchi University School of Medicine.
The Wfs1 −/− mice had a C57BL/6J background . The Agouti yellow mice (C57BL/6JHamSlc-A y ) were obtained from M. Nishimura (Nagoya University Graduate School of Medicine, Nagoya, Japan) and Japan SLC (Hamamatsu, Japan). We used male mice for all experiments. The mice were kept under standard, specific pathogen-free conditions with a constant dark/light cycle and free access to standard mouse chow (MF; Oriental Yeast, Tokyo, Japan) and water. The high-fat diet (rodent diet with 60% energy from fat; D12492) was purchased from Research Diet (New Brunswick, NJ, USA) and was freely accessible. For pioglitazone treatment, mice were fed standard mouse chow with 0.01% pioglitazone (wt/wt) from the age of 4 weeks. Pioglitazone was provided by Takeda Pharmaceutical (Osaka, Japan).
Generation of Wfs1 −/− A y /a mice
To generate A y /a mice lacking the Wfs1 gene, Wfs1 −/− mice were bred with A y /a mice to create the compound heterozygote (Wfs1 +/− A y /a). This heterozygote was bred with Wfs1 +/− a/a, and Wfs1 −/− A y /a, Wfs1 +/+ A y /a, Wfs1 −/− a/a and Wfs1 +/+ a/a (wild-type [WT]) offspring were identified.
The Wfs1 genotype was determined by PCR. We used the sense primer 5′-CCCAGTTCTTGCTTTACCACCAGG-3′ and the anti-sense primers 5′-GCCTTCTTGACGAGTTCTTCTGA-3′ (derived from the neomycin resistance gene) and 5′-ACTTCGTCCAGCACTGGGGTCAG-3′ (derived from the Wfs1 gene). A y /a mice were identified by coat colour.
Body weights were measured weekly. Blood samples were collected from the tail vein. Non-fasting blood glucose was measured by the glucose oxidase method using a GlucoCard device (Arkray, Kyoto, Japan). Serum insulin levels were determined using an insulin ELISA kit (Morinaga Institute of Biological Science, Tokyo, Japan). Serum triacylglycerol was measured by high-performance liquid chromatography at Skylight Biotech (Akita, Japan), according to the procedure described by Usui et al. . Intraperitoneal glucose tolerance test (ipGTT) and insulin tolerance test (ITT) were performed at 8 weeks of age.
Pancreatic insulin content
Pancreases were removed and homogenised in acid/ethanol (0.7 mol/l HCl/ethanol 25:75 plus 0.1% Triton X-100 (vol./vol.)) and left at 4°C for 48 h, with sonication every 24 h. Homogenates were then centrifuged (3,000 g for 15 min) and the insulin content of the acid/ethanol supernatant fraction was measured using insulin ELISA. Protein in tissue extracts was determined using the BCA protein assay reagent.
Immunostaining and morphometry
Formalin-fixed paraffin-embedded sections were de-paraffinised and re-hydrated, then incubated with primary antibodies. For immunofluorescent staining, appropriate FITC-conjugated or Cy3-conjugated anti-IgG was used. Detailed antibody information is given in the Electronic supplementary material (ESM). Immunohistochemical analyses were performed, with at least three animals for each condition being killed for the purpose. For measurement of beta cell area, more than five pancreatic tissue sections per animal were randomly selected and stained with anti-insulin IgG/3,3′-diaminobenzidine tetrahydrochloride and haematoxylin. Microscopic photoimages were taken with a charge-coupled device (CCD) camera and the pancreatic and beta cell areas were each estimated using a computer program (Y. Uehara, S. Saeki and S. Saito S unpublished results).
Tissue preparation and electron microscopy
All animals were anaesthetised with sodium pentobarbital (65 mg/kg, i.p.) and intracardially perfused with 2% glutaraldehyde (vol./vol.) and 4% paraformaldehyde (vol./vol.) in 0.1 mol/l phosphate buffer (pH 7.4) containing 0.2% picric acid (vol./vol.). Pancreatic sections (1 mm thick) were post-fixed with 1% OsO4 (wt/vol.), block-stained with 2% uranyl acetate (wt/vol.), dehydrated, infiltrated with propylene oxide, placed in a mixture of propylene oxide and Epoxy resin (1:1), and flat-embedded on siliconised glass slides in Epoxy resin. Ultrathin sections were made and mounted on to copper grids. To enhance contrast, they were double stained with 2% uranyl acetate (wt/vol.) and 1% lead citrate (wt/vol.), and observed with a Hitachi H-7500 electron microscope (Hitachi High-Technologies, Tokyo, Japan) operating at 80 kV.
Mouse islet isolation and immunoblotting analysis
Pancreatic islets were isolated as described previously . For immunoblotting , approximately 100 islets from two to six mice were pooled, then immediately dissolved in lysis buffer containing 1% SDS (50 mmol/l Tris–HCl [pH 6.8], 1% SDS [wt/vol.], 10% glycerol [wt/vol.] and 50 mmol/l dithiothreitol). Anti-WFS1 (N-terminus) antibody has been described previously . For information on other antibodies, see ESM.
Results are expressed as means ± SE. Differences between means were evaluated using ANOVA or Student’s t test as appropriate. Differences were considered significant at p < 0.05.
Characteristics of the mice
A similar, but milder phenotype was observed in the Wfs1 −/− a/a mice with high-fat diet-induced obesity (ESM Fig. 1). The high-fat diet caused obesity with hyperinsulinaemia in WT mice (Wfs1 +/+ a/a). However, these mice were normoglycaemic (ESM Fig. 1a). In contrast, a high-fat diet induced hyperglycaemia in Wfs1 −/− a/a mice, although less prominently than in Wfs1 −/− A y /a mice (mean non-fasting blood glucose 18.0 ± 2.7 mmol/l, n = 6 vs 36.9 ± 2.4 mmol/l, n = 14 at 24 weeks of age, p < 0.001) (ESM Fig. 1b). Non-fasting insulin levels were lower, although not significantly, in high-fat diet-fed Wfs1 +/+ a/a than in Wfs1 +/+ A y /a mice (402.9 ± 87.5 pmol/l, n = 9 vs 605.7 ± 114.4, n = 7 at age 24 weeks, p > 0.05) (ESM Fig. 1c).
Acceleration of selective beta cell apoptosis
Time courses of beta cell loss in Wfs1 −/− A y /a islets are shown in Fig. 3m–o. In Wfs1 −/− A y /a mice, insulin-positive beta cells were selectively and severely depleted by 24 weeks of age, whereas glucagon-positive alpha cells and somatostatin-positive delta cells were preserved (Fig. 3h, l). This selective beta cell loss was due to apoptotic cell death as indicated by caspase-3 activation in islets of 16-week-old mice (Fig. 3p–s).
We performed ultrastructural analyses of pancreatic beta cells from 12-week-old mice using electron microscopy. The beta cells were distinguished from alpha and delta cells by the appearance of their secretory granules. The beta cell granules had a white halo, not evident in alpha and delta cell granules.
Unfolded protein response in pancreatic islets of obese and Wfs1−/− mice
Pioglitazone prevented beta cell loss and diabetes in Wfs1 −/− A y /a mice
Our observations suggest that beta cell overload markedly accelerates beta cell death and diabetes development in Wfs1 −/− a/a mice. Therefore, we attempted to reduce this beta cell overload, which is likely to be due to obesity-associated insulin resistance, with pioglitazone treatment. Mice were allowed free access to normal chow containing 0.01% pioglitazone immediately after weaning (4 weeks of age). The average dose of pioglitazone was estimated to be 15 mg kg−1 day−1.
Pioglitazone did not suppress the UPR
Electron microscopic examination revealed markedly improved ER appearance in Wfs1 −/− A y /a pancreatic beta cells after pioglitazone treatment (Fig. 9b–e, see also Fig. 5). Without the treatment, almost all beta cells had markedly distended ER (Fig. 9b, c). However, after the treatment, the ER distension was markedly reduced and some beta cells appeared to be normal (Fig. 9d, e), although heterogeneity was observed among the cells.
Here we demonstrated that Wfs1 −/− A y /a mice on a C57BL/6J background develop early-onset insulin-deficient diabetes mellitus as early as 6 weeks of age and that most are overtly diabetic by the 16th week, whereas neither Wfs1-deficient Wfs1 −/− a/a nor obese agouti yellow (Wfs1 +/+ A y /a) mice had developed overt diabetes by the 24th week. In Wfs1 −/− A y /a murine islets, beta cell mass was dramatically decreased due to increased beta cell apoptosis. WFS1 protein clearly plays a pivotal role in beta cell survival.
Our previous study had demonstrated that approximately half of Wfs1 −/− mice develop diabetes when they have the hybrid genetic background of C57BL/6J and 129Sv, while. unexpectedly with the C57BL/6J background, there is no apparent increase in blood glucose levels even at 36 weeks . Because beta cells of Wfs1 −/− mice were also shown to be susceptible to ER stress in ex-vivo experiments and cultured cells [15, 21], we sought to test this in the in vivo model.
The A y /a mouse is a genetic model of mild obesity/insulin resistance with compensatory beta cell hyperplasia  (Figs 1 and 3). Increased demand for insulin biosynthesis and secretion has been thought to cause ER stress in pancreatic beta cells, and here, in fact, we demonstrated, in A y /a murine islets, that ER chaperone expression was apparently increased at 12 weeks of age (Fig. 6). Similar results have been demonstrated in another insulin-resistant mouse model . In addition, WFS1 protein levels were also increased (Fig. 6a). Our group and others have shown WFS1 protein levels to be upregulated by ER stress [19, 20, 35]. These data suggest that increased insulin demand under insulin-resistant conditions produces chronic ER stress in beta cells. In addition, NEFA, which are elevated in these insulin-resistant models, may also contribute to the activation of UPR . In this regard, Lipson et al. recently demonstrated that hyperglycaemia activates inositol-requiring protein-1 (IRE1)α, an ER-resident transmembrane protein kinase regulating UPR in beta cells. Although IRE1α activation by transient hyperglycaemia is beneficial to beta cells, chronically sustained hyperglycaemia causes ER stress and suppresses insulin gene expression . Glucose regulation of the UPR has also been reported .
Dramatically decreased beta cell numbers in Wfs1 −/− A y /a mice suggest that increased insulin demand triggers pancreatic beta cell apoptosis in Wfs1-deficient mice. ER stress induces WFS1 protein production and lack of this protein itself enhances the UPR [15, 19, 20, 21]. This was true in our mice (Fig. 6), with the UPR apparently further enhanced in mildly obese Wfs1 −/− A y /a mice. These data support the hypothesis that beta cell loss in Wolfram syndrome is, at least partly, caused by increased ER stress in beta cells. Although the precise function of ER-resident WFS1 protein remains unknown, our overall results suggest that this protein is likely to protect beta cells from ER stress-induced apoptosis and that Wolfram syndrome is an ER stress-related disease.
In patients with type 2 diabetes, very gradual, but progressive beta cell loss is caused by many factors, both genetic and acquired. One well-established mechanism of acquired beta cell loss is oxidative stress, which appears to be a major mediator of glucotoxicity . We observed 4-hydroxy-2-nonenal-modified protein, an oxidative stress marker, in Wfs1 −/− A y /a murine islets at age 16 weeks (data not shown), suggesting that oxidative stress is also involved in beta cell apoptosis. Oxidative stress is likely to have been secondary to chronic hyperglycaemia because most Wfs1 −/− A y /a mice had developed hyperglycaemia by this age (Fig. 1b). In Wfs1 −/− A y /a mice, the vicious cycle associated with ER stress and oxidative stress probably exacerbates beta cell apoptosis. In this regard, it is again worth emphasising that insulin resistance induces ER stress in pancreatic beta cells . Currently, we do not know the full extent of ER stress involvement in beta cell failure in human type 2 diabetes. However, it is possible that this vicious cycle associated with ER stress and oxidative stress plays an important role in beta cell deterioration .
Pioglitazone treatment almost completely prevented the development of diabetes and beta cell loss in Wfs1 −/− A y /a mice. One simple explanation is that insulin resistance was ameliorated by pioglitazone treatment, reducing ER stress in pancreatic beta cells and preventing beta cell death. However, the mechanism may not be so simple. Unexpectedly, the UPR, represented by ER chaperone expression, was not significantly reduced (Fig. 9a, ESM Figs 2 and 3) to the extent that we expected from the remarkable level of diabetes prevention. Because the UPR is a protective response against ER stress, an apoptotic pathway may have been preferentially suppressed under these conditions. Several pathways are reportedly involved in ER stress-induced apoptosis, including the C/EBP-homologous protein (CHOP), the IRE1–TNF receptor-associated factor 2 (TRAF2)–apoptosis signal-regulating kinase 1 (ASK1), and the caspase 12 pathways . The precise mechanism whereby pioglitazone prevented apoptosis in beta cells awaits determination. The beta cell protection exerted by thiazolidinediones including pioglitazone has been demonstrated in other rodent models [41, 42] and also in humans . Although the precise mechanisms are not fully understood, one possibility is an indirect action through improvements in systemic glucose and lipid metabolism. Another mechanism involves direct actions on pancreatic beta cells . Recent reports have demonstrated that thiazolidinediones directly improve beta cell function , ameliorate lipotoxicity  and prevent beta cell apoptosis [47, 48]. In Wfs1 −/− A y /a mice, direct protective effects, as well as indirect effects, are likely to be exerted. Elucidation of the mechanism whereby pioglitazone directly protects beta cells against apoptosis in Wfs1 −/− A y /a mice would provide insights into the mechanism of beta cell death in patients with Wolfram syndrome, as well as the function of WFS1 protein in beta cells.
This is one of a few good models showing that one genetic defect predisposes beta cells to profound failure upon ER stress induced by systemic insulin resistance [30, 49]. Our findings are important for the understanding of the molecular pathophysiology of Wolfram syndrome. In addition, a common process may be involved in conventional type 2 diabetes patients, whose beta cells decrease very slowly but progressively. In this context, a recent report has confirmed that common variants in the WFS1 gene confer risk of type 2 diabetes . Therefore knowledge from this model would help us to understand the mechanisms of, and to develop a way of preventing beta cell loss in patients with conventional type 2 diabetes mellitus.
We thank M. Nishimura (Nagoya University Graduate School of Medicine) for kindly providing C57BL/6JHamSlc-Ay mice. This study was supported in part by Grants-in-Aid for Scientific Research (grant no16390096, 18390103 and 20390093 to Y. Tanizawa) from the Ministry of Education, Culture, Sports, Science and Technology of Japan, grant no. 16790510 (to K. Ueda) from the Japan Society for the Promotion of Science, grant H16-genome-003 (to Y. Oka and Y. Tanizawa) from the Ministry of Health, Labour and Welfare of Japan, and a grant from the Takeda Science Foundation (to Y. Tanizawa).
Duality of interest
The authors declare that there is no duality of interest associated with this manuscript.
- 31.Wolff GL, Roberts DW, Mountjoy KG (1999) Physiological consequences of ectopic agouti gene expression: the yellow obese mouse syndrome. Physiol Genomics 11:151–163Google Scholar
- 47.Lin CY, Gurlo T, Haataja L, Hsueh WA, Butler PC (2005) Activation of peroxisome proliferator-activated receptor-gamma by rosiglitazone protects human islet cells against human islet amyloid polypeptide toxicity by a phosphatidylinositol 3′-kinase-dependent pathway. J Clin Endocrinol Metab 90:6678–6686PubMedCrossRefGoogle Scholar