Impact of exercise on brain-bone marrow interactions in chronic stress: potential mechanisms preventing stress-induced hypertension

Abstract

We examined the effect of chronic restraint stress and the counteractive effects of daily exercise on the molecular basis of the brain-bone marrow (BM) interactions, by especially focusing on the paraventricular nucleus (PVN) of the hypothalamus. Male Wistar rats were assigned into control, restraint stress, and stress + daily spontaneous exercise (SE) groups. BM and hypothalamic gene expression profiles were examined through the undertaking of RT-PCR and microarrays, respectively. The inflammatory blood cell population was investigated through flow cytometry. Through the use of immunohistochemistry, we examined the presence of BM-derived C-C chemokine receptor type 2 (CCR2)-expressing microglial cells in the rat PVN. The gene expression levels of BM inflammatory factors such as those of interleukin 1 beta and CCR2, and the inflammatory blood cell population were found to be significantly higher in both restrained groups compared with control group. Interestingly, chronic restraint stress alone activated the recruitment of BM-derived CCR2-expressing microglial cells into the PVN, whereas daily spontaneous exercise prevented it. A notable finding was that restraint stress upregulated relative gene expression of hypothalamic matrix metalloproteinase 3 (MMP3), which increases the permeability of the blood-brain barrier (BBB), and that exercise managed to normalize it. Moreover, relative expression of some hypothalamic genes directly involved in the facilitation of cell migration was downregulated by daily exercise. Our findings suggest that daily spontaneous exercise can reduce the numbers of BM-derived CCR2-expressing microglial cells into the PVN through the prevention of stress-induced changes in the hypothalamic gene expression.

NEW & NOTEWORTHY Chronic restraint stress can upregulate MMP3 gene expression in the rat hypothalamus, whereas daily spontaneous exercise can prevent this stress-induced effect. Stress-induced BM-derived inflammatory cell recruitment into the rat PVN can be prevented by daily spontaneous exercise. Stress-induced increase of hypothalamic MMP3 gene expression may be responsible for BBB injury, thereby allowing for BM-derived inflammatory cells to be recruited and to accumulate in the rat PVN, and to be subsequently involved in the onset of stress-induced hypertension.

INTRODUCTION

Cardiovascular reactions, including increased blood pressure (BP) and heart rate (HR), can be caused by acute psychological or emotional stress (1). Prolonged exposure to psychosocial stress promotes the development of stress-induced hypertension through the enhancement of the sympathetic nerve activity (2) and, if left unchecked, can lead to stroke and myocardial infarction (3, 4). However, the mechanism of sympathoexcitation caused by chronic stress remains unknown. The paraventricular nucleus (PVN) of the hypothalamus, a major cardioregulatory autonomic and neuroendocrine region, has been shown to play a critical role in autonomic control of BP (5, 6). Moreover, the inflammatory conditions in the cardiovascular brain center have been identified as critical factors for the induction of hypertension (7–9).

Neuroinflammation is characterized by the infiltration of inflammatory cells (ICs) such as T lymphocyte populations and others into the central nervous system (CNS). Bone marrow is a site for the initiation of T-cell activation responses (10). T-cells, which are involved in the regulation of both innate and acquired immunity, differentiate and mature in the thymus and divide into multiple subpopulations. T-helper (Th)1 and Th2 cells release cytokines and activate other ICs. On the other hand, Treg lymphocytes suppress the proinflammatory actions of the effector Th lymphocytes. Cluster of differentiation (CD)4⁺ Th17 cells have been reported to play a vital role in the pathogenesis of neuroinflammation. Th17 cells increase the CNS infiltration of other ICs such as neutrophils across the blood-brain barrier (BBB), which may trigger inflammatory reactions that occasionally lead to irreversible neuronal damage (11). The recruitment of bone marrow (BM)-derived ICs to brain can be triggered by social stress that contributes to the development of anxiety (12). Herisson et al. (13) have also found a direct local interaction between the brain and the skull BM through the meninges; such an interaction has been shown to enable the myeloid cell migration under aseptic meningitis conditions. In addition, a reported increase in peripheral BM-derived ICs and their extravasation into the brain regions such as the PVN may also contribute to hypertension in spontaneously hypertensive rats (14). In contrast, circulating angiogenic progenitor cells (APCs) from BM play a particularly important role in endothelial repair in the setting of arterial injury following inflammatory and other prohypertensive stimuli (15, 16). The pathophysiology of hypertension may be improved by APC’s ability to repair endothelial damage.

It is currently believed that daily aerobic exercise can prevent hypertension in prehypertensive subjects, as well as reduce the BP in borderline hypertensive subjects and stage 1 hypertensive subjects suffering from essential hypertension (17). In addition, spontaneous momentum (wheel rotation counts) has been shown to promote stress resistance in rodents, characterized by a reduced cardiovascular response to stressor exposure (18, 19). According to the American Heart Association guidelines, exercise is a recommended therapy for lowering BP (20). However, the precise mechanisms of antihypertensive effects of exercise have not been elucidated.

In the current study, we investigated the relationship between chronic restraint stress and brain-BM interaction in rats, and whether daily exercise may modify this interaction. Understanding the interactions between the brain and the BM will further our understanding of mechanisms that lead to the onset of stress-induced hypertension, and how these may be alleviated by exercise.

MATERIALS AND METHODS

Animals

Sixty male Wistar rats (5-wk-old; weighing 104–138 g) were obtained from Japan SLC (Shizuoka, Japan). The animals were housed in a temperature-controlled room with fixed 12-h:12-h dark:light cycle (0600–1800 and 1800–0600). Animals were provided with food and water ad libitum. All experiments were approved by the Ethics Committee for Animal Experiments of Juntendo University (Approval Number: S14), and complied with the guidelines of the Physiological Society of Japan.

Chronic Restraint Stress and Exercise Paradigm

The rats were randomly divided and acclimatized for 1 wk in their regular individual cages in the case of the control (Co) and the stress (St) groups, and in cages equipped with running wheels in the case of the stress + exercise (SE) group, before the restraint stress intervention. The experimental restraint stress model used in this study was as follows (21): rats belonging to the St and SE groups were temporarily restrained in triangular-shaped disposable plastic bags (DecapiCone; Braintree Scientific Inc.) for 1–1.5 h/day, for 5 days/wk, for three consecutive weeks. Body weights (BW) were recorded for individual rats daily in the restraint stress groups (St and the SE groups), and once per week in the Co group. The mean weekly spontaneous momentum (wheel rotation counts) of SE rats was assessed for a week prior to the restraint stress intervention (week −1) and 3 wk during the intervention period (week +1, week +2, week +3, respectively).

Blood Pressure Measurements

Noninvasive BP measurements were performed using a tail-cuff method (THC-31, Softron, Japan) in each rat of experimental groups on the seventh day before and on the first day after the 3-wk intervention.

Fluorescence-Activated Cell Sorting Analysis by Direct Flow Cytometry

Isolation of peripheral blood mononuclear cells (MNCs) was performed as described previously (14). In brief, rats were anesthetized to respiratory arrest with an excess of respiratory isoflurane and then blood was collected via cardiac puncture. MNCs were isolated from diluted blood in MNC isolation buffer at a 1:1 ratio. The diluted blood was gently added to Ficoll-Paque PLUS (Cytiva Life Sciences) at a 2:1 blood-to-Ficoll ratio in a 15-mL conical tube. The tube was then centrifuged at 1,200 rpm, for 15 min, at room temperature (RT). The buffy coat was collected and transferred to a new 15-mL conical tube. Phosphate-buffered saline (PBS) was added to the tube to a volume up to 15 mL, and the tube was centrifuged at 1,200 rpm, for 15 min, at RT, so as to form a pellet. The cells at the bottom of the tube were then washed with 15 mL of sterile PBS before another round of centrifugation took place at 1,200 rpm, for 15 min, at RT. The pellet was resuspended in 400 μL of PBS, then supplemented with 4.5 mL of ammonium chloride (Cat. No. 07850; STEM CELL Technologies) to provoke the lysis of the remaining red blood cells. After incubating for 15 min at RT, the mixture was washed by the addition of MNC isolation buffer to a volume up to 30 mL, and it was centrifuged to form a pellet. This washing process was performed twice and resulted in the formation of a white pellet. The latter was resuspended in 1 mL of MNC isolation buffer, and the MNC samples were then kept at 4°C, overnight, before being used for the undertaking of flow cytometry the next day.

The staining procedure followed the method described by Jun et al. (22). In brief, to profile the levels of the BM and the blood ICs and APCs, the BM MNCs were prepared in a concentration of 0.5–1 × 10⁶ cells/100 μL in a buffer containing PBS, 2% FBS, and 1 mM of EDTA. CD3⁺/CD45⁺, CD4⁺/CD8⁺, and CD4⁺/CD8⁺/CD25⁺ cells were considered to represent T-cells (23–25), whereas CD68⁺ cells were considered as representatives of macrophages (26). Finally, CD90⁺/CD4⁻/CD5⁻/CD8⁻ cells were considered as representatives of APCs with angiogenic and endothelial-reparative properties in rats (26). The following commercially available antibodies were used in this study—RPE-Alexa Fluor 647-conjugated CD8 (Bio-Rad); PE-conjugated CD3, CD4, and CD5 (BD PharMingen); PE/Cy7-conjugated CD25 and CD68 (Novus Biological); PE/Cy7-conjugated CD90 (Biolegend); Pe-Cy5-conjugated CD45 (BD PharMingen); Alexa Fluor 488-conjugated interleukin (IL) 17A (Thermo Fisher Scientific). Individual antibodies were used as controls. Antibodies were added to cell suspension samples and were incubated in the dark for 1 h, at 4°C. After centrifugation and washing, the cells were examined on an Attus NxT Acoustic Focusing Cytometer (Thermo Fisher Scientific) at the Institute of Health and Sports Science & Medicine of Juntendo University. The data were analyzed with the use of the Attune NxT software (v.2.6).

Microarray

Microarray assay was used as before (27) to screen the transcriptomes in male Wistar rat hypothalami to identify differentially expressed (DE) genes in St and SE rats as compared with Co rats and to identify potential candidates that may be related to neuroinflammation and BP regulation. The whole hypothalamus, including many cardioregulatory regions such as the supraoptic nucleus, the PVN, the dorsomedial hypothalamus, and the tuberomammillary nucleus, was chosen for this study. The rat hypothalami were collected and the total RNA of the brain tissue was extracted as described later. In total, 18 total RNA samples (6 rats/group) were analyzed through individual microarrays. Microarray services were provided by Takara Bio (Tokyo, Japan), and a Rat Sureprint G3 Rat GE 8 × 60 K v2 (Agilent) platform was used. Briefly, 100 ng of total RNA from each sample were labeled with Cyanine-3 (Cy3; Low input Quick Amp Labeling Kit; Agilent), and 0.6 µg of Cy3 cRNA were hybridized to the microarrays. Microarrays were then scanned with an Agilent DNA microarray scanner and the data were processed by using the Feature Extraction software. The fold changes (FC) of the gene expression between the three groups (FC_St/Co and FC_SE/Co) were calculated as the ratio of normalized processed signals in St and SE to those in Co; the FC_St/SE (the ratio of normalized processed signals in St to those in SE) was also calculated.

The FC values were calculated through a Log₂-transformation; Log₂(FC_St/Co, FC_SE/Co, and FC_St/SE). We set cut-off levels ( ≥+0.58 or ≤−0.58) to select DE genes (genes that present a downregulation or an upregulation equal to or greater than 1.5-fold) between two of the three groups. The clustered heat map, also known as the “double dendrogram,” was established by uploading the DE gene Log₂ (FC) values into the MultiExperiment Viewer software version 4.9.0.

Functional Annotation Clustering Analysis

DAVID Bioinformatics Resources version 6.8 (DAVID: Database for Annotation, Visualization and Integrated Discovery) was used to identify gene ontology (GO) categories of enriched DE genes in the hypothalamus (Supplemental Fig. S1: https://doi.org/10.6084/m9.figshare.21716627).

Transcriptional Network Analysis

Pathway Studio software version 12.4.0.5 was used to analyze direct interactions and expression regulation between DE genes between all three groups for the retrieval of common targets such as disease or cellular processes of selected genes of interest. The confidence levels of reliability of each interaction were set based on the minimum level of the three citations (27).

RNA Extraction and Reverse Transcription Polymerase Chain Reaction

The detail was described in the Supplemental Materials and Methods (https://doi.org/10.6084/m9.figshare.21716627).

Immunohistochemistry

At end point, rats were anesthetized to respiratory arrest with an excess of respiratory isoflurane and transcardially perfused with heparinized saline followed by 4% paraformaldehyde (PFA). Brains were removed and postfixed with 4% PFA for 48 h, and cryoprotected in 30% sucrose. Series of 50-μm coronal sections that included the PVN were sectioned and incubated them with anti-Iba1 antibody (1:500; Cat. No. 019-19741; Wako) for identification of microglia, and with anti-C-C chemokine receptor type 2 (CCR2) antibody (1:200; Cat. No. NBP2-35334; Novus Biologicals) for identification of BM-derived ICs. Secondary antibodies used were anti-rabbit IgG (H + L) (1:500; Cat. No. A21206; Invitrogen), and anti-mouse IgG (H7L) (1:500; Cat. No. AB150108; Abcam) conjugated with Alexa Fluor 488 and AF 594, respectively. Immunohistochemistry and microglia quantification were performed as before (27). In brief, three coronal sections of each rat brain, corresponding to bregma −1.72 mm, −1.8 mm, and −1.92 mm, were used for this purpose. Images were obtained using EVOS FL Auto 2 microscope. Number of CCR2-expressing microglial cells were brind-counted by a single trained technician using Fiji-ImageJ in ×10 confocal microscopy field.

Statistical Analysis

Results were expressed as means ± SEM for each group. Comparisons between three groups were evaluated by one-way or two-way ANOVA followed by Fisher’s least significant difference (LSD) or Tamhane tests (SPSS v.22.0; SPSS Inc., Chicago, IL). Statistical significance was set at P < 0.05, and denoted in each figure.

RESULTS

Establishment of the Chronic Restraint Stress-Induced Hypertension Model

We first confirmed the establishment of chronic stress-induced hypertension in our model (Fig. 1, A and B). There were no differences of preintervention levels in BW, BP, and HR among the three groups. One day after the 3-wk intervention period, the rat BW (213 ± 15.47 g) was lower (P < 0.001), whereas the BP parameters such as the systolic blood pressure (SBP; 142 ± 2.76 mmHg), the mean arterial pressure (MAP; 116 ± 1.95 mmHg), and the diastolic blood pressure (DBP; 103 ± 1.78 mmHg) were higher (P < 0.001) in the St group than those of the Co group (BW: 252 ± 7.47 g; SBP: 130 ± 2.22 mmHg; MAP: 103 ± 1.36 mmHg; DBP: 89 ± 1.43 mmHg). These same parameters (BW: 199 ± 11.89 g; SBP: 134 ± 3.05 mmHg, MAP: 106 ± 2.26 mmHg, and DBP: 92 ± 2.12 mmHg) were significantly reduced (P < 0.05) in SE when compared with St rats. Two-way ANOVA for repeated measurement revealed a significant relationship between the time factor and the changes in BW (P < 0.001) for all groups during the 4-wk trial. HR (Fig. 1C) was significantly lower in the St group when compared with the Co group after the 3-wk intervention period (P < 0.05). During the first week of the chronic restraint stress intervention, the spontaneous momentum (i.e., wheel rotation counts) was significantly reduced (751 ± 62.8 rounds/day; P < 0.05) when compared with the week −1 of the treatment (901 ± 49.7 rounds/day; Fig. 1D). However, it was significantly higher in the second week (1,224.6 ± 103.8 rounds/day) and in the third week of treatment (2,366.3 ± 245.3 round/day) when compared with the week −1 (P < 0.05 and P < 0.01, respectively) and the week +1 (P < 0.01 and P < 0.01, respectively).

Figure 1.

Physiological parameters under stress and exercise conditions. A: effect of stress and exercise on the body weight of rats during the 4-wk experimental period (n = 14 in each group). B: systolic blood pressure (SBP), diastolic blood pressure (DBP), and mean arterial pressure (MAP) after a 3-wk restraint stress intervention (n = 14 in each group). The SBP, DBP, and MAP levels in the stress (St) group rats were significantly higher than those in the control (Co) and the stress + exercise (SE) group rats. C: heart rate after the stress intervention (n = 14 in each group). Values are presented as means ± standard error of the mean. D: total weekly wheel-running activity for rats of the SE group (n = 14). Rat activity was found to be slightly reduced during the first week of the restraint stress, and then it significantly increased over the experimental period. Notes: *P < 0.05; ***P < 0.001 when compared vs. the Co group; †P < 0.05; ††P < 0.01; †††P < 0.001 when compared vs. the St group; ‡P < 0.05; ‡‡P < 0.01 when compared vs. week −1; §§P < 0.01 when compared vs. week +1; #P < 0.05 when compared vs. week +2; all analyses were performed by using a one-way or two-way ANOVA followed by Fisher’s least significant difference (LSD) or Tamhane tests.

Identification of BM-Derived IC Population Changes

Hypertension is associated with elevated pro-ICs and downregulation of APCs (22, 28). Therefore, we investigated the circulating levels of these cells in our rat groups. We determined the levels of circulating CD4⁻/CD5⁻/CD8⁻/CD90⁺ cells (representing APCs), CD4⁺/CD8⁺, CD4⁺/CD8⁺/CD25⁺, and CD3⁺/CD45⁺ cells (representing subpopulations of T-cells), as well as CD68⁺ cells (representing macrophages/monocytes) (22, 28) in all rats at end point. We detected a 32% decrease in circulating CD4⁻/CD5⁻/CD8⁻/CD90⁺ cells, a 35% increase in circulating CD4⁺/CD8⁺/CD25⁺ cells, and a 19% rise in CD68⁺ cells in the St group when compared with the Co (Fig. 2B). However, daily spontaneous exercise did not improve the levels of these cells in circulation of St group (Fig. 2B). The CD4⁺/CD8⁺ and CD3⁺/CD45⁺ cell populations exhibited no significant differences between the St and Co groups, but they were significantly higher in the SE group compared with the Co group (P < 0.05 and P < 0.05, respectively).

Figure 2.

Chronic restraint stress attenuates peripheral inflammation in the employed model of stress-induced hypertension. A: Subpopulations of CD4⁺/CD8⁺ cells (in red frame) in the peripheral blood were analyzed by flow cytometry. B: Specific inflammatory cell populations increased, including those of the T-cells (CD4⁺/CD8⁺, CD4⁺/CD8⁺/CD25⁺, and CD3⁺/CD45⁺ cells) in the blood of the stress + exercise (SE) group rats, and macrophages/monocytes (CD68⁺) in the blood of the stress (St) and the SE group rats. The angiogenic progenitor cell (APC) population (CD4⁻/CD5⁻/CD8⁻/CD90⁺) was found reduced under both of the examined stress conditions. C: the T-helper 17 cell (CD4⁺/CD17A⁺) population was found increased by both the restraint stress (St) and the stress with exercise (SE). Values are presented as means ± standard error of the mean. Notes: control (Co) group: n = 6; St and SE groups: n = 8; *P < 0.05; **P < 0.01; ***P < 0.001 when compared vs. the Co group; †P < 0.05 when compared vs. the St group; all analyses were performed by using a one-way ANOVA followed by Fisher’s least significant difference (LSD) or Tamhane tests.

In addition, circulating CD4⁺/IL17A⁺ cells were elevated by 182% in the St group, and by 276% in the SE group when compared with the Co group (P < 0.01 and P < 0.001, respectively; Fig. 2C).

Identification of DE Genes in the Hypothalamus by Microarray

We identified a total of 11,149 probes that were differentially expressed in the hypothalamus (P < 0.05), of those 578 probes were expressed with a change that was more than 1.5-fold between two of the three examined groups. Subsequently, we identified 49 DE genes between the St and the Co groups, 551 DE genes between the SE and the Co groups, and 153 DE genes between the St and the SE groups (Supplemental Fig. S1 and Table S1).

Functional Annotation Clustering Analysis

DAVID analysis revealed that 150 GO terms were found to be specifically enriched in the hypothalamus (Supplemental Table S2), and they were mainly related to the development and function of the nervous and cardiovascular systems, regulation of the transcription from the RNA polymerase II promoter, regulation of cell migration, inflammatory responses, and regulation of synaptic transmission. The transcript for the neurotrophic receptor tyrosine kinase 1 (NTRK1), a receptor of tyrosine kinase involved in the development and maturation of the CNS and peripheral nervous systems through the regulation of proliferation, differentiation, and survival signaling pathways, was found to be involved in 50 different GO terms between two of the three examined groups (Supplemental Table S2). The Ntrk1 transcript levels were significantly upregulated in the St and the SE rats when compared with the Co group (Supplemental Table S1). Interestingly, the close interaction between Ntrk1 and nerve growth factor receptor (Ngfr), a member of receptor for tumor necrosis factor (Tnf) family of cytokines, was associated with many cellular processes including cell migration, apoptosis, and neuronal plasticity via the neurotrophin signaling pathway (Supplemental Table S2). In addition, this interaction was linked to the regulation of glutamatergic synaptic transmission, cardiac rhythm, development of cardiovascular system, and detection of external and abiotic stimuli (Supplemental Table S2). We selected Ntrk1 as a candidate gene of interest for transcriptional network analysis due to its differential expression in the hypothalamus, its iteration in the GO terms of various clusters of other DE genes in the St group rat hypothalami, as well as its participation in the regulation of various cellular processes (Supplemental Table S2).

Transcriptional Network Analysis

Pathway Studio was used to explore the molecular interactions and relationships associated with conditions such as hypertension and inflammation, and relevant cellular processes such as cell migration among DE genes in the hypothalami of the St and the SE groups.

Under stress condition, the hypothalamic gene expression profiles of NGFR, NTRK1, matrix metalloproteinase 3 (MMP3), and choline O-acetyltransferase (CHAT) were upregulated (Fig. 3). Enhanced relative expression of these genes was linked with inflammation (Fig. 3A) and cell migration (Fig. 3B) in the brain. In contrast, daily spontaneous exercise downregulated hypothalamic gene expression profiles of most genes investigated in the current study (Fig. 4, Supplemental Table S3). This decrease in relative expression of hypothalamic genes resulted in reduced association with hypertension, inflammation (Fig. 4A), and cell migration (Fig. 4B, Supplemental Table S3).

Figure 3.

Transcriptional network in the hypothalamus comparing the control (Co) and the stress (St) groups. A: the expression of some genes related to a transcriptional network for inflammation was found to be upregulated in the brains of St group rats when compared with that in the brains of the Co group rats. These upregulated genes are predicted to induce neuroinflammation and hypertension. B: the expression of some genes that facilitate cell migration were found to be upregulated in the brains of St group rats when compared with that in the brains of the Co group rats. Different color shades of proteins reflect differences in the magnitude of changes in gene expression. Red indicates that the related gene expression levels were found to be upregulated, while blue indicates that the related gene expression levels were found to be downregulated. Refer to gene abbreviations in Supplemental Table S1 for protein abbreviations in this figure.

Figure 4.

Transcriptional network in the hypothalamus comparing the control (Co) and the stress + exercise (SE) groups. A: the expression of some genes related to a transcriptional network for inflammation was found to be downregulated in the brains of SE group rats when compared with that in the brains of the Co group rats. B: the expression of some genes that facilitate cell migration were found to be downregulated in the brains of SE group rats when compared with that in the brains of the Co group rats. Different color shades of proteins reflect differences in the magnitude of changes in gene expression. Red indicates that the related gene expression levels were found to be upregulated, while blue indicates that the related gene expression levels were found to be downregulated. Refer to gene abbreviations in Supplemental Table S1 for protein abbreviations in this figure.

Stability of Relative Expression Levels of Housekeeping Genes

The expression of 11 rat housekeeping genes was compared between the hypothalami of the rats belonging to the three groups examined. The geNorm method was first used to identify the most stable reference genes in the hypothalamic samples. Supplemental Table S4 presents the average levels of the expression stability measurement (M) value of the candidates. All the M values (0.24–0.47) were well below the default limit of M = 1.5, thereby indicating that all genes were within optimal stability range in the hypothalamic tissues of the three groups examined (29). Beta-2 microglobulin (B2m) and succinate dehydrogenase complex flavoprotein subunit A (Sdha), followed by lactate dehydrogenase A (Ldha), were identified as the most stable genes among the 11 genes examined. By using NormFinder, we found that the most stable reference genes are identical to those previously calculated (Supplemental Table S4A). Based on the results of geNorm and NormFinder, we eliminated the most variable reference gene (ribosomal protein lateral stalk subunit P1: Rplp1). The remaining ten reference genes were also analyzed by using BestKeeper. In conclusion, among the five most stable genes, Sdha and B2m were chosen based on stability of their expression between all hypothalamic samples. Using the same analyses, phosphoglycerate kinase 1 (Pgk1) and actin beta (Actb) were identified as the most stable from the 11 tested genes in BM samples (Supplemental Table S4B).

Increased Proinflammatory Markers in the BM of St Rats

The proinflammatory profiles of BM cells were compared among the three experimental groups by measuring relative expression levels of selected genes. We observed increased relative gene expression levels of IL1 beta (IL1B; FC: 1.48 ± 0.13, P = 0.011 in the St group; FC: 1.95 ± 0.22, P = 0.004 in the SE group), and CCR2 (FC: 1.30 ± 0.08, P = 0.02 in the St group; FC: 1.45 ± 0.08, P = 0.002 in the SE group) in the BM-derived mononuclear cells of both stress groups when compared with those in the Co group (Fig. 5A). The gene expression of interferon gamma (IFNG) was also significantly increased in the St group (FC: 1.75 ± 0.15; P = 0.02) and in the SE group (FC: 2.88 ± 0.45; P = 0.008) compared with the Co group. In addition, relative gene expression of Toll-like receptor 4 (TLR4) was significantly higher in the SE group (FC: 1.50 ± 0.11; P = 0.002), but not in the St group (FC: 1.30 ± 0.10; P = 0.066) when compared with the Co group. Conversely, relative gene expression of nitric oxide synthase 2 (NOS2) was significantly higher in the St group (FC: 1.51 ± 0.15; P = 0.048), but not in the SE group (FC: 1.34 ± 0.10; P = 0.09) when compared with the Co group. These data suggest that individual gene expression profiles of the BM cells may be different between the St and the SE group rats, although they share common proinflammatory characteristics.

Figure 5.

Bone marrow and hypothalamic gene expression profiles. A: bone marrow gene expression profiles (n = 14 in each group). Restraint stress increased the expression of proinflammatory factor genes in the bone marrow such as Ccr2, Il1b, Ifng, and Nos2. Of those, only the abnormal expression of Nos2 was rescued by exercise. B: expression profiles of hypothalamic genes related to cell migration that were identified by the undertaken microarray analysis (n = 6 in each group). C: validation of the microarray data of hypothalamic gene expression profiles by reverse transcription and polymerase chain reaction (RT-qPCR) (n = 6 in each group). B and C: the left five genes listed were found to be increased due to stress (St) as shown in Fig. 3, whereas the right five gene listed were found to be decreased due to stress combined with exercise (SE) as shown in Fig. 4. Values are presented as means ± standard error of the mean. Notes: *P < 0.05; **P < 0.01; ***P < 0.001 when compared vs. the control (Co) group; †P < 0.05; ††P < 0.01 when compared vs. the St group; all analyses were performed by using a one-way ANOVA followed by Fisher’s least significant difference (LSD) or Tamhane tests.

Validation of Hypothalamic Transcriptomes Using Reverse Transcription Polymerase Chain Reaction

Based on the findings of the DAVID and Pathway Studio analyses, relative gene expression of Ntrk1, Ngfr, LIM homeobox 8 (Lhx8), Chat, and Mmp3 transcripts in the hypothalamus—all of which facilitate cell migration (Fig. 3B) and are upregulated in St rats—was validated using reverse transcription polymerase chain reaction (RT-PCR) with housekeeping genes. As observed by microarray analyses (Fig. 5B), the transcript expression of Ngfr, Ntrk1, Lhx8, and Chat in the St group was upregulated when examined by RT-PCR (FC: 1.75 ± 0.25, P = 0.0895; 1.58 ± 0.13, P = 0.022; 1.47 ± 0.07, P = 0.013; and 1.61 ± 0.12, P = 0.007, respectively; n = 6 in each group; Fig. 5C). The microarray data showing upregulation of Ngfr, Ntrk1, Lhx8, and Chat transcripts in the SE rats compared with the Co rats was also confirmed by RT-qPCR (FC: 1.74 ± 0.30, P = 0.175; 1.53 ± 0.20, P = 0.033; 1.43 ± 0.16, P = 0.021; and 1.73 ± 0.23, P = 0.078, respectively; Fig. 5, B and C). As in the microarray analyses, the expression of Mmp3 in the hypothalamus was increased when transcripts were analyzed by RT-PCR (FC: 1.47 ± 0.12, P = 0.002; and 1.59 ± 0.17, P = 0.0098, respectively) in the St group rats; and exercise normalized the relative expression levels of this gene (FC: 1.07 ± 0.07, P = 0.651, and 1.17 ± 0.12, P = 0.459, respectively; Fig. 5, B and C). In addition, we validated five downregulated genes identified by microarray in the SE group; these genes facilitate cell migration and include nerve growth factor (Ngf), aquaporin-1 (Aqp1), high mobility group box 1 protein (Hmgb1), CX3C chemokine receptor 1 (Cx3cr1), and Fas ligand (Faslg) (Fig. 4B). The RT-PCR also confirmed downregulation of relative expression of Hmgb1 in St group (FC: 0.84 ± 0.06, P = 0.013) and of Ngf, Hmgb1, Cx3cr1, and Faslg in SE group (FC: 0.76 ± 0.10, P = 0.035; 0.80 ± 0.03, P = 0.004; 0.86 ± 0.07, P = 0.043; and 0.81 ± 0.05, P = 0.014, respectively) when compared with the Co group (Fig. 5, B and C).

CCR2 Expression in Microglial Cells

Representative images of CCR2 immunohistochemistry in the hypothalami of all examined rat groups is shown in Fig. 6. We performed double-staining for CCR2 and Iba1 to quantify CCR2-expressing microglial cells in the PVN. Colocalization between CCR2 and Iba1 was identified in PVN sections from Co, St, and SE groups (Fig. 6A). We observed an elevated number of CCR2-expressing microglial cells in St group compared with Co group (17.43 ± 3.75; n = 7 vs. 2.88 ± 1.14; n = 8, respectively, P < 0.05). In addition, the number of CCR2-expressing microglia cells in SE group (2.83 ± 0.87; n = 6) was similar to that observed in the Co group (Fig. 6B).

Figure 6.

Stress increases the C-C chemokine receptor type 2 (CCR2) expression in rat paraventricular nucleus (PVN) microglial cells, whereas exercise inhibits this expression. A: representative photomicrographs showing the CCR2 expression in microglial cells identified in the PVN of rats submitted to each of the herein examined treatment conditions. Immunohistochemistry was performed on rat brain coronal sections (bregma: −1.8 mm), while scale bars represent 200 or 50 μm. B: CCR2-expressing activated microglial cells identified in optical fields of stained PVN tissues of rats belonging to each of the examined rat groups. Values are presented as means ± SE of determinations performed in the examined treatment groups (control, Co group: n = 8; stress, St group: n = 7; stress + exercise, SE group: n = 6). Symbols used: ***P < 0.001 when compared vs. the Co group; †††P < 0.001 when compared vs. the St group, all analyses were performed by using a one-way ANOVA followed by Tamhane test.

DISCUSSION

The restraint model used in this study resulted into a reduction of the rat BW gain (Fig. 1A). This finding is consistent with those of previous reports (30, 31), thereby indicating that animals were exposed to both physical and mental stress. Notably, with the exception of the first day of the treatment condition, the combination of restraint stress and exercise did not seem to reduce the BW gain any further than what the restraint stress alone did, thereby suggesting that stress and exercise do not have a cumulative effect on rat BW gain. In addition, restraint stress significantly increased BP (SBP, MAP, and DBP), and the combination of stress and exercise significantly reduced the BP compared with restraint stress alone (Fig. 1C). The fact that the HR of rats in the St and the SE groups was lower than or equivalent to that of the Co group may be the basis for us to eliminate the “white coat syndrome” when measuring the rat BP through the tail-cuff method. The baroreflex triggered by a high level of BP in the stress groups may normalize the HR levels in these same groups. Taken together, the aforementioned findings demonstrate the successful establishment of an animal model of hypertension through the induction of chronic restraint stress.

The proinflammatory status of the BM is characterized by increased inflammatory factors (e.g., cytokines and chemokines) and plays a crucial function in the evolution of hypertension (14, 32). Altered inflammatory responses have been associated with impaired autonomic input to the BM in spontaneously hypertensive rats (28). Mediators and regulators such as IL1, TNF alpha (TNFA), IL6, interferon alpha/beta, and IFNG act on endothelial cells and leukocytes to stimulate early innate immune responses. Moreover, chemokines and their receptors control the migration of leukocytes during inflammation (33). CCR2, a G protein-coupled receptor, plays a main role in monocyte migration pathways from the BM into the peripheral blood (34, 35). CCR2 has both proinflammatory (mediated by antigen-presenting cells and T cells) and anti-inflammatory (mediated by regulatory T cells) effects (36). Our study revealed that the gene expression of proinflammatory cytokines (such as IL1B, and IFNG) and of a factor of immune defense against pathogens (i.e., the inducible nitric oxide synthase; iNOS or NOS2) in the BM was increased as a result of the exposure to restraint stress. In addition, we also found an increased expression of the BM Ccr2 in stress animals. Taken together, these results suggest that restraint stress induces BM inflammation through changes in the BM gene expression profiles. Therefore, restraint stress is predicted to increase the proliferation of BM-derived ICs and the recruitment of these cells into the blood circulation and the rat PVN.

It has been assumed that hematopoietic stem cells (HSCs) release cytokines, whereas immune or other cells release several growth factors within the BM microenvironment (37, 38). Cytokine signals such as IL1 drive the HSC proliferation and differentiate their myeloid development (39–41). In our study, gene expression level of IL1B in BM increased as a result of the exposure to stress. Furthermore, the level of BM-derived ICs in the peripheral blood of restrained rats exhibited an increase in the subpopulations of macrophages/monocytes compared with those of the unrestrained rats. In addition, the percentage of APCs was significantly lower in the restrained rats when compared with that of unrestrained rats. These findings are in agreement with those of previous studies (14, 22, 28, 42). Moreover, Th17 cells were found to be increased by restraint stress. All these changes in the circulating levels of T cells, macrophages/monocytes, and APCs suggest that the stress-induced BM inflammation stimulated an emergency lymphopoiesis and myelopoiesis (43), and enhanced the migration of these cells into the blood and the thymus.

Stress has been demonstrated to activate microglial cells (12, 44). The activated microglial cells have immune functions similar to those of peripheral macrophages under the production of proinflammatory cytokines, chemokines, and prostaglandins (45), and can therefore induce neuroinflammatory signals (46–50). BM-derived peripheral ICs and their extravasation into the brain are also involved in neuroinflammation, and can contribute to the development of neurogenic hypertension (14). Recent studies have shown that restraint stress can damage the BBB in the frontal cortex and hippocampus (51), the amygdala (52), and the hypothalamus (53) of adult rats. Chronic psychological stress can induce the infiltration of BM-derived microglia into the PVN (54). Increased BBB permeability may be considered as one of the mechanisms contributing to autonomic dysfunction in hypertension (55). The concentration of MMP3—an enzyme involved in the degradation of the extracellular matrix (ECM)—is usually low, but can increase in various inflammatory conditions (56). Moreover, MMP3 can induce neuronal death and BBB dysfunction, and can also mediate lipopolysaccharide-induced inflammatory reactions by acting as an inflammatory neurotoxic protein (57). MMP3 can increase the BBB permeability by upregulating the extracellular signal related-kinase (ERK) signaling pathway (58). In this study, we found that the hypothalamic Mmp3 expression was upregulated by stress. Based on the immunohistochemical data of this study, we have suggested that the CCR2-expressing BM-derived ICs were recruited into PVN through a damaged BBB with a MMP3-dependent increased permeability, and were subsequently transformed into activated microglia. However, future studies should clarify which BM-derived ICs were recruited into the rat PVN under these conditions.

The Ntrk1 (encoding TrkA) and Ngfr (encoding p75NTR) expressions were found to be upregulated by stress (59, 60). The Ntrk1 and Ngfr products have been shown to promote cell survival, proliferation, differentiation, inflammation, long-term potentiation of neuronal plasticity, and apoptosis via the neurotrophin and the ERK (mitogen-activated protein kinase, MAPK) signaling pathways. The ERK (MAPK) signaling pathway is also an inflammatory transcriptional pathway. In the TrkA-p75NTR interactions in NGF-induced neuronal responses, p75NTR enhances the TrkA signaling and NGF responses if the TrkA levels are high. In contrast, p75NTR can induce neuronal apoptosis upon high activation if the TrkA activity levels are low (61–63). The activation of p75NTR can also lead to a cell migration increase and an enhancement of the prosurvival effects of TrkA (64), whereas the Ntrk1 knockdown is known to inhibit cell proliferation and migration (65). In the present study, we suggested that the stress-induced upregulation of Ntrk1 and Ngfr expressions in the hypothalamus may play a role in the interactions between the cell migration into the PVN parenchyma and neuroinflammation (Fig. 3A). However, as the expression levels of Ngfr and Ntrk1 remain high in exercised animals, future studies will need to elucidate whether these processes are involved in the observed BP increase in the restrained rats.

It is well known that CCR2 mediates the migration of mature monocytes/macrophages not only from the BM into the blood (34, 35), but also from the blood into tissues in response to inflammation (35). Inflammation was mediated by CCR2 and its ligand (CCL2) through the mobilization of monocytes and macrophages to inflammatory sites (66). The chemokine receptor CCR2 is considered as particularly important for the transport of myeloid cells through inflamed vessels, and the subsequent accumulation in the brain (67). CCR2 is expressed mainly in monocytes, T cells, and dendritic and endothelial cells, whereas the CCR2-expressing monocytes have been suggested to be involved in inflammation (68). Due to its important role in vasculitis and the remodeling occurring in various forms of hypertension, the increased CCR2 expression in monocytes becomes a critical predictor of the presence of hypertension (69). The blocking of CCR2 in experimental rodent models of hypertension decreased the BP and reduced the associated vascular and renal inflammation (70–72). In this study, the bone marrow Ccr2 expression was found to be upregulated in stressed rats, and may have been involved in the recruitment of monocytes/macrophages from the BM into the blood. It may have also caused inflammation of the blood vessels, exaggerated the MMP3-induced BBB damage, and enhanced the BM-derived microglial cell accumulation in the PVN. We have suggested that the increase of the number of activated bone marrow-derived CCR2-expressing microglia in the PVN of stressed rats may play a role in the onset of stress-induced hypertension. We hope to further investigate this matter in further studies.

Physical exercise has been shown to improve physical performance and cardiorespiratory fitness. It has been suggested that exercise can stimulate the immune system (73) through an increase of the hematopoietic progenitor cell quantity (74, 75) or a maintenance of the common lymphoid progenitor populations so as to differentiate them into lymphocytes (cells of the immune system) (76). In our study, lymphocytes (Treg cells) were found to be significantly increased in the exercise-coupled stress rats when compared with those of rats belonging to the Co and the St groups. Physical exercise has also a number of anti-inflammatory effects and can lead to a reduction of the damage and the vascular permeability of BBB, thereby increasing the integrity of the latter (77). A 2-wk period of aerobic training can effectively restore the BBB integrity within the PVN in spontaneously hypertensive rats (78, 79). Temporal changes, such as reductions in the angiotensin II level of the renin-angiotensin system, oxidative stress, proinflammatory mediators, and microglial activation in the PVN may be involved in the exercise-induced improvement of the BBB function that can restore neuronal activity in the PVN (5). According to the findings of the microarray analysis of the hypothalamic gene expression profiles, daily spontaneous exercise can alter the expression levels of many genes involved in the regulation of BBB permeability (Mmp3) or related to inflammation (such as Ngf, Ntrk1, Chat, Mmp3, Aqp1, Hmgb1, Cx3cr1, Faslg, etc.) (Figs. 3 and 4; Supplemental Tables S1 and S3). Many of these genes (i.e., Ngf, Aqp1, Hmgb1, Cx3cr1, and Faslg) are directly involved in the facilitation of cell migration (Fig. 4 and Fig. 5, B and C; Supplemental Table S3). It should be noted that most of them are downregulated by daily spontaneous exercise, suggesting its inhibitory effects on cell migration. Moreover, although we failed to see the protective effects of exercise on the hypothalamic gene expression levels of Ngfr, Ntrk1, and Chat (that were found to be upregulated by stress alone), we did find that daily spontaneous exercise can demote the expression levels of Mmp3 that can remodel the ECM’s basal lamina that forms part of the BBB (58). Long-term physical activity affects the CNS through the reduction of peripheral inflammation-induced CNS infiltration of ICs or by protecting the BBB through the constitution of its tight junctions (77). MMP3 opens the BBB by attacking the basal lamina and the tight junction proteins, thereby facilitating the neutrophil influx (80). We believe that exercise may be involved in the repair of stress-induced BBB damage through the decrease of the Mmp3 expression, thereby preventing the migration of BM-derived CCR2-expressing cells into the PVN. In addition, we found no differences in the BM gene expression profiles and the peripheral cell populations between the St and SE groups. This suggests that daily spontaneous exercise may not impact on the effects of stress on BM and the recruitment of BM-derived ICs into the peripheral blood. Taken together, daily spontaneous exercise training might prevent the stress-induced hypertension through the altering of hypothalamic gene expression, thereby preventing the induction of BBB injury, inhibiting cell recruitment, and reducing inflammation.

Conclusions and Perspectives

Our findings provide new insights into the effects of restraint stress and of exercise on the regulation of cardiovascular homeostasis through brain-bone marrow interactions. A stress-induced increase of the hypothalamic Mmp3 expression might increase the permeability of the BBB to BM-derived CCR2-expressing microglial cells, which in turn may play a role in the onset of stress-induced hypertension. On the other hand, daily exercise may be able to prevent these pathophysiological events through a modification of the Mmp3, Ngf, Aqp1, Hmgb1, Cx3cr1, and Faslg expression in the rat hypothalamus. Further experiments are required to determine the role of CCR2-expressing microglial cells in the BP regulation performed at the PVN. Moreover, whether other cardiovascular-related brain areas than the PVN, such as amygdala and nucleus of the solitary tract, are involved in brain-bone marrow interactions in stress-induced hypertension need to be elucidated. Taken together, our data are supportive of the hypothesis presented in Fig. 7.

Figure 7.

A hypothetical model demonstrating how exercise could prevent the stress-induced effects on blood pressure by blocking the recruitment of bone marrow (BM)-derived inflammatory cells through the blood-brain barrier (BBB) into blood pressure regulating centers such as the paraventricular nucleus (PVN) of the hypothalamus. CCR2, C-C chemokine receptor type 2; IL1B, interleukin 1 beta; MMP3, matrix metalloproteinase 3.

DATA AVAILABILITY

The data sets generated and/or analyzed during the current study are available in the following locations: RNA-microarray data set: NCBI GEO GSE220747; https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE220747.

SUPPLEMENTAL DATA

Supplemental Materials and Methods, Supplemental Tables S1–S4, and Supplemental Fig. S1: https://doi.org/10.6084/m9.figshare.21716627.

GRANTS

The study was financially supported by Japan Society for the Promotion of Science (JSPS) KAKENHI under Grant Numbers JP19KK0251 and JP19K11449. J. Zubcevic is supported by NIH Grant HL152162.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

T.V.N., K.Y., S.S.S.G., and H.W. conceived and designed research; T.V.N., K.Y., K.T., S.S.S.G., and H.W. performed experiments; T.V.N., K.T., S.S.S.G., and H.W. analyzed data; T.V.N., K.Y., K.T., J.Z., and H.W. interpreted results of experiments; T.V.N. prepared figures; T.V.N. drafted manuscript; T.V.N., J.Z., and H.W. edited and revised manuscript; T.V.N., K.Y., K.T., J.Z., S.S.S.G., and H.W. approved final version of manuscript.