Increased Dietary Fat Promotes Islet Amyloid Formation and β-Cell Secretory Dysfunction in a Transgenic Mouse Model of Islet Amyloid

Hemizygous transgenic mice with islet β-cell expression of hIAPP on a C57BL/6xDBA/2 background were generated as previously described (17). Only male transgenic mice were used in this study because in previous studies we observed islet amyloid in 81% of male hIAPP transgenic mice compared with only 11% of female transgenic littermates (20). Transgenic status was determined by PCR using oligonucleotide primers directed against the hIAPP transgene (21). Male nontransgenic littermates were used as controls. The study was approved by the institutional animal care and use committee at the VA Puget Sound Health Care System.

At 6–8 weeks of age, hIAPP transgenic and nontransgenic mice were randomly assigned to one of three dietary groups and followed for 1 year. These diets were chosen to provide a range of dietary fat encompassing the fat content of breeder chow that we had previously shown to be associated with islet amyloid formation (20). Semipurified diets were obtained from Research Diets (New Brunswick, NJ) and contained 15% (low fat), 30% (medium fat), or 45% (high fat) of calories derived from fat. Fat was provided as corn oil and hydrogenated coconut oil, with the ratio of saturated to unsaturated fatty acids being 1:3 in each diet. The progressively increasing amounts of fat were balanced by decreasing amounts of carbohydrate (65, 50, and 35 kcal% in low-, medium-, and high-fat diets, respectively) and constant amounts of protein (20 kcal%). hIAPP transgenic mice were randomly assigned to the dietary groups as follows: low fat (n = 17), medium fat (n = 16), or high fat (n = 15). Similarly, nontransgenic mice were assigned to low (n = 8), medium (n = 7), or high (n = 15) fat diets. Food and water were provided ad libitum throughout the study.

At the beginning of the study, body weight was determined after 4 h of food withdrawal. Mice were then placed on their randomized study diets, and their body weight was determined every 4 weeks throughout the study.

After an 18-h fast, mice were anesthetized with sodium pentobarbital (100 mg/kg body wt i.p.; Nembutal, Abbott Laboratories, Chicago, IL). At 40 min after induction of anesthesia, a baseline blood sample was obtained from the retro-orbital sinus in heparinized capillary tubes. Glucose (1 g/kg) was administered intraperitoneally, and blood samples were obtained at 15, 30, 60, and 120 min after glucose administration. Plasma was separated by centrifugation and stored at −20°C before assay for glucose and immunoreactive insulin (IRI).

Before sacrifice, hIAPP-like immunoreactivity (hIAPP-LI) levels were measured after 4 h of food withdrawal. Mice were anesthetized and blood samples collected as described above. After the mice were killed, a small portion of pancreas was snap-frozen in liquid nitrogen and extracted by homogenization in 50% (vol/vol) isopropanol in water containing 1% (vol/vol) trifluoroacetic acid. Contents of total protein, IRI, hIAPP-LI, and mouse IAPP-LI (mIAPP-LI) were determined in each pancreatic sample.

Plasma glucose was determined using a glucose oxidase method. IRI was measured by a modification of a previously described radioimmunoassay (RIA) (21,22). Pancreatic mIAPP-LI was determined using a previously described RIA for rodent IAPP (21). Pancreatic hIAPP-LI was measured by RIA using an antibody (antibody no. 17863) against human IAPP, high-performance liquid chromatography-purified ¹²⁵I-human IAPP as a tracer, and a human IAPP standard (Peninsula Laboratories, Belmont, CA). The intra- and interassay coefficients of variation for this assay are 8.0 and 12%, respectively. To measure hIAPP-LI directly in plasma, a more sensitive enzyme immunoabsorbance assay was used, with F024 and F002 as the capture and detection antibodies (a gift from Amylin Pharmaceuticals, San Diego, CA) (23).

At the time they were killed, pancreata were fixed in phosphate-buffered paraformaldehyde (4% wt/vol) and embedded in paraffin. Sections (5 μm) were deparaffinized, rehydrated, and stained with thioflavin S (0.5% in water) as previously described (20). Sections were examined for the presence of islet amyloid using a computer-based quantitative method established in our laboratory (24). Briefly, a digital imaging system consisting of a Zeiss Axioplan fluorescent microscope equipped with a Hamamatsu C4880 fast-cooled camera, and M2 imaging software from Imaging Research (St. Catharines, ON, Canada) was used to calculate the area of each individual islet by manually circumscribing the border of the islet image and allowing the imaging system to calculate the area in pixels. Amyloid area was determined as the area of thioflavin S fluorescence above a preset threshold within each individual islet. From these data, three measurements of islet amyloid formation were derived. First, we determined the proportion of mice within each dietary group that developed islet amyloid. Second, we determined the prevalence of islet amyloid in each mouse, defined as the percentage of islets containing amyloid. Third, we quantified the severity of islet amyloid in each mouse as the percentage of islet area occupied by amyloid (∑ amyloid area / ∑ islet area × 100%). Islet amyloid was quantified in an average of 22 islets from different regions of the pancreas for each individual mouse, and these data were used to generate mean values for each dietary group.

Serial pancreatic sections from six mice in each dietary group were immunostained for insulin, glucagon, somatostatin, and pancreatic polypeptide to determine islet composition of β-, α-, δ-, and pancreatic polypeptide cells. Sections were incubated with the following primary antisera: anti-insulin (1:2,000; Sigma), anti-glucagon (14C, 1:50,000; a gift from Dr. Robert McEvoy), anti-somatostatin (AS10, 1:1,000; a gift from Dr. John Ensinck), or anti-pancreatic polypeptide (1:1,000; a gift from Dr. Thomas Paquette) followed by the appropriate Cy-3–conjugated secondary antisera for visualization of primary antibody binding. Islet area and endocrine cell area (fluorescent area caused by antibody binding) was determined in only those islets visible on all four serial sections (average of 12 islets per mouse), using the same imaging technique as described above.

The incremental insulin response during the intraperitoneal glucose tolerance test (IPGTT) was calculated by subtracting the 18-h fasting IRI level (time 0 of the IPGTT) from that 30 min after glucose administration. Area under the glucose curve was calculated using the trapezoidal method. Data are expressed as the means ± SE. Body weight data were analyzed using a mixed linear model (Procedure MIXED; SAS Institute, Cary, NC) with time on study, diet, and transgenic status as independent variables. Comparisons between the three dietary groups were performed using the Jonckheere-Terpstra nonparametric test for ordinal variables, a test that takes into account the dose effect of increasing dietary fat (25). Two-group comparisons were performed using unpaired t tests, and correlation analyses were performed by simple regression. A P value ≤0.05 was considered significant.

Baseline body weight was determined at 6–8 weeks of age, before mice were placed on their respective diets. There were no differences among any of the groups, except for the hIAPP transgenic mice that subsequently received the high-fat diet, which were significantly lighter than the other dietary groups (P < 0.05). As expected, mice in all six dietary groups gained weight throughout the course of the study (P < 0.001) (Fig. 1). Increased dietary fat was associated with greater body weight (high > medium > low) in both nontransgenic (Fig. 1A) and hIAPP transgenic mice (P < 0.01) (Fig. 1B). There was no effect of genotype (nontransgenic versus hIAPP transgenic) on body weight. The incremental body weight (body weight at sacrifice − baseline body weight) was significantly increased with increased dietary fat in both nontransgenic mice (16.4 ± 4.2, 21.4 ± 4.6, 32.4 ± 3.9 g for low, medium, and high, respectively, P < 0.005) and hIAPP transgenic mice (16.3 ± 1.2, 22.3 ± 1.5, and 33.4 ± 2.7 g for low, medium, and high, respectively, P < 0.001).

Samples of pancreas from all mice were examined for the presence of islet amyloid at the end of the study. As expected, because endogenous murine IAPP is not capable of forming amyloid fibrils, none of the nontransgenic mice developed islet amyloid. In contrast, in hIAPP transgenic mice, intake of low-, medium-, or high-fat diets for 1 year was associated with the development of islet amyloid. The percentage of hIAPP transgenic mice within each dietary group that developed amyloid was similar: 71, 88, and 87% for low-, medium-, and high-fat diets, respectively. Because this measure of islet amyloid formation reveals no information regarding the extent (prevalence and severity) of islet amyloid formation in each individual mouse, we performed a more detailed, quantitative analysis of islet amyloid formation.

First, the prevalence of islet amyloid (number of amyloid-containing islets/number of islets scored per mouse × 100%) was determined for all hIAPP transgenic mice in each dietary group. Increased dietary fat was associated with a significant increase in the prevalence of islet amyloid (P < 0.05) (Fig. 2A). The effect of dietary fat to increase the prevalence of islet amyloid formation was dose dependent, with a stepwise increase in amyloid prevalence from the low- to high-fat diets (34 ± 8 to 45 ± 8 to 58 ± 10%). Second, the severity of islet amyloid (Σ amyloid area / Σ islet area × 100%) was quantified. This analysis demonstrated that the high-fat diet was associated with a greater severity of islet amyloid compared with the low- and medium-fat groups (P = 0.05) (Fig. 2B). The effect of increased dietary fat to increase islet amyloid severity appeared to have a threshold. No difference in islet amyloid severity existed between the low- and medium-fat groups, whereas a striking increase in severity was observed between the medium- and high-fat groups (0.8 ± 0.5 and 1.0 ± 0.5% in the low- and medium-fat groups, respectively, compared with 4.7 ± 2.5% in the high-fat group).

Mean islet size increased with increased dietary fat in nontransgenic mice (2.7 ± 0.3, 4.7 ± 1.1, and 6.3 ± 1.0 × 10⁴ μm² for low, medium, and high, respectively, P < 0.005) and was positively correlated with body weight (r² = 0.38, P < 0.005). In sharp contrast, in the hIAPP transgenic mice, there was no corresponding change in islet size with increased dietary fat (3.3 ± 0.3, 3.5 ± 0.3, and 4.3 ± 0.6 × 10⁴ μm² for mice receiving the low-, medium-, or high-fat diets, respectively), and islet size was not correlated with body weight (r² = 0.07). Thus, dietary fat-induced obesity is associated with a compensatory increase in islet size in nontransgenic mice, whereas this adaptation does not occur in the hIAPP transgenic mice.

In addition to a compensatory increase in islet size, the proportion of insulin-positive area within islets from the nontransgenic mice was maintained among the dietary groups (73.0 ± 1.3, 74.3 ± 2.0, and 74.2 ± 2.6% for low, medium, or high fat, respectively) (Fig. 3A). In contrast, in the hIAPP transgenic mice, the increase in islet amyloid severity observed with increased dietary fat was associated with a decrease in the proportion of insulin-positive area per islet (72.1 ± 1.9, 71.8 ± 1.2, and 67.5 ± 2.0% for low, medium, or high fat, respectively, P < 0.05) (Fig. 3B). The proportion of insulin-positive area was negatively correlated with islet amyloid severity in islets examined from hIAPP transgenic mice fed the medium-fat (r² = 0.14, P < 0.01) and high-fat (r² = 0.24, P < 0.001) diets but not the low-fat diet (r² = 0.001). Thus, the presence of the hIAPP transgene, hIAPP production, and/or islet amyloid contributes to the decrease in the proportion of islet area comprised of β-cells. The proportion of glucagon-, somatostatin-, and pancreatic polypeptide-positive areas were similar among mice fed increasing amounts of dietary fat in both nontransgenic and hIAPP transgenic mice (Table 1).

Pancreas samples were removed at the time they were killed and analyzed for IRI, hIAPP-LI, and mIAPP-LI content (Table 2). Increased dietary fat was not associated with changes in pancreatic content of IRI or murine IAPP (mIAPP) in either nontransgenic or hIAPP transgenic mice. Pancreatic content of hIAPP was similar among hIAPP transgenic mice on the different diets and, as expected, was not detectable in nontransgenic mice. However, both IRI and mIAPP content in hIAPP transgenic mice fed the high-fat diet were significantly lower than in nontransgenic mice fed the same diet (P < 0.05 for both IRI and mIAPP), consistent with the decreased proportion of insulin-positive area in islets from high fat-fed hIAPP transgenic mice. The pancreatic IRI content for nontransgenic mice fed the low- and medium-fat diets were similar to those seen in the hIAPP transgenic mice. Finally, the ratio of mIAPP to IRI did not differ with increased dietary fat in nontransgenic or hIAPP transgenic mice, and in hIAPP transgenic mice, the ratios of hIAPP to mIAPP and hIAPP to IRI were similar among all dietary groups. These data were in agreement with those published by us and others (10,26) and suggest that production of IAPP is not dissociated from that of insulin when dietary fat is increased.

In agreement with the pancreatic content data, plasma hIAPP levels 4 h after food withdrawal in hIAPP transgenic mice were not different among dietary groups (80 ± 16, 83 ± 15, and 76 ± 10 pmol/l for low, medium, and high, respectively). As expected, plasma hIAPP was undetectable in nontransgenic mice.

To relate the effects of increased dietary fat and islet amyloid formation to changes in glucose-stimulated insulin secretion and glucose tolerance, all mice underwent an IPGTT after an 18-h fast when they had received the special diets for 1 year. Significant increases in fasting plasma glucose were observed with a high-fat diet in both nontransgenic and hIAPP transgenic mice (nontransgenic: 6.0 ± 0.3, 6.6 ± 1.0, and 7.2 ± 0.4 mmol/l, P < 0.05; hIAPP transgenic: 6.7 ± 0.3, 6.8 ± 0.3 and 7.9 ± 0.6 mmol/l for low-, medium-, and high-fat diets, respectively, P = 0.05). Fasting plasma glucose levels were not different between nontransgenic and hIAPP transgenic mice fed the same diet. Fasting plasma IRI was elevated with increased dietary fat in nontransgenic mice (112 ± 63, 459 ± 242, and 706 ± 150 pmol/l, P < 0.001) and hIAPP transgenic (181 ± 48, 280 ± 41, and 539 ± 160 pmol/l, P < 0.005) with no statistically significant differences between nontransgenic and hIAPP transgenic mice fed the same diet. The IRI and glucose responses during the IPGTT were used to calculate measures of glucose-stimulated insulin secretion (the IRI response 30 min after glucose administration [ΔI_0–30]) and glucose tolerance (expressed as the area under the glucose curve). Increased dietary fat was not associated with changes in glucose-stimulated IRI release in nontransgenic mice (Fig. 4A). In contrast, there was a marked reduction in IRI release in hIAPP transgenic mice fed increased dietary fat (P = 0.05) (Fig. 4B). Thus, there was a significant decrease in IRI secretion in hIAPP transgenic mice fed the high-fat diet compared with nontransgenic mice on the same diet (P < 0.05), likely due to hIAPP production and secretion and/or islet amyloid formation. No change in glucose tolerance was observed with increased dietary fat in either nontransgenic mice (1,429 ± 96, 1,291 ± 78, and 1,343 ± 76 mmol/l · 120 min for low, medium, and high, respectively) or hIAPP transgenic (1,540 ± 84, 1,517 ± 90, and 1,551 ± 104 mmol/l · 120 min for low, medium, and high, respectively). However, hIAPP transgenic mice fed the high-fat diet were significantly more glucose intolerant than nontransgenic mice receiving the same diet (P < 0.05).

In the present study, we have demonstrated that prolonged dietary fat feeding in hIAPP transgenic mice is associated with a significant increase in both the prevalence and severity of islet amyloid. This increase in islet amyloid was associated with an impairment in glucose-stimulated insulin secretion that was not observed with increased dietary fat alone in a group of nontransgenic mice.

We found that hIAPP transgenic mice receiving the low-fat (15 kcal%) diet had a low prevalence and a low severity of islet amyloid. Moderate dietary fat (30 kcal%) was associated with an increase in the prevalence of islet amyloid formation, whereas the severity remained low. In contrast, in the high-fat (45 kcal%) dietary group, a further increase in prevalence of islet amyloid was accompanied by a fivefold greater severity of amyloid deposition. This marked increase in the severity of islet amyloid occurring only with the highest prevalence of amyloid is in keeping with our recent observation in hIAPP transgenic mice that the severity of islet amyloid only increases exponentially once involvement of nearly all islets has occurred (24). This finding in our hIAPP transgenic mice is in agreement with human data, which demonstrated that large increases in islet amyloid severity occur only when the vast majority of islets are affected (6). An underlying cause of this exponential increase in islet amyloid formation in our hIAPP transgenic mice could be obesity-induced insulin resistance and the resulting increased secretory demand on the β-cell. It is also possible that the presence of a yet-unidentified dietary fat-derived amyloid enhancing factor, the concentration of which could have increased with increased dietary fat, might also have contributed to the increased islet amyloid deposition observed in this study.

Increased dietary fat was associated with a greater than twofold increase in mean islet size in nontransgenic mice. Islet size was strongly correlated with body weight, in keeping with a compensatory increase in islet size as part of an adaptive response to the obesity-induced increased demand for insulin. A similar finding has been made in most (27,28) but not all rodent studies (29). The reason for the discrepancy in islet size between the latter study and our own may be related to differences in weight gain. Over the 12 months of our study, our high fat-fed nontransgenic mice gained an average of 32.4 g, compared with only 9.6 g after 6 months of high fat-feeding in the study by Westermark et al. (29). We also observed no change in the proportion of insulin-, glucagon-, somatostatin-, or pancreatic polypeptide-positive areas per islet in nontransgenic mice, suggesting that normal islet endocrine cell composition was maintained when islet size increased. This finding is consistent with other studies in nondiabetic humans (30) and rodents (31–33) that also reported an increase in β-cell mass in response to obesity. In sharp contrast to the findings in nontransgenic mice, islet size did not increase in the hIAPP transgenic mice receiving increased dietary fat, suggesting that islet amyloid and/or the presence of hIAPP prevented adequate adaptation in islet size in response to diet-induced obesity. This lack of a compensatory increase in islet size was accompanied by a decrease in the proportion of insulin-positive area per islet that correlated with islet amyloid severity, suggesting that islet amyloid formation is associated with β-cell loss. This loss of β-cells has also been described in diabetic humans (30,34). A marked reduction in pancreatic insulin (and mIAPP) content was also observed in the hIAPP transgenic fat-fed mice, providing additional evidence for a lack of islet adaptation to diet-induced obesity. Glucose-stimulated insulin secretion in hIAPP transgenic mice fed a high-fat diet was significantly reduced compared with hIAPP transgenic mice fed the low- and medium-fat diets and to nontransgenic mice on all diets (P < 0.05). These findings are consistent with combined effects of increased dietary fat together with hIAPP production and/or islet amyloid to impair β-cell function.

Collectively, these findings provide insight into β-cell adaptation to dietary fat-associated obesity and the failure thereof in hIAPP transgenic mice. In nontransgenic mice, increased dietary fat is associated with an increase in islet size, the maintenance of the proportion of islet area comprised of β-cells, and increased pancreatic insulin content, suggesting that nontransgenic islets are able to adapt to diet-induced obesity. In contrast, in hIAPP transgenic mice, increased dietary fat is associated with morphologic abnormalities comprising enhanced islet amyloid formation, a lack of an adaptive increase in islet size, and a decrease in the proportion of insulin-positive cells. In addition, decreased pancreatic insulin content and a functional impairment (manifest as a reduction) in glucose-stimulated insulin secretion also occur. This suggests that dual defects, one functional and one morphologic, underlie the impairment in β-cell function seen in hIAPP transgenic mice on a high-fat diet and result in β-cell dysfunction compared with nontransgenic mice. Similarly, morphologic and functional β-cell defects are both found in human type 2 diabetes (35,36).

Fasting plasma glucose levels were increased in the both hIAPP transgenic and nontransgenic mice on the high-fat diet. Interestingly, the hIAPP transgenic and nontransgenic mice fed the low- and medium-fat diets did not differ in their glucose area under the curve after intraperitoneal glucose administration. In contrast, hIAPP transgenic mice fed the high-fat diet were significantly more glucose intolerant than nontransgenic mice on the same diet, suggesting that in the presence of the hIAPP transgene and relatively small amounts of islet amyloid, changes in glucose tolerance may occur. Within mice from the same genotype, however, we did not observe changes in glucose tolerance with increased dietary fat. This lack of change in glucose tolerance was somewhat unexpected in view of the decreased insulin release, particularly in the hIAPP high fat-fed mice. This contrasts with other studies that have shown the development of marked glucose intolerance in mice fed increased dietary fat (37,38). This discrepancy may be related to the genetic background strain of the mice, a factor known to be an important determinant of susceptibility to glucose intolerance and other metabolic abnormalities (39–41). Many observations of dietary fat-induced impaired glucose tolerance and/or diabetes have been made in C57BL/6 mice (37,41), whereas the current study was performed in C57BL/6xDBA/2 mice. Alternatively, it is possible that the contribution of insulin-independent glucose disposal, which is a major component of glucose clearance in mice (42), masked any effect of dietary fat on insulin dependent glucose disposal in the present study.

We have demonstrated that increased dietary fat is associated with increased islet amyloid deposition. Although the mechanism(s) by which this occurred is not known, several possibilities could be considered. Whereas IAPP and insulin are usually cosecreted from the islet β-cell in a fixed molar ratio, the production and secretion of IAPP can be increased independent of insulin (29,33,43,44), resulting in an increase in both plasma levels and pancreatic content of IAPP and a change in the IAPP-to-insulin ratio (45,46). In the present study, the presence of the hIAPP transgene was associated with decreased insulin and mIAPP content in high fat-fed mice compared with nontransgenic controls, but no changes in the pancreatic IAPP-to-insulin ratio were observed with increased dietary fat in either hIAPP or nontransgenic mice. Therefore, whereas diet-induced changes in absolute insulin and IAPP content and/or release may have an impact on islet amyloid formation, the effect of dietary fat to enhance islet amyloid formation in our hIAPP transgenic mice does not appear to occur through a disproportionate increase in IAPP production or secretion. Our observations contrast with those in two other studies of nontransgenic mice fed high-fat diets, which reported disproportionate changes in IAPP mRNA, plasma levels, and pancreatic content compared with insulin (29,33). This disparity may be related to the experimental paradigms that differed in factors such as the genetic background and sex of the mice as well as the saturated fat content of the diets and the duration of their administration. The difference in fat composition suggests that an increased quantity of saturated dietary fat may influence IAPP and insulin production differentially. Furthermore, although the fatty acid composition of the three diets used in the present study was identical, it was different from the fatty acid composition of the diets used in these other studies. Because data from in vitro studies suggest that saturated fatty acids have deleterious effects on β-cell function (47), it is possible that increasing the saturated fatty acid content of the diet above the 25% we used could have further increased islet amyloid formation.

Changes in IAPP processing from its precursor proIAPP have been proposed as a potential mechanism contributing to islet amyloid formation in type 2 diabetes (48). There is evidence that the effect of free fatty acids to alter β-cell function in vitro includes both changes in proinsulin biosynthesis (49,50) and impaired processing of proinsulin to insulin in both human islets (47) and the MIN6 islet β-cell line (51). Although we did not examine proIAPP processing, proIAPP is converted to IAPP by the action of the same enzymes responsible for proinsulin processing (52–54), and it is therefore possible that reduced processing of proIAPP may have occurred in the present study and contributed to islet amyloid formation.

An important observation from the current study is that increased dietary fat and the associated islet amyloid together result in β-cell loss and impaired insulin secretion. The mechanism by which islet amyloid contributes to decreased β-cell mass and function is thought to be through β-cell cytotoxicity. In cell culture, hIAPP is toxic to β-cells, inducing cell death by apoptosis and necrosis (55–58). It is conceivable that in the present study dietary fat had a direct effect on the β-cell to induce islet amyloid formation, and the toxic effect of hIAPP and/or islet amyloid together led to decreased β-cell mass and a decline in β-cell function. The increase in the incidence of type 2 diabetes is closely linked to an increase in obesity in the world’s populations (59). In turn, this increase in obesity is partly caused by an increase in dietary fat intake (60). The results of the present study in an islet amyloid-susceptible animal model demonstrate that increased dietary fat increases both the prevalence and the severity of islet amyloid, decreases the proportion of islet area comprised of β-cells, and reduces pancreatic insulin content and β-cell secretory function. These findings provide a potential mechanism by which increased dietary fat may contribute to the β-cell loss and dysfunction seen in type 2 diabetes.

This work was supported by the Medical Research Service of the Department of Veterans Affairs, National Institutes of Health Grants DK-17047 and DK-50703, and the American Diabetes Association. R.L.H and S.A. were supported by Juvenile Diabetes Foundation fellowships, F.W. by a McAbee Fellowship from the Diabetes Research Council, J.V. by a fellowship from the Spanish Ministry of Science and Technology, and M.C. by a fellowship from the Belgian American Educational Foundation.

We thank Maggie Abrahamson, Robin Vogel, Jira Wade, Ruth Hollingworth, Yuli McCutchen, and Caj Fernstrom for excellent technical support.