Reduction of acute mild stress corticosterone response and changes in stress‐responsive gene expression in male Balb/c mice after repeated administration of a Rhodiola rosea L. root extract

Abstract Rhodiola rosea L. (R. rosea) is an adaptogenic plant increasing body resistance to stress. Its efficacy has been evidenced mainly in chronic stress models, data concerning its effect in acute stress and underlying mechanisms being scarce. The objective was to investigate the effect of repeated doses of a R. rosea hydroethanolic root extract (HRE) on hypothalamic pituitary adrenal response in a murine model of acute mild stress and also the mechanisms involved. Stress response was measured in Balb/c mice having received by gavage HRE (5 g/kg) or vehicle daily for 2 weeks before being submitted to an acute mild stress protocol (open‐field test then elevated plus maze). Corticosterone was measured in plasma from mandibular vein blood drawn before and 30, 60, and 90 min after initiation of the stress protocol. Mice were sacrificed at 90 min, and the hippocampus, prefrontal cortex, and amygdala were excised for high‐frequency RT‐PCR gene expression analysis. At 30 min after acute mild stress induction, corticosterone level in mice having received the HRE was lower than in control mice and comparable to that in nonstressed mice in the HRE group. HRE administration induced brain structure‐dependent changes in expression of several stress‐responsive genes implicated in neuronal structure, HPA axis activation, and circadian rhythm. In the acute mild stress model used, R. rosea HRE decreased corticosterone level and increased expression of stress‐responsive genes, especially in the hippocampus and prefrontal cortex. These findings suggest that R. rosea HRE could be of value for modulating reactivity to acute mild stress.


| INTRODUC TI ON
Stress is the physiological reaction to environmental threats or pressure and can be self-driven or of external origin (Anghelescu, Edwards, Seifritz, & Kasper, 2018). It is manifested by a wide variety of physical and psychological symptoms. If persistent and left untreated, stress can result in serious health problems including burnout, depression, post-traumatic stress disorder, anxiety, and cardiovascular, gastrointestinal, neurological, and musculoskeletal diseases. Stress appears to be a particular problem in our modern society. Work-related stress is experienced by all sections of society, being estimated to affect 22% of the European workforce (Milczarek & Gonzales, 2009). The World Health Organization has called stress "the health epidemic of the 21st century," recognizing its substantial impact on personal life and also its social and economic consequences (Anghelescu et al., 2018;Subhani et al., 2018).
Stress management strategies include nonpharmacological approaches, such as cognitive behavioral therapy and relaxation, but recourse to pharmacological treatment is standard if stress and its symptoms become harmful. Anxiolytics and antidepressants, associated with known risks of adverse effects and dependency, are generally indicated for more severe situations. Several plants, including chamomile, melissa, and rhodiola, have been shown to be valuable for managing stress and its consequences, with fewer adverse effects and a lower risk of dependency (Sarris, McIntyre, & Camfield, 2013). Rhodiola rosea L. (rosenroot or golden root), manifesting adaptogenic properties, is among those most widely used (Anghelescu et al., 2018;Kasper & Dienel, 2017). Extracts of adaptogenic plants can normalize body functions and reinforce systems compromised by stress (Anghelescu et al., 2018). They have no specific pharmacological properties and act by increasing resistance to a broad spectrum of adverse expressions of stress. Preclinical in vivo and ex vivo studies in animal models and experiments on cell lines have highlighted several biochemical and pharmacological stressreducing properties of R. rosea extracts (Abidov, Crendal, Grachev, Seifulla, & Ziegenfuss, 2003;Olsson, von Scheele, & Panossian, 2009;Panossian, Hambardzumyan, Hovhanissyan, & Wikman, 2007;Panossian, Hovhannisyan, Abrahamyan, Gabrielyan, & Wikman, 2009). In clinical studies, various extracts of R. rosea were found to be effective and safe, improving mental work capacity, concentration, task performance, fatigue, burnout symptoms, and overall mood, besides reducing stress level and self-reported mild anxiety (Cropley, Banks, & Boyle, 2015;Darbinyan et al., 2000;Edwards, Heufelder, & Zimmermann, 2012;Kasper & Dienel, 2017;Panossian, Wikman, Kaur, & Asea, 2009;Punja, Shamseer, Olson, & Vohra, 2014). R. rosea was approved by the European Medicines Agency Stress response typically begins with activation of the hypothalamus-pituitary-adrenal (HPA) axis, one of the main stress response pathways, and the production of corticosteroids (Anghelescu et al., 2018;Subhani et al., 2018). Acute or chronic stress produces characteristic changes in the HPA axis, including an increase in cortisol in humans and corticosterone in rodents, as well as a reduction in the sensitivity of the HPA axis to feedback down-regulation (Anghelescu et al., 2018;. Chronic stress results in persistent elevation of cortisol or corticosterone levels, which may lead to fatigue, depression, and other symptoms (Anghelescu et al., 2018). The reduction in stress-induced damage by R. rosea is characterized by a decrease in or the prevention of hormonal changes characteristic of stress, including cortisol or corticosterone release, as shown in humans suffering from chronic stress following administration of the standardized R. rosea root extract SHR-5 during 28 days (Olsson et al., 2009) and in rabbits subjected to acute stress after 7 days of SHR-5 administration (Panossian et al., 2007). HPA axis modulation by R. rosea extracts also involves the inhibition of stress-induced protein kinases and nitric oxide in animals . The HPA axis is not the only target of R. rosea. For instance, R. rosea extracts stimulated energy metabolism in rodents via the activation of ATP synthesis in mitochondria (Abidov et al., 2003) and might protect against neurodegenerative brain diseases through antioxidative and anti-inflammatory mechanisms (Lee et al., 2013;Zhang, Zhu, Jin, Yan, & Chen, 2006).
Investigations of the molecular mechanisms underlying central corticosteroid action following a stress event led to the identification of genetic pathways and, in particular, stress-responsive genes (Hunter et al., 2016;Kohrt et al., 2016). Modification of target gene transcription, the so-called genomic action of corticosteroids, is therefore most likely one of the main mechanisms underlying corticosteroid action in the brain (Gray, Kogan, Marrocco, & McEwen, 2017). These genomic effects can occur within 15-30 min after the activation of corticosteroid receptors and may last for less than an hour or up to several days, depending on the duration of exposure to the hormone and the type of stress (Dong, Poellinger, Gustafsson, & Okret, 1988;. These stress-responsive genes are divided into several functional classes according to their implication in energy metabolism, signal transduction, neuronal structure, vesicle dynamics, neurotransmitter catabolism or cell adhesion, their encoding of neurotrophic factors and their receptors, and their involvement in the regulation of glucocorticoid signaling (Andrus et al., 2012;Datson, Morsink, Meijer, & de Kloet, 2008;Datson et al., 2012;Hunter et al., 2016). The effects of R. rosea extracts on these stress-responsive genes are unknown. Furthermore, all the data on R. rosea reported so far have been obtained following intense stress, either acute or chronic. Characterizing the effects of R. rosea on the HPA axis and stress-responsive gene transcription under acute mild stress conditions would contribute to a better understanding of how extracts of this adaptogenic plant act to prevent the negative effects of stress.
The purpose of this study was therefore to evaluate, in a murine model of acute mild stress, the effects on the HPA axis of repeated administration of a hydroethanolic root extract (HRE) of R. rosea, phytochemically characterized by high-performance thin-layer chromatography (HPTLC) and ultra-high-performance liquid chromatography coupled with mass spectrometry (UHPLC-MS). Corticosterone secretion and stress-responsive gene expression were determined in the prefrontal cortex (PFC), amygdala, and hippocampus, the main structures implicated in stress management.

| Preparation of the R. rosea HRE
The R. rosea HRE was obtained according to the patented process WO2001056584A1 by crushing frozen fresh roots of R. rosea and leaching with 20%-70% (v/v) ethanol. The extract was then concentrated under reduced pressure to evaporate ethanol. The salidroside titer was adjusted within the range of 0.7-1.4 mg/ml by adding glycerin to the concentrated extract. The batch of HRE used in this study (16H321), containing 83% glycerin, had a salidroside content of 1.02 mg/ml and a dry drug: dry genuine extract ratio of 17:1. This glycerin-containing HRE corresponds to the standardized extract of R. rosea marketed in France under the brand name "Extrait de plante fraîche standardisé (EPS) R. rosea" (PiLeJe Laboratoire, France).

| LC/MS analysis of the R. rosea HRE
UHPLC analysis was performed on an Ultimate 3000 RSLC UHPLC system (Thermo Fisher Scientific Inc., MA, USA) coupled to a quaternary rapid separation pump (Ultimate autosampler) and a rapid separation diode array detector. Compounds were separated on an Uptisphere Strategy C18 column (25 × 4.6 mm; 5 μm; Interchim, Montluçon, France), maintained at 40°C. The mobile phase was a mixture of 0.1% (v/v) formic acid in water (phase A) and 0.1% (v/v) formic acid in acetonitrile (phase B). The gradient of phase A was 100% (0 min), 80% (10 min), 73% (35 min), 0% (40-50 min), and 100% (51-60 min). The flow rate was 0.8 ml/min and the injection volume 10 µl. The UHPLC system was connected to an Orbitrap mass spectrometer (Thermo Fisher Scientific Inc., MA, USA) operating in negative electrospray ionization mode. Source operating conditions were as follows: 3 kV spray voltage for negative mode; 320°C heated capillary temperature; 400°C auxiliary gas temperature; sheath, sweep, and auxiliary gas (nitrogen) flow rate 60, 17.5, and 3.5 arbitrary units, respectively; and collision cell voltage between 20 and 50 eV. Full scan data were obtained at a resolution of 35,000 whereas MS 2 data were obtained at a resolution of 17,500. Data were processed using Xcalibur software (Thermo Fisher Scientific Inc., MA, USA).
The constituents of the R. rosea HRE were identified according to their retention times and mass spectral data and by comparison with authentic standards, if available, or otherwise with published data.

| HPTLC analysis of R. rosea HRE
Standards were diluted in methanol at a concentration of 0.5 mg/ ml for rosavin and 0.1 mg/ml for salidroside (Sigma Aldrich, Saint Louis, USA). One mL of the R. rosea HRE (without added glycerol) was diluted in 3 ml of a mixture of 50% ethanol and water (50/50:

| Animals and experimental design
Seven-week-old male Balb/c mice, a highly stress-sensitive strain (Janvier, Le Genest-Saint-Isle, France), were housed under a normal 12-hr light/dark cycle (07 hr-19 hr) with food (AO4 diet; Safe, Augy, France) and water available ad libitum in a controlled environment (22 ± 1°C, 40% of humidity). The mice were handled daily for 1 week before the start of the experiment to minimize stress reactions to manipulation. During the following 2 weeks, they received each morning a supplement comprising either R. rosea HRE (a 5 g/kg solution containing 80% glycerin, i.e., 4 g/kg; test group, n = 8) or glycerin F I G U R E 1 Experimental protocol in adult Balb/c mice alone (4 g/kg; control group, n = 8) administered by gavage using a V0105040 feeding probe (ECIMED, Boissy-Saint-Léger, France). The two groups received the same amount of glycerin. The volume of supplementation was adapted to the weight of each mouse. At the end of this period, the mice were subjected to an acute mild stress protocol and anxiety-like behavior was evaluated. Blood was drawn from the mandibular vein before initiation of the stress protocol (at t0 min) and then at t30 min and t60 min. Mice were sacrificed at t90 min, and brain structures (hippocampus, hypothalamus, and amygdala) and plasma were excised and frozen at −80°C ( Figure 1).

| Induction of acute mild stress
On the last day of supplement administration, half the mice in each group were subjected to acute mild stress. The stress protocol consisted in subjecting the mice to an open-field (OF) test for 10 min immediately followed by an elevated plus maze (EPM) test for 5 min (see the following sections for details; Figure 1). Experiments were performed in the morning, one hour after gavage, under conditions of dim light and low noise. Both tests induce mild stress in animals by subjecting them to anxiogenic conditions (Treit, Menard, & Royan, 1993).

| Evaluation of anxiety-like behavior
Anxiety-like behavior was evaluated after induction of acute mild stress as previously reported by Dinel et al. (2011). Mouse behavior was videotaped and scored using "Smart" software (Noldus, Wageningen, Netherlands).

| OF test
Mice were exposed to an unfamiliar square (40 × 40 cm) OF from which escape was prevented by surrounding walls (16 cm high). The apparatus was virtually divided into 4 central squares defined as the central area (anxiogenic) and 12 squares along the walls, defined as the periphery. Each mouse was placed in the central area and allowed to freely explore the OF for 10 min. Parameters recorded to evaluate anxiety-like behavior comprised the number of entries into the central area and the percentage of time spent in this area (Dinel et al., 2011).

| Assessment of RNA expression using Fluidigm microfluidic arrays
One microgram of total RNA was obtained from each brain area as described in Dinel et al. (Dinel, Andre, et al., 2014) and was reverse-transcribed with SuperScript III reverse transcriptase control specific amplification for each primer. Then, the raw data of the qPCR were analyzed using GenEx software (MultiD analyses AB, Freising, Germany) in order to choose the best reference gene for normalizing mRNA expression and to measure the relative expression of each of the 93 genes analyzed in the group receiving the HRE and the control group. GAPDH was found to be the best reference gene in this experiment and was therefore used for normalization of gene expression.

| Bivariate statistical analysis
All data were expressed as the mean value ± SEM (standard error of the mean). A p-value of 0.05 was considered as significant. Data were analyzed using a one-way ANOVA (one factor: supplementation) or a two-way ANOVA with supplementation (HRE, control), and stress (stress; no stress) as between factors followed by a Bonferroni post hoc analysis when interaction was significant (GraphPad software, La Jolla, US). Heatmaps were obtained using the Permut Matrix program (Caraux & Pinloche, 2005).

| Principal component analysis (PCA)
PCA was used to assess the gene expression pattern under stress conditions in the group receiving R. rosea HRE and the control group.
The PCA is a dimension reduction technique that clusters data into principal components (PC) maximizing the variance of the data considered. These PCs are uncorrelated linear combinations of the initial variables which can be interpreted as a pattern. PCA generates factor loadings which reflect the correlation of each variable with the PC and attributes a PC score for each individual. We selected the number of components using the Cattell criterion. Statistical analyses were performed using the XLSTAT program (Addinsoft, Paris, France).

| R. rosea HRE did not impact behavior in acute mild stress protocol
As expected, we did not observed any significant effect of the diet

| R. rosea HRE modulated corticosterone secretion consecutive to acute mild stress
Corticosterone was measured in plasma prepared from blood samples drawn before the induction of acute mild stress and 30, 60, and 90 min after the start of the stress protocol. At t0, mice having received R. rosea HRE exhibited a significantly higher plasma corticosterone level (110.8 ng/ml) than mice given the control supplement (glycerin alone, 31.31 ng/ml) (t = 2.789, p < .01; Figure 3a). Interestingly, ND2 and ITPR1 expressions were similarly increased F I G U R E 3 Corticosterone secretion in adult mice having received a R. rosea HRE or glycerin (control) supplement for 2 weeks by daily gavage before the induction of acute mild stress (a) and at t30 (b), t60 (c), and t90 min (d) after initiation of the stress protocol. Glycerin versus HRE: *p < .05, **p < .01; glycerin stress versus HRE stress: $$, p < .01. HRE, hydroethanolic root extract TA B L E 1 Stress-responsive genes studied by high-frequency RT-qPCR in the prefrontal cortex, hippocampus, and amygdala  Phylogenetic analysis based on Pearson's correlation was performed for the three brain structures studied ( Figure 5). The heatmap generated demonstrated that gene regulation depends on the group considered (HRE-supplemented or control), especially as regards the PFC. However, we did not observe any real gene clusters.

| D ISCUSS I ON
The objective of this study was to evaluate the effect on the HPA axis of chronic administration of a R. rosea HRE in a murine acute mild stress model by measuring corticosterone secretion and assessing cerebral expression of stress-responsive genes.

| R. rosea HRE decreased stress-induced corticosterone secretion
In the acute mild stress model used in this study, Balb/c mice were consecutively subjected to an OF and an EPM test. We chose to use Balb/c mice as studies have shown this strain to be highly stress-sensitive compared with other strains (Moloney, Dinan, & Cryan, 2015).

Both tests used in this study induce stress in animals by placing them in anxiogenic environments: an open place in the OF test and open
arms in the EPM test (Treit et al., 1993).
The basal level of corticosterone was higher in mice receiving R.
rosea HRE than in control mice receiving a supplement containing glycerin alone. This difference might be explained by the organoleptic characteristics and higher viscosity of the HRE compared with glycerin alone, which could have created additional stress during administration of these supplements (Hoggatt, Hoggatt, Honerlaw, & Pelus, 2010). Even if the percentage of increase was important, the level of corticosterone in mice having received the R. Rosea HRE was far below levels obtained after a stress, even in low reactive mice (Mattos et al., 2013). Moreover, we did not observe any behavioral difference in anxiety-like tests between glycerin-and R. Rosea HREtreated mice.
Thirty minutes after acute mild stress induction, control mice presented, as expected, an increase in corticosterone secretion, whereas mice receiving R. rosea HRE did not. At t60 and t90, the percentage corticosterone increase was comparable between stress-free and stressed mice. We hypothesize that the effect of experimentally induced acute mild stress was masked by that of gavage. We nevertheless observed that at both times, mice having received R. rosea HRE presented a lower percentage increase in corticosterone as compared to the control group. This result implies that administration of R. rosea HRE resulted in better regulation of stress homeostasis, characterized by more effective control of corticosterone increase that probably led to more efficient restoration of corticosterone level to the basal value.
At the intracellular level, high corticosteroid levels impact the balance between trophic and atrophic factors within neurons (Liu et al., 2017). For instance, glucocorticoids have been shown to inhibit cell proliferation in the dentate gyrus by reducing the proliferation of granule cell precursors (Gould & Tanapat, 1999;Saaltink & Vreugdenhil, 2014). Moreover, chronic stress results in persistent inhibition of granule cell production and changes in the structure of the dentate gyrus, raising the possibility that stress alters hippocampal function through this mechanism (Gould & Tanapat, 1999).
By preventing the substantial increase in corticosterone level, R.
rosea extracts could prevent this negative impact of corticosteroids.
Our results confirm those of previous studies demonstrating the impact of R. rosea extracts on inhibition of the HPA axis, as illustrated notably by the serum level of corticosteroids in rats (Cifani  Xia, Li, Wang, Wang, & Wang, 2016). The antistress properties of R. rosea extracts have been attributed to their interference with both the HPA axis and the sympathoadrenal system (Panossian, Hovhannisyan, et al., 2009;Panossian & Wagner, 2005;Panossian, Wikman, & Wagner, 1999). However, all these results were obtained in animals subjected to intense acute or chronic stress. In this study, we demonstrated for the first time that a specific R. rosea extract affects HPA axis reactivity even under conditions of mild stress of short duration. The dampening of corticosterone secretion could be due to a decrease in stress reactivity amplitude or to better control of the glucocorticoid pathway.

| R. rosea HRE upregulated the expression of functional stress-responsive genes
One of the main mechanisms of action of corticosteroids in the brain is their genomic effect, resulting in modification of target gene transcription. Corticosteroid-mediated transcriptional changes within the brain have been studied by means of large-scale gene expression profiling (Datson et al., 2008(Datson et al., , 2012Hunter et al., 2016;Kohrt et al., 2016). The resulting gene expression profile showed a highly dynamic transcriptional response to glucocorticoid receptor activation throughout a specific time window, shifting from exclusively downregulation of genes 1 hr after glucocorticoid receptor activation to both up-and down-regulation after 3 hr (Morsink, Steenbergen, et al., 2006). We investigated the impact of R. rosea HRE, 1h30 after the induction of acute mild stress, on the expression of stress-responsive genes (Datson et al., 2008(Datson et al., , 2012Hunter et al., 2016) in the PFC and amygdala, structures involved in the regulation of stress, as well as in the hippocampus, a medial temporal lobe structure implicated in the formation of stable memories and highly susceptible to stress (Kim & Diamond, 2002).
Interestingly, most genes modulated in the PFC, amygdala, and hippocampus by R. rosea HRE belong to four main functional groups of genes implicated in the functioning of neuronal structures, glucocorticoid signaling, circadian rhythm, and mood regulation, respectively.
Genes affecting the actin cytoskeleton were modulated by the HRE in all three brain structures studied, but acute mild stress affected their expression only in the hippocampus. The actin cytoskeleton is involved in the morphology of dendritic spines, and changes in actin cytoskeletal configurations have been postulated to influence long-term potentiation, affecting synaptic transmission (Meng et al., 2002;Smart & Halpain, 2000). Under stress, these mechanisms are dysregulated and the connectivity between the various brain structures is impaired (Christoffel, Golden, & Russo, 2011). Several studies have demonstrated that stress induces adverse changes in the morphology and strength of hippocampal excitatory synapses, inducing a generalized atrophy of dendrites and spines in the PFC (Goldwater et al., 2009;Sandi et al., 2003;Stewart et al., 2005;Wellman, 2001 Finally, R. rosea HRE modulated the expression of SIRT2, a gene implicated in mood regulation. Adverse changes in SIRT2 expression have been reported in mood disorders, with a decrease in SIRT2 expression consecutive to a chronic stress. Treatment with the antidepressant fluoxetine reversed the stress-induced changes in SIRT2 (Liu et al., 2015). By upregulating SIRT2 expression in the hippocampus and PFC,

R. rosea HRE could act like an antidepressant. Previous research has
demonstrated that salidroside, one of the active substances of R. rosea HRE, prevented the development of depression-like behavior as effectively as fluoxetine (Zhu et al., 2015). The antidepressant effect of R.
rosea extracts might be mediated by their impact on SIRT2 expression.
Other genes were regulated by R. rosea HRE but their modulation depended more strongly on the brain structure considered. In the amygdala, the R. rosea HRE and acute mild stress interaction damped the expression of ND2, a mitochondrial membrane respiratory chain gene, suggesting an essential role of mitochondrial activity as an adaptive response to stress, as previously proposed (Vishnyakova et al., 2016). In the PFC, BHLHB2, a gene implicated in neurotrophic factor activity and neuronal excitability, was upregulated by R. rosea The new data presented here nevertheless suggest that R. rosea HRE could be of value in modulating reactivity to acute mild stress. section.

E TH I C A L S TATEM ENTS
Animal husbandry and experimental procedures were in accordance with the EU Directive 2010/63/EU for animal experiments and were approved by the national ethical committee for the care and use of animals (approval ID A13169).