Dietary iron restriction or iron chelation protects from diabetes and loss of beta-cell function in the obese (ob/ob lep-/-) mouse

Abstract

Iron overload can cause insulin deficiency, but in some cases this may be insufficient to result in diabetes. We hypothesized that the protective effects of decreased iron would be more significant with increased beta-cell demand and stress. Therefore, we treated the ob/ob mouse model of type 2 diabetes with an iron-restricted diet (35 mg/kg iron) or with an oral iron chelator. Control mice were fed normal chow containing 500 mg/kg iron. Neither treatment resulted in iron deficiency or anemia. The low-iron diet significantly ameliorated diabetes in the mice. The effect was long lasting and reversible. Ob/ob mice on the low-iron diet exhibited significant increases in insulin sensitivity and beta-cell function, consistent with the phenotype in mouse models of hereditary iron overload. The effects were not accounted for by changes in weight or feeding behavior. Treatment with iron chelation had a more dramatic effect, allowing the ob/ob mice to maintain normal glucose tolerance for at least 10.5 wk despite no effect on weight. Although dietary iron restriction preserved beta-cell function in ob/ob mice fed a high-fat diet, the effects on overall glucose levels were less apparent due to a loss of the beneficial effects of iron on insulin sensitivity. Beneficial effects of iron restriction were minimal in wild-type mice on normal chow but were apparent in mice on high-fat diets. We conclude that, even at "normal" levels, iron exerts detrimental effects on beta-cell function that are reversible with dietary restriction or pharmacotherapy.

Fig. 1.

Glucose tolerance and weights in mice fed normal (nl) chow or a low-iron diet. Intraperitoneal glucose tolerance testing was performed on ob/ob mice on nl chow at age 4–6 wk, after which time they were maintained on that diet or placed on a lower-iron diet (35 mg/kg). The mice were retested 4 wk later. Wild-type C57BL6 mice, the background strain of the ob/ob mice, were also tested at the same age, 8–10 wk. A: results of intraperitoneal glucose tolerance testing. B: fasting insulin levels. C: homeostasis model assessment of insulin resistance (HOMA-IR). D: homeostasis model assessment of β-cell function (HOMA-B). E: weights of the mice at the 2nd glucose tolerance test. F: values for HOMA-IR were plotted as a function of weight for the individual mice studied. The relation of HOMA-IR to weight for the ob/ob mice on low iron was significant (r² = 0.35, P = 0.006), but there was no such relationship for the ob/ob mice on higher iron. Shown are means ± SE; n = 12–20/group. In A, all glucose values differ significantly among the 3 groups of ob/ob mice; P < 0.001. In B and C, *P < 0.01 compared with prediet levels in the ob/ob mice on normal chow; ‡P < 0.01 compared with ob/ob mice on normal chow.

Fig. 2.

Reversibility and duration of effects of low iron on glucose tolerance. Ob/ob mice aged 4 wk had glucose tolerance testing performed, after which they were randomized to either the low-iron or nl chow diets. After retesting 4 wk later, the diets of the 2 groups were reversed, with repeat glucose tolerance testing performed 7, 17, and 90 days after the switch. A: areas under the glucose curve (AUC_G) for the 2 groups over time (n = 5–12/group, *P < 0.05 for the difference between the 2 groups at the indicated times). B: The AUC_G values for the groups at the end of the study compared with wild-type mice of the same age on the 2 diets (*P < 0.01 for the low-iron group compared with the same strain on nl chow). C: weights of the mice at the end of the study (*P < 0.01 for the low-iron group compared with the same strain on nl chow).

Fig. 3.

Mice on powdered diets containing 35 mg/kg iron exhibit better glucose tolerance than mice fed the same powdered diet supplemented with carbonyl iron to 500 mg/kg. Ob/ob mice (10/group) were placed on the powdered diet containing 35 mg/kg iron, with or without added carbonyl (elemental) iron, to total 500 mg/kg. Glucose tolerance testing was performed at 6 wk (A), at which time the mice were also weighed (B). Individual glucose values are significantly different at 30, 60, and 120 min.

Fig. 4.

Maintenance of insulin secretion in the face of obesity in the ob/ob mice on low-iron chow. From the intraperitoneal glucose tolerance test studies described in Fig. 3, the insulin level 30 min after glucose challenge and the AUC_G values were calculated for mice on 500 mg/kg iron chow (○) and mice on reduced (35 mg/kg) iron chow (■). Insulin levels (solid lines) are presented as ng/ml. The AUC_G values were g·min⁻¹·dl over the 120-min study. The ratios of insulin to AUC_G (dashed lines) were multiplied by 100 to allow visualization on the same graphing scale. Shown are means ± SE; n = 5–12/group. For the insulin values, error bars are within the data points and thus not visible. *P < 0.001 for low-iron group vs. normal chow; ‡P < 0.01 within groups for change from previous value.

Fig. 5.

Effects of iron chelation on glucose tolerance in ob/ob mice. At 30 days of age, mice on normal chow containing 500 mg/kg iron (5/group) were treated with 10 mg/kg of an iron chelator mixed in 100 mg of peanut butter delivered once a day. Control mice were treated with vehicle alone. Inspection verified consistently complete ingestion of the peanut butter in both groups. Glucose tolerance was assessed at 3.5, 6.5, and 10.5 wk after therapy was initiated. A: AUC_G was calculated after intraperitoneal glucose tolerance testing. B: fasting glucose values. C: weight of the mice over the treatment period. D–F: at the 6.5-wk test, insulin values were obtained at fasting (D) and at 30 min to allow calculation of the ratio of stimulated insulin to the AUC_G (E); HOMA-IR values were also calculated (F). Low fasting glucose values in the treated mice precluded the obtaining of meaningful HOMA-B measurements. G: at the end of 10.5 wk of therapy, the mice were studied in a calorimeter. H: mice were treated for 1 wk with the chelator, at which time they were euthanized, and mitochondrial oxygen consumption was determined in permeabilized cardiac muscle fibers; n = 5 mice/group. *P < 0.05–0.001; see text for specific probability values.