Fe is an essential micronutrient, required for adequate erythropoietic function, oxidative metabolism, enzymatic activities and cellular immune responses(Reference Muñoz, Villar and García-Erce1). Increased Fe requirements, limited dietary supply and blood loss may lead to Fe-deficiency anaemia(Reference Kaushansky2). In addition, the same properties that make Fe essential also make this element potentially harmful; in excess, Fe is highly toxic due to its ability to form highly dangerous reactive oxygen species, and it has been implicated in several pathological conditions(Reference Halliwell, Gutteridge and Packer3, Reference Aust and Eveleigh4). The physiological functions and the integrity of the cellular structures can be altered due to the reactivity of these free radicals. The organism has an enzymatic machinery that works to eliminate and neutralise these harmful species. In several pathological conditions, the balance between pro-oxidants and antioxidants can be altered depending on the oxidative stress as well as on the system of antioxidant defence(Reference Vives Corrons, Miguel-Garcia and Pujades5–Reference Díaz-Castro, Alférez and López-Aliaga8).
On the other hand, antioxidant defence depends not only on dietary Fe concentrations, but also on the type, amount and quality of fat in the diet(Reference Brandsch, Ringseis and Eder9). Our research group has previously reported that goat milk fat has a higher nutritional quality than cow milk fat(Reference Alférez, Barrionuevo and López-Aliaga10) and improves Fe metabolism(Reference Nestares, Díaz-Castro and Alférez11). Nevertheless, in the scientific literature, to date, no studies are available about the influence of goat milk consumption on antioxidant defence mechanisms and lipid peroxidation.
Therefore, the present study was carried out to assess the effect of cow or goat milk-based diets, either with normal Fe content or Fe-overload, on the antioxidant defence and lipid peroxidation of control and anaemic rats, during the course of chronic Fe repletion in animals with experimentally induced Fe-deficiency anaemia.
Materials and methods
Animals
All animal care procedures and experimental protocols were approved by the Ethics Committee of the University of Granada in accordance with the European Community guidelines. A total of 184 male Wistar albino breed rats (21 d of age and weighing about 40–47 g), purchased from the University of Granada Laboratory Animal Service (Granada, Spain), were used in the present study.
Experimental design
At the beginning of the study, rats were divided into two groups: a control group receiving a normal-Fe diet (44·7 mg/kg by analysis)(Reference Reeves, Nielsen and Fahey12) and an anaemic group receiving a low-Fe diet (6·3 mg/kg by analysis). Anaemia was induced experimentally during 40 d by a technique developed previously by us(Reference Pallarés, Lisbona and López-Aliaga13).
During the course of the study, animals were kept at an automatically controlled temperature (22–23°C), humidity (55–65 %) and a 12 h light–12 h dark cycle (09.00–21.00 hours). Diet intake was controlled, pair-feeding all animals (80 % of the average intake) and double-distilled water was available ad libitum. On day 40 of the study, twelve rats per group (n 12 for the control group and n 12 for the anaemic group) were anaesthetised by intraperitoneal injection with sodium pentobarbital (Sigma Diagnostics, St Louis, MO, USA), totally bled out and blood aliquots with EDTA were analysed to measure haematological parameters; other blood samples were centrifuged (1500 g, 4°C, 15 min) to measure thiobarbituric acid-reactive substances (TBARS) and superoxide dismutase (SOD) activity. The remaining blood sample was centrifuged without anticoagulant to separate erythrocytes from the serum and for subsequent analysis of Fe, ferritin and total iron-binding capacity (TIBC). Liver and brain were removed and washed with ice-cold saline solution (0·9 %, w/v, NaCl). Liver, brain and erythrocyte cytosolic fractions were freshly prepared the same day by successive differential centrifugations with hypotonic haemolysis according to the method of Hanahan & Ekholm(Reference Hanahan and Ekholm14), preserving these cytosolic fractions at − 80°C for further analyses of antioxidant enzymes, catalase (CAT) and glutathione peroxidase (GPx). Protein contents in the cytosolic fractions were measured following the method described by Lowry et al. (Reference Lowry, Rosenburgh and Farr15).
After the induction of anaemia (day 40 of the study), the remaining animals (n 160) were then placed on an experimental period in which the control and anaemic groups were further fed for 30 or 50 d with cow or goat milk-based diet, either with normal Fe content (45 mg/kg) or Fe-overload (450 mg/kg), to induce chronic Fe overload(Reference Raja, Simpson and Peters16). The milk-based diets were prepared with skimmed cow (Holstein breed) or goat milk (Murciano-granadina breed) powder (20 % protein) and 10 % fat, obtained by purification from cow or goat milk fat. Fe contents in the diets by analysis were as follows: normal-Fe diet – 41·6 mg/kg (cow milk-based diet) and 42·3 mg/kg (goat milk-based diet); Fe-overloaded diets – 471·0 mg/kg (cow milk-based diet) and 473·2 mg/kg (goat milk-based diet).
On days 70 and 90 of the study, animals were totally bled out by cannulation of the abdominal aorta (n 80 on day 70 and n 80 on day 90) and plasma was obtained for TBARS determination. The remaining blood sample was collected to obtain serum and separate erythrocytes; later, liver and brain were also removed. Liver, brain and erythrocyte cytosolic fractions were freshly prepared and SOD, CAT and GPx activities were subsequently measured as described previously (day 40 of the study).
Iron determination in the diets
After mineralisation by a wet method in a sand bath (J.R. Selecta, Barcelona, Spain), Fe concentrations in the experimental diets were determined by atomic absorption spectrophotometry (PerkinElmer Analyst 1100B spectrometer with WinLab32 for AA software; PerkinElmer, Ueberlingen, Germany). To calibrate the measurements, samples of lyophilised bovine liver (certified reference material BCR 185; Community Bureau of References, Brussels, Belgium) were used to determine Fe recovery (Fe value = 210 (sem 5·8) mg/kg, means with their standard errors of the mean of five determinations, certified value = 214 (sem 5·0) mg/kg).
Haematological test
All the haematological parameters studied were measured using an automated haematology analyser (Sysmex K-1000D; Sysmex, Tokyo, Japan).
Serum ferritin
Serum ferritin concentration was determined using the Rat Ferritin ELISA Kit (Biovendor Gmbh, Heidelberg, Germany). The absorbance of the reaction was read at 450 nm using a microplate reader (Bio-Rad Laboratories, Inc., Hercules, CA, USA). Colour intensity developed was inversely proportional to the concentration of serum ferritin.
Serum iron, total iron-binding capacity and transferrin saturation
To calculate the rate of transferrin saturation, serum Fe concentration and TIBC were determined using Sigma Diagnostics Iron and TIBC reagents (Sigma Diagnostics). The absorbance of samples was read at 550 nm on a microplate reader (Bio-Rad Laboratories Inc.). The percentage of transferrin saturation was calculated from the following equation:
Antioxidant enzyme activity in the cytosolic fractions of liver, brain and erythrocytes
GPx activity was measured by the method of Flohé & Günzler(Reference Flohé and Günzler17), which is based on the formation of oxidised glutathione catalysed by GPx. Oxidised glutathione is reduced by an excess of glutathione reductase and NADPH. The subsequent oxidation of NADPH to NADP+ was monitored spectrophotometrically (Thermo Spectronic, Rochester, NY, USA) at 340 nm. Cumen hydroperoxide was used as the substrate for the reaction. SOD activity was determined according to the method of Crapo et al. (Reference Crapo, McCord and Fridovich18), based on its inhibition in the reduction of cytochrome c, measured in a spectrophotometer (Thermo Spectronic) at 550 nm. In this sense, one unit of SOD activity is defined as the amount of enzyme required to produce 50 % inhibition of the rate of reduction of cytochrome c. CAT activity was determined following the method described by Aebi(Reference Aebi19), monitoring at 240 nm in a spectrophotometer (Thermo Spectronic) the H2O2 decomposition, as a consequence of the catalytic activity of CAT. The activity was calculated from the first-order rate constant K (1/s).
Thiobarbituric acid-reactive substances measurement
The extent of lipid peroxidation was evaluated on plasma by measuring the concentration of TBARS according to the methods of Yagi(Reference Yagi20) and Ohkawa et al. (Reference Ohkawa, Ohishi and Yagi21). The reaction product was measured by spectrophotometric analysis (Thermo Spectronic) at 532 nm. The assay was calibrated using tetraethoxypropanone (Sigma-Aldrich, Taufkirchen, Germany) as a malondialdehyde source. The results are expressed as nmol TBARS/mg protein.
Statistical analysis
Differences between the control and anaemic groups were tested for statistical significance with Student's t test. Variance analysis by one-way ANOVA methods was used to compare the different diets supplied to the animals. Individual means were tested by pairwise comparison with Tukey's multiple comparison test when the main effects and their interactions were significant. The level of significance was set at P < 0·05. All data are reported as mean values with their standard errors.
Results and discussion
After Fe deprivation (5 mg/kg of diet) during 40 d, all the haematological parameters in the anaemic group were different from their counterpart control group, due to progressive Fe depletion from body stores. The haematological parameters in control rats were as follows: serum Fe 1370 (sem 120) μg/l; Hb concentration 125·5 (sem 2·4) g/l; erythrocytes 7·0 (sem 0·16) × 1012/l; haematocrit 39·2 (sem 0·77) %; mean corpuscular volume (MCV) 55·5 (sem 0·2) fl; serum ferritin 82·6 (sem 2·6) μg/l; transferrin saturation 47·1 (sem 7·0) %; platelets 735 (sem 25·5) × 109/l; TIBC 2825 (sem 199) μg/l. In the anaemic group, the haematological parameters were as follows: serum Fe 710 (sem 57) μg/l; Hb concentration 74·9 (sem 2·2) g/l; erythrocytes 6·19 (sem 0·19) × 1012/l; haematocrit 27·1 (sem 0·44) %; MCV 39·2 (sem 0·6) fl; serum ferritin 50·3 (sem 1·3) μg/l; transferrin saturation 3·6 (sem 0·3) %; platelets 1354 (sem 65·6) × 109/l; TIBC 17 787 (sem 735) μg/l. All these parameters were statistically different between control and anaemic rats (P < 0·001). Taking into account the long-term Fe restriction in the diet, all these findings were expected and feature the severe Fe deficiency induced.
Antioxidant enzymatic defence during iron deficiency
Under our experimental conditions, Fe deficiency had no influence on enzymatic mechanisms of antioxidant defence and lipid peroxidation at day 40 of the study. Several authors have stated that GPx activity and TBARS in the case of anaemia are similar to that of normal cells(Reference Díaz-Castro, Alférez and López-Aliaga8, Reference Acharya, Punchard and Taylor22, Reference Isler, Delibas and Guclu23), which coincides with the results obtained in the present study.
The drastic decrease in serum Fe found in anaemic rats at the end of the anaemia induction period (P < 0·001) involves a minor Fe trafficking in the organism, diminishing the production of free radicals via Fenton and Haber–Weiss chemistry(Reference Tanaka and Kiyosawa24), which can explain why SOD activity in plasma is within normal levels in Fe-deficient animals (2·62 (sem 0·23) U/mg protein in control and 2·36 (sem 0·25) U/mg protein in anaemic animals). CAT is a ferric enzyme containing the haem group; therefore, Fe deficiency could impair the activity of Fe-dependent enzymes(Reference Hess, Zimmermann and Arnold25). Nevertheless, under our experimental conditions, CAT activity (Fig. 1(a)) in anaemic rats was not modified with respect to their counterpart control group, probably as a result of a compensatory mechanism, in which, in spite of there being less Fe available to act as a cofactor in the enzymatic molecule, the generation of free radicals is lower due to dietary Fe restriction; therefore, the resultant net antioxidant activity did not have any modification. In addition, SOD, CAT (Fig. 1(a)) and GPx (Fig. 1(b)) are involved in the reduction of harmful peroxides that can damage PUFA, preventing lipid peroxidation; therefore, TBARS levels in plasma did not differ in both experimental groups (3·12 (sem 0·53) nmol/mg protein in the control group and 3·20 (sem 0·45) nmol/mg protein in the anaemic group).
Fig. 1 (a) Catalase (CAT) and (b) glutathione peroxidase (GPx) activity in the liver, brain and erythrocyte cytosolic fractions of control (
) and anaemic (□) rats fed a normal-iron or a low-iron diet during anaemia induction. Values are means, with their standard errors represented by vertical bars. The data were analysed by Student's t test with repeated measures and no differences between the groups were found.
Antioxidant enzymatic defence during chronic iron repletion with milk-based diets
As a result of the regular cell metabolism, the anion superoxide (O2∙ −) is converted into H2O2 by the ubiquitous enzyme SOD. Later, H2O2 produced in this reaction is converted into harmless compounds by the activities of CAT and GPx; therefore, SOD is the first line of enzymatic defence against free radicals(Reference Ghone, Kumbar and Suryakar26). The increased activity of SOD in the groups fed the cow milk-based diet reveals that these animals feature a higher generation of radicals O2∙ − than those fed the goat milk-based diet, either with normal Fe content or Fe overload (P < 0·001), in all the organs studied (Table 1), which indicates an increase in the production of H2O2, directly correlated with the rate of neutralisation of radicals O2∙ − (Reference Sahin and Gümüşlü27). Alférez et al. (Reference Alférez, López-Aliaga and Barrionuevo28) and Díaz-Castro et al. (Reference Díaz-Castro, Alférez and López-Aliaga29) demonstrated that goat milk consumption improves Zn bioavailability, a mineral with antioxidant capacity(Reference Zago and Oteiza30). In addition, this type of milk has a better lipid quality than cows' milk(Reference Alférez, Barrionuevo and López-Aliaga10); therefore, the generation of free radicals is lower when goat milk is supplied, in comparison with cows' milk, which can be linked to the higher rate of lipid peroxidation achieved in animals fed cows' milk, as will be discussed later. It is also noteworthy that anaemia does not have an influence in the activity of SOD during Fe repletion because, as it has been previously mentioned, Fe deficiency induces modifications in Fe homeostasis, decreasing mitochondrial aconitase activity and preventing the mitochondrial release of oxidants(Reference Chen, Schalinske and Eisenstein31).
Table 1 Superoxide dismutase (U/mg protein) in the cytosolic fractions of liver, brain and erythrocytes from control and anaemic rats fed the cow or goat milk-based diets, with normal iron content or iron overload, on days 70 and 90 of the study
(Mean values with their standard errors)
a,b Mean values within a row with unlike superscript letters were significantly different among the groups of control rats (P < 0·05; Tukey's test).
A,B Mean values within a row with unlike superscript letters were significantly different among the groups of anaemic rats (P < 0·05; Tukey's test).
* Mean values were significantly different for the corresponding group of rats fed with normal Fe content (P < 0·05; Student's t test).
† Mean values were significantly different for the corresponding group of control rats (P < 0·05; Student's t test).
The slightly modified activity of CAT (Table 2) and GPx (Table 3) in animals fed the cow milk based-diet indicates that the remarkable increase observed in SOD, which is shown to be connected to an increment of the formation of hydroperoxides and CAT and GPx, is insufficient to neutralise and scavenge this high output of H2O2(Reference Ghone, Kumbar and Suryakar26). Fe overload increases the production of reactive oxygen species, which is linked to the higher levels of SOD in order to remove the excessive production of O2∙ −. The lower levels of GPx found in the liver and erythrocytes in a situation of Fe overload could be due to the reduction of the enzyme in the process of neutralisation of free radicals generated in animals fed the cow milk-based diet.
Table 2 Catalase (U/ml cytosol) in the cytosolic fractions of liver, brain and erythrocytes from control and anaemic rats fed the cow or goat milk-based diets, with normal iron content or iron overload, on days 70 and 90 of the study
(Mean values with their standard errors)
a,b Mean values within a row with unlike superscript letters were significantly different among the groups of control rats (P < 0·05; Tukey's test).
A,B Mean values within a row with unlike superscript letters were significantly different among the groups of anaemic rats (P < 0·05; Tukey's test).
* Mean values were significantly different for the corresponding group of rats fed with normal Fe content (P < 0·05; Student's t test).
Table 3 Glutathione peroxidase (mmol/mg protein per ml) in the cytosolic fractions of liver, brain and erythrocytes from control and anaemic rats fed the cow or goat milk-based diets, with normal iron content or iron overload, on days 70 and 90 of the study
(Mean values with their standard errors)
a,b Mean values within a row with unlike superscript letters were significantly different among the groups of control rats (P < 0·05; Tukey's test).
A,B Mean values within a row with unlike superscript letters were significantly different among the groups of anaemic rats (P < 0·05; Tukey's test).
* Mean values were significantly different for the corresponding group of rats fed with normal Fe content (P < 0·05; Student's t test).
The liver is an organ particularly vulnerable to oxidative stress, which is reflected in the increase in SOD activity in hepatic cytosol (Table 1), since the consumption of diets with high Fe content leads to chronic overload(Reference Feder, Penny and Irrinki32–Reference Esposito, Breuer and Sirankapracha35), which produces a high expression of hepcidin. This overexpression of hepcidin controls the Fe levels, directly acting reciprocally with ferroportin 1, driving to the internalisation and degradation of ferroportin 1 when Fe levels are high, consequently blocking the liberation of Fe in the hepatocytes and making high concentrations of Fe accumulate in this organ(Reference Weinstein, Roy and Fleming36, Reference Ben-Assa, Youngster and Kozer37). This hepatic Fe overload promotes the generation of free radicals; therefore, the up-regulation of SOD would be used to neutralise this high rate of generation of reactive oxygen species.
The organism does not possess any totally effective way of excreting Fe; therefore, the regulation of the absorption of this mineral in the duodenum plays a critical role in its homeostasis(Reference Siah, Ombiga and Adams38). The Fe released in the circulation is bound to transferrin and is transported to the use (mainly erythroid system) and storage (liver) organs. Under our experimental conditions, the erythrocyte is a cell type with some unique features that seems to be affected by Fe overload. An increase in SOD activity was observed in the control group fed the cow milk-based diet (P < 0·001) in all organs (Table 1), and a decrease in GPx activity (P < 0·001; Table 3) was recorded in control and anaemic animals. This finding indicates that the increase observed in SOD activity induces a concomitant augmentation in the formation of hydroperoxides and CAT is unable to scavenge these reactive oxygen species, needing an increase in GPx activity to remove the excess of these intermediary products highly harmful to the cells(Reference Ghone, Kumbar and Suryakar26). The lower levels of GPx in all the groups fed the Fe-overloaded diets found in the present study can be due to the reduction of the enzyme in the process of neutralisation of free radicals generated, as mentioned previously. The beneficial nutritional characteristics of goats' milk can explain the lower effect in the erythrocytes induced by the Fe overload in animals fed this type of diet.
Fe is a necessary cofactor in many metabolic processes in the central nervous system, including oxidative phosphorylation, myelin synthesis, neurotransmitter production, NO metabolism and oxygen transport. It plays an important role in electron transfer and is a cofactor for a large number of enzymes, including a number of key enzymes of neurotransmitter biosynthesis in the brain(Reference Crichton, Danielsson, Geisser, Crichton, Danielsson and Geisser39). The brain is unique among all the organs of the body, hidden behind a relatively poorly permeable vascular barrier, which limits the access to plasma nutrients, such as Fe(Reference Crichton, Dexter and Ward40). The mechanism of Fe transport into the brain is still uncertain. Endothelial cells make up the blood–brain barrier and express transferrin receptor 1 on the luminal side of the capillaries. These receptors do not modify their expression of transferrin receptor 1 in Fe deficiency or overload(Reference Moos, Rosengren Nielsen and Skjørringe41); therefore, we can conclude that the nervous system is relatively independent of Fe variations in the organism, at least during the period of the present study, because, as we have examined, it is slightly affected by both Fe overload and anaemia in view of the practically unaltered activities of antioxidant enzymes in both periods of the study.
On days 70 and 90 of the study, an increase in TBARS levels was found in animals fed the Fe-overloaded diets, compared with those fed the normal-Fe diets (Table 4). The generation of superoxide ions, one of the more harmful free radicals, is eight times higher in a situation of Fe overload than in the absence of this condition(Reference Scott, VanDer Berg and Repka42). In addition, Linpisarn et al. (Reference Linpisarn, Satoh and Maikami43) found very high levels of TBARS in the plasma and tissues of rats after a single dose of intravenous Fe. Later, other authors(Reference Vives Corrons, Miguel-Garcia and Pujades5) have reported an increase in TBARS production after the stimulus induced by H2O2. Taking into account the effect of Fe as a catalyst of Fenton and Haber–Weiss chemistry and its influence on lipid peroxidation, this result was expected. Nevertheless, this increase is more pronounced in animals receiving the cow milk-based diet that in those fed the goat milk-based diet (P < 0·001 and P < 0·05, respectively). This lower trend to prevent lipid peroxidation in animals fed the goat milk-based diet can be due to the beneficial nutritional characteristics of this type of milk that increases the bioavailability of antioxidant minerals, such as Zn(Reference Alférez, López-Aliaga and Barrionuevo28, Reference Díaz-Castro, Alférez and López-Aliaga29). Several metals may interact with Zn in biological systems, and this interaction could have beneficial effects, preventing undesirable Fe-mediated damage, so Zn can replace Fe and reduce oxidation in the plasma membrane. Moreover, Zn could avoid lipid peroxidation in erythrocytes(Reference Girotti, Thomas and Jordan44) because this mineral competes with Fe for binding sites on the cell membrane(Reference Zago and Oteiza30).
Table 4 Thiobarbituric acid-reactive substances (nmol/mg protein) in the plasma of control and anaemic rats fed the cow or goat milk-based diets, with normal iron content or iron overload, on days 70 and 90 of the study
(Mean values with their standard errors)
a,b Mean values within a row with unlike superscript letters were significantly different among the groups of control rats (P < 0·05; Tukey's test).
A,B Mean values within a row with unlike superscript letters were significantly different among the groups of anaemic rats (P < 0·05; Tukey's test).
* Mean values were significantly different for the corresponding group of rats fed with normal Fe content (P < 0·05; Student's t test).
On the other hand, the better nutritive utilisation of goat milk fat(Reference Alférez, Barrionuevo and López-Aliaga10) provides a lower substrate for lipid peroxidation and consequently decreases the generation of free radicals in animals consuming this type of milk, explaining once more the lower TBARS levels found in the groups consuming the goat milk-based diets.
TBARS levels increased at day 90 compared with day 70 (P < 0·001), which could be due to the fact that all the processes of lipid peroxidation are intimately related to the age. The generation of free radicals and subsequently the related processes of lipid peroxidation are more significant as the life of the animal increases(Reference Jacob, Janner and Jahn45), which explains why TBARS are higher at the end of the experimental period (day 90; Table 4), since at this stage, the animal has evolved from a juvenile condition, where growth is very active (the phase of rapid physical and mental development), to the adult stage of maturity, where development has stopped(Reference Sisk and Zehr46, Reference McCutcheon and Marinelli47). In addition, as was expected, the lipid peroxidation processes are higher in a situation of chronic Fe overload, due to the increased generation of free radicals and the consequent damage to fatty acids(Reference Aust and Eveleigh4).
In summary, habitual goat milk consumption, a natural food with excellent nutritional characteristics, has positive effects on enzymatic antioxidant defence, even in a situation of Fe overload, which limits the processes of lipid peroxidation in comparison with cow milk consumption. Therefore, it would be recommendable to include this type of milk in the habitual diet, in the general population and in those affected by nutritional Fe-deficiency anaemia, especially in those consuming oral supplements of Fe, given the high prevalence of this pathology worldwide and because Fe overload is a common consequence of its treatment.
Acknowledgements
This study was funded by the Inter-Ministerial Commission of Science and Technology Research Project no. AGL-2006-02301/ALI. The authors are grateful to Lactalis Iberia, S.A. (Lugo, Spain) for providing the dairy products used in the present study, the University of Granada for the personal support of J. D.-C. and Elisa Alcover for her efficient administrative support. M. S. C. designed and conducted the study, analysed the data and wrote the manuscript; J. D.-C., I. L.-A., M. J. M. A., T. N. and L. J. P.-S. performed the experiments, data collection, statistical analysis and wrote the manuscript. M. R. L.-F. and M. L. O. collaborated in the enzyme analysis. All authors declare that they have no conflicts of interest.
