Abstract
Dietary insult in early life can affect the development and future function of the endocrine pancreas. We maintained pregnant non-obese diabetic (NOD) mice on a low protein (LP, 8% protein versus control, 20%) diet from conception until the weaning of pups at day 21. Serum insulin and pancreatic insulin content were reduced in LP-fed NOD offspring at 8 weeks, as were serum interferon γ and pancreatic tumor necrosis factor α, while the number of pancreatic islets demonstrating peri-insulitis, and the degree of invasiveness were reduced. To determine if LP caused early morphometric changes in the pancreas, we measured mean islet area at days 3 and 21. Mean islet size did not differ with diet, but by 8 weeks of age LP-fed NOD females exhibited a significantly reduced islet number and mean islet area, and a lower fractional area of pancreas occupied by both α- and β-cells than control-fed mice. The onset of diabetes was delayed in NOD mice of both genders fed LP diet. The mechanism is likely to involve both altered β-cell morphology and function and changes in cytotoxic cytokines.
Introduction
Type 1 diabetes originates with the autoimmune-mediated destruction of pancreatic β-cells, and is characterized by a change in cytokine secretion from a helper T-cell (Th) 2 phenotype with relatively low levels of interferon γ (IFNγ) and tumor necrosis factor α (TNFα) but high levels of interleukin 10 (IL-10), to a Th1 phenotype where these ratios are reversed (Pearl-Yafe et al. 2007). Apoptosis of β-cells is mediated by a number of intracellular pathways, including the generation of nitric oxide, in response to the actions of IFNγ, TNFα and IL-1β. The non-obese diabetic (NOD) mouse is a well-studied model of spontaneous type 1 diabetes in which predominantly the female animals develop an insulitis between 5 and 8 weeks of age, and most will become clinically diabetic by age 20–30 weeks (Delovitch & Singh 1997). Environmental factors such as viruses and dietary antigens have been shown to precipitate autoimmune diabetes both clinically and in animal models such as the NOD mouse and BioBreeding (BB) rat (Lefebvre et al. 2006). However, food antigens can also interact with the gut immune system, resulting in a Th1 cytokine pattern of expression within the Peyer's patches of young NOD mice (Chakir et al. 2005, Barbeau et al. 2007).
Examples of food antigens that can advance the onset of diabetes in NOD mice include wheat flour-based diets and soya, while limiting exposure reduced the incidence of diabetes (Flohé et al. 2003, Schmid et al. 2004). Feeding of wheat proteins resulted in intestinal enteropathy and higher mucosal levels of proinflammatory cytokines (Maurano et al. 2005), while gluten-free diets were protective against diabetes (Funda et al. 1999). The amount of dietary carbohydrates was not important for the timing of diabetes in the BB rat and the NOD mouse (Scott et al. 1989).
The type of dietary protein has a major impact on the incidence of diabetes in the NOD mouse with meat meal or casein resulting in a high rate of onset while casein hydrolysate, a denatured form, or lactalbumin-based diets being relatively protective (Elliott et al. 1988, Beales et al. 2002). Elliott et al. (1988) postulated that the presence of cow's milk casein in commercial mouse chow may be a dietary trigger of diabetes in the NOD mouse if it is introduced at weaning. Although casein hydrolysate was protective against spontaneous diabetes in NOD females the ability of T-cells from these animals to adoptively transfer diabetes was not changed (Hermitte et al. 1995), suggesting that the effects might be due to the alterations in islet biology rather than on T-cell selection and activation. Less attention has been given to the quantity of protein within the diet, although a relatively high-protein diet accelerated the onset of diabetes in the pregnant NOD mouse (Schneider et al. 1996).
Reduced dietary protein content has developmental effects on the pancreatic β-cells of the rat in early life, involving reductions in both islet size and β-cell mass resulting in impaired insulin release and glucose tolerance (Snoeck et al. 1990, Petrik et al. 1999, Boujendar et al. 2002, 2003, Hales & Ozanne 2003), but the effects on the mouse pancreas have not been described. The deficits in rats were reversible following supplementation of a maternal low protein (LP) diet with the amino-acid, taurine (Arany et al. 2004). The effects of a LP diet in early life on susceptibility to autoimmune diabetes in rodent models have not been described. Given the greater apoptotic rate seen in β-cells of neonatal rats exposed to LP (Boujendar et al. 2002) it might be anticipated that the onset of diabetes would occur earlier in the NOD mouse. However, total calorie restriction of the pregnant NOD mouse delayed the onset of diabetes in the female offspring and reduced insulitis at 12 weeks of age (Oge et al. 2007). A human correlate would be the lower incidence of type 1 diabetes seen in individuals born of low birth weight or subjected to micro-nutritional deficiencies (Dahlquist et al. 1996, Fronczak et al. 2003).
The aim of this study was therefore to examine the effects of a LP diet on neonatal pancreatic development in NOD mice, the incidence of insulitis and the age of onset of diabetes.
Materials and Methods
Animals
Male and virgin female NOD mice were purchased from Robarts Research Institute (London, ON, Canada) at 4–6 weeks of age. BALB/c mice of the same age were purchased from Charles River (Montreal, PQ, Canada) and were bred at the Lawson Health Research Institute, London, ON, Canada. The animals were housed in temperature-controlled rooms with a 12 h light:12 h darkness cycle and were given food and water ad libitum. At 8 weeks of age, breeding pairs were caged individually and females were checked daily for the presence of sperm in the vaginal smear. When pregnancy was confirmed, the animals were randomly divided into two groups and fed immediately with a defined diet. The control animals (C) were fed with a 20% (w/w) protein diet. The second group received an 8% (w/w) isocaloric protein diet (LP). Diets were purchased from Bioserv (Frenchtown, NJ, USA). Equal amount of calories were given in both diets by increasing the carbohydrate content in the 8% protein diet. Protein restriction was implemented by reducing casein, which is the main protein source in the 20% diet. The proportion of the amino-acid methionine was also reduced in the LP diet. For each experimental group 7–10 litters of animals were used. The body weight, litter size, and gender of the offspring were recorded at birth. At 21 days of age, all pups from both C and LP groups were weaned and fed with the 20% protein diet until the end of the study. Weight gain of pregnant animals was monitored on a daily basis and for all offspring each week until 8 weeks of age. All procedures were performed with the approval of the animal ethics committee of the University of Western Ontario, and in accordance with the guidelines of the Canadian Council on Animal Care.
BALB/c or NOD mice fed with either LP versus C diet were killed at postnatal day 3 by decapitation and at day 21 by CO2 asphyxiation. At 8 weeks of age, males and females from litters of NOD mice were fasted overnight and blood drawn by lancing the tail vein for blood glucose measurement using a 2 μl volume on a hand held glucometer (Ascencia Dex2, Bayer), and some were killed by CO2 asphyxiation. Their littermates were monitored every week for the onset of diabetes determined by polyuria on two consecutive days followed by a confirmation of glucosuria (6–14 mmol/l) using urine glucose strips (Diastix, Ames, Toronto, Canada). Diabetes was confirmed after an overnight fast when blood glucose was in excess of 11 mM, after which animals were killed by CO2 asphyxiation. Non-diabetic mice were killed at 45 weeks of age. At postnatal days 3 and 21 and at 8 weeks, the pancreas was dissected and cut longitudinally and fixed in 10% formalin for further analysis. Serum was collected for insulin and cytokine determinations.
Immunohistochemistry
Pancreata were fixed in 10% formalin and embedded in paraffin using a standardized protocol, and sections of 5 μm cut and mounted on Superfrost-plus slides (Fischer Scientific, Toronto, ON, Canada). Dual immunohistochemistry was performed to localize glucagon and insulin. Glucagon presence was identified by a modified avidin-biotin peroxidase method and insulin by an alkaline-phosphatase method. In brief, slides were incubated overnight at 4 °C in a humidified chamber with mouse anti-glucagon antibody in a 1:2000 dilution (Sigma) followed by biotinylated horse anti-mouse IgG (Vector Laboratories, Burlingame, CA, USA) for 2 h at room temperature. Slides were washed and incubated with extravidin peroxidase (Sigma) for 30 min at room temperature. Immunoreactivity was then visualized with diaminobenzidine solution (BioGenex, San Ramon, CA, USA). Slides were then incubated overnight at 4 °C in a humidified chamber with rabbit anti-insulin antibody at a 1:200 dilution (Santa Cruz, CA, USA) followed by biotinylated goat-anti rabbit IgG (Vector) for 2 h at room temperature. Slides were washed and incubated with Vectastain ABC-AP reagent (Vector) for 30 min at room temperature. Immunoreactivity was then visualized with Vector Red AP substrate Kit I (Vector). Tissue sections were counter-stained with Carrazi's hematoxylin and mounted under glass coverslips with Eukit (Ruth Wagener Ent, Newmarket, ON, Canada). To establish specificity of the antisera, controls included substitution of the primary antibody with non-immune serum, omission of the secondary antibody, and pre-absorption of the antibody with an excess of homologous antigen.
Morphometrical analysis
The tissue sections were analyzed by light microcopy at a magnification of ×25 or ×400. Analyses were performed with Northern Eclipse (version 6.0) morphometric analysis software (Empix Imaging, Mississauga, ON, Canada). For each animal, five sections per pancreas were examined, taken at 50–60 section intervals to represent the entire tissue. For each section of pancreas, the number of islets with or without insulitis was calculated; as well as the total tissue area, the area of each islet present, the area within each islet occupied by α- or β-cells, and the area of insulitis present within each pancreas section. Insulitis was quantified in NOD mice at 8 weeks of age. At this time, this consisted of predominantly peri-insulitis and occasional non-aggressive insulitis with lymphocytic infiltration in <50% of the islet.
Hormone and cytokine measurement
Blood samples from each animal were collected by killing after overnight fast, and serum was separated and stored at −20 °C. Circulating insulin measurement was performed with a sensitive rat insulin RIA kit (Linco Research Inc., St Charles, MO, USA). Following extraction of pancreas tissue with RNA Trizol the total pancreatic protein was recovered from the protein phase as described by the manufacturer (Life Technologies, BRL). Total insulin content was assayed by RIA and expressed per μg tissue protein. IFNγ, TNFα, IL-4, and IL-10 were measured within serum or pancreatic tissue using ELISA assays for mouse cytokines (BD Biosciences, Mississauga, ON, Canada) according to the instructions of the manufacturers. The sensitivities of detection were as follows: IFNγ (1 pg/ml), TNFα (5 pg/ml), IL-4 (4 pg/ml), IL-10 (0.05 pg/ml). Intra- and inter-assay coefficients of variation were between 4–10% and 6–13% respectively.
Auto-antibody detection
A micro-assay for insulin auto-antibody was performed as described by Yu et al. (2000). In brief, [125]I insulin was incubated with serum (5 μl diluted 1:5) with or without unlabeled insulin for 3 days at 4 °C. Protein A (50%) and Protein G Sepharose (50 μl; Pharmacia Biotech Inc.) were added to the incubation and antibody complexes separated on 96-well filtration plates (Millipore, Nepean, ON, Canada). After washing, scintillation liquid was added to each well and radio-labeled insulin quantified by liquid scintillation counting. The presence of auto-antibodies against glutamic acid decarboxylase 65 (GAD65) in serum from NOD mice was similarly determined (Koczwara et al. 2003).
Statistical analysis
Data are presented as mean±s.e.m. for all offspring when considering weight gain or the time of onset of diabetes. For analyses of circulating insulin or cytokines, or pancreatic histology, two or three animals per litter were examined representing both genders and a total number of animals between 15 and 30. Animals selected were those closest to the mean litter body weight at that age. For histological assessments, five pancreatic tissue sections were used per organ. This number of sections was shown to yield equivalent mean value estimation to more extensive sectioning accounting for the entire pancreas. Growth curves of pregnant mice or offspring were compared by two way repeated measures of ANOVA followed by a Holm-Sidek post hoc analysis. Serum cytokine levels were compared by Kruskal-Wallis non-parametric test followed by Dunn's post test. The analysis of the cumulative incidence of diabetes was performed by the comparison of survival curves with a Log rank test. For other results, group means were compared by t-test or two-way ANOVA according to the number of data sets. Upon significant interactions on the two-way ANOVA, differences between individual group means were analyzed by Duncan's test. In all cases, differences were considered statistically significant at the level of P<0.05.
Results
No differences in food intake were found between LP or control diets for NOD mice within pregnancy or lactation. However, the body weight gain of NOD mice during pregnancy was significantly reduced by the day of birth in animals fed LP diet compared with C-fed animals (Fig. 1A). Despite this, the mean birth weight of the offspring did not change with maternal diet (Fig. 1B and C). Mean litter size and gender ratio were also not altered by diet and the number of surviving offspring after 6 days did not differ (Table 1). Mean birth weight for a non-diabetic prone mouse strain, BALB/c, was similar to that of NOD animals for both control and LP diet, as were the mean litter size, gender ratio, and the weights of offspring at 21 days (not shown). The weight gain of the female NOD offspring between birth and 8 weeks of age was similar between diets, but male offspring fed LP diet until weaning subsequently gained significantly less weight than those receiving C diet.
Change in body weight (A) of pregnant NOD mice from conception to the day of birth (day 20) following feeding a control (open circles) or LP diet (closed circles). Mice fed LP diet showed a significantly reduced gain in weight (*P<0.001). Changes in body weight with age for female and male NOD offspring following feeding with control (open circles) or LP (closed circles) diet are shown in (B) and (C) respectively. Male offspring from mice given LP diet had a significantly reduced weight gain between 4 and 8 weeks of age (*P<0.001). Mean values (±s.e.m.) are shown for 10 pregnant mice and 20–30 offspring followed until 8 weeks (2–3 animals per litter).
Citation: Journal of Endocrinology 201, 2; 10.1677/JOE-09-0002
Mean litter size (±s.e.m.) at birth, number of pups surviving after 6 days, and gender ratio for offspring of non-obese diabetic (NOD) mice fed either control of low protein (LP) diet from conception until weaning
| Pups born | Pups survived | |||
|---|---|---|---|---|
| Control | LP | Control | LP | |
| Number of pups | 8.71±0.78 | 7.50±0.82 | 5.71±0.52 | 5.50±0.89 |
| Gender ratio (F/M) | 1.09±0.11 | 1.31±0.36 | 1.04±0.19 | 1.10±0.40 |
Values represent 10 litters of animals.
To determine if there were developmental changes in the size of islets in the NOD mouse that pre-dated insulitis, animals were examined at postnatal days 3 and 21 in comparison with BALB/c mice. Mean islet area in NOD, C-fed mice did not differ with gender but was significantly lower than in the BALB/c strain mice at day 21 (Table 2). Also at day 21, the mean area of islets occupied by insulin-immunoreactive β-cells was lower in NOD females (53±2%, P<0.05) than in BALB/c female mice (67±1%). No such differences were seen in males (NOD 65±5%, BALB/c 75±3%). Exposure to LP diet did not significantly alter mean islet area in NOD or BALB/c mice at either age (Table 2). This suggests that β-cell mass might already be lower in NOD female mice compared with a non-diabetic strain prior to the development of insulitis, and that this was not significantly reduced further by LP exposure.
Mean islet area (μm2±s.e.m.) at postnatal days 3 and 21 for BALB/c or non-obese diabetic (NOD) mice. Animals were the offspring of mice fed either control of low protein (LP) diet from conception until weaning
| Mean islet area | ||||
|---|---|---|---|---|
| Day 3 | Day 21 | |||
| Control | LP | Control | LP | |
| BALB/c male | 6244±933 | 5621±363 | 10 171±339 | 9579±280 |
| BALB/c female | 8965±348 | 7583±277 | 11 835±923 | 9847±665 |
| NOD male | 4864±1073 | 5119±622 | 6049±847* | 5398±787* |
| NOD female | 6012±1811 | 5395±448 | 6256±715* | 5005±641* |
Values represent 10–15 animals derived from five separate litters of animals. *P<0.05 versus BALC/c mice of the same gender.
When islet histomorphometry was analyzed at 8 weeks of age, LP-fed NOD female mice had a significantly lower number of islets per unit area of pancreas and a lower mean islet area (Fig. 2A and B). The percentage areas of pancreas occupied by insulin-immunopositive β-cells or glucagon-immunoreactive α-cells were reduced by half in LP compared with C-fed mice (Fig. 2C and D). In males, there was no significant reduction in islet number or in mean islet area (C, 5235±1862 μm2; LP, 3910±651 μm2).
The number of islets (mean±s.e.m.) per sectional area of pancreas (A), mean area of islets (B), and the percentage area of pancreas occupied by glucagon-immunoreactive α-cells (C) or insulin-immunoreactive β-cells (D) in female offspring of control (open bars) or LP-fed (LPD, closed bars) NOD mice at age 8 weeks. *P<0.05 versus control fed mice (n=25 animals, 2–3 per litter).
Citation: Journal of Endocrinology 201, 2; 10.1677/JOE-09-0002
At 8 weeks of age, serum insulin levels in the NOD offspring were significantly lower in animals that received LP diet compared with C diet regardless of gender (Fig. 3A), as was the pancreatic insulin content (Fig. 3B). Animals not yet diabetic at 45 weeks of age had lower circulating insulin levels than at 8 weeks and this did not differ with gender or diet. In non-diabetic NOD mice, mean blood glucose values were similar between males and females at 8 and 45 weeks of age in animals fed either LP or C diet (not shown). At the time of onset of diabetes, there was also no significant difference in serum insulin with diet or gender (Fig. 3A).
Serum insulin levels in female and male offspring (A) of control (open bars) or LP-fed (LPD, closed bars) NOD mice at age 8 weeks, for animals not yet diabetic at 45 weeks, and at the time of onset of diabetes. The pancreatic content of insulin in female animals is shown in (B) at 8 weeks of age. *P<0.05, **P<0.005 versus control fed, aP<0.05, aaP<0.001 vs 8 weeks for same diet. Mean value (±s.e.m.) are shown for 15–20 mice (2–3 animals per litter).
Citation: Journal of Endocrinology 201, 2; 10.1677/JOE-09-0002
Since cytotoxic cytokine presence is associated with autoimmune destruction of β-cells in the NOD mouse, we measured the circulating levels of IFNγ, TNFα and IL-4 in pre-diabetic animals and at the diagnosis of diabetes. Circulating levels of IL-4, thought to be protective against the onset of diabetes in the NOD mouse, were significantly greater in NOD females, but not males, fed LP diet at 8 weeks of age (Fig. 4). IL-4 levels were generally lower in non-diabetic animals at 45 weeks, and following the onset of diabetes, and did not differ with diet. At 8 weeks of age, circulating levels of IFNγ were significantly lower in LP-fed diabetes-prone female offspring (Fig. 4), but not so in males. At diabetes onset serum IFNγ was lower in both female and male mice in animals that had received LP diet. This did not occur in mice of either sex that had not become diabetic by 45 weeks. Serum levels of TNFα were substantially increased at the onset of diabetes compared with non-diabetic female mice, regardless of diet, and less so in males (Fig. 4). However, the pancreatic content of TNFα in 8-week female offspring was significantly lower in LP-fed mice (control 94±6 μg/mg protein, LP 42±2 μg/mg, P<0.05 versus control; mean±s.e.m., n=5). Circulating IL-10 levels did not significantly differ with gender or diet at 8 weeks of age or at the onset of diabetes (not shown). These findings suggest a selective change occurred in the levels of certain cytotoxic cytokines occurred in both serum and pancreas of pre-diabetic animals previously exposed to LP diet.
Serum levels of IL-4, IFNγ or TNFα (logarithmic scale) in female or male offspring of control (open circles) or LP-fed (closed circles) NOD mice at age 8 weeks, for animals not yet diabetic at 45 weeks, and at the time of onset of diabetes. Individual values are shown for 15–20 mice (2–3 animals per litter). *P<0.05 versus control fed animals.
Citation: Journal of Endocrinology 201, 2; 10.1677/JOE-09-0002
Antibodies against insulin were not detected in serum of any diabetic or pre-diabetic NOD mouse. Prior to diabetes, GAD65 antibody was present in the majority of female and in ∼50% of male NOD mice both following C diet at 8 weeks of age, and this did not differ with LP diet. GAD65 antibody was detectable in all animals once diabetic, but not in those that had not become diabetic by 45 weeks of age, regardless of diet.
The number of pancreatic islets showing evidence of peri-insulitis at 8 weeks of age in NOD female mice was substantially reduced in animals that had received LP diet (Fig. 5A, C and D). Similarly, when the area of pancreas occupied by insulitis was quantified it was significantly reduced following LP diet (Fig. 5B). Male animals showed a similar percentage of islets with peri-insulitis at 8 weeks of age to females and this did not differ with diet (C, 13±5%; LP 8±3%).
Percentage of islets (mean±s.e.m.) that showed associated peri-insulitis (A) and the percentage of pancreas area accounted for by insulitis (B) in female offspring of control (open bars) or LP-fed (LPD, closed bars) NOD mice at age 8 weeks. *P<0.05 versus control fed mice (n=25 animals, 2–3 per litter). The immunohistochemical localization of insulin (red) and glucagon (brown) is shown within islets without (C) or with (D) peri-insulitis (arrow) in a control-fed animal. Magnification bar 100 μm.
Citation: Journal of Endocrinology 201, 2; 10.1677/JOE-09-0002
The cumulative incidence of diabetes was compared for NOD males and females previously given LP or control diet (Fig. 6). By 45 weeks of age, 86% of female NOD mice that had received C diet had become diabetic, but only 63% of LP-fed animals. Of the diabetic animals, the age at which 50% of these had become diabetic was 23.3 weeks following C diet but 30.8 weeks following LP diet. Only 60% of male animals receiving C diet were diabetic by 45 weeks and 15% of animals receiving LP diet.
Kaplan–Meier plots of the onset of diabetes in NOD female or male mice with increasing age until 45 weeks of age for control-fed (broken line, n=25) or LP-fed (closed line, n=40) animals. 95% confidence limits are shown. P<0.05 for LP-fed versus control-fed females, P<0.01 for males.
Citation: Journal of Endocrinology 201, 2; 10.1677/JOE-09-0002
Discussion
Nutritional sufficiency in early life can influence the incidence and time of onset of autoimmune diabetes in animal models (Pedersen et al. 1999, Oge et al. 2007). Altered dietary protein content has been shown to exacerbate diabetes, a high-protein diet accelerated the onset of disease (Schneider et al. 1996) while a protein-rich high-fat diet increased β-cell loss by apoptosis in pre-diabetic NOD mice (Linn et al. 1999). We report here that a LP but isocalorific diet can delay the onset of diabetes in the NOD mouse. However, it is not clear if the altered ontogeny of disease is primarily mediated through changes in the development and function of the pancreatic β-cell, by changes in T-cell trafficking or cytokine release, or both. It must also be considered that the deficiency of calories due to protein in the diet was balanced by additional carbohydrate, which may also have contributed to the delay in autoimmune diabetes, although a previous report using NOD mice suggested that carbohydrate content of diet did not change the onset of diabetes (Scott et al. 1989).
The provision of LP diet during gestation and lactation in the normal rat resulted in only a mild reduction in birth weight, but severe changes in pancreatic phenotype of the offspring (Snoeck et al. 1990). We found here that in the NOD mouse, feeding of the same LP diet limited maternal weight gain during pregnancy but did not significantly alter birth weight of the offspring, mean litter size, infant survival or gender ratio. The subsequent weight gain of the offspring from LP-fed animals was retarded, but only in males, who have a lower incidence of diabetes. Thus, the general body growth rate of the diabetes-prone females was unaltered. Feeding of LP diet to both NOD and BALB/c mice in early life did not significantly reduce birth weight or neonatal islet morphology showing that the phenotype in the offspring in this species is noticeably different from that reported in rat.
Regardless of diet, we observed that the mean islet area in both female and male NOD animals was significantly less than that found in BALB/c mice at day 21, with a reduced percent contribution of β-cells to islets in NOD compared with BALB/c females. NOD offspring have a higher islet endocrine mass in the first week of life compared with C57BL/6 or BALB/c strains, due in one report to greater numbers of small islets (Palegri et al. 2001) and in another to a greater mean area of larger islets (Geutskens et al. 2004). These differences were less apparent by weaning, but islets from NOD animals had a greater percentage composition of glucagon-positive α-cells suggesting that they were relatively immature (Palegri et al. 2001). We did not detect a greater mean area of islets in NOD versus BALB/c mice on day 3, but it is possible that this was missed since no further tissues were examined until day 21. The extracellular matrix composition of pancreas in neonatal NOD mice showed increased deposition of fibronectin around the islets (Geutskens et al. 2004). Such fibronectin-rich sites are associated with the localization of macrophages during insulitis, which might contribute to the precipitation of diabetes.
Substantial remodeling of rodent islets occurs between birth and weaning with β-cell turnover by apoptosis and islet replacement through neogenesis (Petrik et al. 1998). The rate of β-cell apoptosis at this time was elevated in the female NOD mouse, leading to a relatively reduced β-cell mass (Trudeau et al. 2000). When neonatal β-cell remodeling was prevented in the NOD mouse by trans-placental exposure to bafilomycin, a vacuolar-ATPase inhibitor, this was associated with an accelerated onset of diabetes, suggesting that disruption of β-cell ontogeny renders the adult cells more susceptible to cytokine-induced death (Hettiarachchi et al. 2008).
The mean pancreatic islet size was not significantly different in offspring of control versus LP-fed NOD mice on postnatal days 3 or 21, prior to the commencement of insulitis. This is in agreement with the intrauterine exposure of NOD mice to a reduced calorie diet where β-cell mass in the offspring was unaltered at postnatal day 14 relative to control-fed animals (Oge et al. 2007). However, by 8 weeks of age, female offspring exposed to LP diet had significantly reduced mean islet number and size compared with control-fed mice. This was not likely to be due to autoimmune loss of β-cells since the extent of insulitis was much reduced in LP-fed mice, and the smaller islet size was associated with a reduction in both β- and α-cell components. No such changes were seen in male offspring suggesting that there might be a differential developmental deficiency in the islets of the female NOD mouse.
The higher rate of developmental β-cell turnover and cell death reported in the neonatal female NOD mouse (Trudeau et al. 2000) could act as an initiator for T-cell activation and later insulitis. This may result in a higher metabolic stress on individual β-cells resulting in additional cell damage and auto-antigen release. It has been shown that at 4–8 weeks of age, NOD and NOD/SCID mice present a paradoxical but transient hyperinsulinemia. This is manifested as a lower glycemia that is preceded by transient, perinatal β-cell hyperactivity, particularly in the females (Homo-Delarche 1997, Rosmalen et al. 2002). The beneficial effects of prior LP exposure on preventing insulitis may be related to a reduction in β-cell metabolic stress. We found that at 8 weeks of age serum insulin levels in the offspring were significantly lower in animals that received LP diet regardless of gender, as was the pancreatic insulin content. This might be expected to correlate with a reduced presence of islet auto-antibodies. However, antibodies to insulin were not detectable in these mice while anti-GAD65 antibodies were present in all females at 8 weeks of age, and some males, regardless of diet, suggesting that a primary effect of LP diet on the delay in diabetes is likely not due to altered GAD65 antibody presence.
It seems unlikely that β-cells in NOD females that received LP diet were more resistant to cytokine-induced damage than those of control-fed mice, leading to less auto-antigen release, since islets from rats fed LP diet in early life were more susceptible to cytokine-induced apoptosis (Merezak et al. 2004). However, the balance of cytotoxic cytokines might be altered in the pancreas of the LP-fed NOD mouse resulting in less insulitis and a delay in β-cell loss. A direct effect of LP diet on the development of the immune system is likely to contribute to the delayed onset of diabetes seen in these studies. The maternal immune system is known to be influenced by gestational nutritional deficiency, with a reduced immune function in the offspring (Chandra 1975). Lifelong food restriction reduces antigen-triggered memory T-cell responses and increases levels of the anti-inflammatory IL-2, while lowering levels of IL-6 and TNFα. Restriction of energy intake increases circulating corticosterone levels, which in turn enhances negative selection of potentially autoreactive CD4–CD8 cells in thymus (Wilder 1995). In our studies, female animals fed LP diet in early life had lower circulating levels of IFNγ at 8 weeks of age, and also at the onset of diabetes. Serum levels of TNFα were not altered, but the pancreatic content of TNFα was significantly lower in LP-fed female NOD mice, consistent with a reduction in the number of islets demonstrating insulitis, and the degree of insulitis, where this was present. While the onset of autoimmune diabetes is more rapid in the female NOD mouse, the ability of LP diet in early life to delay the onset was seen in both sexes, implying that effects of diet on subsequent cytotoxic cytokine levels are likely to be gender independent.
In summary, we found that feeding a LP diet to NOD mice throughout gestation and until weaning selectively decreased levels of cytotoxic cytokines, reduced insulitis, and delayed the onset of diabetes. In females, LP diet was also associated with a reduction in the number and size of pancreatic islets, a lower pancreatic insulin content, and lower serum insulin by 8 weeks of age. The mechanism by which LP diet delays the onset of diabetes is likely to involve both immune alterations and changes in β-cell development.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding
This work was supported by the Canadian Institutes of Health Research (2005) and the Juvenile Diabetes Research Foundation (2003).
Acknowledgements
We would like to thank Brenda Strutt for her expertise and technical support and the animal care and support workers for their hard work and assistance.
References
Arany E, Strutt B, Romanus P, Remacle C, Reusens B & Hill DJ 2004 Taurine supplement in early life altered islet morphology, decreased insulitis and delayed the onset of diabetes in non-obese diabetic mice. Diabetologia 47 1831–1837.
Barbeau WE, Bassaganva-Riera J & Hontecillas R 2007 Putting the pieces of the puzzle together – a series of hypotheses on the etiological pathogenesis of type 1 diabetes. Medical Hypotheses 68 607–619.
Beales PE, Elliot RB, Flohe S, Hill JP, Kolb H, Pozzilli P, Wang G-S, Wasmuth H & Scott FW 2002 A multi-centre, blinded international trial of the effect of A1 and A2 β-casein variants on diabetes incidence in two rodent models of spontaneous type 1 diabetes. Diabetologia 45 1240–1246.
Boujendar S, Reusens B, Merezak S, Ahn MT, Arany E, Hill D & Remacle C 2002 Taurine supplementation to a low protein diet during foetal and early postnatal life restores a normal proliferation and apoptosis of rat pancreatic islets. Diabetologia 45 856–866.
Boujendar S, Arany E, Hill D, Remacle C & Reusens B 2003 Taurine supplementation during fetal life reverses the vascular impairment caused to be endocrine pancreas by a low protein diet. Journal of Nutrition 133 2820–2825.
Chakir H, Lefebvre DE, Wang H, Caraher E & Scott FW 2005 Wheat protein-induced proinflammatory T helper 1 bias in mesenteric lymph nodes of young diabetes-prone rats. Diabetologia 48 1576–1584.
Chandra RK 1975 Antibody formation in first generation offspring of nutritionally deprived rats. Science 190 289–290.
Dahlquist G, Bennich SS & Källén B 1996 Intrauterine growth pattern and risk of childhood onset insulin dependent (type I) diabetes: population based case–control study. BMJ 313 1174–1177.
Delovitch TL & Singh B 1997 The nonobese diabetic mouse as a model of autoimmune diabetes: immune dysregulation gets the NOD. Immunity 7 727–738.
Elliott RB, Reddy SN, Bibby NJ & Kida K 1988 Dietary prevention of diabetes in the non-obese diabetic mouse. Diabetologia 31 62–64.
Flohé SB, Wasmuth HE, Kerad JB, Beales PE, Pozzilli P, Elliott RB, Hill JP, Scott FW & Kolb H 2003 A wheat-based, diabetes-promoting diet induces a Th1-type cytokine bias in the gut of NOD mice. Cytokine 21 49–54.
Fronczak CM, Barón AE, Chase HP, Ross C, Brady HL, Hoffman M, Eisenbarth GS, Rewers M & Norris JM 2003 In utero dietary exposures and risk of islet autoimmunity in children. Diabetes Care 26 3237–3242.
Funda DP, Kaas A, Bock T, Tlaskalova-Hogenova H & Buschard K 1999 Gluten-free diet prevents diabetes in NOD mice. Diabetes/Metabolism Research and Reviews 15 323–327.
Geutskens SB, Homo-Delarche F, Pleau J-M, Durant S, Drexhage HJ & Savino W 2004 Extracellular matrix distribution and islet morphology in the early postnatal pancreas: anomalies in the non-obese diabetic mouse. Cell and Tissue Research 318 579–589.
Hales CN & Ozanne SE 2003 For debate: fetal and early post-natal growth restriction lead to diabetes, the metabolic syndrome and renal failure. Diabetologia 46 1013–1019.
Hermitte L, Atlan-Gepner C, Payan MJ, Mehelleb M & Vialettes B 1995 Dietary protection against diabetes in NOD mice: lack of a major change in the immune system. Diabetes & Metabolism 21 261–268.
Hettiarachchi KD, Zimmet PZ, Danial NN & Myers MA 2008 Transplacental exposure to the vacuolar-ATPase inhibitor bafilomycin disrupts survival signaling in beta cells and delays neonatal remodeling of the endocrine pancreas. Experimental and Toxicologic Pathology 60 295–306.
Homo-Delarche F 1997 Beta cell behaviour during the pre-diabetic stage. Part II. Non-insulin-dependent and insulin dependent diabetes mellitus. Diabetic Medicine 23 473–505.
Koczwara K, Schenker M, Schmid S, Kredel K, Ziegler AG & Bonifacio E 2003 Characterization of antibody responses to endogenous and exogenous antigen in the nonobese diabetic mouse. Clinical Immunology 106 155–162.
Lefebvre DE, Powell KL, Strom A & Scott FW 2006 Dietary proteins as environmental modifiers of type 1 diabetes mellitus. Annual Review of Nutrition 26 175–202.
Linn T, Strate C & Schneider K 1999 Diet promotes beta-cell loss by apoptosis in prediabetic nonobese diabetic mice. Endocrinology 140 3767–3773.
Maurano F, Mazzarella G, Luango D, Stefanile R, D'Arienzo R, Rossi M, Auricchio S & Troncone R 2005 Small intestine enteropathy in non-obese diabetic mice fed a diet containing wheat. Diabetologia 48 931–937.
Merezak S, Reusens B, Renard A, Goosse K, Kalbe L, Ahn MT, Tamarit-Rodriguez J & Remacle C 2004 Effect of maternal low-protein diet and taurine on the vulnerability of adult Wistar rat islets to cytokines. Diabetologia 47 669–675.
Oge A, Isganaitis E, Jimenez-Chillaron J, Reamer C, Faucette R, Barry K, Przybyla R & Patti ME 2007 In utero undernutrition reduces diabetes incidence in non-obese diabetic mice. Diabetologia 50 1099–1108.
Palegri C, Rosmalen J, Durant S, Throsby M, Alves V, Coulaud J, Esling A, Pleau J-M, Drexhage H & Homo-Delarche F 2001 Islet endocrine-cell behavior from birth onward in mice with non-obese diabetic genetic background. Molecular Medicine 7 311–319.
Pearl-Yafe M, Kaminitz A, Yolcu ES, Yaniv I, Stein J & Askenasy N 2007 Pancreatic islets under attack: cellular and molecular effectors. Current Pharmaceutical Design 13 749–760.
Pedersen CR, Hagemann I, Bock T & Buschard K 1999 Intermittent feeding and fasting reduces diabetes incidence in BB rats. Autoimmunity 30 243–250.
Petrik J, Arany E, McDonald TJ & Hill DJ 1998 Apoptosis in the pancreatic islet cells of the neonatal rat is associated with a reduced expression of insulin-like growth factor-II which may act as a survival factor. Endocrinology 139 2994–3004.
Petrik J, Reusens B, Arany E, Remacle C, Hoet JJ & Hill DJ 1999 A low protein diet alters the balance of islet cell replication and apoptosis in the fetal and neonatal rat, and is associated with a reduced pancreatic expression of insulin-like growth factor II. Endocrinology 140 4861–4873.
Rosmalen JG, Leenen PJ, Pelegri C, Drexhage H & Homo-Delarche F 2002 Islet abnormalities in the pathogenesis of of autoimune diabetes. Trends in Endocrinology and Metabolism 13 209–214.
Schmid S, Koczwara K, Schwinghammer S, Lampasona V, Ziegler AG & Bonifacio E 2004 Delayed exposure to wheat and barley proteins reduces incidence in non-obese diabetic mice. Clinical Immunology 111 108–118.
Schneider K, Laube H & Linn T 1996 A diet rich in protein accelerates diabetes manifestation in NOD mice. Acta Diabetologia 33 236–240.
Scott FW, Elliot RB & Kolb H 1989 Diet and autoimmunity: prospects of prevention of type I diabetes. Diabetes, Nutrition & Metabolism 2 61–73.
Snoeck A, Remacle C, Reusens B & Hoet JJ 1990 Effect of a low protein diet during pregnancy on the fetal rat endocrine pancreas. Biology of the Neonate 57 107–118.
Trudeau JD, Dutz JP, Arany E, Hill DJ, Fieldus WE & Finegood DT 2000 Neonatal β-cell apoptosis: a trigger for autoimmune diabetes? Diabetes 49 1–7.
Wilder RL 1995 Neuroendocrine–immune system interactions and autoimmunity. Annual Review of Immunology 13 307–338.
Yu L, Robles DT, Abiru N, Kaur P, Rewers M, Kelemen K & Eisenbarth GS 2000 Early expression of anti-insulin autoantibodies of humans and the NOD mouse: evidence for early determination of subsequent diabetes. PNAS 97 1701–1706.
