Exercise Modalities Improve Aversive Memory and Survival Rate in Aged Rats: Role of Hippocampal Epigenetic Modifications

Abstract

We aimed to investigate the effects of aging and different exercise modalities on aversive memory and epigenetic landscapes at brain-derived neurotrophic factor, cFos, and DNA methyltransferase 3 alpha (Bdnf, cFos, and Dnmt3a, respectively) gene promoters in hippocampus of rats. Specifically, active epigenetic histone markers (H3K9ac, H3K4me3, and H4K8ac) and a repressive mark (H3K9me2) were evaluated. Adult and aged male Wistar rats (2 and 22 months old) were subjected to aerobic, acrobatic, resistance, or combined exercise modalities for 20 min, 3 times a week, during 12 weeks. Aging per se altered histone modifications at the promoters of Bdnf, cFos, and Dnmt3a. All exercise modalities improved both survival rate and aversive memory performance in aged animals (n = 7–10). Exercise altered hippocampal epigenetic marks in an age- and modality-dependent manner (n = 4–5). Aerobic and resistance modalities attenuated age-induced effects on hippocampal Bdnf promoter H3K4me3. Besides, exercise modalities which improved memory performance in aged rats were able to modify H3K9ac or H3K4me3 at the cFos promoter, which could increase gene transcription. Our results highlight biological mechanisms which support the efficacy of all tested exercise modalities attenuating memory deficits induced by aging.

Introduction

Preclinical and clinical studies have shown that physiological aging affects cognitive skills, such as short- and long-term memory [1–5]. Epigenetic mechanisms and the consequent loss of transcriptional activity have been associated to age-related impairments [6] and age-related illnesses, such as Alzheimer’s disease and various cancers [7–9]. Increased histone deacetylase (HDAC) activity and decreases in global H4 acetylation in hippocampus related to memory declines were demonstrated in aged rats [4, 10]. Age-related histone methylation changes have also been observed in animal models, such as decrease in global H3K9 methylation in the hippocampus [11]. Along with known age-related changes in histone modifications, changes in gene expression have been reported through the brain for several genes of interest [4]. Notably, genes related to learning and memory, including brain-derived neurotrophic factor (BDNF), cFOS, and DNA methyltransferase 3a (DNMT3A), undergo substantial expression changes within the brain during the aging process [12–14], all of which were related to aging-induced cognitive impairments. However, histone modifications on promoters of these genes are rarely evaluated as a potential molecular mechanism.

Several lines of evidence indicate that regular exercise is beneficial during aging [15, 16], reducing mortality rates and improving brain functions as motor performance and cognition both in humans and animal models [17–21]. However, little is known regarding the benefits of beginning an exercise routine in an aged model and if different exercise modalities provide greater benefits than others in an age-specific manner. It has been presumed that some of exercise modality, such as aerobic, balance, and resistance, have greater influence on some parameters than on others, such as aerobic training is related to improvement of aerobic capacity and resistance training is associated with improved muscle mass [22]. Interestingly, the American Heart Association and American College of Sports Medicine recommend a combination of aerobic, resistance, and balance exercise for elderly patients in order to contribute to a healthy independent lifestyle, improving functional capacity, and cognitive performance [18, 23].

Different exercise modalities were able to improve memory in rodents during adulthood [24–26]. For example, 8 weeks of progressive resistance exercise improved aversive memory and increased peripheral and hippocampal IGF-1 levels [24]. To our knowledge, only aerobic modalities have been studied in normal aging models; 2 weeks of a moderate aerobic protocol reversed age-induced aversive memory impairment [4]. There are few studies investigating resistance, acrobatic, or combined exercise effects in memory tasks using animal models during aging process.

Epigenetic mechanisms have been proposed as modulators of the protective effects of exercise on memory. Previous data have shown an age-dependent effect of aerobic exercise on hippocampal epigenetic parameters using animal models; global histone acetylation, its enzymatic system, site-specific histone acetylation, and epigenetic landscapes at Bdnf gene promoter were studied. Daily aerobic exercise protocol during 2 weeks (20 min/day) improved aversive memory performance and hippocampal H4 acetylation in aged rats [4]. Additionally, this protocol increased acetylation of specific lysine residues (H4K12 and H3K9) in hippocampus of 20–21-month-old Wistar rats [27]. Aerobic exercise also induced hippocampal H3 acetylation at Bdnf gene promoter of 3-month-old rats, a crucial gene for learning, memory, and synaptic plasticity [28]. Besides, a single aerobic exercise session decreased both DNMT3b and DNMT1 levels in hippocampus of young adult rats, without any effect in the aged group [29]. Although the epigenetic mechanisms have not been investigated, resistance and acrobatic modalities increased Bdnf expression in the hippocampus [30, 31]. However, there is a dearth of information regarding the epigenetic modulation exerted by exercise modalities in the aging hippocampus.

Therefore, the aim of this study was to investigate the potential protective effects of different exercise modalities, aerobic, resistance, acrobatic, and their combination, on age-related declines in aversive memory and identify exercise-induced epigenetic mechanisms, specifically histone modifications, associated with aging and the protective effects of exercise within the hippocampus. We have focused on the promoter regions of three genes important for synaptic plasticity, neuronal activity, and neuronal development: Bdnf, cFos, and Dnmt3a [13, 32].

Materials and Methods

Animals

Male Wistar (n = 76) rats aged 2 and 22 months were maintained under standard conditions (12-h light/dark, 22 C ± 2 °C), with food and water ad libitum. Animals were provided by the Centro de Reprodução e Experimentação de Animais de Laboratório (CREAL) at the Universidade Federal do Rio Grande do Sul (UFRGS) and housed three per cage (Plexiglass cages, dimensions 40 × 33.3 × 17 cm). The NIH “Guide for the Care and Use of Laboratory Animals” (NIH publication no. 80–23, revised 1996) was followed in all experiments. The Local Ethics Committee approved all handling and experimental conditions (no. 29818).

Exercise Protocols

Rats were randomly divided into five groups (n = 6–9): sedentary, aerobic, acrobatic, resistance, and combined. Experimental groups, except sedentary, were subjected to 20-min exercise sessions, three times a week in alternate days, for 12 weeks (experimental design in Fig. 1). Animals were habituated to each exercise modality by exposure to the different exercise apparatuses for 1 week prior to exercise protocol commenced. Sedentary animals were exposed to “sham” exercise protocols in that they were placed on a non-functional treadmill, acrobatic or resistance apparatus. Neither electric shock nor physical prodding was used in this study. All the procedures took place between 14:00 and 17:00.

Fig. 1.

Experimental design. I.A. inhibitory avoidance

Aerobic Protocol

Aerobic training was performed on a motorized rodent treadmill (AVS Projetos, São Paulo, Brazil) with individual Plexiglas lanes. The peak oxygen uptake (VO₂) was indirectly measured in all animals prior to training. Each rat ran on a treadmill at a low initial speed, and the speed was increased at a rate of 5 m/min every 3 min until the point of exhaustion (i.e., failure of the rat to continue running). The time to fatigue (in min) and workload (in m/min) were obtained as indices of exercise capacity, which, in turn, were taken as VO₂max [33, 34]. The aerobic training consisted of running sessions at 60% of VO₂max [4]. The animals initially refusing to run were encouraged by gently tapping on their backs. VO₂max was assessed every 3 weeks in order to increase treadmill speed according to the animal’s adaptation to exercise, maintaining training at 60% of VO₂max.

Acrobatic Protocol

Acrobatic exercises were employed according to Jones et al. [35]. Animals were required to complete five activities six times each: [1] transverse a horizontal ladder (100 cm of diameter, 3 cm spaced rungs), [2] walk through an obstacle course (barriers 5 to 21 cm high), [3] cross a seesaw, [4] transverse a narrow bar (90 cm of diameter and 10 cm of width), and [5] transverse a rope (100 cm of diameter and 5 cm of width). The obstacles were changed, bars and rope were narrowed, and the space between the rungs of the ladder was increased every 3 weeks in order to increase the training difficulty [36].

Resistance Protocol

Resistance exercise training was performed as adapted from Gil and Kim [29], using a climbing ladder (height 1 m, inclination of 85°), which was designed to make the rats ascend towards a dark chamber (20 × 20 × 20 cm) with weight attached to their tails. Animals scaled the ladder in series of 8 repetitions with weight attached to their tails, with 2 min of rest. The weight for training was determined using 1 RM (repetition maximum) test performed prior to training, and animals climbed the ladder twice with 50% of their body weight attached to their tails. After successful completion of the task, 30 g were added for another trial (climbed ladder 2× plus 2 min of rest). This was repeated until animals were unable to climb the ladder and the weight was recorded as the maximum overload. Animals started the experiment climbing with 50% of the maximum overload attached to their tails; every 3 weeks, 10% of the maximum overload was added, until they reached the maximum of 80% of the overload in the last 3 weeks [29].

Survival Rate

Survival was tracked from the beginning of the study (22 months of age) until 25 months of age at which point animals were submitted to behavioral test and euthanized.

Memory Paradigm: Inhibitory Avoidance

We used single-trial step-down inhibitory avoidance conditioning as a model of fear-motivated memory in which the animals learned to associate a location in the training apparatus (a depressed area indicated by a grid floor) with an aversive stimulus (footshock). In the training trial, rats were placed on a platform and immediately after stepping down on the grid received a 0.6 mA, 3.0 s footshock prior to removal from the apparatus. The test trial took place 24 h after the training trial, in which the rats were placed in the platform and no footshock was delivered; latencies to step down were recorded and used as a measure of memory retention. Retention test latency measurements were cut off at 180 s. All animals were subjected to inhibitory avoidance test session 30 min after the last exercise session. The general procedures for inhibitory avoidance behavioral training and the retention test were described in a previous report [4, 37].

Preparation of the Samples

All rats were euthanized by decapitation 1 h after the last exercise session. The whole hippocampi were quickly dissected, immediately snap-frozen in liquid nitrogen, and then stored at −80 °C until the ChiP analysis.

qChIP

ChIP analysis of the frozen hippocampus (n = 4–5) was performed as previously described by Peña et al. [38] with minor modifications. Samples were cross-linked in 1% formaldehyde for 12 min at room temperature. The crosslinking reaction was stopped by adding glycine (2 M). Samples were homogenized in cold PBS with protease inhibitors (PIs), incubated in cold cell lysis buffer (5 mM PIPES pH 8, 85 mM KCl, 0.5% NP40) with PIs, and sonicated in cold nuclear lysis buffer (50 mM Tris-HCl pH 8, 10 mM EDTA, 1% SDS) with PIs at 4 °C for 30–40 cycles in a Bioruptor (Diagenode®) until average fragment size was 100–150 bp as assessed by BioAnalyzer (Agilent). Samples were frozen in aliquots at −80 °C until immunoprecipitation. Histone modifications (acetylation and methylation) were assessed using the following antibodies from Abcam: anti-acetyl(ac) H3K9ac (15823), H4K8ac (15823), anti-methyl(me) H3K9me2 (1220), and H3K4me3 (8580). Mouse and rabbit isotype control antibodies (37355 and 172730, respectively) were included as negative controls.

Magnetic beads (Dynabeads M-280 Sheep anti-Mouse/ Rabbit IgG) were blocked (4% BSA in PBS) and incubated with antibody at 4 °C for 6 h. One thousand nanograms of chromatin from each sample was incubated with antibody-bead mix overnight (16 h) for immunoprecipitation. Prior to antibody incubation, 10% aliquots were removed from each sample as input. Chromatin samples were then washed 1× each with low salt buffer (0.1% SDS, 1% Triton X-100, 2 mM EDTA pH 8, 150 mM NaCl, 20 mM Tris-HCl pH 8), high salt buffer (0.1% SDS, 1% Triton X-100, 2 mM EDTA pH 8, 500 mM NaCl, 20 mM Tris-HCl pH 8), LiCl buffer (150 mM LiCl, 1% NP40, 1% NaDOC, 1 mM EDTA, 10 mM Tris-HCl pH 8), and TE buffer + NaCl (50 mM NaCl, 10 mM Tris-HCl pH 8, 1 mM EDTA). Samples were then eluted in buffer (1% SDS, 100 mM NaHCO₃) and reverse cross-linked overnight at 65 °C. After RNAse and proteinase K incubations, DNA was purified using the PCR Purification Kit (Qiagen® 28106) and stored at 4 °C until qPCR analysis (Applied Biosystems QuantStudio®).

Primers for quantitative chromatin immunoprecipitation (qChIP) (Integrated DNA Technologies®) were designed using Primer Blast (NCBI) to amplify promoter regions (Table 1). qChIP results, including IgG controls, were calculated as percent of input. All antibodies showed significant enrichment over their comparable IgG control (Supplemental Data 1).

Table 1.

Gene name, accession number, forward primer sequence, and reverse primer sequence of primer pairs used in quantitative PCR amplification

Rat qChIP primers
Gene	Accession no.	Forward sequence	Reverse sequence
Bdnf (Exon 1)	NC_005102.4	GGCAGTTGGACAGTCATTGG	GGGCGATCCACTGAGCAAAG
cFos	NC_005105.4	CCGACTCCTTCTCCAGCATG	GCGGCAGGTTTACTCTGAGT
Dnmt3A	NC_005105.4	TGGTGCCCAGGAAAGGAAAG	TGAGGCACCAATCTGAACCC

Statistical Analysis

All data were evaluated for normal distribution and homogeneity of variance using Kolmogorov–Smirnov and Levene tests. The influence of age on inhibitory avoidance paradigm and epigenetic markers was evaluated by Student’s t test comparing only sedentary adult and sedentary aged animals. The exercise modalities effects on inhibitory avoidance paradigm and epigenetic markers were evaluated by one-way ANOVA followed by Tukey’s post hoc tests. Survival rate was evaluated by Fisher’s exact test. In all tests, p ≤ 0.05 was considered to indicate statistical significance.

Results

Aging and Exercise Modalities Affects Aversive Memory Performance

First, baseline differences in YA vs aged sedentary controls were analyzed (Fig. 2). Student’s t test revealed that aged rats underperformed in the inhibitory avoidance paradigm (p = 0.03) compared to YA animals (Fig. 2a). Given the age-related changes observed between control groups, we analyzed the effects of exercise modalities within YA and aged animals separately. One-way ANOVA showed a main effect of exercise modalities on step-down latency in YA (F_(4;41) = 17.3; p < 0.0001) and aged (F_(4;37) = 3.63; p = 0.015) rats (Fig. 2b, c). Tukey’s post hoc analysis revealed that acrobatic (p = 0.005 and p = 0.022), aerobic (p = 0.002 and p = 0.031), and combined (p < 0.001 and p = 0.016) exercise enhanced memory performance in YA rats, and beyond these modalities, resistance exercise (p = 0.049) also ameliorated the age-related deficiencies observed in aged rats compared to sedentary animals. The p values refer to YA and aged rats, respectively.

Fig. 2.

Effect of aging and exercise modalities on step-down latency in inhibitory avoidance paradigm. Columns represent mean ± SD (n = 7–10). a Effect of aging on step-down latency. Student’s t test *values significantly different from the 5-month-old groups. b, c Effect of exercise on step-down latency in young adult and aged rats respectively. One-Way ANOVA; ^#values significantly different from the respective sedentary group

Exercise Improved Survival Rate in Aged Animals

Survival rate of aged animals was tracked from the beginning of the study (22 months of age) until the last exercise session (25 months of age). Late-life exercise increased rate of survival (96.7%) compared with the control group (45%) as evidenced by Fisher’s exact test (p = 0.002).

Aging Modulates Epigenetic Marks at the Bdnf, cFos, and Dnmt3a Gene Promoters in Hippocampus

We next evaluated the age-related alterations in histone modifications associated with 3 genes (Bdnf, cFos, and Dnmt3a) important for brain development, synaptic plasticity, and cognition [13, 28, 32] in the sedentary controls. Age-related changes were observed in histone methylation at Bdnf (Fig. 3a), cFos (Fig. 3b), and Dnmt3a (Fig. 3c). Specifically, H3K4 trimethylation (H3K4me3) was decreased at the cFos promoter but increased (p = 0.007) at the Dnmt3a and Bdnf promoters (p = 0.04 and p = 0.001) in aged vs YA animals. Age-related effects were observed for H3K9 dimethylation (H3K9me2) at the Bdnf promoter (p = 0.05). Fewer age-related effects were observed with regard to histone acetylation. While no effects were observed in H3K9 acetylation (H3K9ac), H4K8 acetylation (H4K8ac) was increased at the Bdnf promoter (p = 0.002) and decreased at the cFos promoter (p = 0.01) in aged animals.

Fig. 3.

Effect of aging at the Bdnf, cFos, and Dnmt3a promoters in hippocampus of Wistar rats. a Effect of aging on acetylated (H4K8 and H3K9) and methylated (H3K9me2 and H3K4me3) chromatin. b Effect of aging on acetylated (H4K8 and H3K9) and methylated (H3K4me3 and H3K9me2) chromatin. c Effect of aging on acetylated (H3K9) and methylated (H3K4me3 and H3K9me2) chromatin. Columns represent mean ± SD (n = 4–5), and data are presented as percent input. Student’s t test ^*values significantly different from the 5-month-old groups

Exercise Modalities Influenced H4K8ac and H3K4me3 at the Bdnf Promoter in Hippocampus in an Age-Dependent Way

We hypothesized that the protective effects of exercise in the aged animals would be associated with changes in histone modifications at the Bdnf, cFos, and Dnmt3a promoters. Therefore, we analyzed changes in histone modifications in hippocampus after exercise in the YA and aged animals separately. The tested exercise modalities altered epigenetic parameters at the Bdnf promoter in an age-specific manner (Fig. 4a, b). One-way ANOVA showed the effect of exercise (F_(4;20) = 4.43; p = 0.013); resistance (p = 0.029), acrobatic (p = 0.039), and aerobic (p = 0.027) modalities increased H4K8ac in YA rats. Meanwhile, in aged animals, exercise modalities decreased H3K4m e3 ( F_(4;20) = 12.014; p < 0.0001) in aerobic (p = 0.012) and resistance (p = 0.012) groups.

Fig. 4.

Effect of exercise at the Bdnf promoter in hippocampus of Wistar rats. a Effect of exercise on H4K8ac chromatin. b Effect of exercise on H3K4me3 chromatin. Columns represent mean ± SD (n = 4–5), and data are presented as percent input. One-way ANOVA ^#values significantly different from the respective sedentary group

Exercise Modalities Altered H3K9ac or H3K4me3 at the cFos Promoter in Hippocampus in an Age-Dependent Way

One-way ANOVA demonstrated the effect of exercise on H3K9ac (Fig. 5a; F_(4;21) = 5.17; p = 0.008) and H4K4me3 (Fig. 5b; F_(4;21) = 3.92; p = 0.018) at the cFos promoter in a modality-dependent way only in aged rats. H3K9ac was induced in hippocampus by acrobatic (p = 0.04) and by aerobic (p = 0.05) protocols. Concurrently, acrobatic (p = 0.05) and combined (p = 0.03) increased H3K4me3 in this gene region.

Fig. 5.

Effect of exercise at the cFos promoter in hippocampus of Wistar rats. a Effect of exercise on H3k9ac chromatin. b Effect of exercise on H3K4me3 chromatin. Columns represent mean ± SD (n = 4–5) and data are presented as percent input. One-way ANOVA ^#values significantly different from the respective sedentary group

Resistance and Combined Protocols Decreased H4K8ac at the Dnmt3a Promoter in Hippocampus in an Age-Dependent Way

We observed an effect of exercise (F_(4;20) = 4.517; p = 0.014) on H4K8ac at the Dnmt3a promoter in adult animals in a modality-dependent way (Fig. 6). Resistance (p = 0.048) and combined (p = 0.012) groups exhibited lower H4K8ac levels in hippocampus compared to the sedentary group in YA animals. In contrast, there was no effect of exercise on aged animals.

Fig. 6.

Effect of exercise on H4K8ac chromatin at the Dnmt3apromoter in hippocampus of Wistar rats. Columns represent mean ± SD (n = 4–5) and data are presented as percent input. One-way ANOVA ^#values significantly different from the respective sedentary group

Discussion

The central purpose of this study was to compare the efficacy of different exercise modalities starting later in life on functional parameters, such as survival rate and aversive memory performance, as well as on hippocampal epigenetic marks at three important genes for memory formation, Bdnf, cFos, and Dnmt3a [13, 28, 32]. In addition, our experimental design allows us to study the age-induced epigenetic modifications on these genes. This work supports the hypothesis that aging process induces epigenetic modifications of histones at promoters of genes related to learning and memory, which can be attenuated, at least partially, by exercise.

Aging Impacts Aversive Memory Performance and Epigenetic Marks

The present work showed age-related deficits in memory performance, in agreement with previous reports [4, 39, 40]. It is important to describe that the impact of aging on BDNF, cFOS, and DNMT3A expression have been determined in brain areas previously [13, 28, 41]. In order to identify potential molecular mechanisms underlying this impairment, we first examined histone modifications associated with promoters of genes related to learning and memory in the hippocampus: Bdnf, cFos, and Dnmt3a [13, 28, 32].

We observed lower H3K9me2 and increased H4K8ac and H3K4me3 at the Bdnf promoter in hippocampus of aged rats. Taken that H3K9me2 is a repressive mark related to gene silencing [42], while H4K8ac and H3K4me3 are associated to transcriptional activation [43], our results suggest a chromatin landscape aiding Bdnf transcription in healthy aged animals. Previous literature suggests that expression of a Bdnf precursor, proBDNF, was augmented in the hippocampus of 24-month-old Wistar rats [12] and 22–24 months old C57BL/6 mice [44], which is in agreement our findings. Considering that proBDNF processing into mature BDNF occurs at different rates over the lifespan [12, 45], it is possible to suggest that aged hippocampus has higher proBDNF levels. Considering that this immature protein has been proposed to play a role in neuronal apoptosis [46], inhibiting neuronal migration [47], leading to memory impairment [44], and inducing hippocampal long-term depression (LTD), proBDNF may oppose the known effects of its mature counterpart on memory [48].

Regarding the cFos promoter, our results demonstrate reductions in H3K4me3 and H4K8ac, both associated with transcriptional activation, in aged animals [43]. These data corroborate previous research demonstrating regulation of cFos expression by aging [13, 49]. In this context, age-related reduction in Cfos protein levels was found in brain regions such as hippocampus, hypothalamus, striatum, cerebral cortex, and cerebellum of rats, probably owing to reduced neuronal activity during senescence. In addition, decreased cFos expression was observed in hippocampus of middle-aged mice with visuospatial memory impairment [50]. This report may be related to ours and reinforce the evidence that cFos expression has an important role in memory impairment during aging. Since Ono et al. did not show alterations in cFos gene DNA methylation in the whole brain [51], our results support histone modifications as possible epigenetic mechanisms mediating age-related alterations in cFos expression.

Age-related effects in hippocampal Dnmt3a promoter regulation were also observed. DNMT3A is an important enzyme for de novo DNA methylation, which is strongly associated with gene repression [17, 52]. Also, the Dnmt3a gene is necessary for normal memory formation [53–55]. A previous report by Morris et al. demonstrated that Dnmt3a knockout mice have synaptic alterations as well as learning deficits in several associative and episodic memory tasks [55]. Here, we observed an increase in hippocampal H3K4me3 at the Dnmt3a promoter associated to age that can be related to increased Dnmt3a expression and consequently increased DNA methylation. Our finding is in agreement with recent literature questioning the paradigm of global DNA hypomethylation in aging [56], since it has been observed that levels of 5-methylcytosine (5-mC), 5-hydroxymehtylcytosine (5-hmC), and DNMT3A are augmented in brain tissue, especially in hippocampus [41, 57, 58]. Our data may be related to findings obtained by Chouliaras et al., who observed enhancement of Dnmt3a expression in 24-month-old mice [57]. Furthermore, Daniele et al. demonstrated that DNA methylation in peripheral blood was significantly higher in elderly individuals [59]. This work contributed to the understanding of how aging process regulate hippocampal epigenetic histone marks at three important genes for memory formation, Bdnf, cFos, and Dnmt3a.

Exercise Modalities Improved Survival Rate and Aversive Memory Performance

All exercise modalities improved both survival rate and aversive memory performance in aged animals. Understanding how exercise modalities produce cognitive health benefits in an aging population is vital to our ability to producing individualized exercise prescription that respect the needs and characteristics of the older individuals and help to increase adherence to exercise programs.

Previous studies demonstrated that exercise (specifically aerobic protocols) is able to increase lifespan in disease model [20, 60, 61] and healthy rodents [62], even when started late in life. Here, our results add evidences to those findings, since all exercise modalities impacted similarly on survival rate in aged rats. This interesting data indicates that exercise results in a healthier global status, probably owing improvements on—including gene expression involved with—immune responses, metabolism, angiogenesis, and other physiological processes as observed in previous clinical studies [63–67].

A growing body of evidence suggests that aerobic exercise is able to prevent and/or improve aging-induced memory declines [27, 68–70]. Here, all tested modalities, aerobic, resistance, acrobatic, and combination, reversed partially age-related memory declines. Besides, exercise modalities, except resistance exercise, improved aversive memory performance in YA rats. Previously, we showed that daily aerobic exercise, 20 min/day for 2 weeks, improved aversive memory performance in both YA and aged rats [4, 27]. Now, our findings demonstrate a similar effect on memory performance can be obtained with a long-term duration protocol of aerobic exercise, with different frequency of training, e.g., 12 weeks, three times a week, during 20 min.

Interestingly, our resistance protocol was unable to improve aversive memory performance in YA rodents. In contrast, this modality (5 times a week) improved passive avoidance task and spatial memory performance in 3-month-old rats [24, 71], suggesting that YA animals require greater training frequencies to achieve memory improvements. Our findings are in accordance with previous studies evaluating resistance exercise effects in humans and animal models [72, 73].

The American College of Sports and Medicine recommends a combination of exercise modalities for elderly populations to improve general health [15]. However, our combined modality did not show better outcomes than other modalities. Previously, cognitive effects of combined modalities have been evaluated in YA rodents. Zarrinkalam et al. compared endurance, resistance, and combined exercise (resistance and endurance) and showed selective effects of these modalities on spatial and aversive memory in YA rats submitted to a morphine addiction model, which leads to memory impairment [74]. The results here reported showed the combined exercise benefits on aversive memory performance in both tested ages, providing rational basis for the clinical recommendation during life span.

Taken together, our results indicate that all studied exercise modalities are able to impact in general health, decreasing mortality, and also are beneficial to memory during aging process. This data may support individualized prescription of these modalities, according individual needs, goals, and health conditions, which could increase the adherence to exercise in elderly.

Exercise Modalities Affect Histone Marks at the Bdnf, cFos, and Dnmt3a Promoters

In addition to the functional findings, we also investigated how exercise impacted on the hippocampal epigenetic histone marks. An age-dependent effect on the Bdnf promoter was observed, since aerobic and resistance aged groups had decreased hippocampal Bdnf promoter H3K4me3 and also ameliorated memory age-induced impairment, suggesting that these protocols can attenuate aging effects, and could lead to a reduction in gene transcription and consequently reduced levels of proBDNF. Conversely, YA rats submitted to resistance, acrobatic, and aerobic modalities had increased hippocampal Bdnf promoter H4K8ac, an active epigenetic histone mark, which may indicate an increase in Bdnf transcription. Previous reports have described the impact of several exercise modalities on Bdnf expression. Aerobic exercise increased Bdnf expression in YA [28] and adolescent rodents [75]. Klintsova and colleagues [30] also observed increased levels in BDNF expression in motor cortex and cerebellum in YA rats after acrobatic training [30]; however, for the first time, the effects of acrobatic exercise on hippocampal Bdnf epigenetic regulation have been described. In addition, it is possible to suggest that longer training periods are necessary for resistance training to modulate Bdnf expression, since our protocol of 12 weeks increased hippocampal Bdnf promoter H4K8ac, while other during 4 or 8 weeks did not alter hippocampal Bdnf expression in YA rats [24, 31]. Surprisingly, our combined modality did not regulate epigenetically the Bdnf promoter, which can be related to previous reports demonstrating that multicomponent exercise was unable to impact peripheral blood BDNF levels in humans [76]. Taken that resistance modality did not affect cognitive outcomes in YA animals even with increased Bdnf H4K8ac, while combined exercise improved memory performance without any effect on this histone modification, it is possible to infer that cognitive effects of exercise cannot be attributed exclusively to Bdnf H4K8ac modulation.

Regarding cFos promoter, it is important to note that acrobatic, aerobic, and combined exercise improved memory performance and concomitantly affected hippocampal active epigenetic histone marks, specifically H3K9ac and/or H3K4me3, at cFos promoter in aged rats. Aerobic exercise has been pointed out as capable of attenuating cognitive decline and epigenetic repression promoted by age [27]. In this study, this modality induced an age-dependent effect, since we observed increased hippocampal H3K9ac at the cFos promoter only in aged groups, which could be related to augmented gene expression. This result suggests that aerobic exercise alters this mark in adverse situations, such as during aging. In accordance, swimming increased hippocampal H3K9ac and cFos expression in animals submitted to an isoflurane-induced memory impairment paradigm without any effect in animals not exposed to isoflurane [77].

Acrobatic modality induced two epigenetic modifications, H3K9ac and H3K4me3, at the cFos promoter in aged animals, which would suggest transcriptional activation. In accordance, previous studies reported increased cFos expression in motor cortex after acrobatic training in rodents [78]. In addition to the management of the postural instability and improve balance [79] in elderly people, this work brings a new perspective about acrobatic exercise effects on aversive memory in aging and highlights the involvement of hippocampal epigenetic regulation.

Our study is the first report of the impact of combined exercise, including aerobic, resistance, and acrobatic training, in animal models. Our combined protocol had similar effects as acrobatic and aerobic protocols on memory performance, and concomitantly was able to induce epigenetic changes, specifically H3K4me3, in hippocampal at the cFos promoter. The relevance to clinical practice is based on the fact that the spent time with each component is reduced in combined modality, allowing individuals with some limitations to achieve full beneficial response as well.

Evidence shows that DNA methylation is an important mechanism by which exercise affects gene expression. In this study, age- and modality-associated effects were observed in epigenetic marks at the Dnmt3a promoter, given that resistance and combined protocols decreased hippocampal H4K8ac in YA rats, which might decrease its expression. Moreover, exercise modalities did not induce any effect in aged animals. Together, we could exclude the involvement of epigenetic modulation of exercise at the Dnmt3a promoter in hippocampus on functional improvement observed during aging. Although the aerobic protocol used here (12 weeks, 3 times a week) did not modify any studied epigenetic marks (H3K9ac, H3K4me3, H3K9me2, and H4K8ac) at the Dnmt3a promoter, previously our research group and others showed exercise modulation of hippocampal DNMT levels in YA rodents [11, 75, 80]. In this context, Elsner and colleagues (2013) observed modulation of DNMT3B levels after a single session of aerobic exercise in 3 month-old rats, without any effect in aged animals [11]. Moreover, Abel and colleagues (2013) observed decreased Dnmt3a, Dnmt3b, and Dnmt1 expression in adolescent mice exercised with voluntary wheel running [75]. Comparing these data with our results, it is possible to suggest an age-dependent component in this response.

Conclusions

Results of the current study support the hypothesis that exercise training starting in late adulthood impacts age-related cognitive declines, since resistance, acrobatic, aerobic, and combined modalities were able to improve aversive memory performance in aged rats and that epigenetic mechanisms at hippocampal Bdnf, cFos, and Dnmt3a promoters may be involved with both aging process and exercise effects. These insights provide a substantial basis for rational prescription of exercise modalities in the elderly population, not only for metabolic and cardiovascular improvements, but also for cognitive benefits, what can be exploited as a potential therapeutic strategy to reduce memory disorders during aging.