Abstract
Keywords
INTRODUCTION
Metal dyshomeostasis, especially endogenous metal ions such as copper (Cu), iron (Fe), manganese (Mn), and zinc (Zn) or the exogenous contaminant aluminum (Al), has attracted recent interest for the etiology of a variety of neurodegenerative conditions [1]. Maintaining transition metal homeostasis is known to be important in a wide variety of biological functions, such as antioxidant defense mechanisms. However, it is still unclear how metals such as Al, Cu, Fe, and Mn, contribute to oxidative stress and protein aggregation leading to neurodegenerative diseases [2]. A number of possible mechanisms have been proposed, including the promotion of α-synuclein aggregation and fibril formation, the activation of microglial cells leading to inflammation, and impaired production of metalloproteins [2]. It has also been suggested that Al may be involved in the formation of neurofibrillary tangles in the brain [3]. An unusual aspect of the biochemistry of this non-redox active metal is its pro-oxidant activity, which might be explained by formation of the Al superoxide semireduced radical ion AlO2. 2 + [3]. In addition, it has been suggested that Al stabilizes Fe as Fe2 +, a very active component of the Felton equation [1].
Yang et al. [4] found that serum Fe, Zn, Mn, and Cu concentrations were significantly reduced after administering Al to rats, possibly because the Al load caused an increase in the transport of Fe, Mn, Cu, and Zn into hepatocytes. It has been reported that Al would be neurotoxic through the induction of metal imbalance [5] and that Al exposure causes oxidative stress and inflammatory events [6]. However, to the best of our knowledge, the link between metal imbalance and oxidative and inflammatory events has never been tested.
Silicon (Si) has been recognized to exert health benefits with regard to skeletal and neurological functions and status [7]. Very recently our group found that organic Si protected neuroblastoma cells from H2O2 aggression [8, 9] and improved liver antioxidant status of aged rats fed cholesterol-enriched diets [10]. It has been proposed that Si may reduce Al bioavailability by blocking gastrointestinal uptake [11] and impeding renal reabsorption [12]. Bioavailable Si is mainly found in fiber-rich foods, whole grains, and beer. Bioavailable Si is mainly found in the form of silicic acid or orthosilicic acid [13], while Si in beer is present chiefly in a monomeric form [14]. A previous paper [15] demonstrated that beer intake affected the kinetics of Al uptake and excretion,possibly due to an interaction between Al and Si in the digestive tract. Therefore, Si in the form of silicic acid may lower Al bioavailability and hence protect against Al toxicity by preserving the brain antioxidant status [6]. Given all these premises, this paper hypothesizes that Al administration negatively affects levels of various metals in the brain, some of them involved in antioxidant/pro-oxidant reactions, thus causing oxidative stress and deleterious inflammatory effects. In addition, the simultaneous incorporation of Al and Si, as silicic acid or beer, arrests those negative effects.
The goals of the present study were to evaluate in mouse brain a) the effect of Al exposure on Al and other metals (i.e., Cu, Fe, Mg, Mn, Si, and Zn); and b) the correlations between those metal levels and some oxidation and inflammation markers. In addition, the paper proposes the design of a statistical model that would allow us to elucidate additional mechanisms whereby the combined effects of Al and other metals induce or reduce brain toxicity.
MATERIAL AND METHODS
Reagents
Aluminum nitrate [Al(NO3)3 · 9H2O] (Aldrich, CAS 7784-27-2) and silicic acid, Si(OH)4 (Fluka Chemie, Buchs, Switzerland) were used.
Animals and treatments
Aluminum nitrate was dissolved in deionized water to a final concentration of 450 μg/mL, equivalent to 100 mg aluminum nitrate/kg bodyweight/day, or 13 mg Al/kg bodyweight/day. This dose was already used by Colomina et al. [16] in rats. Silicic acid was dissolved in 0.9% saline solution to 50 mg/mL and 2% ethanol. A commercial beer (Mahou, Madrid, Spain) containing 5.5% alcohol by volume and 24.56±2.45 μg/g Si and 0.40±0.12 μg/g Al was selected [15, 17].
Six-week-old male NMRI mice (weighing approximately 30 g) were obtained from the Animal Research Center, Universidad de Alcalá, homologated by the Spanish Competent Authority, Register Number ES280050001165 implementing the Spanish Royal Decree 53/2013 and the European Directive 63/2010/EU. The protocol employed in this experiment was approved by the Ethics Committee of the Universidad de Alcalá (Spain). All procedures were performed in compliance with the Directive 63/2010/EU for the protection of animals inresearch.
The animals were randomly divided into four groups (n = 12 per group) and experimentally treated for three months. A control group (Group 1), receiving only deionized water, was used as the negative control. Group 2 received only aluminum nitrate (Aluminum group or positive control); Group 3 (Al+silicic acid group) received aluminum nitrate and the silicic acid solution; and finally, group 4 (Al+beer group) was given the same dose of aluminum nitrate and beer.
Groups 2–4 were orally administrated by including aluminum nitrate solution in the feeding bottles while silicic acid solution and beer [equivalent to moderate to high consumption in humans (1L/day)] were administrated by gavage. Silicon amount administrated in both groups (Al+beer and Al+silicic acid) was around 9 μg/day. This methodology for long-term exposure of mice to Al is based on previous studies from our lab [15] and others [16].
At the end of the treatment period, animals were euthanized, after anesthesia with halothane (5% vol/vol), by heart puncture and total and terminal bleeding. Brain tissue samples were taken for subsequent analysis.
The right hemibrain was used to determine mineral content, while the left hemibrain was washed with saline solution, minced, and homogenized (10%, w/v) separately in ice-cold 1.15% KCl–0.01M sodium, potassium phosphate buffer (pH 7.4) in a Potter–Elvehjem type homogenizer. The homogenates were centrifuged at 10,000 g for 20 min at 4°C, and the resultant supernatants used for biochemical analysis.
Analytical methods
The mineral levels of total brain were measured after ashing of the organic matter according to the method proposed by Granero et al. [18]. Each sample (2.5 mL) was digested with 2 mL of 65% nitric acid (Suprapur, Merck, Darmstad, Germany) in Teflon bombs for 8 h at room temperature, and subsequently heated at 100°C for 12 h. After cooling, solutions were filtered and made up to 25 mL with deionized water.
Si content of the brain was measured by means of inductively coupled plasma atomic emission spectrometry ICP-OES (Perkin Elmer Optima model 3200 RL, Boston, MA, USA), using Si emission lines of 254.611 nm and 212.412 nm. Brain levels of Al were determined by means of inductively coupled plasma mass spectrometry (Perkin Elmer Elan model 6000, ICP-MS), using the only Al isotope, 27Al. The emission lines and isotopes used are free of spectral interferences in these matrix types. Multi-element analysis of Mn in digested samples was performed with ICP-MS spectrometry, while the other elements (Cu, Mg, Fe, Zn) were analyzed with ICP-OES spectrometry. Elements with the highest isotopic abundance, free from isobaric and polyatomic interferences, were selected as analytical mass to perform determinations by ICP-MS spectrometry. When ICP-OES spectrometry was employed, possible interferences and selection of analytical lines were checked selecting three of the most sensitive spectral lines.
Validation of the methods, based on the ICP-OES and ICP-MS techniques, was performed according to EURACHEM guidelines [19] with regard to accuracy, precision, sensitivity, and linearity, using the experimental setting that provided the optimal conditions. According to these guidelines, intra-assay and inter-assay imprecision, measured as variation coefficients, should be below 5 and 10%, respectively. Sensitivity of the determination of each chemical element was expressed by the slope of the linear regression equation. Linearity was assessed by the correlation coefficients of calibration curves and was considered acceptable when r≥0.9995. Detection limits were calculated on the basis of the 3 s criterion for ten replicate measurements of blank solutions subjected to the same treatment as the samples.
The accuracy of the instrumental methods was validated by replicating all samples and by taking measurements of reference material (lobster hepatopancreas, NRC Canada TORT 2) every 10 samples. Quantification was based on the most abundant isotope of each element free of analytical interferences. The mean recovery rates were between 90% and 95%. More details on validation criteria have been previously reported [20, 21].
Biochemical assays
Lipid peroxidation was measured by following the formation of malonyldialdehyde (MDA), according to the presence of thiobarbituric acid reactive substances (TBARS) in the brain homogenates [22] using an Uvikon 930 spectrophotometer (San Diego, CA, USA). Concentrations were calculated using a standard curve obtained with MDA. TBARS values were expressed as μmol of MDA per mg protein. Protein was determined by the Bradford method [23] using bovine serum albumin as standard.
RT-PCR real time analysis
Total RNA was extracted from frozen brain samples following the guanidinium thiocyanate/phenol reagent method [24]. Reverse transcription and amplification using the Titan system involved the preparation of master-mix 1 and 2 on ice. Mix 1 was comprised of dNTPs, primers, dithiothreitol (DTT), extracted RNA (1 μg), and sterile pre-chilled deionized water. Mix 2 consisted of RT-PCR buffer, enzyme mix (AMV reverse transcriptase, Taq DNA polymerase), and sterile pre-chilled deionized water. All reagents were thawed, vortexed briefly, and centrifuged before setting up the reactions. Twenty-five microliters each of master-mix 1 and 2 was added to a 0.2 mL PCR tubes kept on ice. This was vortexed and centrifuged briefly to collect the sample at the bottom of the tube, and RT was carried out at 50°C for 30 min. β-actin cDNA was used as an internal control.
The expression of Catalase (CAT), CuZn superoxide dismutase (CuZn-SOD), Glutathione peroxidase (GPx) enzymes and tumor necrosis factor alpha (TNFα) in brain were tested.
The sequences of the primers used were as follows:
The number of PCR cycles was adjusted to avoid saturation of the amplification system: 94°C for 30 s, 55°C for 45 s, and 72°C for 30 s (30 cycles) for CAT; 95°C for 30 s, 5°C for 1 min, and 72°C for 30 s (30 cycles) for CuZn-SOD; 94°C for 30 s, 56°C for 45 s, and 72°C for 60 s (30 cycles) for GPx; 94°C for 60 s, 59°C for 60 s, and 72°C for 60 s (35 cycles) for TNFα; 58°C for 45 s and 72°C for 30 s (24 cycles) for β-actin with a final elongation at 72°C for 10 min. Amplification products were visualized on 1.8% agarose gels containing ethidium bromide (1 μg/mL): CAT product, 395 bp; CuZn-SOD product, 383 bp; GPx product, 697 bp; TNFα product, 281 bp; β-actin product 630 bp. A 100 bp DNA ladder was used as marker. The products were quantified by laser densitometry.
Principal component analysis
Principal component analysis (PCA) was applied to understand possible associations between minerals and the pro-oxidant/antioxidant or inflammatory markers in brain. To avoid overlapping and a very complex model, only parameters significantly affected by treatments were included. Ten different variables were tested; five corresponding to mineral content (Zn, Cu, Mn, Si, and Al), and five to pro-oxidant/antioxidant or inflammatory markers (CuZn-SOD, CAT, GPx, TNFα, and TBARS). PCA create new variables PCs containing the initial variables which correlate among them.
Statistical analysis
All analyses were performed in triplicate. Data were expressed as means±SE. Statistical analyses were performed using the SPSS statistical software package (version 15.0) and the SAS (version 19.0). The Kolmogorov-Smirnov test was used for assessing normal data distribution. ANOVA one way followed by the T2 of Tamhane test were used for group statistical comparison. Pearson product moment-correlations were performed to study the relationship between minerals and TBARS, TNFα and the different expressions antioxidant enzymes. p values < 0.05 were considered statistically significant.
RESULTS
Brain tissue aluminum and trace metals levels
Table 1 presents the concentrations of Al, Cu, Fe, Mg, Mn, Si, and Zn in mouse brain from the different groups. Aluminum nitrate exposition significantly increased brain Al and Si contents but significantly decreased (p < 0.05) Cu, Mn, and Zn levels. Under aluminum nitrate exposition, beer or silicic acid, significantly lowered Al and Si levels and normalized those of Cu, Mn, and Zn in brain.
Brain TBARS levels and TNFα and antioxidant enzymes expression
Brain tissue levels of TBARS, measured as the lipid peroxidation end product MDA, were significantly higher (p < 0.001) in the Al group than in control one. Brain levels of TBARS in the Al+silicic acid and Al+beer groups were significantly lower (p < 0.001) than those of the Al one (Fig. 1).
The relative TNFα expression of the Al group significantly increased versus control (p < 0.001). TNFα expression was significantly reduced in the Al+silicic acid and Al+beer groups versus the Al group and they were not different (p < 0.05) from that of control animals (Fig. 1).
The expression of CAT and CuZn-SOD was significantly lower (p < 0.001) while that of GPx higher (p < 0.001) in the Al group versus the control counterpart. Enzyme expressions were normalized in the brains of the Al+silicic acid and Al+beer groups, as their values did not differ from those of the control group (p > 0.05) (Fig. 1).
Metal levels, TBARS levels, and gene expression relationships
Table 2 shows the Pearson product-moment correlations between mineral and TBARS levels and the TNFα and antioxidant enzymes expression. Brain Al correlated negatively and significantly with Cu (p < 0.01), Mn and Fe (both p < 0.05), and Zn (p < 0.001) but positively with Si (p < 0.001). Cu was positively correlated with Fe (p < 0.01), Mg (p < 0.001), Mn (p < 0.01), while Fe positively and significantly correlated with Mn (p < 0.001) and Zn (p < 0.01). Zn showed positive correlation (p < 0.001) with Mg. TBARS levels and TNFα expression were positively correlated with Al (both p < 0.001) and Si (p < 0.05 and p < 0.01, respectively). Cu (p < 0.01) and Zn (p < 0.05) negatively correlated with TBARS (at least p < 0.05) and TNFα (bothp < 0.001).
CAT gene expression showed significant positive correlations with Cu (p < 0.001), Fe (p < 0.05), Mg (p < 0.05), Mn (p < 0.05), and Zn (p < 0.001), while CuZn-SOD gene expression positively correlated with Cu (p < 0.01) and Zn (p < 0.001). Both CAT and CuZn-SOD expressions negatively correlated with Al (p < 0.001) and Si (p < 0.001) whereas GPx expression negatively correlated with Cu (p < 0.01), Fe (p < 0.01), Mn (p < 0.05), and Zn (p < 0.01) but positively with Al (p < 0.001) and Si(p < 0.001).
PCA results
PCA was conducted to ascertain possible relationships between mineral and TBARS contents and TNFα and antioxidant enzymes expressions. In order to avoid obtaining very complicate models, only parameters that appear significantly correlated with most of the other parameters tested and that were significantly affected by treatments were selected. Thus, ten variables were evaluated (Zn, Cu, Mn, Si, and Al; CuZn-SOD, CAT, GPx, TNFα, and TBARS). PCA created two components which explained the 71.2% of the total variance in the data set (Fig. 2). First component (PC-1) accounted for 59.48% of the variance. It strongly and positively correlated with GPx, TNFα, Al, and TBARS. Their values indicated in Table 3 were close to 1. Si was found relatively close to them with a 0.532 of PC-1. On the other hand, there were CAT and CuZn-SOD, with values close to –1, located on the left side of the plot (Fig. 2), which correlated negatively with the previous ones. The second component (PC-2) explained 11.73% of the total variance. It positively correlated with Zn and Cu, both appearing negatively correlated also with PC-1 (Fig. 2). Mn was in the middle of the two components, with a positive value of PC-2 but a negative valueof PC-1.
The PCA connected the brain pro-oxidant markers with brain Al content, and to a lesser extent with that of Si. On the contrary, Zn and Cu were closer to antioxidants enzymes. The PCA scatter plot (Fig. 3) indicates the position of the studied groups (Control, Aluminum, Al+silicic acid, Al+beer) in a graph compiled with the two PC-1 and PC-2. In addition, it shows how the Aluminum group was located on the positive part of PC-1, and with more positive than negative value with respect to 0 at the PC-2. The control group appears oppositely in the negative part of both PC-1 and PC-2. Finally, Aluminum and Al+silicic acid groups were both located in the negative part of PC-1 as the control group, but with more positive values of PC-2 than thecontrol.
DISCUSSION
Our results clearly show that aluminum nitrate administration induced changes in mineral contents and inflammatory and oxidant/antioxidant status of the brain. These negative effects were significantly reversed by the conjoint administration of aluminum nitrate with silicic acid or beer.
Aluminum can produce toxicity to the central system nervous and contribute to some neurodegenerative diseases, including Alzheimer’s disease. Non-evident toxic clinical signs (e.g., muscle weakness, bone diseases) were observed in the present study, but in fact the brains of these mice presented clear signs of neurotoxicity as already reported [25]. The lack of clinical signs appears to be chiefly a result of the aluminum nitrate dose used, which was lower than the dose (20–35 mg Al element/kg/day) employed in other studies [26, 27]. In addition, there is the low bioavailability of aluminum nitrate reported in other studies [28], and also demonstrated in a parallel study in mice [16] in which Al fecal excretion (in μg/g feces) was 120 times higher in the Aluminum group than in the control.
The role of metal ions (especially Zn, Cu, and Fe) in neurobiological processes is a topic of growing interest. They participate in many essential activities, and their deficiencies can be lethal. Loss of Fe, Cu, and Zn can lead to neurological disease, but conversely accumulation or abnormal interactions with proteins, lipids, or nucleic acids can also contribute to neurological disease [29]. Yang et al. [4] found that Fe, Zn, Mn, and Cu concentrations decreased in rat serum after aluminum nitrate administration, possibly because the Al load caused liver uptake and increased storage of those minerals. Thus, it seems possible that the Cu, Mg, and Zn decrease observed in the brain following aluminum nitrate administration could be a consequence of the lower plasma content of these metals. Other authors have suggested that Al and other metals (e.g., Fe) compete for a common gastrointestinal absorption mechanism [30, 31], which would at least partially explain the lower brain metal levels observed in the present paper.
In contrast with the other minerals, brain Si levels increased in aluminum nitrate treated animals, even in the absence of Si supplement, displaying brain levels as high as in Al+silicic acid or Al+beer groups, and suggesting some Al-Si affinity. In this regard, Noremberg et al. [32] also found such an increase of Si brain levels in rats exposed to Al. Although we previously suggested that this could be due to increased blood-brain barrier permeability [15], we are now reconsidering that explanation in the light of the present results, since Cu, Mn, and Zn significantly decreased in the brains of these mice. With the latest findings, our new hypothesis posits the formation of aluminosilicate complexes similar to those found in the senile plaques of Alzheimer’s disease patients [33]. However, this hypothesis needs to be definitively tested in future studies analyzing the aluminosilicate content in mouse brains or even in brain like-plaque structures under identical experimental conditions.
The conjoint administration of aluminum nitrate and silicic acid or beer partially blocked the metal imbalance induced by aluminum nitrate. Such a blocking effect can be at least partially explained by the impact of silicic acid/beer on aluminum nitrate absorption. In fact, fecal Al was 34% and 76% higher in the Al+beer and Al+silicic acid groups, respectively, than in the Aluminum group [16]. Domingo [34] suggests the role of silicon in preventing oral Al absorption and retention in mammals. In addition, the lower amounts of Al in the brain of Al+silicic acid and Al+beer mice also appear to be related to the intestinal blocking effect mentioned above. Brain Si concentrations in Al+silicic acid and Al+beer mice were lower than those of the Aluminum group; however, they were higher than in the control group. These results suggest that when both available Al and Si levels are high, there is less Al absorption and the levels of both metals in the brain are low; however, when availability of Al is high and Si low, there is high Al absorption and easy access to the brain, inducing time-course Si retention in the form of aluminosilicate. The positive correlations between Al and Si support the potential mechanisms already discussed. The negative correlations found between brain Al and Cu, Fe, Mg, Mn, and Zn levels corroborate the palliative effect of Si on the metal imbalance induced by Al. These results emphasize the importance of promoting increased levels of Si in people exposed to Al, or, more importantly to adjust dietary Si levels to Al intake.
Changes observed in TBARS, TNFα, and enzymes expression in the Aluminum group suggest pro-inflammatory and pro-oxidant effects due to the metal imbalance induced by aluminum nitrate administration. Cu is required for mitochondrial respiration, neurotransmitter biosynthesis and as a cofactor for antioxidant enzymes [35]. Activities of the Cu-dependent proteins CuZn-SOD [36] and cytochrome C oxidase [37] have been reported to be lower in AD patients than in controls. These results suggest that there is alteration of Cu homeostasis in AD and that such alteration can lead to redox imbalance by altering the functioning of important enzymes like CuZn-SOD and ceruloplasmin [38]. Therefore, the decrease in Cu levels caused by Al could at least partially explain the decreased expression of CuZn-SOD observed in rats treated with Al [4]. Silicon administration, as organic silicon or beer, restores brain levels of Cu and regulates expression of CuZn-SOD. Silicon, especially in supra nutritional amounts, has been reported to facilitate the absorption, retention, and/or utilization of copper [39] and magnesium [40]. These data indicate that some of the metabolic effects attributed to silicon may be manifested through a silicon-facilitated increase in copper utilization [41]. The high positive correlation observed between brain Cu levels and gene expression of CuZn-SOD, as well as the negative correlation of Al levels with this enzyme, support previous comments.
Zn nutritional deficiency is common in advanced age, and a recent report indicates that Zn deficiency in AβPP transgenic mice increased the volume of amyloid plaques [41]. The data reported by Cuajungco and Fagét [42] suggest a protective role for Zn2 + in AD, where plaques form as the result of a more robust Zn2 + antioxidant response to the underlying oxidative attack. Alterations in zinc metabolism and homeostasis have been reported in Parkinson’s and Alzheimer’s diseases as well as in transient forebrain ischemia, seizures, and traumatic brain injury, but little is known regarding aged brain [43]. Present results suggest that Al poisoning causes a significant decrease in brain Zn levels and that the effect is reversed when Si is administered as silicic acid or beer. Therefore, Si could somehow mitigate the damage induced by Al, restoring Zn homeostasis and antioxidant and inflammatory status expression (CAT, CuZn-SOD, GPx, TNFα, TBARS) in the brain.
Fe is required to support a high brain respiratory rate as well as for myelination, gene expression, and neurotransmitter synthesis [44]. According to Yang et al. [4], Al readily occupies the binding sites of Fe, reducing the interactions between Fe2 + and iron transporters. This increase in free Fe2 +, and hence the production of a large amount of active oxygen free radicals, could result in cell membrane lipid peroxidation, accelerated cell apoptosis, necrosis, and mitochondrial function damage that clearly link to the impairment of antioxidant and inflammation status observed in Al mice [1]. Contini et al. [45] suggested that the dysregulation of intestinal Fe absorption observed under Al exposure resulted in a 26% increase of lipid peroxidation (TBARS).
The fact that brain Al levels were diminished after Al+silicic acid or Al+beer administration suggests that the Al-Fe2 + interaction, and hence the active oxygen free radicals linked to Al, may be reduced, explaining the observed improvement in inflammation and antioxidant status. In fact, the concentration of brain Fe correlated significantly and positively with gene expression of catalase and negatively with that of GPx and TNFα.
Although there have not yet been reports on direct interaction between Mn and AD pathology, our results showed a significant reduction in Mn levels in the brains of mice treated with aluminum nitrate. The administration of Si as silicic acid or beer succeeded in restoring Mn levels in brain, suggesting that silicic acid and beer improved antioxidant protection against brain disorders caused by exposure to aluminum nitrate.
As noted, aluminum nitrate provoked oxidative stress and dysregulation in antioxidant enzyme expression, either directly or by inducing metal imbalance. Such alterations were blocked and reversed by the conjoint administration of Al+silicic acid or Al+beer. This outcome was readily observable in the PCA scatter plot which located Al+silicic acid and Al+beer groups close to control mice and far from the Aluminum group, suggesting a reversion of Al pro-oxidant effects.
Conclusion
Chronic aluminum exposure leads to brain metal imbalance inducing brain inflammatory and pro-oxidant status impairment. This study is the first to show that silicic acid and beer help to restore depleted levels of Cu, Fe, Mn, and Zn in the brain and to suggest that silicon, in turn, improves oxidation and inflammatory markers negatively affected by aluminum administration.
Footnotes
ACKNOWLEDGMENTS
This study was supported by the Asociación de Cerveceros Españoles, project XF20 and by the Spanish Ministerio de Economía y Competitividad, project AGL 2014-53207-C2-2-R. We also acknowledge the pre-doctoral fellowship granted to Alba Garcimartín by BES-2012-054752(FPI).
