Thirty four-week-old C57BL/6J male mice (Jackson Laboratories, Bar Harbor, ME, USA) were randomly assigned to six dietary treatment groups (n = 5). These diets were 10% kcal derived from fat + 36.9 mg Fe/kg diet (control fat control iron, CF/CI, Cat# D13010403), 10% kcal derived from fat + 529 mg Fe/kg diet (control fat high iron, CF/HI, Cat# D13010405), 10% kcal derived from fat + 3.6 mg Fe/kg diet (control fat low iron, CF/LI, Cat# D13010401), 45% kcal derived from fat + 36.9 mg Fe/kg diet (high fat control iron, HF/CI, Cat# D13010404), 45% kcal derived from fat + 529 mg Fe/kg diet (high fat high iron, HF/HI, Cat# D13010406), and 45% kcal derived from fat + 3.6 mg Fe/kg diet (high fat low iron, HF/LI, Cat# D05101905) (Research Diets Inc., New Brunswick, NJ, USA). Other major ingredients in the experimental diets, expressed as grams per kilogram diet, include casein (200 g/kg), l-Cystine (3 g/kg), cornstarch (high fat: 452.2 g/kg, low fat: 72.8 g/kg), sucrose (172.8 g/kg), cellulose (50 g/kg), soybean oil (25 g/kg), mineral mix S18708 (10 g/kg), and vitamin mix V10001 (10 g/kg). Diets and de-ionized water were given to mice ad libitum. Mice were housed individually in a temperature-controlled room. The room temperature was maintained at 25 ± 1°C, with each dark cycle occurring between 7 p.m. and 7 a.m. daily.

Dietary intake and mice body weight were measured weekly. During the 24th week, mice from each treatment group (n = 5), except mice fed with HF/CI diet, were recorded for stereotypical behaviors. Mice fed with HF/LI diet developed ulcerative dermatitis and were euthanized at the end of 16th week after stereotypical behaviors were recorded. The rest of the treatment groups were euthanized at the end of the 24th week_. Blood, liver, and brain tissues were collected. Brains were dissected into the hippocampus, midbrain, striatum, and thalamus. Tissues were snap frozen in liquid nitrogen and stored at −80°C. Hematocrit levels were measured using a micro-capillary centrifuge (Model MB, IEC, Needham Heights, MA, USA) by spinning samples at 10,000 rpm for 10 min. Hemoglobin concentrations were measured at 540 nm following the manufacture’s protocol (Sigma #9008-020).

All studies were conducted in an American Association for Laboratory Animal Care-accredited facility following protocols approved by the Institution of Animal Care and Use Committee (IACUC) at the University of North Carolina at Greensboro (UNCG). The procedures were performed by the principles and guidelines established by the National Institutes of Health for the care and use of laboratory animals.

Liver triglyceride (TG) concentrations were determined by a colorimetric assay (41). Briefly, liver tissues (100–300 mg) were weighed and placed into ethanolic potassium hydroxide solution (1 part of 100% ethanol:2 parts of 30% KOH). The mixture was incubated at 55°C overnight and then centrifuged at 14,000 rpm at 4°C for 5 min. The supernatant was mixed with 1M magnesium chloride (MgCl₂) and incubated on ice for 10 min. The mixture was centrifuged at 14,000 rpm at 4°C for 5 min. The supernatant was taken to measure TG content using free glycerol reagent (Cat#F6428, Sigma-Aldrich, St Louis, MO, USA) and glycerol standards (Cat#G7793, Sigma-Aldrich, St Louis, MO, USA) following the manufacturer’s protocol.

Brain samples were sonicated in the cold radioimmunoprecipitation assay buffer (1% non-idet-P40, 0.1% sodium dodecyl sulfate, 0.5% sodium deoxycholate in the phosphate saline buffer, pH = 7.5) containing protease inhibitors. Separate aliquots of each sonicated sample were assayed for protein concentration [Pierce Bicinchoninic Acid (BCA) Protein Assay Thermo Fisher Scientific Inc., Rockford, IL, USA] and iron content. Samples used for the iron assay were digested in ultrapure nitric acid (1:10 ratio) for 48 h in a sand bath (60°C). Aliquots (20 μl) of digested homogenate were further diluted with 2% nitric acid for analysis. Iron concentrations were measured using graphite furnace atomic absorption spectrometry (Varian AA240, Varian, Inc., USA). Bovine liver (NBS Standard Reference Material, USDC, Washington, DC, USA) containing 184 μg Fe/g was digested in ultrapure nitric acid and used as an internal standard for analysis. All samples and controls were run in triplicates. Iron contents were expressed as micrograms of iron per milligram of protein.

Brain and liver samples from each treatment were homogenized in RIPA buffer with protease inhibitor in a 1:10 weight:volume ratio. Samples were spun at 15,000 and 14,000 rpm, respectively. Aliquots of the homogenates were used to determine protein concentration using the BCA Protein Assay. Equal amounts of lysate proteins were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), transferred onto nitrocellulose membranes, and immunoblotted with primary antibodies. Ferritin heavy chain (FtH) and light chain (FtL) primary antibodies (abcam) were diluted in 1:500 to determine the brain or liver protein expression. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (Novus Biological) and β-Actin (abcam) were all diluted in 1:1000. Mouse monoclonal or polyclonal secondary antibodies (Bio-Rad laboratories) were used in 1:3000 dilutions. The protein bands were visualized using SuperSignal West Pico chemiluminescent substrate Western blotting detection reagents (thermo scientific) and X-ray film (GeneMate).

Real-time PCR gene expression assay was conducted for FtH gene (Mm00850707_g1). Twenty nanograms of FtH cDNA were used in the TaqMan^® Universal Master Mix II (Cat#4440040, Life Technologies, Carlsbad, CA, USA), and the assay was performed according to the manufacturer’s protocol. Control and target assays were validated on excess sample tissue (n = 5). The 18s gene assay (Hs99999901_s1) was selected as the appropriate endogenous control. Relative gene expression was quantified using the delta delta C_t method.

Behavioral analysis was conducted using the Clever Home Cage Scan (HCS) system (Clever Systems Inc., Reston, VA, USA) (42, 43). The HCS system utilized video images from the home cage acquired at 30 frames/s. Based on the sequential postures of the mice and position of body parts in space, behaviors were assigned using pre-trained data sets as a reference. The computer software categorized these behaviors into different categories (44). The agreement between behaviors identified by the HCS and mice actual behavior was ≥90% (42). During 24 weeks of the dietary treatments, each mouse was placed in an individual shoebox cage with food, water, and minimal bedding. Before recording, mice were acclimated for 24 h in the recording room to ensure that any behavior alterations captured were treatment effects. After acclimation, mouse activities, such as total distance traveled (TDT), inactivity, sleeping, grooming, sniffing, foraging, feeding, and twitching, were recorded by video surveillance for 24 h. These data were exported to Microsoft Excel 2007, Prism 5, and SPSS for graphs and statistical analysis.

Multivariate analysis (two-way ANOVA) was used to analyze the effect of dietary fat and iron, respectively, as well as the interaction effect of dietary iron and fat, on all data outcomes (P < 0.05). When there is no interaction effect of dietary iron and fat, one-way ANOVA was used to test the significant effect of high-fat diet at all iron levels of all data outcomes (P < 0.05). The Post Hoc analysis was used to test the effect of dietary iron at both fat levels on all data outcomes (P < 0.05). When there is an interaction effect of dietary iron and fat, an independent t-test was used to compare the differences between control fat group and high-fat group at the same iron level of all data outcome (P < 0.05). All statistical analyses were done in SPSS.

The average dietary intake per mice per week was converted to energy intake shown in Figure 1. Dietary fat and iron, respectively, had a significant effect on the energy intake (P < 0.05). Mice fed with high fat diet had significantly higher energy intake than mice fed with the control fat diet at all iron levels (HF vs. CF, P < 0.05).

The effects of dietary fat and iron on body weight were shown in Figure 2. Dietary fat and iron, respectively, had a significant effect on body weight (P < 0.05). There was an interaction between dietary fat and iron on body weight (P < 0.05). Mice fed with high fat diets, at both control and high iron levels, had a significant higher body weight compared with their control fat fed pairs (HF/CI vs. CF/CI, HF/HI vs. CF/HI, P < 0.05). Mice fed with low iron diet, at both control and high fat levels, had a decreased body weight compared with control iron pairs (CF/CI vs. CF/LI, HF/CI vs. HF/LI, P < 0.05).

Liver TG concentration was measured to evaluate the effect of high-fat diet (Figure 3). High-fat diet significantly increased the liver TG concentration at the control iron diet level (CF/CI vs HF/CI, P < 0.05). There was an interaction between dietary iron and fat on TG content (P < 0.05). Low iron diets significantly decreased TG concentrations at both control and high fat diet levels (CF/CI vs. CF/LI, HF/CI vs. HF/LI, P < 0.05).

The impacts of dietary iron were measured through hemoglobin, hematocrit, dietary iron intake, and liver FtL expression (Figure 4). Dietary iron had a significant effect on hematocrit and hemoglobin contents (Figures 4A,B). At both control and high-fat diet levels, mice fed with low iron diets exhibited a significantly lower hematocrit and hemoglobin contents compared with control iron groups (CF/CI vs. CF/LI, HF/CI, vs. HF/LI, P < 0.05, Figures 4A,B). The average iron intake per mouse per week wash shown in Figure 4C. At both fat levels, the high iron intake was approximately 10 times higher than that in mice fed with the control iron diet, and the low iron intake was about 10 times lower than that in mice fed with the control iron diet (P < 0.05). The impact of dietary iron was further confirmed by liver FtL expression in Figure 4D, which showed that low iron diets significantly decreased the FtL expression at both control and high-fat levels.

The effects of dietary fat and iron on brain regional iron content were shown in Figure 5. Both dietary fat and iron had significant effects on brain iron contents, but in a regionally specific manner. For example, the striatal iron content was decreased by the high-fat and low iron diets, respectively, and there was no interaction between dietary iron and fat in this region (HF/CI vs. CF/CI, P < 0.05, Figure 5). An interaction of dietary fat and iron was found in the hippocampus, midbrain, and thalamus (P < 0.05). The HF/HI diet significantly increased brain iron content in the hippocampus compared with control fat fed mice (CF/HI vs. HF/HI, P < 0.05). The CF/LI diet significantly decreased the brain iron contents in the midbrain, striatum, and thalamus compared with its control (CF/LI vs. CF/CI, P < 0.05).

Ferritin heavy chain is an iron-storage protein functioning as an indicator of the cellular iron status. The protein and mRNA expressions of FtH in the hippocampus, midbrain, striatum, and thalamus were studied (Figures 6 and 7). In Figure 6, low iron diet decreased the FtH protein expression at either control or high-fat level in all regions studied. In Figure 7, no dietary interaction between fat and iron was found on the FtH mRNA expression in all regions studied. Dietary fat did not alter FtHmRNA expression level. Only the dietary iron status significantly decreased the FtH mRNA expression in thalamus.

The overall mouse behavior profile was shown in Figure 8. The stereotypical behaviors, such as inactivity, grooming, locomotion, and feeding, were included in the profile. Inactivity took the largest portion of the profile in all treatment groups. There was an interaction between fat and iron on inactivity (Figure 9, P < 0.05). The high-fat diet significantly increased the percentage of inactivity at the control iron level (CF/CI vs. HF/CI, independent t-test, P < 0.05). Both high and low iron diet increased the percentage of inactivity at control fat level (CF/CI vs. CF/HI, CF/CI vs. CF/LI, P < 0.05). Total distance traveled was analyzed in accordance with the results of inactivity. There was also an interaction between iron and fat on the total distance traveled (Figure 10, P < 0.05). High-fat diet decreased the total distance traveled at high iron level (CF/HI vs. HF/HI, P < 0.05). Low iron diet decreased the total distance traveled at the control fat level (CF/CI vs. CF/LI, P < 0.05).

The correlations between brain iron contents and the stereotypical behaviors were examined in all regions studied. A positive correlation between midbrain iron content and the sleeping time fraction was found at the high-fat diet level (r = −0.6, P < 0.05, Figure 11). No other significant correlations were found between brain regional iron contents and stereotypical behaviors.