A high-salt/high fat diet alters circadian locomotor activity and glucocorticoid synthesis in mice

Salt is an essential nutrient; however, excessive salt intake is a prominent public health concern worldwide. Various physiological functions are associated with circadian rhythms, and disruption of circadian rhythms is a prominent risk factor for cardiovascular diseases, cancer, and immune disease. Certain nutrients are vital regulators of peripheral circadian clocks. However, the role of a high-fat and high-salt (HFS) diet in the regulation of circadian gene expression is unclear. This study aimed to investigate the effect of an HFS diet on rhythms of locomotor activity, caecum glucocorticoid secretion, and clock gene expression in mice. Mice administered an HFS diet displayed reduced locomotor activity under normal light/dark and constant dark conditions in comparison with those administered a normal diet. The diurnal rhythm of caecum glucocorticoid secretion and the expression levels of glucocorticoid-related genes and clock genes in the adrenal gland were disrupted with an HFS diet. These results suggest that an HFS diet alters locomotor activity, disrupts circadian rhythms of glucocorticoid secretion, and downregulates peripheral adrenal gland circadian clock genes.


Introduction
High salt intake is a prominent lifestyle-related risk factor for hypertension and cardiovascular diseases [1]. A reduction in salt intake at the population level has been considered one of the top five interventions to prevent such non-communicable diseases (NCDs) [2]. To reduce NCDs, the World Health Organization (WHO) intends to reduce salt intake by 30% as one of their nine global targets [3]. adrenal gland were extracted at Zeitgeber time (ZT) 1, 5, 9, 13, 21 (ZT0 was defined as the time of lights on) (n = 18 for each dietary group; n = 3 for each time points).

Assessment of locomotor activity rhythm
Each mouse (N = 12 for each group) was subjected to a wheel-running activity as previously reported [29]. Each mouse was housed in a separate cage (183 × 340 × 148 mm; CL-0135, CLEA Japan, Tokyo Japan) with a running wheel (12 cm diameter, SANKO, Osaka, Japan). The cages were placed in light-impermeable, ventilated boxes, wherein the light intensity at the bottom of the cage was 200-300 lx. The number of wheel revolutions was determined using a magnet-sensor-activated signal between a button magnet on the running wheel and a magnet relay (59070-010, Littelfuse, Inc., Chicago, IL, USA) fixed on a sidewall of the cage, and was fed into a computer every minute. A chronobiology kit (Stanford Software Systems, Naalehu, HI, USA) and ClockLab software (version 2.72, Actimetrics, Wilmette, IL, USA) were used to curate and visualise the activity data. Periods of wheel-running activity for 15 d under the constant dark (DD) condition were determined using chi-square periodograms. Daily activities and durations of the active phase for 7 d in a normal LD cycle and 15 d in DD were quantified using the activity profile function in ClockLab.

Assessment of glucocorticoid secretion and quantitative RT-PCR (reverse transcription polymerase chain reaction) analysis
Glucocorticoid secretion and qualitative RT-PCR analyses were conducted independent of wheel-running recordings (N = 18 for each dietary group; N = 3 for each ZT). Animals were kept in their LD cycles for four weeks and euthanized at Zeitgeber time (ZT) 1, 5, 9, 13, 17, and 21 (ZT0 was defined as the time of lights on) for the 24-h study. Animals were subjected to isoflurane anaesthesia and caecal contents and the adrenal gland were extracted at ZT 1,5,9,13,17, and 21 and frozen in liquid nitrogen. Samples were prepared in accordance with the manufacturer's instructions (https://www. arborassays.com/assets/steroid-solid-extraction-protocol.pdf) [30]. Caecum corticosterone levels were analysed using a corticosterone chemiluminescent immunoassay kit in accordance with the manufacturer's instructions (Arbor Assays, Ann Arbor, MI, USA).
Total RNA was prepared from adrenal grand tissues at ZT 1,5,9,13,17,and 21. Total RNA was extracted from the tissue samples using the RNeasy Mini Kit (Qiagen, Hilden, Germany). cDNA was synthesized from total RNA with the Prime Script RT Reagent Kit (Takara, Shiga, Japan). For the real-time PCR, gene-specific primers (Table 1) and an SYBR Green Real-Time PCR Master Mix (Takara, Shiga, Japan) were used, and the products were detected using PIKO Real (Thermo Fisher Scientific, Waltham, MA, USA). Relative mRNA expression levels were normalised to that of Actb in the same cDNA.

Statistical analysis
Data are represented as mean ± SEM values. Differences in the locomotor activity, degree of gene expression, caecum corticosterone levels and results of fitted cosine wave were evaluated using the ANOVA and Tukey's post hoc test. Circadian rhythms were statistically analyzed using a fitted cosine wave procedure (Acro program by R.Refinetti: http://www.circadian.org/ softwar.html) [31]. A p-value < 0.05 indicated a statistically significant difference. Data analysis was performed using IBM SPSS Statistics Version 25.0 (IBM Corp. Armonk, NY, USA).

Activity rhythms were not advanced; however, their amplitudes decreased with a HFS diet
We examined the effect of the HFS diet on locomotor activity rhythms through wheel-running activity in mice. No differences in locomotor activity patterns were observed among groups

PLOS ONE
during LD and DD conditions, as revealed actogram analysis (Fig 2A), α/ρ ratio at DD (activity time/rest time) (Fig 2C), and period ( Fig 2D). Locomotor activity assessed through wheel running was significantly reduced in HFS mice compared to C or HF fed mice (p < 0.001) (Fig 2A and 2B). This was consistent in both LD and DD conditions. The α/ρ ratio at LD was significantly reduced in HF and HFS fed mice compared with C fed mice.
Average food intake was not different between HF and HFS in LD and DD but it was significantly different between C and HF in LD period (S1A and S1B Fig). In C diet, food intake during LD and DD was significantly different (S1A and S1B Fig) but it was not different in HF and HFS diet.

Rhythms of glucocorticoid levels were disrupted with an HFS diet
Since glucocorticoid levels are important to orchestrate circadian oscillations, we investigated glucocorticoid rhythms in caecum content ( Fig 3A). Caecal glucocorticoid levels displayed nocturnal rhythms, maximum expression at the beginning of the dark period (ZT13) in all groups. However, only the HFS group displayed a bimodal pattern with two peaks per 24-h cycle maximum expression in the beginning of the dark (ZT13) and the end of the dark (ZT21) phase. Low amplitude levels at ZT5 (middle of the light period) (p = 0.023) and ZT17 (middle of the dark period) (p = 0.035) were significantly higher in the HFS group than in the HF group. MESOR (mean statistics of rhythm) of caecal glucocorticoid was significantly lower in C than HF fed mice (p = 0.039) but it was not significantly different between HF and HFS ( Table 2).

Expression rhythm of glucocorticoid synthesis-related genes in the mouse adrenal gland was disrupted by an HFS diet
Since the adrenal gland is a key tissue orchestrating circadian oscillations and synthesising corticosteroids (stress hormones), we focused on the adrenal gland. Expression levels of corticosterone synthesis-related genes ( Fig 3B) were lower in the HFS group than in the other groups ( Fig 3C). StAR, a glucocorticoid rate-limiting enzyme, was significantly downregulated in the HFS group rather than the HF group at ZT1 (beginning of the light period) (p = 0.048). Furthermore, first and rate-limiting steroid biosynthesis enzyme, Cyp11a were downregulated at ZT1 (p = 0.002) (Fig 3C). MESOR of Cyp11a was also significantly lower in HFS fed mice than HF fed mice (p = 0.012) ( Table 2). Only HF group significantly upregulated Cyp11b than control group (Fig 3C). StAR expression levels peaked at 4 h in the HFS group (ZT9 in C and HF compared with ZT5 in HFS diet fed mice) (Fig 3C).

The expression rhythm of clock genes in mouse peripheral tissues was altered with an HFS diet
We performed a 24-h investigation of the expression of clock genes Bmal1, Clock, Per1, Per2, and Cry2 in the adrenal gland. The HFS diet significantly cumulatively downregulated Cry2 in the adrenal grand (p = 0.027) (Fig 4E). Also, cosinor analysis found that MESOR of Cry2 was significantly lower in HFS than HF fed mice (p = 0.027) ( Table 3). Furthermore, Per1 and Per2 were cumulatively downregulated; however, their expression levels did not significantly differ (Fig 4C and 4D). Expression levels of Bmal1 in the end of the dark period (ZT21) and Clock in the beginning of the dark period (ZT13) differed significantly in the adrenal gland (p = 0.045 and p = 0.022, respectively) (Fig 4A and 4B).

Discussion
This study examined the effects of high-fat and high-salt intake on circadian rhythms in mice. Although no differences in the period of circadian rhythm of locomotor activity were observed, the α/ρ ratio at LD was significantly reduced in HF and HFS fed mice compared with C fed mice. Since it is known that the α/ρ is positively correlated with the circadian period [32,33], these results suggest that HF and HFS diet may have an effect on the circadian period. However, it is possible that difference of food intake between C and HF may be confounded factors of this result. Furthermore, our results show that an HFS diet reduced wheel-running activity in both LD and DD conditions. In contrast, Oike et al. revealed that high salt administration did not alter locomotor activity including activity levels [18]. This study used mice of the same strain as previously reported [18], but used an HFS diet of 8% w/w salt and 60% kcal fat in this study relative to their study's 4% w/w and 10% kcal, respectively. These differences may account for the differences observed in those studies.
We investigated a potential mechanism explaining lower activity levels in the HFS group than HF group. Firstly, a disrupted association between the central and peripheral clock may have reduced activity levels. The adrenal gland links the central and peripheral clocks through glucocorticoids [34]. However, some circadian rhythms in the expression of clock genes were retained in adrenalectomized mice, and over half of the genes did not show circadian rhythmicity, suggesting that glucocorticoids are one of the links between the central and peripheral clock [35]. Previous studies have reported that high salt intake (4%) for 4 weeks further promotes circadian expression of clock genes in peripheral tissues such as the liver, kidneys, and lungs [36]. Furthermore, our study shows that circadian clock gene expression levels were altered in the adrenal gland. Indeed, a recent study investigated a hypertensive rat model (spontaneously hypertensive rats; SHR) and revealed an abnormal adrenal circadian clock that potentially affects the transcriptional regulation of clock-controlled genes and steroid hormone secretion by the adrenal gland [20].
The adrenal peripheral clock regulates the autonomous circadian rhythm of glucocorticoid secretion by causing rhythmic steroid production [34]. StAR is a CLOCK-BMAL1-regulated gene and is a rate limiting enzyme for glucocorticoid synthesis [37,38]. Although Bmal1 and Clock expression levels were different between the HFS and HF groups in the dark period, a clear association between circadian clock gene expression and StAR expression in the adrenal gland was not observed. However, in another first and rate limiting enzyme for glucocorticoid synthesis genes Cyp11a [39], significantly lower MESOR was found in HFS than in HF diet fed mice. It is possible that negative feedback of glucocorticoid reduced Cyp11a gene expression level in HFS than HF.
Furthermore, adrenal Cry2 cumulative expression levels decreased in the HFS group. It was confirmed by the result which the MESOR of Cry2 was significantly lower in HFS than HF fed mice. In mice, the loss of Cry1 and/or Cry2 resulted in glucose intolerance and constitutively high levels of circulating corticosterone, suggesting reduced suppression of the hypothalamicpituitary-adrenal (HPA) axis coupled with increased glucocorticoid transactivation in the liver [40]. Cry1 and Cry2 are genetically associated with a glucocorticoid response element (GRE) [40]. Cry2 expression levels peaked in the HFS group at ZT9 and ZT17, which was advanced

PLOS ONE
by 4 h in comparison with caecal corticosterone peaks. These results suggest that disruption of adrenal circadian Cry2 expression potentially leads to differences in circulating levels of corticosterone owing to the GRE. Circulating levels of corticosterone may affect locomotor activity levels. A previous study reported that modulation of circadian glucocorticoid oscillation through an enhancement in its amplitude leads to anxiolytic-like behaviour [41]. The present results show that an HFS diet decreases locomotor activity in both light and dark periods with abnormal corticosterone circulation rhythms. Further studies are needed to clarify whether abnormal circulation of corticosterone decrease locomotor activity under normal conditions without anxiety intervention. In addition, it is well known that excessive salt intake affects the renin-angiotensin-aldosterone system (RAAS) [42,43]. RASS is closely related with HPA and circadian release of glucocorticoids is regulated by HPA, hypothalamic suprachiasmatic nucleus (SCN) signaling, and peripheral adrenal clocks interaction [44]. Previous study shows that hypertensive TGR (mREN-2) 27 (TGR) rats which is a model of upregulated renin-angiotensin system found changes in the clock gene expression in the area postrema [45]. Since RAAS is also located in adrenal gland [46], potential relationship between RAAS and circadian clock are needed to clarify in the future.
Downregulation of clock genes in the HFS group may be associated with reduced locomotor activity. A previous study suggested that circadian gene expression patterns in the mouse central clock predict the levels of locomotor activity [36]. They reported an association between lower circadian gene expression levels and lower locomotor activity [36]. Furthermore, aging is associated with altered locomotor activity rhythms, including decreased amplitudes [47][48][49][50][51]. The HFS diet induced similar effects with aging in terms of decreased amplitudes; however, other changes such as increased fragmentation, shortened or lengthened free-running periods, slower re-entrainment following LD cycle shifts, and altered light

PLOS ONE
sensitivity were not observed. Although we did not examine circadian gene expression patterns in the brain, our data suggest that circadian genes were downregulated in peripheral tissues in the HFS group, wherein locomotor activity levels were lower under both LD and DD conditions. However, in same time, it is possible that the phenotypes of this study in peripheral adrenal clock gene expression may simply be a result of reduced locomotor activity. Further studies are needed to clarify this possible associations. This study has several strengths. Firstly, this study was conducted under experimental conditions that potentially reduce confounders such as environmental and genetic factors. For instance, temperature and humidity or salt sensitivity would be possibly be those factors. Secondly, we focused on novel outcomes such as the association between circadian rhythms, and high-fat and high-salt intake. The present results suggest that an HFS diet, a prominent public health concern worldwide, potentially affects physical activity levels.
However, this study has the following limitations. Firstly, alterations in feeding pattern that is confounded by the diet and also the switch between LD and DD could influence the overall measures described here, including glucocorticoids and gene expression. Secondly, we analysed the circadian rhythm of glucocorticoid secretion using caecal contents. A previous study compared plasma and faecal corticosterone levels and reported that although the phase was advanced in plasma, followed by faecal corticosterone levels, the circadian rhythm of glucocorticoid secretion was similar in rats [52]. Herein, owing to the high sensitivity of corticosterone to various stressors and the need for its assessment every 4 h, we analysed caecum corticosterone levels; however, the similarity in caecal and faecal corticosterone levels remains unclear in mice. Thirdly, although we investigated locomotor activities under LD and DD conditions,

PLOS ONE
we did not investigate central clock gene expression patterns. Fourthly, this study is based on gene expression analysis in a single inbred strain. Especially, gene expression levels of StAR were not well corresponding to the caecum corticosterone levels. It is possible that mRNA levels were not well correlated with protein levels or caecum corticosterone levels were correlated with other factors were unknown. Therefore, the results must be interpreted with caution. Further studies are required to explore the potential mechanisms underlying the association between high-fat and high-salt intake and lower activity levels.

Conclusions
This study shows that an HFS diet is associated with lower locomotor activity and disrupted circadian rhythms of glucocorticoid synthesis in mice. Further studies are required to clarify the mechanisms underlying these associations and to explore their occurrence in humans.