|Home | About | Journals | Submit | Contact Us | Français|
These experiments reveal for the first time that microRNAs mediate oxidant regulated expression of a mitochondrial tricarboxylic acid cycle gene (mdh2). mdh2 encoded malate dehydrogenase (MDH) is elevated by an unknown mechanism in brains of patients that died with Alzheimer’s disease (AD). Oxidative stress, an early and pervasive event in AD, increased MDH activity and mRNA level of mdh2 by 19% and 22%, respectively, in a mouse hippocampal cell line (HT22). Post-transcriptional events underlie the change in mRNA because Actinomycin D did not block the elevated mdh2 mRNA. Since microRNAs regulate gene expression post-transcriptionally, the expression of miR-743a, a microRNA predicted to target mdh2, was determined and showed a 52% reduction after oxidant treatment. Direct interaction of miR-743a with mdh2 was demonstrated with a luciferase based assay. Over-expression or inhibition of miR-743a led to a respective reduction or increase in endogenous mRNA and MDH activity. The results demonstrate that miR-743a negatively regulates mdh2 at post-transcriptional level by directly targeting the mdh2 3′ UTR. The findings are consistent with the suggestion that oxidative stress can elevate the activity of MDH through miR-743a, and provide new insights into possible roles of microRNA in oxidative stress and neurodegeneration.
Malate dehydrogenase (MDH, EC220.127.116.11) encoded by mdh2 is the final enzyme in the mitochondrial tricarboxylic acid (TCA) cycle. It catalyzes the inter-conversion of L-malate and oxaloacetate using nicotinamide adenine dinucleotide (NAD) as a cofactor to generate reducing equivalents (Dupourque & Kun 1969). MDH activity is elevated in brains of patients that died with Alzheimer’s disease (AD) (Bubber et al. 2005, Op den Velde & Stam 1976). Whether direct modifications of MDH at protein and/or gene levels contribute to the increased MDH activity in AD is unknown.
Oxidative stress, an early event in AD (Nunomura et al. 2001, Pratico et al. 2002), can directly modify DNA, RNA, lipids and proteins, which in turn lead to altered gene expression and enzyme activity. For example, oxidized RNA nucleoside 8-hydroxyguanosine is significantly increased in neurons from the frontal cortex of patients with familial AD bearing a mutation gene encoding presenilin-1 or amyloid-β protein precursor (Nunomura et al. 2004). Oxidative stress increases expression and activity of β-secretase in Ntera 2/D1 (NT2) neurons (Tamagno et al. 2002). On the other hand, the activity of the mitochondrial enzyme α-ketoglutarate dehydrogenase complex (KGDHC), a key TCA cycle enzyme, is reduced in AD brains (Butterworth & Besnard 1990, Gibson et al. 1988) and other neurodegenerative diseases (Albers et al. 2000, Butterworth et al. 1993, Gibson et al. 2003). The oxidative stress associated with AD may contribute to its reduction, as numerous oxidants reduce KGDHC activity regardless of whether they are added to cells or generated internally in cells (Butterworth et al. 1986, Gibson & Shi, Jeitner et al. 2005, Kumar et al. 2003, Sheu et al. 1998, Shi & Gibson 2007). Succinate dehydrogenase (SDH), another TCA cycle enzyme, is elevated in brains of patients with AD (Bubber et al. 2005) and in mouse cerebral cortex during chronic hypoxia (Caceda et al. 2001). Oxidative stress regulates expression of sdha, gene encoding one of the four subunits of SDH, through induction of nuclear respiratory factor-1 (NRF-1) (Au & Scheffler 1998, Miranda et al. 1999, Piantadosi & Suliman 2008). The induction of sdha by oxidative stress is likely to contribute to the increased SDH activity in AD. The current experiments tested whether oxidative stress may also mediate the elevation of MDH activity through induction of expression of mdh2.
Gene regulation can occur at multiple levels including transcriptional and post-transcriptional modifications. MicroRNAs (miRNAs) are small (~22 nucleotides) non-coding RNA molecules that act as key regulators of many cellular processes through post-transcriptional regulation (Bartel 2004, He & Hannon 2004). More than one third of human genes may be regulated by miRNAs. MiRNAs negatively regulate gene expression by degradation and translational inhibition of their target mRNAs (Ambros 2003, Pasquinelli et al. 2005). Dysregulation of miRNAs in human diseases including neurodegenerative diseases implicates their critical roles in the pathogenesis of these diseases (Gao 2008, Hebert & De Strooper 2009, Hebert et al. 2008, Ma & Weinberg 2008, Nelson et al. 2008, Saugstad, Schonrock et al., Sonntag, Yokota 2009). For example, loss of microRNA cluster miR-29a/b-1 in sporadic AD correlates with increased BACE1/beta-secretase expression (Hebert et al. 2008). The results suggest that loss of specific miRNAs can contribute to increased BACE1 and Aβ levels in sporadic AD.
The mechanism underlying dysregulated miRNAs in the diseases is unclear. Oxidative stress may initiate abnormalities of miRNAs in neurodegenerative diseases. Although roles of oxidative stress are well established in inducing transcriptional changes of genes that are implicated in these diseases (Shi & Gibson 2007, Maruichi et al., Piantadosi & Suliman 2008, Udelhoven et al.), few studies have examined oxidative stress mediated changes in miRNAs and their biological implications in neurodegenerative diseases. Two studies from one group show same small subset of miRNAs is increased in hippocampus from patients with AD and also in cell culture models of oxidative stress induced by aluminum and iron sulfate (Lukiw 2007, Lukiw & Pogue 2007). This is the first study that directly links oxidative stress with miRNA and neurodegenerative diseases. A very recent study also reveals that tert-butyl hydroperoxide induces diverse changes in 75 miRNAs, and the up-regulated (or down-regulated) miRNAs are associated with the decreased (or increased) expression of predicted targeted mRNAs (Wang et al.). However, the direct interaction between miRNAs and their predicted target mRNAs has not been experimentally validated.
The present study first tested whether oxidative stress leads to increased MDH activity in a mouse hippocampal line. In addition, the underlying mechanism of elevated MDH under oxidative stress was explored by determining whether oxidative stress alters gene expression of MDH and levels of the miRNAs that are predicted to target mdh2, the gene encoding MDH, and whether the miRNAs that are altered by oxidative stress mediate the activation of MDH.
All chemicals were purchased from Sigma (St. Louis, MO) unless indicated otherwise.
A mouse hippocampal HT22 line (purchased from Salk Institute for Biological Studies; San Diego, CA) was maintained at 37°C in a humidified incubator under 10% CO2 and 90% air in the complete medium [500 ml DMEM supplemented with fetal bovine serum (10%), 100 unit penicillin and 100 μg streptomycin (Invitrogen; Carlsbad, CA)]. Cells were typsinized when they reached 50% confluence and were seeded into 6-well plates at a density of 1×105 cells/well. Two days later, cells were treated with H2O2 (20 μM) in DMEM for 8 hr in the absence or presence of the inhibitor of transcription actinomycin D (Act D, 8 μM). Cells were harvested immediately after treatment for assay of MDH activity and isolation of total RNA and miRNAs (see below sections for details).
Male B6.129S7 mice (5 month) from The Jackson laboratory (Bar Harbor, Maine) were housed with constant temperature (22 ± 2 C), humidity (50 ± 5 %) and illumination (12 h light/dark cycles). The Institutional Animal Care and Use Committee of Weill Medical College of Cornell University approved all procedures with the animals. The mice were sacrificed and brains were removed and pulverized in liquid nitrogen into fine powders and aliquoted into Eppendorf tubes and stored at −80°C. Six mice were used.
MDH activity was measured by an oxaloacetate-dependent NADH oxidation assay as described previously (Bubber et al. 2005) with slight modifications. Briefly, treated cells were rinsed twice with 1 ml PBS buffer, and then 300 μl lysis buffer (Gibson et al. 1988) was added to each well. The cell suspensions were passed through a 1 ml syringe with a 27 gauge needle ten times. Twenty μl of blank (lysis buffer), standards or samples was added to wells of a 96-well plate followed by addition of 140 μl of 100 mM potassium phosphate buffer (pH 7.2) and 20 μl of 20 mM oxaloacetic acid. The reaction was initiated by adding 20 μl of 10 mM NADH to the mixture. The plate was read at an excitation wavelength of 340 nm and an emission wavelength of 460 nm (cut off 455 nm) at 30 °C for 25 min using a SPECTRAmax GEMINI XS fluorescence micro-plate reader (Molecular Devices; Sunnyvale, CA). The rate of NADH oxidation was corrected with an NADH standard curve and normalized to protein contents as determined by a Bio-Rad protein assay based on the Bradford dye-binding method (Bio-Rad; Hercules, CA) using BSA as standard.
Total RNAs were isolated from the treated cells and mouse brain using a mirVana miRNA Isolation Kit (Applied Biosystems; Foster City, CA) following the Total RNA Isolation Procedure. The reverse transcription of single-strand cDNA for gene expression was performed with 100 ng total RNA using Transcriptor First Strand cDNA Synthesis kit (Roche; Indianapolis, IN). Real-time PCR of mdh2 or sdha (gene encoding succinate dehydrogenase subunit A) or ogdh (gene encoding KGDHC subunit E1k) was performed using pre-designed TaqMan® gene expression assays (Applied Biosystems; Foster City, CA) with slight modification of previously described methods (Shi et al. 2007). In brief, each amplification mixture (20 μl) contained 9μl of cDNA template, 10μl of TaqMan® Fast Universal PCR Master Mix, 1μl of Taqman Gene Expression Assay (20x). Thermal cycler (Applied Biosystems 7500 Fast Real-Time PCR system) conditions were 95°C for 20 sec, and 40 cycles of 95°C for 3 sec and 60°C for 30 sec. All samples were normalized for beta-2-microglobulin (b2m) expression.
miRNAs were isolated from the same treated cells and mouse brain using the mirVana miRNA Isolation Kit (Applied Biosystems; Foster City, CA) following the Enrichment Procedure for Small RNAs. For real-time PCR analysis of miR-743a and miR-378, the reverse transcription of single-strand cDNA was performed with 10 ng total RNA using the TaqMan MicroRNA Reverse Transcription Kit (Applied Biosystems; Foster City, CA). Real-time PCR amplification mixture (20 μl) contained 9μl of cDNA template, 10μl of TaqMan 2x Universal PCR Master Mix, 1μl of TaqMan MicroRNA Assay (20 x). Thermal cycler conditions were 95°C for 10 min, and 40 cycles of 95°C for 15 sec and 60°C for 60 sec. Sno-202 was included as endogenous control for miRNA. A comparative Ct (the threshold cycle of PCR at which amplified product was first detected) method was used to compare the mRNA levels in samples from treated to that of the control.
Computational identification of the putative miRNAs that target mdh2 was performed using Microcosm Targets Version 5 (http://www.ebi.ac.uk/enright-srv/microcosm/htdocs/targets/v5/) and TargetScan 5.1 (http://www.targetscan.org/).
Co-transfection of different combination of a dual-luciferase-based reporter with or without mdh 3′UTR and miRNA or anti-miRNA (Table 1) was used to test whether miR-743a can specifically target mdh2 mRNA. The reporter contains firefly luciferase, Renilla luciferase and the unique miRNA target site at 3′UTR of mdh2. When the miRNA is expressed, it binds to the 3′UTR of its target gene and results in repression of luciferase gene expression. Therefore, reduction in luciferase activity represents the direct interaction between miRNA and 3′UTR of miRNA target gene. The dual-luciferase reporter with mdh2 3′UTR (Luc+mdh) was constructed by cloning the 3′ UTR from the mouse mdh2 gene into a vector containing the firefly luciferase under the control of an SV 40 promoter and Renilla luciferase under the control of a CMV promoter (pEZX-MT01, GeneCopoeia; Rockville, MD). Renilla luciferase was co-expressed with the firefly luciferase for normalization. The same reporter without the 3′UTR of mdh2 (Luc-mdh) was used as a negative control. HT22 cells in 6-well plates were transfected with different combination of Luciferase reporter (0.75 μg/well) and miRNA or anti-miRNA (12.5 nM) (Applied Biosystems/Ambion; Austin, TX) (Table 1) using Lipofectamine 2000 (Invitrogen; Carlsbad, CA). A Luc-Pair miR Luciferase Assay Kit (GeneCopoeia; Rockville, MD) was used to measure the activities of firefly and Renilla luciferase sequentially from a single sample. The firefly luciferase luminescence is elicited by one reagent, while a second reagent simultaneously quenches the firefly luciferase and elicits Renilla luciferase luminescence. Both firefly luciferase and Renilla luciferase activities were measured 24 hours after transfection with LMaxII384 (Molecular Devices; Sunnyvale, CA). In brief, media in the 6-well plates were removed and 400 μl of Working Solution I (lysis and firefly luciferase buffer, firefly luciferase substrate) were added to each well. After incubation for 10 min, three aliquots of 100 μl of cell lysate were transferred to three wells of a white well 96-well plate for measures of the firefly luminescence. After measurement, another 100 μl of Working Solution II (Renilla luciferase buffer, Renilla luciferase substrate) were added to each well already containing 100 μl of Working Solution I. After incubation for 10 min, the Renilla luminescence was measured. Firefly luciferase activity was normalized with Renilla luciferase activity in the same well by calculating the ratio of luminescence from the firefly luciferase to the Renilla luciferase. The luciferase activity in cells co-transfected with Luc-mdh and negative control miRNA, or Luc-mdh and negative control for anti-miRNA was regarded as 100%.
HT22 cells were seeded in 6-well plates at a density of 2×105 cells/well the day before transfection. Transfection with either miR-743a (12.5 nM) or anti-miR-743a (12.5 nM) (Applied Biosystems/Ambion; Austin, TX) was performed using Lipofectamine 2000 (Invitrogen; Carlsbad, CA). Cells were harvested 24 hr post-transfection for MDH activity and mRNA measurement.
Hydrogen peroxide (H2O2) is a common oxidant that models the oxidant changes in brains from patients that died with AD. Increased MDH activity has been shown in AD brain in regions with extensive oxidative stress (Bubber et al. 2005). Whether the increased MDH in AD can be mimicked by H2O2 in a mouse hippocampal HT22 cells was tested. An 8 hr treatment of HT22 cells with 20 μM H2O2 increased the MDH activity by 19% (Figure 1A), and also increased the mRNA level of MDH (mdh2) by 22% as compared to the untreated control (CON) (Figure 1B).
To test the mechanism of H2O2 induced activation of MDH, HT22 cells were treated with H2O2 in the presence of actinomycin D (Act D), an inhibitor of transcription. Act D did not block the activation of mdh2 induced by H2O2 (Figure 2A). We also included a positive control to validate the Act D experiment. The induction of sdha, the gene encoding one of the subunits of mitochondrial complex II or succinate dehydrogenase, is known to occur at transcriptional level (Au & Scheffler 1998, Miranda et al. 1999, Piantadosi & Suliman 2008). Under the same experimental conditions that had no effect on mdh2, Act D did block the oxidant induced activation of sdha. This suggests a transcriptional regulation of sdha. (Figure 2B), which is consistent with previous findings (Au & Scheffler 1998, Miranda et al. 1999, Piantadosi & Suliman 2008). The result suggested that H2O2 mediated activation of mdh2 likely occurs at post-transcriptional level.
Several microRNAs were predicted to target mdh2 by microRNA prediction software including Microcosm Targets Version 5 and MicroCosm Targets Version 5. The miR-743a was the one of the microRNAs with the highest score as predicted by these two programs and the alignment with mdh2 3′UTR was shown in Figure 3A. Taqman MicroRNA qPCR Assay was used to detect whether miR-743a exists in mouse brain and HT22 cells (a mouse hippocampal line) since it was not detected by Northern analysis in mouse brain (Watanabe et al. 2006). The results showed that miR-743a is a low abundant microRNA. MiR-743a is about 0.5% of a positive control (miR-378) in mouse brain, and is about 0.6% of miR-378 in mouse hippocampal HT22 cells. miR-378 is expressed in a number of cell lines and is involved in the expression of vascular endothelial growth factor. We next tested whether the mRNA level of miR-743a was altered by H2O2. The same treatment that altered activity and mRNA levels of MDH (20 μM H2O2 for 8 hr) reduced the mRNA level of miR-743a by 52% as compared to untreated control (Figure 3B).
To test for direct effects of miR-743a on the mdh2 transcript, a dual luciferase reporter with mdh2 3′ UTR (Luc+mdh) or without mdh2 3′ UTR (Luc-mdh) was used. Co-transfection of miR-743a with Luc+mdh into HT22 cells reduced the luciferase activity by 37% as compared to cells cotransfected with a negative control miRNA and Luc-mdh (luciferase reporter contains no mdh 3′UTR) (Figure 4). Co-transfection of miR-743a with Luc-mdh into HT22 cells did not alter the luciferase activity (Figure 4).
Anti-miR-743a inhibited miR-743a and antagonizes the repressive action of miR-743a on mdh2. Cotransfection of anti-miR-743a with Luc+mdh into HT22 cells increased the luciferase activity by 16% as compared to cells cotransfected with a negative control for anti-miRNA and Luc-mdh (luciferase reporter contains no mdh 3′UTR) (Figure 5). Cotransfection of anti-miR-743a with Luc-mdh into HT22 cells did not alter the luciferase activity (Figure 5).
We then tested whether over-expression or inhibition of miR-743a alters endogenous mRNA and activity levels of MDH in HT22 cells. Transfection of miR-743a into HT22 cells reduced endogenous MDH mRNA levels by 23% (Figure 6A), and MDH activity by 31% as compared to cells transfected with a negative control miRNA (Figure 6B), while no effect on expression of ogdh, gene encoding one of the subunits of KGDHC (Figure 6C) was observed. Transfection of anti-miR-743a increased the endogenous MDH mRNA level by 16% (Figure 7A), and MDH activity by 21% as compared to cells transfected with a negative control for anti-miR (Figure 7B), while no effect on expression of ogdh was observed (Figure 7C).
Oxidative stress is pervasive in AD brains (Calingasan et al. 1999) and may underlie many of the other changes. Enzymes related to oxidative stress and antioxidant pathways are generally induced in AD (Shi & Gibson 2007). For example, both mRNA and protein of heme oxygenase-1 (HO-1) are induced in cerebral cortex and vessels of AD (Premkumar et al. 1995). The gene expressions of Mn-SOD, Cu, Zn-SOD, catalase (CAT), glutathione peroxidase (GSH-Px), and glutathione reductase (GSSG-R) are elevated in brains from AD patients (Aksenov et al. 1998). HO-1 and Mn-Cu/Zn SOD are profoundly increased in PC-12 cells upon exposure of oxidative stress (Pappolla et al. 1998). The current study shows that H2O2 increased the MDH activity and mRNA in HT22 cells and suggests that the elevated MDH that occurs in AD brain may also result from oxidative stress.
The Act D results suggest that the up-regulation of mdh2 by H2O2 is likely to occur at post-transcriptional level, as Act D did not block the up-regulation of mdh2 by H2O2. Under the same condition, Act D completely blocked the induction of sdha by H2O2, suggesting a transcriptional regulation of sdha. This data is consistent with previous findings that sdha is transcriptional regulated by nuclear respiratory factor-1 (NRF-1), and NRF-1 is induced by ROS (Au & Scheffler 1998, Miranda et al. 1999, Piantadosi & Suliman 2008). Thus, the consistency of our data on sdha with previous studies also validated the Act D experiment on mdh2. In addition, our data suggest that increased sdha expression in response to oxidative stress may underlie the increased SDH activity in AD brain.
Although dysregulated miRNAs have been well demonstrated for other diseases such as cancer (Ma & Weinberg 2008) and normal neuronal function, surprisingly little research has been done related to the TCA cycle in neurodegenerative diseases. The importance of miRNAs in a number of physiologically relevant processes in the nervous system is firmly established (Gao 2008, Hebert & De Strooper 2009, Hebert et al. 2008, Nelson et al. 2008, Saugstad, Schonrock et al., Sonntag, Yokota 2009) including a role in neurite outgrowth and synapse formation (Hebert & De Strooper 2009). miRNAs have been shown to be particularly abundant in the brain (Corbin et al. 2009). Changes in miRNA profiles in neurodegenerative diseases have been supported by a number of studies. For example, loss of miR-29a/b-1 in sporadic AD correlates with increased expression of β-secretase (Hebert et al. 2008). Altered miRNA expression profiles are found in schizophrenia and Down syndrome; 16 miRNAs are differentially expressed in prefrontal cortex of patients with schizophrenia (Perkins et al. 2007). Five miRNAs are over-expressed in fetal brain and heart specimens from individuals with Down syndrome (Kuhn et al. 2008). Altered miRNA expressions are also found in cellular and animal models of neurodegenerative diseases (Lee et al., Schonrock et al.). These findings demonstrate that miRNA dysregulation may be important in the pathogenesis of neurodegenerative disorders.
The causal factor of miRNA dysfunction in neurodegenerative diseases is largely unknown and the current results suggest oxidative stress is a critical factor. A few recent studies have linked oxidative stress with altered miRNA expression in cellular models (Lukiw & Pogue 2007, Wang et al.). Possible post-transcriptional modification of mdh2 by H2O2 demonstrated by the Act D experiment led us to hypothesize that miRNAs may mediate oxidant induced elevation of mdh2 mRNA in HT 22 cells. miR-743a was predicted by at least two common computational programs to target mdh2 and its expression showed a 52% reduction after oxidant treatment. This is consistent with results showing that miR-743a is down-regulated by tertbutyl-hydroperoxide in auditory cells (Wang et al.). miRNAs generally negatively regulate gene expression by targeting 3′UTR of the target gene (Fabian et al., Zhang & Su 2009). A direct interaction between miR-743a and mdh2 3′UTR was demonstrated by the response of the luciferase activity and the endogenous MDH to over-expressing or inhibition of miR-743a. The luciferase reporter analyses showed a reduction or increase in the luciferase activity upon cotransfection of miR-743a or anti-miR-743a with Luc+mdh, respectively. Overexpression or inhibition of miR-743a leads to a respective reduction or increase in endogenous mRNA and activity levels of MDH. That inhibition of miR-743a by anti-miR-743a leads to increased endogenous mRNA and activity levels of MDH is consistent with H2O2 mediated inhibition of miR-743a and activation of mRNA and activity levels of MDH. Taken together, the current study clearly demonstrated that a direct interaction of miR-743a with mdh2 3′UTR, and oxidative stress induced MDH activity is likely mediated by the downregulation of miR-743a.
The long term goal of these studies is to understand the increased MDH in brains from AD patients. Oxidative stress is a prominent feature of AD and has been shown to alter miRNA expression in cellular models in a limited number of studies (Lukiw & Pogue 2007, Wang et al.). We demonstrate here that oxidative stress reduces the expression of miR-743a, which leads to increased MDH activity in a mouse hippocampal line as an experimental model. The finding provides novel insights into the mechanism underlying the elevated MDH in response to oxidative stress. However, the regulatory role of miR-743a on MDH may have limited direct relevance to AD in humans since the 3′UTR for mdh is not conserved between mice and humans. Limited data is available on the regulation of TCA cycle genes at the transcriptional and post-transcriptional level in any biological system and nothing is known about the regulation in human brain. Studying the low abundance micoRNAs is particularly challenging for the human neurons. Whether a similar mechanism underlies elevated MDH in AD needs to be tested in human neuronal lines in the future.
The functional significance of the increase in MDH is unknown. The malate-aspartate shuttle translocates electrons produced during glycolysis across the semipermeable inner membrane of the mitochondrion to the electron transport chain via NADH to generate ATP. The shuttle system is required since the mitochondrial inner membrane is impermeable to NADH. To circumvent this, malate carries the reducing equivalents across the membrane. Our data on elevated MDH under oxidative stress suggest a contribution of the functioning of the malate-aspartate shuttle and changes in the redox state of the neuron under oxidative stress. Another possibility is that under conditions in which the electron transport chain blocked (ie oxidative stress), that MDH works in the NADH to NAD direction to provide NAD to KGDHC and other dehydrogenases.
In summary, the results show that H2O2 up-regulates both gene expression and activity of MDH in a hippocampal cell line (HT22). Our data also show that miR-743a negatively regulates mdh2 at the post-transcriptional level by directly targeting the mdh2 3′ UTR. Moreover, H2O2 induced up-regulation of MDH is likely mediated by the down-regulation of miR-743a. The present study, for the first time, reveals an important role of miRNAs in mediating regulation of MDH under both normal and oxidative stress conditions. Further studies on regulatory mechanisms of miRNA expression during oxidative stress will help us to better understand the interactions between miRNAs and their target genes.
This work was supported by the Alzheimer’s Association New Investigator Grant NIRG-08-89776 and NIH PPG AG14930.
The authors would like to thank Dr. Rajiv Ratan, MD, PhD for his advice on Actinomycin D experiments and Hui Xu, B. Med. for her technical assistance with MDH activity assay.