Abstract
The altered metabolism of iron impacts the brain function in multiple deleterious ways during normal aging as well as in Alzheimer’s disease. We have shown in this study that chelatable iron accumulates in the aged rat brain along with overexpression of transferrin receptor 1 (TfR1) and ferritin, accompanied by significant alterations in amyloid-β (Aβ) peptide homeostasis in the aging brain, such as an increased production of the amyloid-β protein precursor, a decreased level of neprilysin, and increased accumulation of Aβ42. When aged rats are given daily the iron chelator, deferasirox, over a period of more than 4 months starting from the 18th month, the age-related accumulation of iron and overexpression of TfR1 and ferritin in the brain are significantly prevented. More interestingly, the chelator treatment also considerably reverses the altered Aβ peptide metabolism in the aging brain implying a significant role of iron in the latter phenomenon. Further, other results indicate that iron accumulation results in oxidative stress and the activation of NF-κB in the aged rat brain, which are also reversed by the deferasirox treatment. The analysis of the results together suggests that iron accumulation and oxidative stress interact at multiple levels that include transcriptional and post-transcriptional mechanisms to bring about changes in the expression levels of TfR1 and ferritin and also alterations in Aβ peptide metabolism in the aging rat brain. The efficacy of deferasirox in preventing age-related changes in iron and Aβ peptide metabolism in the aging brain, as shown here, has obvious therapeutic implications for Alzheimer’s disease.
INTRODUCTION
Alzheimer’s disease (AD), which accounts for the majority of dementia cases above the age of 65 years, is fast becoming a huge disease burden all over the world, and an estimated 24 million people suffer from this disease as described in a published study in 2012 [1]. The sporadic variety of the disease is by far the most common accounting for nearly 95% of the AD population [2]. Apart from the hallmark neuropathological features of neuritic plaques, amyloid deposits, and neurofibrillary tangles, the AD brain shows extensive areas of neurodegeneration and synaptic loss in entorhinal cortex, amygdala, frontal, temporal, parietal, and occipital association cortices along with hippocampus and parahippocampal gyrus [3, 4]. The etiopathogenesis of sporadic AD is uncertain, but abnormal accumulation and oligomerization of amyloid-β protein (Aβ42 in particular), mitochondrial dysfunction, oxidative stress, inflammation, metal dyshomeostasis, and various metabolic alterations are key elements of the pathologic process [5–12]. The metal dysregulation is very characteristic, and increased amount of Cu, Zn, and Fe in the brain are likely triggers for the onset of sporadic AD [13]. In particular an increased accumulation of iron has been demonstrated in AD brain primarily in the regions of amyloid and neuritic plaques and also more diffusely within neurons as well as around neurofibrillary tangles by postmortem biochemical analysis or antemortem imaging studies [8, 15]. Further, the mutant allele of human hemochromatosis gene, responsible for the inherited disorder known as hemochromatosis with enhanced tissue depositions of iron, can accelerate the onset of sporadic AD by several years [8, 16]. It has been suggested that iron can affect AD pathology in multiple ways such as induction of oxidative stress, translational upregulation of intraneuronal amyloid-β protein precursor (AβPP) production, aggregation of Aβ42, and conversion of Aβ42 to a more toxic species by forming metal chelate complex [14, 17–19]. The translational upregulation of AβPP synthesis is particularly important because of the presence of multiple regulatory sites in 5’-UTR of AβPP mRNA. The 5’-UTR of AβPP mRNA has multiple regulatory sites such as interleukin-1 responsive acute box element, iron-responsive element (IRE), transforming growth factor β responsive element, etc., which could control AβPP production under in vivo conditions making it an attractive drug target for decreasing Aβ peptide load in AD brain [20]. The iron responsive element binding protein (IREBP) occupies the RNA stem-loop structure of IRE at the 5’-UTR of AβPP mRNA preventing the initiation of translation, but when cellular iron is elevated, the complex of IREBP-Fe is dissociated from the mRNA triggering the synthesis of AβPP [21, 22]. Although this proposed involvement of iron in AD pathogenesis is generally accepted, several questions have not been critically examined. First of all, the reasons for increased iron accumulation in AD brain are not known. Secondly, it is not established with certainty that increased iron in the brain is primarily responsible for altered AβPP metabolism and consequent Aβ42 accumulation during AD pathogenesis. However, both these issues should be addressed in order to understand the pathogenic mechanisms of sporadic AD and to identify new drug targets for this disease.
Aging is the most dominant risk factor for AD, and the aged brain manifests iron accumulation and Aβ accumulation and deposition similar to AD brain albeit in a much subdued fashion [11, 23–27]. Some studies have attempted to elucidate the mechanisms of increased Aβ deposition in the aging brain of normal and transgenic AD models, but it is not known to what extent iron accumulation impacts this process [28–32]. We hypothesize that understanding the mechanism of iron dysregulation and its relationship with Aβ metabolism in the aging brain would provide important clues to similar phenomena in the AD brain. For this purpose, we have analyzed the iron content, the expression levels (both protein and mRNA) of transferrin receptor and ferritin, and different parameters of Aβ metabolism in the aging rat brain and also examined how all these could be modulated by dietary administration of an iron chelator to the aged rats. The iron chelator used in this study, deferasirox, is presently used in different clinical conditions to prevent iron overload, and this compound can be administered orally without much toxic effects [33, 34]. Although the present study has employed the aging brain to examine the iron-Aβ relationship, the results, nevertheless, can provide new clues to the understanding of AD pathogenesis in the context of metal dyshomeostasis.
MATERIALS AND METHODS
Chemicals
All common chemicals were of analytical grade and obtained from Sisco Research Laboratory (India). DL-dithiothreitol (DTT), phenylmethanesuphonyl fluoride (PMSF), ethylenediaminetetraacetic acid (EDTA), ethylene glycol-bis(2-aminoethylether)-N,N,N,N-tetraacetic acid (EGTA), protease inhibitor cock-tail, bovine serum albumin (BSA), sodium dodecylsulphate (SDS), N-(2-Hydroxyethyl) piperazine-N′-(2-ethanesulfonic acid) (HEPES), 2,4-dinitropheny-lhydrazine (DNPH), ferene, ascorbic acid, ammonium acetate, polyclonal anti-dinitrophenyl (anti-DNP) primary antibody, N-dansyl-Ala-Gly-D-nitro-Phe-Gly, and rabbit AβPP primary antibody (A8717) were purchased from Sigma Chemical Co. (USA). 2-Thiobarbituric acid was purchased from Acros Organics (Belgium). Goat anti rabbit IgG HRP conjugate was purchased from Bangalore Genei (India). We have used proprietary preparation of deferasirox from a particular company used for human medication. RNA extraction kit (PureLink RNA minikit) was obtained from Ambion (USA) while cDNA synthesis kit (iScript cDNA synthesis kit) and SYBR Green master mix were purchased from Roche Diagnostics (India). Primary antibody for NF-κB (p65 subunit) was obtained from Santa Cruz Biotechnology, Inc. Transferrin receptor 1 (TfR1) primary antibody and ECL detection kit were obtained from Pierce (Thermo Scientific, Rockford, IL, USA). Beta secretase assay kit and Aβ42 ELISA kit were purchased from Biovision (USA) and Covance (USA), respectively. Ferritin H (FtH) ELISA kit was purchased from Cloud Clone Corp. (USA). Pierce BCA protein assay kit (Thermo Scientific, Rockford, IL, USA), DNA molecular weight markers (100 bp DNA ladder plus) and protein molecular weight markers were purchased from Fermentas (USA) and Biobharati (India), respectively.
Animals and treatment
Albino rats of Wistar strain were kept on laboratory chow with water ad libitum in 12 h light/12 h dark cycles and maintained as per the guidelines of the Animal Ethical Committee of the institute. For experiments, the animals were divided into three groups: young (4–6 months, Y), aged (22–24 months, A), and aged (22–24 months) with chelator supplementation (CT). Starting from 18 months onwards, the chelator supplementation was given with the diet daily at a dose of 25 mg/Kg body weight to the CT group of animals and continued until they were used for various experiments between 22 and 24 months.
Gene expression analysis by real time polymerase chain reaction
RNA from rat brain cortex was extracted using a commercial kit and following the manufacturer’s protocol. The extracted RNA was checked for purity by the 260 nm/280 nm ratio as well as by electrophoresis on 2% agarose gels. The reverse transcription was performed (80 ng total RNA) as per the manufacturer’s protocol. The qPCR analysis of the cDNA samples was performed in triplicate using SYBR Green on a real-time PCR machine (Light cycler, Roche) with a reaction volumes of 20 μl each containing 5 pmol each of forward and reverse primers for the target genes (AβPP, TfR1, or FtH) or the reference β-actin gene. The average quantification cycle (Cq) value from the replicate reactions was calculated for each sample (target and reference). The gene expression was quantified by relative quantification method following the procedure proposed by Pfaffl [35]. The primers used were: β actin-Forward: 5 CCACACCCGCCACCAGTTCG 3, Reverse: 5 CCCATTCCCACCATCACACC 3; AβPP-Forward: 5 AGAGGTCTACCCTGAACTGC 3, Reverse: 5 ATCGCTTACAAACTCACCAAC 3; TfR1: Forward: 5 ATA CGT TCC CCG TTG TTG AGG 3, Reverse: 5 GGC GGA AAC TGA GTA TGG TTG A 3; FtH: Forward: 5 GCC CTG AAG AAC TTT GCC AAA T 3, Reverse: 5 TGC AGG AAG ATT CGT CCA CCT 3.
Spectrophotometric measurement of iron level
Chelatable iron levels were measured by a method previously described [36]. Briefly, brain cortical tissue (approx. 0.2 g) was dissected, homogenized in 50 mM phosphate-buffered saline (PBS) (10% w/v), and subsequently centrifuged at 25,000 g for 25 min. To 900 μl of diluted brain supernatant (1:3), an aliquot (90 μl) of iron-detection reagent (6.5 mM ferene, 2.5 M ammonium acetate, and 1 M ascorbic acid dissolved in water) was added and the mixture further incubated for 30 min at 37°C. The absorbance of the reaction mixture was read at 550 nm. The iron content of the sample was calculated by comparing its absorbance to that of a range of standard concentrations of iron (12.5–800 ng/100 μl).
Western blot analysis
The brain tissue was processed differently for immunoblotting of different proteins. Brain cortices were dissected out cleanly on ice and homogenized in a lysis buffer (0.5% phenyl methylsulfonyl fluoride plus protease inhibitor cocktail (5 μl/100 μg tissue) in Tris-buffered saline-Tween 20 (TBS-T). After centrifugation at 10,000 g for 30 min at 4°C, the supernatant was collected and used for the immunolotting of TfR1. For immunoblotting of AβPP, a membrane enriched brain tissue fraction was used. Briefly, the rat brain cortex (0.5 mg) was homogenized in 10× volume of homogenizing buffer (225 mM mannitol, 75 mM sucrose, 5 mM HEPES, 1 mM EGTA, 1 mg/ml BSA, pH 7.4) containing the protease inhibitor cocktail (5 μl/100 μg tissue) and centrifuged at 1 000× g for 3 min. The supernatant was centrifuged at 1, 00, 000× g for 1 h at 4°C and the pellet suspended in the lysis buffer (50 mM Tris, pH 8.0, 150 mM NaCl, 5 mM EDTA, 1% SDS, 0.25% deoxycholate, 0.25% NP-40) and kept for 5 min on ice. The extract was subsequently used for the immunoblottting of AβPP.
NF-κB (p65) was detected and quantified by western blotting in the nuclear and cytosolic extracts obtained from the rat cerebral cortex. The cerebral cortex was dissected out from the brain, homogenized in 1 ml ice-cold buffer (for 0.3 g of cortical tissue) containing 10 mM HEPES (pH 7.9), 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM DTT, 0.5 mM PMSF plus the protease inhibitor cocktail (5 μl/100 μg tissue) and centrifuged at 800 g for 2 min at 4°C. The supernatant (cytosolic extract) was collected, and the nuclear pellet resuspended in 1.5 ml buffer containing20 mM HEPES (pH 7.9), 0.4 M NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM DTT, 1 mM PMSF, and the protease inhibitor cocktail (5 μl/100 μg tissue) and kept overnight at 4°C. The nuclear lysate was spun at 25,000 g for 5 min, to obtain the supernatant (nuclear extract) [37].
The protein concentration in each sample was measured with a Pierce BCA protein assay kit. Aliquots of the extract (25 μg of protein) were separated by sodium dodecylsulphate polyacrylamide gel electrophoresis (SDS-PAGE) under reducing condition on a 10% resolving gel, and the separated proteins electroblotted onto nitrocellulose membranes by using standard blotting protocols [38]. The membranes were blocked in 5% non-fat milk in phosphate-buffered saline-Tween 20 (PBST, 25 mM PBS, pH 7.6, 137 mM NaCl, 0.1% Tween-20) for 2 h at room temperature, and then incubated with primary antibodies against TfR1 (1:2500), AβPP (1:3000), and NF-κB (1:2000), for 3 h at 37°C. After washing with PBS-T three times, the blots were incubated in secondary-antibody-conjugated horseradish peroxide (1:5000) for 90 min at room temperature, using the enhanced chemiluminescence method. The film was digitized and analyzed by Gel Quant imaging and quantitation software. The band intensities from each sample were normalized with respect to the corresponding γ-actin loading control.
Measurement of β-secretase activity
β-secretase enzyme activity was measured by a fluorometric assay using a commercial kit that employed the cleavage of a secretase specific peptide substrate. Two reporter groups EDAN (fluorophor) and DABCYL (quencher) were conjugated to the peptide substrate. In the uncleaved form, the fluorescence emission from EDANS was quenched by the DABCYL moiety. The cleavage of the peptide by the secretase separated EDANS from DABCYL allowing the emission of a fluorescence signal (λex 345 nm/λem 500 nm). The brain cortical tissue from Y, A, or CT rats was homogenized with five volumes of extraction buffer provided with the kit, kept in ice for 5 min, centrifuged for 10 min at 10,000 g at 4°C and the supernatant used for the assay. An aliquot of the supernatant (20–60 μg protein) was mixed with 100 μl reaction buffer and 2 μl of substrate in a total of 200 μl reaction volume. The reaction mixture was incubated for 1 h at 37°C in the dark and fluorescence intensity measured (λex 345 nm/λem 500 nm) against appropriate sample and dye blanks. The specific β-secretase activity was expressed as fluorescence unit per 100 μg of protein. The specificity of the assay procedure in our samples was validated by the use of an inhibitor specific forβ-secretase and a source of purified β-secretase, which were supplied with the kit.
Measurement of Aβ1 - 42 and ferritin heavy chain by sandwich Elisa
The brain cortical level of Aβ42 was analyzed by chemiluminescence based enzyme linked immunoassay using a commercial kit. Freshly isolated cortical tissue was homogenized in 5× volume of lysis buffer (50 mM Tris, pH 8.0, 150 mM NaCl, 5 mM EDTA, 1% SDS, 0.5% Triton X-100, protease inhibitor cocktail (5 μl/100 μg tissue) followed by sonication with 5 bursts (8 s each with 30 s interval) in ice-cold condition. The tissue homogenate was centrifuged for 1 h at 45,000× g at 4°C, and the supernatant used for Aβ42 measurement in duplicate following manufacturer’s protocol. The chemiluminescence signal was detected on a plate luminometer (ELX –800, Biotek, USA). The standard curve was generated by using synthetic Aβ42 peptide over the concentration range of 0–77.2 pg/ml, and the concentration of Aβ42 in tissue extract was expressed as ng Aβ42/mg of protein of cortical extract.
The brain cortical tissue level of ferritin H (FtH) was analyzed by colorimetric enzyme linked immunoassay using a commercial kit (Cloud-Clone Corp., USA). Freshly isolated cortical tissue (0.5 mg) was homogenized in 5 ml of ice cold PBS (0.01 M, pH 7.0–7.2), followed by two freeze-thaw cycles. The tissue homogenate was centrifuged for 5 min at 5,000× g at 4°C, and the supernatant used for FtH measurement in duplicate following manufacturer’s protocol.
Absorbance at 450 nm was measured on a microplate reader (Sunrise TECAN, Germany). The standard curve was generated by using the stock standard FtH peptide (800 ng/ml) supplied with the kit, and the concentration of ferritin in tissue extract was expressed as ng FtH/ ml of cortical extract.
Measurement of neprilysin activity
The neprilysin activity was measured fluorometrically as adapted from a published procedure [39]. Briefly, the brain was quickly removed, cortices dissected out, washed extensively with cold saline solution and homogenized with five volumes of 50 mM Tris buffer, pH 7.4, containing 1% triton X100. The homogenate was kept in ice and sonicated with 5 bursts (8 s each with 30 s interval). The extract (200–400 μg protein) was incubated in a total volume of 200 μl for 1 h at 37°C in the presence of the synthetic substrate N-dansyl-Ala-Gly-D-nitro-Phe-Gly (500 μM) without or with DL-thiorphan (1 mM) which is a specific inhibitor of neprilysin. After the end of the incubation the reaction was stopped by heating at 95°C for 5 min and centrifuged for 5 min at 5,000 g at room temperature. The fluorescence intensity of the supernatant was measured at λex 342 nm/λem 562 nm keeping appropriate sample and dye blanks, and the DL-thiorphan sensitive readings were taken to calculate the enzyme activity. However, this assay method might not differentiate the true neprilysin activity from other closely related endopeptidases even in the presence of DL-thiorphan [40].
Protein carbonyl detection
Protein carbonyls were detected by derivatizing the proteins with DNPH, followed by SDS–PAGE, electroblotting, and probing of the derivatized proteins first with anti-DNP antibody and subsequently with alkaline phosphatase-conjugated secondary antibody. The bands were developed by a chromogenic detection technique using BCIP/NBT [37]. For the control of loading and transfer, identical amounts of protein from different samples were loaded in individual wells in duplicate sequence. After electroblotting, the protein samples in one-half of the membrane were immunostained, and the remaining half containing the duplicate sample lanes was stained with amido black. Visual inspection and densitometric scanning of the amido black-stained membrane were carried out to verify equal loading and transfer. This procedure was adapted from an earlier published method [37].
Statistical analysis
Statistical significance was calculated by one - way ANOVA followed by Tukey’s post-test to assess differences between groups. A p value ≤0.05 was considered to be statistically significant. Each value is expressed as the mean±SEM of 6 observations.
RESULTS
Iron and transferrin receptor 1 in aged brain: Effect of deferasirox
As evidenced in Fig. 1, the level of chelatable iron was increased in the brain cortical extract by nearly 2.3 times in aged rats compared to that in young, but the phenomenon was prevented significantly in aged rats treated with deferasirox. The mRNA expression level of TfR1 gene in the aged (A) rat brain was also increased by 2.5 folds than in the young (Y) brain which, however, was markedly prevented in the brains of deferasirox-treated aged (CT) rats (Fig. 2a). Likewise, the protein expression level of TfR1, as shown by immunoblotting, was conspicuously increased in the aged than in the young brain (Fig. 2b). The densitometric analysis indicates an almost 5-fold increase in the TfR1 band intensity in the aged group (Fig. 2c). The age-related increase in the protein expression level of TfR1 was, however, conspicuously reversed in deferasirox-treated aged rats (Fig. 2b, c).
Deferasirox inhibits age-dependent oxidative stress and NF-κB activation in rat brain
A marked increase in protein carbonyls, the marker of oxidative protein damage, was noticed in the aged rat brain cortical extract by immunoblotting, which was significantly prevented by deferasirox treatment of the aged rats (Fig. 3a). The amido black stained portion of the membrane (Fig. 3b) containing the duplicates of different samples used for immunostaining (Fig. 3a) shows that the protein loading in each lane was identical. The NF-κB (p65 subunit) band in the nuclear extract from the brain cortices was much more intense in the aged group (nearly 2.8 fold by densitometric analysis) than in the young indicating an increased nuclear translocation and activation of NF-κB in the aged brain, but the latter phenomenon was significantly inhibited in the deferasirox-treated aged animals (Fig. 4a, b). There was no difference in the cytosolic p65 content among different samples (Fig. 4a, b).
Elevated ferritin in aged rat brain: Effects of deferasirox
As presented in Fig. 5a, a statistically significant increase (more than 2-fold) in the mRNA expression level of FtH gene in the cerebral cortex was observed in aged (A) compared to that in young (Y) rats, but the increase was markedly prevented in the deferasirox-treated aged (CT) animals (Fig. 5a). In accordance with the increase in FtH mRNA expression level, FtH protein level in the brain, measured by ELISA, was found to increase, almost 3 times, in aged (A) rats compared to young (Y) ones. However, the deferasirox treatment (CT) significantly lowered the brain FtH level of aged (A) rats (Fig. 5b).
Effect of deferasirox on age-related changes in AβPP in rat brain cortex
The mRNA expression levels of AβPP gene in the cerebral cortex of young (Y), aged (A), and deferasirox-treated aged (CT) rats did not show any statistically significant difference among the groups, but the expression level of AβPP appeared to be marginally higher in the aged (A) animals (Fig. 6a). In contrast, the protein expression level of AβPP in the cortex, determined by immunoblotting, was found to be significantly higher (nearly 3-fold) in aged (A) than in young (Y) rats (Fig. 6b, c). The age-dependent increase in cortical AβPP level, however, was reversed substantially in the deferasirox-treated (CT) animals (Fig. 6b, c). The polyclonal AβPP antibody used by us detected several isoforms of AβPP with approximate molecular weights of 160 kD, 110 kD, and 97 kD, respectively (data not shown). The prominent 110 kD band, presented in Fig. 6b, was densitometrically analyzed for all our experimental purposes.
Deferasirox prevents age-related changes of β-secretase and neprilysin activities and Aβ42 content in rodent brain
The activity of β-secretase, measured in the cerebral cortical extract, was increased by nearly 50% in aged (A) compared to that in young (Y) rats and the deferasirox-treatment (CT) marginally prevented this (Fig. 7a). On the other hand, the activity of neprilysin in the brain cortex was decreased by 48% in aged (A) compared to young (Y) rats which was markedly prevented when the aged rats were treated with deferasirox (CT) (Fig. 7b). A marked increase (2.4 fold) in cerebral cortical content of Aβ42 was also observed in aged in comparison to that in young (Y) rats, which again was prevented remarkably in deferasirox-treated aged (CT) rats (Fig. 8).
DISCUSSION
This study has attempted to elucidate the underlying mechanisms of altered iron accumulation in aged brain and its correlation with Aβ homeostasis. First of all, we have analyzed several parameters related to iron transport and storage in the aged brain and how these are affected by the treatment of the aged animals with the iron chelator deferasirox. Secondly, we have attempted to explore how the treatment with deferasirox affects the Aβ homeostasis in the aged rat brain.
The uptake and trafficking of iron in the brain through different compartments and its efflux from the CNS is a highly coordinated phenomenon that involves several transport proteins, receptors, enzymes, and storage proteins [41, 42]. The entry of iron in to the brain is regulated at the blood-brain barrier, where transferrin dependent uptake through transferrin receptor mediated endocytosis plays a crucial role. The subsequent trafficking of iron through brain interstitial fluid, neurons, glia, and cerebrospinal fluid is mediated by both vesicular and non-vesicular transport involving many different transport proteins and enzymes, e.g., transferrin, transferrin receptor, divalent metal transporter 1, ferroportin, hepcidin, hephaestin, etc. [42, 43]. The age-related increase in brain iron concentration in different regions and in animals and human beings is well documented, but the underlying mechanism is not clear [11, 44–47]. Our present study also confirms an elevated chelatable iron level in aged rat brain (Fig. 1). Since TfR1 mediated endocytosis is the major mechanism of iron uptake across the blood-brain barrier and other brain compartments, we have examined the mRNA and protein expression levels of TfR1. Our results show that both mRNA (Fig. 2a) and protein levels of TfR1 are increased in aged rat brain (Fig. 2b, c), we suggest that the age-dependent increase in brain iron content could be, at least partially, the result of overexpression of TfR1. The increased intracellular content of iron may, in turn, regulates the level of TfR1 post-transcriptionally through iron-responsive element binding proteins at 3’ UTR producing a feed-back inhibition [48]. After deferasirox treatment, the chelatable iron content of the aged brain is diminished (Fig. 1) which is expected since deferasirox is known to decrease the tissue iron load and can also reach the brain from the peripheral circulation [34, 50]. More interestingly, in this study we have also observed that deferasirox decreases the TfR1 expression (mRNA and protein) in the aged brain which could also contribute to the iron-lowering effect of deferasirox. The trigger for the overexpression of TfR1 in aged brain could be the increased reactive oxygen species (ROS)acting through redox-sensitive transcription factors like NF-κB and HIF which are implicated in the regulation of TfR expression [51, 52]. In support of this, our results show an increased level of protein carbonyls (Fig. 3a), a hallmark of oxidative stress, and enhanced NF-κB translocation to the nucleus in the aged brain(Fig. 4a, b). Deferasirox, because of its iron-chelating property, prevents the metal-catalyzed oxidative stress and consequent activation and nuclear translocation of NF-κB (Figs. 3a, 4a, b), and this may explain the downregulation of TfR1 by deferasirox in the agedbrain.
Ferritin stores iron intracellularly in the inactive form by forming a Fe3 +-ferritin complex. The intracellular level of ferritin is regulated at the post-transcriptional level by iron through IRE present at 5’-UTR of ferritin mRNA and at the transcriptional level by oxidants [48, 53–55]. There are scattered reports that document an increased content of ferritin in the aged brain, especially in the cerebral cortex [44, 56]. This age-dependent increase in brain ferritin content may be construed as mediated through transcriptional andpost-transcriptional mechanisms by the increased oxidative stress and iron present in the aged brain, and the process may be considered as a compensatory mechanism to shunt excess reactive iron in an inactive form inside the cells to prevent metal-catalyzed oxidative damage. In agreement with this presumption, we have also noticed an increased level of ferritin (mRNA and protein) in the aged brain, which is reversed when aged rats are treated with deferasirox that decreases the excess iron accumulation and oxidative stress in the aged brain (Fig. 5a, b).
Among the multiple consequence of excess iron accumulation in the aged brain, we have particularly explored the one leading to altered Aβ homeostasis. Like ferritin mRNA, the AβPP mRNA is also post-transcriptionally upregulated by intracellular iron through the binding of IREBP at the 5’-UTR [22, 57]. Our earlier study has also shown that iron causes a marked increase in AβPP content in SHSY5Y cells without a corresponding change in gene expression implying a post-transcriptional upregulation [58]. Thus, the increased accumulation of iron in the aged brain could lead to an increased rate of translation of AβPP mRNA and a consequent accumulation of AβPP. Our results precisely confirm this by showing a significant rise in AβPP content of the aged brain but a relatively marginal increase in the AβPP mRNA expression (Fig. 6a-c). On treatment with deferasirox, the iron content of the aged brain is decreased with a consequent decrease in the AβPP accumulation (Fig. 6b, c). The key enzyme of Aβ42 synthesis from AβPP is β-secretase, which is increased in the aging brain of multiple species [59]. This is also confirmed in our study, but deferasirox-treatment only marginally prevents the phenomenon implying that iron does not have a major role to play in the age-dependent increase in the brain β-secretase activity (Fig. 7a). The promoter region of β-secretase gene contains an NF-κB response element, and thus the former is likely to be upregulated following NF-κB activation by ROS [60, 61]. The increased oxidative stress and activation of NF-κB in the aged brain, as noticed in this study, could thus account for the observed increase in β-secretase activity in this condition (Fig. 7a). However, this possibility is apparently ruled out because deferasirox fails to prevent the increased activity of β-secretase in the aged brain aged brain in a significant manner even when it considerably abolishes the associated increased oxidative stress and NF-κB activation. Our present results do not offer any clear mechanism for the age-related increase in β-secretase activity in the aged rat brain. On the other hand, the age-dependent decrease in neprilysin, an important degrading enzyme of Aβ42, is very significantly reversed by iron chelation in our study (Fig. 7b). The inhibition of neprilysin could be related to increased iron-dependent oxidative stress in the aged brain, since the treatment of aged rats with deferasirox substantially abolishes the increase in the brain neprilysin activity (Fig. 7b). Oxidative inactivation of different enzymes in aged brain has been suggested in many studies, and an accumulation of oxidatively-modified neprilysin also occurs in the brain in aging and AD [62–64]. Thus our results tend to imply that the increase in AβPP synthesis and decrease in neprilysin activity would lead to an increased content of Aβ42 in the aged brain. This is clearly confirmed from our results that also show that the iron chelator deferasirox is a very effective drug to prevent Aβ accumulation in the aged brain (Fig. 8).
In summary, this study has shown that iron dysregulation and oxidative stress work in concert to alter the Aβ homeostasis that culminates in the accumulation of Aβ42 in the aged rat brain. The efficacy of the orally administered metal chelator deferasirox in preventing the accumulation of Aβ peptide in the aged brain is also clearly established from this study. However, several other kinds of age-related alterations in the brain such as mitochondrial dysfunction, increased levels of proinflammatory cytokines, or synaptosomal ionic alterations have been reported from our earlier studies, which are all remarkably prevented by the prolonged dietary administration to the aged rats of a combination of N-acetylcysteine, α-lipoic acid, and α-tocopherol having strong antioxidant and multiple non-antioxidant actions [37, 65]. In particular, we have reported increased ROS production, NF-κB activation, and higher levels of proinflammtory cytokines in the aged brain, which is an interesting observation given the fact that NF-κB could be activated both by ROS and proinflammatory cytokines leading to overexpression of a large set of genes including those for proinflammatory cytokines, antioxidant and pro-oxidant enzymes [37, 66]. In the present study, we have observed that the oral administration of deferasirox to aged rats prevents NF-κB activation in the brain, and thus it might be pertinent to investigate in a later study whether this treatment also would prevent the rise in the levels of proinflammatory cytokines in the aging brain. Another limitation of the present study is the absence of behavioral data to show the age-dependent impairment of learning and memory and any possible reversal of the process by deferasirox treatment of aged rats. Further, it is also not established from our data whether the age-related accumulation of Aβ42 in the brain contributes to the behavioral deficits in the aged rats. However, in our earlier published studies we have reported the deficits of memory and learning in aged rats which is significantly prevented by the dietary administration of the same combination of N-acetylcysteine, α-lipoic acid, and α-tocopherol [65]. Thus, it would be interesting also to investigate later if deferasirox mimics this combination formulation in preventing the aging associated memory deficits in rats.
We have used a prolonged treatment schedule with deferasirox in this study without any apparent toxicity or any significant difference in the body weights and survival rates of the animals between the two groups. Although the present results have been obtained in the context of non-pathological brain aging, the results have obvious implications in AD also. The use of metal chelators as a potential therapy for AD has been advocated by many primarily because accumulations of Cu, Zn, and Fe occur in AD brain especially in amyloid plaques and multiple toxic actions of metal-peptide chelate are noted in experimental conditions [67–69]. Unlike Zn, Cu and Fe are redox-active transition metals that can lead to oxidative damage in AD brain, and thus the interactions of Cu2 + and Fe3 + with AβPP or Aβ42 have been the subject of extensive investigations. The binding sites of Cu2 + in both AβPP and Aβ42 have been analyzed in considerable details, and it has been documented that Cu2 + not only catalyzes Aβ42 aggregation but the bound metal also initiates redox cycling reactions to potentiate the toxic effects of the peptide [70–72]. Despite some uncertainty and lesser number of published reports, it is suggested that the binding of Fe3 + to Aβ42 occurs in a similar fashion with similar consequences as noted in the case of Cu2 + binding [73–75]. In addition, the role of iron in the translational regulation of the intracellular level of AβPP through IRE at 5’-UTR of mRNA has been well worked out, which clearly justifies the use of iron chelators as disease-modifying drugs for AD [13, 20]. In fact, the availability of IRE, acute box element, and other regulatory units at 5’-UTR of AβPP mRNA has provided us an opportunity for developing AD therapeutics using various small molecules such as metal chelators or non-steroidal antiinflammatory agents or compounds directly binding to stem-loop structure of IRE [13, 22]. In contrast to various in vitro or cell based studies investigating the role of iron in the translational control of AβPP synthesis, the present work has verified the multiple effects of iron in enhancing the synthesis and inhibiting the degradation of Aβ42 in an in vivo model of brain aging, and thus we suggest that iron chelators would be effective drugs to decrease the amyloid burden of the AD brain also. It is important to mention that some metal chelators like deferoxamine, clioquinol, etc., have produced limited benefit in several clinical trials [68, 77]. Deferasirox, unlike deferoxamine, is effective after oral administration, and the drug has been cleared by the FDA for use in different clinical conditions to prevent iron-overloading of tissues [33, 79]. The present study suggests that the drug would be useful for AD as well and should be explored further in this respect.
