International Journal of Molecular Sciences Article Chronic Mild Stress Modified Epigenetic Mechanisms Leading to Accelerated Senescence and Impaired Cognitive Performance in Mice Dolors Puigoriol-Illamola 1,2 , Mirna Martínez-Damas 3 , Christian Griñán-Ferré 1,2,† and Mercè Pallàs 1,2,†, * 1 Pharmacology Section, Department of Pharmacology, Toxicology and Therapeutic Chemistry, Faculty of Pharmacy and Food Sciences. University of Barcelona, Av Joan XXIII 27-31, 08028 Barcelona, Spain; dolo.puigoriol@gmail.com (D.P.-I.); christian.grinan@ub.edu (C.G.-F.) 2 Institute of Neuroscience, University of Barcelona (NeuroUB), Campus Mundet, Edifici de Ponent Passeig de la Vall d’Hebron, 171 08035 Barcelona, Spain 3 Institute of Biomedical Research (IIBO), Universidad Nacional Autónoma de México (UNAM), Subdirección de Enseñanza e Investigación, División de Investigación Biomédica, Centro Médico Nacional 20 de Noviembre, 04510 Ciudad de México, CDMX, Mexico; damas.87@gmail.com * Correspondence: pallas@ub.edu † These authors contributed equally.  Received: 20 January 2020; Accepted: 6 February 2020; Published: 10 February 2020  Abstract: Cognitive and behavioural disturbances are a growing public healthcare issue for the modern society, as stressful lifestyle is becoming more and more common. Besides, several pieces of evidence state that environment is crucial in the development of several diseases as well as compromising healthy aging. Therefore, it is important to study the effects of stress on cognition and its relationship with aging. To address these queries, Chronic Mild Stress (CMS) paradigm was used in the senescence-accelerated mouse prone 8 (SAMP8) and resistant 1 (SAMR1). On one hand, we determined the changes produced in the three main epigenetic marks after 4 weeks of CMS treatment, such as a reduction in histone posttranslational modifications and DNA methylation, and up-regulation or down-regulation of several miRNA involved in different cellular processes in mice. In addition, CMS treatment induced reactive oxygen species (ROS) damage accumulation and loss of antioxidant defence mechanisms, as well as inflammatory signalling activation through NF-κB pathway and astrogliosis markers, like Gfap. Remarkably, CMS altered mTORC1 signalling in both strains, decreasing autophagy only in SAMR1 mice. We found a decrease in glycogen synthase kinase 3 β (GSK-3β) inactivation, hyperphosphorylation of Tau and an increase in sAPPβ protein levels in mice under CMS. Moreover, reduction in the non-amyloidogenic secretase ADAM10 protein levels was found in SAMR1 CMS group. Consequently, detrimental effects on behaviour and cognitive performance were detected in CMS treated mice, affecting mainly SAMR1 mice, promoting a turning to SAMP8 phenotype. In conclusion, CMS is a feasible intervention to understand the influence of stress on epigenetic mechanisms underlying cognition and accelerating senescence. Keywords: stress; epigenetics; senescence; cognition; age-related cognitive decline; Alzheimer’s disease; SAMP8; SAMR1; oxidative stress; inflammation; autophagy 1. Introduction Aging is a multifactorial process characterized by a progressive loss of physiological integrity, resulting in an impaired function and increased vulnerability to death. Physiological brain aging involves cognitive impairment, which includes decreased learning and memory abilities and slower Int. J. Mol. Sci. 2020, 21, 1154; doi:10.3390/ijms21031154 www.mdpi.com/journal/ijms Int. J. Mol. Sci. 2020, 21, 1154 2 of 23 responses to different stimulus [1]. Indeed, normal aging differs from pathological aging and this might be explained by the lifestyle. Environmental factors drive epigenetic modifications, resulting in phenotypic differences, which alter almost all tissues and organs. Although the brain is one of the most affected structures, such modifications lead to a progressive decline in the cognitive function and create a favourable context for the development of neurodegenerative diseases [2]. Aging is the most significant risk factor for most chronic diseases, such as age-related cognitive decline and Alzheimer’s disease (AD) [3]. In fact, AD is the most prevalent dementia and its progression is influenced by both genetic and environmental factors [4]. Several studies define that a good environment is essential to enhance learning and cognitive abilities, besides the continued presence of stressors being associated with the opposite effects. Epigenetics refers to potentially heritable and environmentally modifiable changes in gene expression mediated via non-DNA encoded mechanisms [5]. Recent evidence has demonstrated significant associations between epigenetic alterations and stress [2,6]. Potential threats cause a course of action releasing numerous transmitters and hormones throughout our body, particularly catecholamines and glucocorticoids. The interaction of glucocorticoids and adrenergic systems in specific brain regions has proved an essential mediating mechanism for a wide variety of actions displayed by stress on cognition [7]. Firstly, stress response allows body adaptation. However, when stress is prolonged it has been shown to decrease synaptic plasticity [8], alter hippocampal volume [9,10] and neurotransmitters [11], and may lead to AD [12]. In particular, chronic stress is associated with increased Aβ deposits and hyperphosphorylated Tau [13]. Oxidative stress (OS) and inflammation are deeply involved in age-related deleterious disorders. Along with the aging process, several factors, such as a naturally decreased capacity of the antioxidant enzymes system, create an imbalance between antioxidant mechanisms and reactive oxygen system (ROS)–production equilibrium, accumulating ROS beyond the detoxifying capacity of the antioxidant system resulting in OS, eventually causing cellular damage that can no longer be repaired by internal mechanisms, and finally causing dysfunction of the system [14]. In addition, one of the major changes during aging is the dysregulation of the immune response, leading to chronic systemic inflammatory state [15]. Overall, OS imbalance, mitochondrial dysfunction and the inflammatory response have been linked to accelerated aging and faster progression of neurodegenerative diseases [14,16]. Moreover, the accumulation of dysfunctional and damaged cellular proteins and organelles occurs during aging, resulting in a disruption of cellular homeostasis and progressive degeneration and increases the risk of cell death [17]. Autophagy is particularly important in the maintenance of homeostasis, as it participates in the elimination of disrupted proteins and molecules and helps to maintain a clean cellular environment. It has been widely described that there are alterations in autophagy during aging as well as in age-related cognitive decline, AD and other dementias [18–20]. Moderating all these detrimental components is key in the promotion of cell survival and longevity. The senescence-accelerated mouse (SAM) models include senescence-accelerate prone (SAMP) and senescence-accelerated resistant (SAMR) mice [21]. SAMP8 strain has been widely used as an aging model for the study of brain aging and age-related pathologies. Its cognitive impairment is associated with alterations in both hippocampal structure and activity, causing oxidative stress imbalance and driving to AD neuropathology, such as tau- and amyloid-related alterations [1,22]. On the contrary, SAMR1 strain manifests a normal aging process and is considered as the control strain for SAMP8. Understanding the epigenetic modifications that a stressful environment triggers in neurodegeneration mechanisms is essential to develop novel therapeutics for age-related cognitive decline. This study aims to determine the effects on brain function of a stressful lifestyle in an animal model with accelerated senescence, SAMP8, and in the resistant senescence strain, SAMR1. Int. J. Mol. Sci. 2020, 21, 1154 3 of 23 2. Results 2.1. Epigenetic Modulation Triggers Compacted Chromatin after Chronic Mild Stress In attempting to systematically address which are the epigenetic modifications caused by chronic mild stressful stimuli, we studied the major epigenetic marks that include histone modifications, DNA methylation and non-coding RNAs. Regarding histone modifications, we studied global histone 3 (H3) and histone 4 (H4) acetylation, phosphorylation of histone 2A (H2A.X) protein levels, methylation of H3 acetylated at Lys9 protein levels, as well as several deacetylases gene expression and/or protein levels, such as HDAC2, Sirt1, Sirt2, and Sirt6. CMS decreased H3 acetylation in senescence-resistant mice, similarly to SAMP8 control group, and increased in SAMP8 mice (Figure 1A). Considering H4, no statistically significant differences were determined, although a tendency to diminish in SAMR1 was observed (Figure 1B). On one hand, in accordance to H3 acetylation levels, there was a huge difference in histone deacetylase 2 (HDAC2) protein levels between SAMR1 and SAMP8 control groups, being higher in SAMP8 (Figure 1C). On the other hand, CMS modified HDAC2 protein levels, particularly decreased in SAMP8 and increased in SAMR1 compared to their littermates. Other deacetylases enzymes are of the sirtuin (Sirt) family. Not only gene expression (Figure 1D,E), but also protein levels of Sirt1, Sirt2 and Sirt6 were clearly decreased in SAMP8 mice compared to their control strain (SAMR1). In addition, CMS reduced Sirt1 (Figure 1D) and Sirt2 (Figure S1) gene expression and protein levels (Figure S1) in SAMR1 mice, but not in SAMP8. Moreover, histones can suffer different modifications such as phosphorylation or methylation, among others. In our hands, p-H2A-X protein levels did not differ between SAMR1 and SAMP8 control groups. Methylation of H3K9 was also higher in SAMP8 (Figure 1G). Of note, CMS increased the phosphorylated form of histones p-H2A-X and methylation of H3K9 in SAMP8 in reference to their control littermates, as well as compared to stressed SAMR1 (Figure 1F,G). Considering DNA methylation, global 5-mC DNA was slightly lower in SAMP8 compared to SAMR1. CMS significantly reduced global methylation both in SAMP8 and SAMR1 (Figure 1H). DNA methylation occurs through DNA MethylTransferase (DNMT) enzymes, of which DNMT1, DNMT3A and DNMT3B are the best characterized. As described, DNMTs were significantly higher expressed in SAMP8 compared to SAMR1 control group in CMS animals (Figure 2B for Dnmt3; data not shown for Dnmt1 and Dnmt2). 5-mC can be further oxidized to 5-hydroxymethyl-cytosine (5-hmC) due to ten-eleven translocase (TET) enzymes. SAMP8 mice had lower 5-hmC levels than SAMR1 mice (Figure 1I). Accordingly, TET2 protein levels were lower in SAMR1 than in SAMP8. CMS reduced TET2 protein levels in SAMR1 up to levels observed in SAMP8 control group (Figure 1J). miRNA also play an important role in regulating gene expression. Thereby, we decided to perform a microarray with 84 miRNAs related with neurodegeneration and aging in order to evaluate the influence of our experimental conditions (strain and CMS) those miRNA expression. Considering these results (Tables S3 and S4), we proceeded to validate the relative expression of those modified miRNA in a significant manner regarding strain and CMS. miR-29c is involved in neural proliferation regulation by Dnmt3. Although no statistical differences were seen, it seems that SAMP8 has higher relative expression than SAMR1, and CMS contributed to increasing it (Figure 2A,B). Those changes correlate with the tendencies seen in miR-29c-3p relative expression. miR-431-5p, miR-298-5p, miR-98-5p, and miR-140-5p expression were lower in SAMP8 than in SAMR1 (Figure 2C,E,G,I). Those miRNA are related to the inhibition of neurodegenerative pathways; its expression diminution correlated with changes in effector protein expression as β-catenin, β-site APP Cleaving Enzyme 1 (BACE1), soluble amyloid precursor protein-β (sAPPβ), and A Disintegrin and Metalloproteinase Domain-containing protein 10 (ADAM10) (Figure 2D,F,H,J). Furthermore, CMS seemed to reduce miR-431-5p expression in SAMR1 group, although not statistical differences were observed (Figure 2C,D). miR-181a-5p expression was higher in SAMP8 than SAMR1 (Figure 2K). Mammalian Target Of Rapamycin Complex 1 (mTORC1) is targeted by this miRNA, regulating cell growth, proliferation and survival. Likewise, mTORC1 protein levels were significantly higher Int. J. Mol. Sci. 2020, 21, 1154 4 of 23 inInt.SAMP8 than J. Mol. Sci. 2020,in 21,SAMR1 mice x FOR PEER (Figure 2L). Stressful stimuli only increased miR-181a-5p relative REVIEW 4 of 23 expression in SAMR1 mice compared to their control littermates but not in SAMP8 (Figure 2K). SAMR1-stressed Correlating group showed with miRNA higher expression, the mTORC1 protein group SAMR1-stressed levels showed than SAMR1 highercontrol mTORC1 (Figure 2L). protein BCL2 protein levels are controlled by miR-106b-5p. SAMP8 mice showed higher levels than SAMR1 control (Figure 2L). BCL2 protein levels are controlled by miR-106b-5p. SAMP8 relative expression in comparison mice showed higher with relative SAMR1 expression animals, andin CMS reducedwith comparison its expression only inand SAMR1 animals, SAMP8 CMSstrain (Figure reduced its 2M), which correlated with BCL2 protein levels (Figure 2N). expression only in SAMP8 strain (Figure 2M), which correlated with BCL2 protein levels (Figure 2N). Figure 1. Cont. Int. J. Int. J. Mol. Mol. Sci. Sci. 2020, 2020, 21, 21, x1154 FOR PEER REVIEW 55 of 23 of 23 Figure 1. Epigenetic marker changes after CMS. Representative Western Blot for the ratio of Acetylated at Lys9 H3 Figure 1. protein levels Epigenetic and quantification marker changes after (A),CMS. the ratio of AcetylatedWestern Representative at Lys 12 Blot H4 protein for thelevels ratioand of quantification(B), Acetylated at Lys9 H3 protein levels and quantification (A), the ratio of Acetylated at Lys 12and HDAC2 protein levels and quantification (C), the ratio of p-H2A.X protein levels H4 quantification protein levels(F), andthequantification(B), ratio of H3K9me2HDAC2 protein levels andlevels protein quantification (G), and TET2 and quantification (C),protein levels the ratio of and quantification (J). Relative gene expression of Sirt1 (D) and Sirt6 (E). Global 5-methylated cytosine p-H2A.X protein levels and quantification (F), the ratio of H3K9me2 protein levels and quantification (H) and 5-hydroxymethylated cytosine levels (I). Gene expression levels were determined by real-time (G), and TET2 protein levels and quantification (J). Relative gene expression of Sirt1 (D) and Sirt6 (E). PCR. Values in bar graphs are adjusted to 100% for protein levels of SAMR1 control (R1 Ct). Values are Global 5-methylated cytosine (H) and 5-hydroxymethylated cytosine levels (I). Gene expression mean ± Standard error of the mean (SEM); (n = 4 for each group). * p < 0.05; ** p < 0.01; *** p < 0.001; levels were determined by real-time PCR. Values in bar graphs are adjusted to 100% for protein **** p < 0.0001. levels of SAMR1 control (R1 Ct). Values are mean ± Standard error of the mean (SEM); (n = 4 for each group). * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001. Int. J.Int. Mol. Sci. Sci. J. Mol. 2020, 21, 21, 2020, 1154x FOR PEER REVIEW 6 of2323 6 of Figure 2. Cont. Int. Int. J. Mol. J. Mol. Sci.Sci. 2020, 2020, 21,21, x FOR PEER REVIEW 1154 7 of 2323 7 of Figure 2. Validation of a representative subset of miRNA involved in brain aging, neurodegeneration Int. J. Mol. Sci. 2020, 21, 1154 8 of 23 Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 8 of 23 and autophagy. Relative expression Figure 2. Validation of miR-29c-3p, of a representative miR-431-5p, subset of miRNA miR-298-5p, involved miR-98-5p, in brain aging, miR-140-5p, neurodegeneration miR-181a-5p, and miR-106b-5p in the hippocampus of SAMR1 and SAMP8 mice and autophagy. Relative expression of miR-29c-3p, miR-431-5p, miR-298-5p, miR-98-5p, miR-140-5p, (A,C,E,G,I,K,M). RelativemiR-181a-5p, gene expression of Dnmt3a (B) and miR-106b-5p andhippocampus in the representative Westernand of SAMR1 Blot for β-Catenin SAMP8 protein mice (A, C, E, G, I,levels and K, M). quantification Relative(D), BACE1 gene protein expression levels and of Dnmt3a quantification (B) and (F),Western representative sAPPβ Blot protein levels andprotein for β-Catenin quantification levels and quantification (H), ADAM10 (D),and protein levels BACE1 protein levels quantification and quantification (J), mTORC1 (F), sAPPβ protein levels protein levels(L), and quantification and and BCL2 protein levels and quantification (N). Gene expression levels were determined by real-timeand quantification (H), ADAM10 protein levels and quantification (J), mTORC1 protein levels PCR. quantification Values in bar graphs are(L),adjusted and BCL2 to protein 100% for levels andlevels protein quantification of SAMR1 (N). Gene (R1 control expression levels Ct). Values arewere mean determined by real-time PCR. Values in bar graphs are adjusted to 100% for protein ± Standard error of the mean (SEM); (n = 6 for each group in miRNAs validation; n = 4 for each group levels of SAMR1 control (R1 Ct). Values are mean ± Standard error of the mean (SEM); (n = 6 for each group in in WB and real-time PCR studies). * p <0.05; ** p < 0.01. miRNAs validation; n = 4 for each group in WB and real-time PCR studies). * p <0.05; ** p < 0.01. 2.2. Chronic Mild Stress Attenuates Antioxidant Defence Mechanisms 2.2. Chronic Mild Stress Attenuates Antioxidant Defence Mechanisms As mentioned before, CMS modulated cell oxidative stress mechanisms favouring an oxidative As mentioned before, CMS modulated cell oxidative stress mechanisms favouring an oxidative state. SAMP8 animals showed higher H2 O2 levels in the hippocampus than SAMR1 groups, as well as state. SAMP8 animals showed higher H2O2 levels in the hippocampus than SAMR1 groups, as well increased protein levels of the antioxidant enzymes superoxide dismutase 1 (SOD1), catalase (CAT) and as increased protein levels of the antioxidant enzymes superoxide dismutase 1 (SOD1), catalase glutathione peroxidase 1 (GPX1), suggesting impaired mechanisms to fight against OS (Figure 3A–E). (CAT) and glutathione peroxidase 1 (GPX1), suggesting impaired mechanisms to fight against OS CMS induced (Figure a3A–E). diminution in Nuclear CMS induced factor erythroid a diminution 2-related in Nuclear factorfactor 2 (NRF2) erythroid protein 2-related levels factor only in 2 (NRF2) SAMR1,protein but reduced antioxidant levels only in SAMR1, enzymes studied but reduced in bothenzymes antioxidant strains. In addition, studied CMS in both produced strains. a slight In addition, but notCMS significant producedtendency a slight to butincrease aldehyde not significant oxidase tendency 1 (Aox1)aldehyde to increase gene expression, oxidase 1 a(Aox1) pro-oxidant gene expression, a pro-oxidant enzyme, in both strains (Figure 3F). enzyme, in both strains (Figure 3F). Figure 3. Changes in cellular oxidative stress after CMS. Representative OS measured as hydrogen Int. Int. J.J. Mol. Mol. Sci. 2020, 21, Sci. 2020, 21, 1154 x FOR PEER REVIEW 99 of of 23 23 Figure 3. Changes in cellular oxidative stress after CMS. Representative OS measured as hydrogen peroxide concentration in homogenates of hippocampus tissue (A). Representative Western Blot for peroxide concentration NRF2 protein in homogenates levels and quantification of (B), hippocampus tissuelevels GPX1 protein (A). Representative Western and quantification (C),Blot for SOD1 NRF2 protein levels and quantification (B), GPX1 protein levels and quantification (C), SOD1 protein levels and quantification (D), and Catalase protein levels and quantification (E). Relative protein levels and quantification gene expression of Aox1(D), (F).and Catalase Values protein in bar graphslevels are and quantification adjusted to 100% for(E).protein Relativelevels gene expression of SAMR1 of Aox1 (F). Values in bar graphs are adjusted to 100% for protein levels control (R1 Ct). Gene expression levels were determined by real-time PCR. Values of SAMR1 control are (R1 mean Ct).± Gene expression Standard error oflevels were(SEM); the mean determined by real-time (n = 4 for PCR. each group). Values * p< < 0.01.± Standard error of the arep mean 0.05; ** mean (SEM); (n = 4 for each group). * p< 0.05; ** p < 0.01. 3.3. Inflammatory 2.3. Inflammatory Activation Activation is is Induced Induced after after Application Application of of Chronic Chronic Mild MildStress Stress SAMP8 mice SAMP8 miceshow showhigher higherprotein protein levels levels of NF-κB of NF-κB compared compared toSAMR1 to the the SAMR1 control control group.group. CMS CMS seems to promote higher cytokines expression in SAMR1 than in seems to promote higher cytokines expression in SAMR1 than in SAMP8, as shown in the ratio SAMP8, as shown in the ratio Interleukin(Il)-6/Il-10 and tumor necrosis factor (Tnf) α levels (Figure 4A–C), Interleukin(Il)-6/Il-10 and tumor necrosis factor (Tnf) α levels (Figure 4A–C), although no statistical although no statistical differences differences were found. were found. In accordance, In accordance, Glial fibrillarGlial acidicfibrillar proteinacidic (Gfap)protein (Gfap) geneincreased gene expression expression in increased in the SAMR1 CMS group, and there is a slight increase in the the SAMR1 CMS group, and there is a slight increase in the SAMP8 CMS group in comparison to SAMP8 CMS group in comparison respective to respective controls (Figurecontrols 4D). (Figure 4D). Figure 4. CMS induced changes in inflammatory markers. Representative Western Blot for NF-κB protein levels and quantification (A). Relative gene expression of Il-6/Il-10 ratio, Tnf-α and Gfap (B–D). Figurein4.bar Values CMS induced graphs changestoin100% are adjusted inflammatory for proteinmarkers. Representative levels of SAMR1 Western control (R1 Blot expression Ct). Gene for NF-κB protein levels levels were and quantification determined by real-time(A). Relative PCR. Valuesgene expression are mean of Il-6/Il-10 ± Standard ratio, error of Tnf-α (SEM); the mean and Gfap (B– (n = 4 D). Values in bar graphs for each group). * p < 0.05. are adjusted to 100% for protein levels of SAMR1 control (R1 Ct). Gene expression levels were determined by real-time PCR. Values are mean ± Standard error of the mean (SEM); (n = 4 for each group). * p < 0.05. Int. J. Mol. Sci. Sci. 2020, 21, x1154 2020, 21, FOR PEER REVIEW 10 of 10 of 23 2.4. Chronic Mild Stress Modulates APP Processing as Well as Alters Microtubule Stabilization Through 2.4. Chronic Mild Stress Hyperphosphorylation of Modulates Tau ProteinAPP Processing as Well as Alters Microtubule Stabilization Through Hyperphosphorylation of Tau Protein Observing the results, SAMP8 showed lower ADAM10 protein levels than SAMR1 animals. Observing the results, SAMP8 showed lower ADAM10 protein levels than SAMR1 animals. CMS CMS promoted the β-amyloid amyloidogenic pathway in SAMR1 mice, increased sAPPβ protein promoted the β-amyloid amyloidogenic pathway in SAMR1 mice, increased sAPPβ protein levels and levels and BACE1 protein levels, and there was a tendency to increase Aβ-precursor gene BACE1 protein levels, and there was a tendency to increase Aβ-precursor gene expression as well as expression as well as reduce ADAM10 protein levels (Figures 2F,H,J and 5A). reduce ADAM10 protein levels (Figure 2F,H,J and Figure 5A). SAMP8 had higher Ser396 p-Tau levels (Figure 5B,C) than SAMR1. Similarly, GSK-3β SAMP8 had higher Ser396 p-Tau levels (Figure 5B,C) than SAMR1. Similarly, GSK-3β phosphorylation at Ser9 protein level was lower in SAMP8 in reference to SAMR1 (Figure 5D). CMS phosphorylation at Ser9 protein level was lower in SAMP8 in reference to SAMR1 (Figure 5D). CMS increased tau hyperphosphorylation at Ser396 and Ser404 compared to control groups; surprisingly, increased tau hyperphosphorylation at Ser396 and Ser404 compared to control groups; surprisingly, there is a trend towards CMS reducing GSK-3β protein levels (Figure 5D). there is a trend towards CMS reducing GSK-3β protein levels (Figure 5D). Figure 5. AD hallmarks changes after CMS. Relative gene expression of Aβ-precursor (A). Representative Figure 5. Western AD Blot for hallmarks the ratio ofchanges pS396-Tau after CMS. protein Relative levels gene expression and quantification of ratio (B), the Aβ-precursor (A). of pS404-Tau Representative Western Blot for the ratio of pS396-Tau protein levels and quantification protein levels and quantification (C), and the ratio of pS9-GSK-3β protein levels and quantification (B), the ratio of pS404-Tau (D). protein levels Gene expression levels were and quantification determined by(C), and the real-time ratio PCR. of pS9-GSK-3β Values protein in bar graphs levels and are adjusted to 100% for protein levels of SAMR1 control (R1 Ct). Values are mean ± Standard error of the meangraphs quantification (D). Gene expression levels were determined by real-time PCR. Values in bar (SEM); are=adjusted (n 4 for eachtogroup). 100% for * pprotein < 0.05; levels of SAMR1 ** p < 0.01; *** p