Targeted disruption of the Chop gene delays endoplasmic reticulum stress–mediated diabetes (original) (raw)
Diabetic phenotype of mice with Ins2 mutation. To examine the phenotypic consequences of Ins2 mutation in the Akita mouse, we monitored blood glucose levels in each genotype (Figure 1). Ins2C96Y/C96Y mice developed hyperglycemia at 2 weeks of age, and blood glucose levels continued to increase up to 700 mg/dl at 10 weeks of age. At 2 weeks of age, Ins2WT/C96Y mice had euglycemia. They developed diabetes between 4 and 8 weeks of age. Thus, the severity and the onset of disease correlated well with number of the mutant allele. All Akita mice developed diabetes (100% penetration) and showed the simultaneous onset of disease in each genotype. There was no gender difference in onset of the disease. However, the increase in blood glucose levels was more marked in male than in female mice, both Ins2C96Y/C96Y and Ins2WT/C96Y.
Development of diabetes in Akita mice. Mice were provided food ad libitum, and blood glucose was measured between 9:00 a.m. and 10:00 a.m. at indicated age. (a) Male. (b) Female. Wild-type Ins2WT/WT mice (open circles), heterozygous Ins2WT/C96Y mice (filled circles), and homozygous Ins2C96Y/C96Y mice (filled squares) are represented. Data are shown as mean ± SD (n = 10).
Induction of ER stress–associated genes in the pancreas of Akita mice. The Ins2 mutation in Akita mice disrupts a disulfide bond between the A and B chains of insulin, and can induce a drastic conformational change in the molecule. We speculated that this misfolded insulin could cause ER stress. To check ER stress in the Akita mice, Northern blot analysis for insulin and ER stress–associated proteins in pancreas was done (Figure 2). In Ins2WT/WT mice, insulin mRNA gradually increased with age (data not shown). At 3 weeks of age, insulin mRNA was higher in Ins2WT/C96Y mice than in Ins2WT/WT mice, and was higher yet in Ins2C96Y/C96Y mice. At 6 weeks of age, insulin mRNA was higher in Ins2WT/C96Y mice than in controls, but was similar between Ins2WT/C96Y mice and Ins2C96Y/C96Y mice. In contrast, at 9 weeks of age, the mRNA was slightly higher in Ins2WT/C96Y mice than in controls, and was much lower in Ins2C96Y/C96Y mice than in controls. This decrease in insulin mRNA in Ins2C96Y/C96Y mice appears to be due to a loss of β cells (see below). Of note is that the onset of diabetes is associated with an increase in insulin mRNA. This suggests that insulin mRNA is induced to compensate for the deficiency in secretion of active insulin. Chop/GADD153, a transcription factor that plays a role in growth arrest and cell death, and Bip/GRP78, an ER chaperone, are induced by ER stress. Chop mRNA was much higher in Ins2C96Y/C96Y mice than in controls at 3 weeks of age. It was also much higher than that in controls in both Ins2WT/C96Y mice and Ins2C96Y/C96Y mice at 6 weeks of age, and in Ins2WT/C96Y mice at 9 weeks of age. The profile of fluctuation in Chop mRNA was similar to that of insulin mRNA. Bip mRNA and Chop mRNA changed in a similar manner. These results indicate that β cells in both heterozygous and homozygous mutant mice are subjected to ER stress, and that the ER stress is closely associated with the induction of insulin mRNA.
Northern blot analysis for insulin, Chop, and Bip mRNAs in the pancreas of male Akita mice. (a) Chemiluminescent images for insulin, Chop, and Bip mRNAs at the indicated ages are shown. Total RNAs (2.0 mg) from the pancreas of each genotype were subjected to Northern blot analysis. The results at 3, 6, and 9 weeks of age were obtained under essentially identical conditions. (b) Quantification of the results obtained in a, shown as mean ± SD (n = 3).
Overexpression of insulin 2C96Y leads to apoptosis in MIN6 cells. Induction of Chop and postnatal hyperglycemia in Ins2C96Y mice suggests that the mutant insulin causes ER stress and leads to apoptosis in β cells via Chop induction. To test this hypothesis, we overexpressed wild-type and Cys96Tyr mutant insulin 2 in MIN6 cells (Figure 3). We generated fusions of mouse insulin and EGFP for real-time monitoring of apoptosis (Figure 3a). When cells were transfected with the p_Ins2WT_-EGFP plasmid, fluorescence was observed in transfected cells 24 hours after transfection, and apparently increased with time up to 60 hours (Figure 3a). When cells were transfected with the p_Ins2C96Y_-EGFP plasmid, fluorescent cells were morphologically similar to those expressing _Ins2WT_–EGFP at 24 hours. However, cells expressing _Ins2C96Y_–EGFP became round and detached from dishes at 36 hours and thereafter. These results indicate that cells expressing _Ins2C96Y_–EGFP undergo cell death.
Apoptosis in MIN6 cells overexpressing Ins2C96Y–EGFP and Ins2C96Y. (a) Cells were transfected with p_Ins2WT_-EGFP or p_Ins2C96Y_-EGFP. At the indicated times after transfection, cells were observed under a fluorescence microscope. Original magnification: ×100. (b) Cells were cotransfected with pEYFP-ER and either pcDNA-Ins2WT or pcDNA-Ins2C96Y. Forty-eight hours after transfection, cells were observed under a fluorescence microscope. Original magnification: ×200. (c) Cells were cotransfected with pEGFP and either pcDNA-Ins2WT or pcDNA-Ins2C96Y. The transfected cells and apoptotic cells were visualized by GFP fluorescence and Hoechst 33258 staining, respectively. pcDNA-_Ins2WT_–transfected cells were not apoptotic (arrowheads), whereas pcDNA-_Ins2C96Y_–transfected cells were apoptotic (arrows). Original magnification: ×400. (d) Cells were transfected with pcDNA-Ins2WT or pcDNA-Ins2C96Y. DNA was extracted, electrophoresed in 2% agarose gel, stained with SYBR Green I, and visualized by UV transillumination. (e) Cells were cotransfected with pEGFP and either pcDNA-Ins2WT or pcDNA-Ins2C96Y. The transfected cells and apoptotic cells were visualized by GFP fluorescence and annexin V staining, respectively. Original magnification: ×200.
To exclude artificial effects of fusion proteins, cells were cotransfected with an EYFP-ER expression plasmid and either the pcDNA-Ins2WT plasmid or the pcDNA-Ins2C96Y plasmid. Cells expressing Ins2WT showed ER-like fluorescence, whereas cells expressing Ins2C96Y became round and detached from dishes (Figure 3b). Staining with Hoechst 33258 revealed that cells expressing Ins2C96Y undergo apoptosis (Figure 3c). Transfection with pcDNA-Ins2C96Y plasmid led to formation of DNA ladders, which is characteristic of apoptotic cells (Figure 3d). Annexin V staining revealed that cells expressing Ins2C96Y undergo apoptosis (Figure 3e). Considering these results together, we conclude that overexpression of Ins2C96Y causes ER stress and leads to Chop-mediated apoptosis.
Characteristics of double-mutant mice with Ins2 mutation and Chop disruption. To further investigate the involvement of Chop in ER stress–mediated diabetes, we generated double-mutant mice with mutation in the Ins2 gene and disruption in the Chop gene. Progeny were obtained by crossing Chop–/– mice with Ins2WT/C96Y mice. Figure 4a shows typical examples of genotyping. Nine genotypes of mice were generated: Ins2WT/WTChop+/+, Ins2WT/WTChop+/–, Ins2WT/WTChop–/–, Ins2WT/C96YChop+/+, Ins2WT/C96YChop+/–, Ins2WT/C96YChop–/–, Ins2C96Y/C96YChop+/+, Ins2C96Y/C96YChop+/–, and Ins2C96Y/C96YChop–/–, all of which had the same genetic background (C57BL/6). As shown in Figure 1, Ins2WT/C96Y and Ins2C96Y/C96Y mice developed hyperglycemia by 8 weeks of age. Therefore, we measured body weights, blood glucose levels, and pancreatic insulin content of each genotype at 8 weeks of age (Figure 4b). Ins2C96Y/C96Y mice of all Chop genotypes showed loss of body weight, severe hyperglycemia, and reduced pancreatic insulin content. There was no statistically significant difference among Chop+/+, Chop+/–, and Chop–/– mice. Ins2C96Y/C96Y mice were often perinatally lethal due to defective insulin secretion in response to glucose (data not shown). In contrast, among the Ins2WT/C96Y mice, the Chop–/– mice did not develop hyperglycemia and had significantly higher body weights and pancreatic insulin content than did Chop+/– and Chop+/+ mice. There was little difference in these parameters between Chop+/– and Chop+/+ mice. These results are consistent with our previous finding that pancreatic islets from Chop–/– mice are much more resistant to nitric oxide than are those from Chop+/+ and Chop+/– animals (11). These results indicate that disruption of the Chop gene delays the onset of diabetes in Ins2WT/C96Y mice but not in Ins2C96Y/C96Y mice.
Characteristics of the double-mutant mice with Ins2 mutation and Chop disruption. (a) Genotyping of double-mutant mice. Representative genotyping of the Ins2 gene by RFLP; the Chop gene from nine mutant lines is shown by PCR. Ins2 exon 3 was amplified by PCR using genomic DNA. The Ins2C96Y mutation in Akita mice disrupts an _Fnu4H_I site in exon 3 of Ins2. Digestion with Fnu4H_I did not change the size of the PCR product from the mutated allele (280 bp), but it decreased that of the wild-type allele to 140 bp. Primers for wild-type and mutated Chop mice are described in Methods. The left lane shows 100-bp DNA ladder markers (top panel) and λDNA/Hind III + φ×174DNA/Hae III fragment markers (middle and bottom panels). (b–d) Phenotypic characterization of double-mutant mice at 8 weeks of age. (b) Body weight. (c) Morning blood glucose. (d) Pancreatic insulin content. Data are shown as mean ± SD. Significant differences among the genotype of the CHOP gene were evaluated by the Student t test: *P < 0.001 versus Chop+/+. ‡_Ins2C96Y/C96YChop+/+ mice died within days of birth. Therefore, †data are shown as only one mouse and ‡data are shown as means ± ranges (n = 2).
There was no statistically significant difference in body weight among Chop+/+, Chop+/–, and Chop–/– male mice. The female Chop–/– mice were seen to have higher body weight than female Chop+/+ animals had. In addition, the development of obesity and fat deposits in the Chop–/– mice after 16 weeks of age appeared to be accelerated (data not shown). It has been suggested that Chop inhibits adipose differentiation by interfering with the accumulation of adipogenic C/EBP isoforms (17, 18). Therefore, we speculate that targeted disruption of the Chop gene results in obesity by increasing adipocyte differentiation. In this study, the Chop–/– mice, generated on a C57BL/6 strain, were backcrossed only two generations from the original 129/Sv × C57BL/6 mice. Therefore, it is also possible that the 129/Sv background influences body weight.
Disruption of the Chop gene protects islet cells of Ins2WT/C96Y mice from apoptosis. The effect of disruption of the Chop gene in apoptosis of β cells in Ins2WT/C96Y mice was examined histologically (Figure 5). At 4 weeks of age, Chop induction was observed in Ins2WT/C96YChop+/+ islets. Distribution of Chop-positive cells was similar to that of insulin-positive cells, suggesting that Chop-positive cells are β cells. In Ins2WT/C96Y mice, the size of islets decreased markedly, and the mass of insulin-positive cells diminished. Apoptotic (TUNEL-positive) cells were clearly observed in islets. In contrast, in Ins2WT/C96YChop–/– mice, islet size was fairly well retained, and the number of TUNEL-positive cells was much smaller. The number of TUNEL-positive cells per islet in Ins2WT/C96YChop+/+ mice was 10.50 ± 10.73 (n = 20), whereas that in Ins2WT/C96YChop–/– mice was 1.10 ± 1.74 (n = 20). In Ins2C96Y/C96Y mice, disruption of the Chop gene failed to protect islets from apoptosis (data not shown). These data indicate that the hyperglycemia in Ins2C96Y mutant mice was primarily due to a specific decrease of β cells through ER stress–mediated apoptosis, and that disruption of the Chop gene rescues β cells of Ins2WT/C96Y mice from apoptosis.
Immunohistochemical detection of Chop and apoptosis in islets of male Ins2WT/WTChop+/+, Ins2WT/C96YChop+/+, and Ins2WT/C96YChop–/– mice. (a) Immunostaining of insulin and Chop. The pancreata from Ins2WT/C96YChop+/+ and Ins2WT/C96YChop–/– male mice at 4 weeks of age were fixed and then costained for insulin (red) and Chop (green). Phase-contrast images and fluorescence images of the same fields are shown. Original magnification: ×100. (b) Apoptosis detection by the TUNEL method. The pancreata from Ins2WT/WTChop+/+, Ins2WT/C96YChop+/+, and Ins2WT/C96YChop–/– male mice were obtained and costained for insulin (red) and TUNEL-positive cells (green). Phase-contrast images and fluorescence images of the same fields are shown. Original magnification: ×200.
Disruption of the Chop gene delays the onset of diabetes in Ins2WT/C96Y mice. To determine whether disruption of the Chop gene would block or delay the onset of diabetes in Ins2WT/C96Y mice, we monitored blood glucose levels in Ins2WT/C96YChop+/+, Ins2WT/C96YChop+/–, and Ins2WT/C96YChop–/– mice (Figure 6). Both Ins2WT/C96YChop+/+ and Ins2WT/C96YChop+/– mice developed hyperglycemia beginning at 4 weeks of age. Blood glucose levels in these mice increased up to about 800 mg/dl, at which point mice were 20 weeks of age. In contrast, blood glucose in Ins2WT/C96YChop–/– mice remained unchanged (less than 250 mg/dl) until the mice reached 8 weeks of age. At that point levels began to increase; at 20 weeks the blood glucose level in these mice was about 500 mg/dl. Thus, disruption of the Chop gene delayed the onset of diabetes in Ins2WT/C96Y mice by 8–10 weeks; hyperglycemia was somewhat overcome but was not prevented.
Development of diabetes in male Ins2WT/C96YChop+/+, Ins2WT/C96YChop+/–, and Ins2WT/C96YChop–/– mice. Mice were provided food ad libitum, and blood glucose was measured between 9:00 a.m. and 10:00 a.m. Ins2WT/C96YChop+/+ mice (filled squares), Ins2WT/C96YChop+/– mice (filled circles), and Ins2WT/C96YChop–/– mice (open circles) are represented. Data are shown as mean ± SD (n = 7).