Decompression Mechanism of Radish Seed in Prehypertension Rats through Integration of Transcriptomics and Metabolomics Methods

Radish seed (RS), the dried ripe seed of Raphanus sativus L., is widely used in traditional Chinese medicine (TCM) to reduce blood pressure. However, the molecular and pharmacological mechanisms underlying its therapeutic effects are still unclear. In this study, we analyzed the effects of RS in a rat model of prehypertension and assessed the mechanistic basis by integrating transcriptomics and metabolomics. RS administration significantly reduced blood pressure in prehypertensive male Wistar rats, negatively regulated endothelin-1, increased nitric oxide levels, and reduced the exfoliation of endothelium cells. In vitro vascular ring experiments further confirmed the effects of RS on vascular endothelial cells. Furthermore, we identified 65 differentially expressed genes (DEGs; Padj < 0.05 and fold change (FC) > 2) and 52 metabolites (VIP > 1, P < 0.05 and FC ≥ 2 or ≤0.5) in the RS intervention group using RNA-seq and UPLC-MS/MS, respectively. A network of the DEGs and the metabolites was constructed,q which indicated that RS regulates purine metabolism, linoleic acid metabolism, arachidonic acid metabolism, circadian rhythm, and phosphatidylinositol signaling pathway, and its target genes are Pik3c2a, Hspa8, Dnaja1, Arntl, Ugt1a1, Dbp, Rasd1, and Aldh1a3. Thus, the antihypertensive effects of RS can be attributed to its ability to improve vascular endothelial dysfunction by targeting multiple genes and pathways. Our findings provide new insights into the pathological mechanisms underlying prehypertension, along with novel targets for the prevention and treatment of hypertension.


Introduction
Prehypertension has been defned by the 2020 Global Hypertension Practice Guidelines as blood pressure (BP) in the range of 130-139/85-89 mmHg [1]. Prehypertensive patients have a greater risk of developing myocardial infarction, incident stroke, and cardiovascular disease (CVD) due to pathological changes in the blood vessels and humoral nervous system [2]. Te renal renin-angiotensin system, sympathetic tone [3], immune cell infltration [4], and endothelial dysfunction (ED) play an important role in the pathogenesis of hypertension [5]. ED is characterized by elevated endothelin (ET)-1 and decreased NO bioavailability, which impairs vascular reactivity [6]. ET-1mediated vasoconstriction is increased in prehypertensive adults, and its activity may increase with age [7]. Furthermore, prehypertension is associated with impaired NOmediatedendothelium-dependent vasodilation [8], and NO supplementation can lower BP in prehypertensive individuals [9]. Since prehypertension can progress to hypertension and increase the risk of cardiovascular events and death [10], pharmacological intervention is recommended in prehypertensive individuals [11].
Radish seed (RS), the dried ripe seed of Raphanus sativus L., has been used as food and medicine for thousands of years in China. It is prescribed for the treatment of peptic diseases, respiratory diseases, hypertension, and cardiac diseases [12,13]. We identifed its main active compounds, sinapine thiocyanate and glucoraphanin [14,15], which had a good efect on the regulation of blood pressure [16,17]. RS extract mediates antihypertensive and anti-infammatory efects [18]. Te curative efect of RS against prehypertension and the pharmacological mechanisms are not completely clear, which limits its clinical applications.
Transcriptomics and metabolomics are increasingly being used to study the mechanisms of various drugs [19][20][21][22]. Transcriptomics refers to the complete set of RNA transcripts, including messenger, ribosome, and noncoding RNAs, that are produced by cells or organisms under specifc conditions and refects the gene expression profle associated with physiological and pathological conditions. In recent years, it has been used to identify markers of hypertensionrelated diseases [23,24]. Metabolomics is the quantitative analysis of metabolites within cells, tissues, and organisms and is used to identify metabolic changes associated with diferent physiological and pathological states, as well as drug interventions [25,26]. Tis approach has been used to elucidate the pharmacological mechanisms of traditional Chinese medicine (TCM) formulations [17,27]. While transcriptomics provides genotypic information, metabolomics represents the actual phenotype of the organism since metabolites are the fnal results of the transcriptional program. Terefore, the integration of transcriptomics and metabolomics can elucidate the complex molecular mechanisms and regulatory networks underlying disease progression and drug action, while obviating the unreliability of single omics sequencing. In this study, we integrated metabolomics and transcriptomics to determine the relationship between diferentially expressed genes (DEGs) and diferential metabolites (METs) associated with RS intervention in prehypertension in order to identify potential pharmacological mechanisms.

Chemicals and Drugs.
Mass spectrometry-grade methanol, acetonitrile, and formic acid were acquired from Termo Fisher Scientifc Inc.
(Loughborough, United Kingdom). RS was purchased from Jinan Jianlian TCM Pharmacy (Batch No. 210415, Shandong, China) and authenticated and conformed to the standards of the Pharmacopoeia of the People's Republic of China. To prepare the RS extract, 3 kg of RS was reconstituted with 60 liters of water and extracted twice. Te extraction yield was 13.4% (g/g). Details of the three major compounds in the RS extract are shown in Table S1 and Figure S1 in the Supplementary Material. Te extract was dried in an oven and dissolved in distilled water to a fnal concentration of 1 g/ml. Valsartan (Lot X2882, Beijing Novartis Pharma Ltd., China) was also dissolved in distilled water to a fnal concentration of 1.44 mg/ml. N′-nitro-L-arginine (L-NNA) (Lot BCCD9665) was obtained from Sigma Aldrich (St. Louis, MO, United States), and N G -Nitro-L-arginine Methyl Este (L-NAME) (Lot N1101A) was purchased from Meilunbio (Dalian, Liaoning, China).

Animals.
Fifty healthy male Wistar rats (body weight: 150-180 g) were purchased from Vital River Laboratory Animal Technology Co., Ltd., Beijing, China (Animal Qualifcation Certifcate No. SCXK (Jing) 2016-0006). All animal experiments were approved by the Animal Care and Ethics Committee of Shandong University of Traditional Chinese Medicine (No. YYLW2021000015). Te animals were housed in a standard laboratory environment (temperature 22 ± 2°C, humidity 55 ± 5%, and 12-h light/dark cycle) and provided water and food ad libitum. Prehypertension was modelled by injecting the rats daily intraperitoneally with L-NNA (7.625 mg/kg) twice a week for three weeks [18], whereas the control group (C) received the same volume of physiological saline. Te BP was measured after injections to confrm prehypertension (systolic BP is approximately 160 mmHg). Te successfully modelled animals were randomly divided into untreated prehypertension (PHT), RS-treated (R), and valsartan-treated (V) groups, and the respective saline doses of 2.5 g/kg/day for RS and 14.4 mg/kg/day for valsartan were given via the intragastric route for six weeks.

BP Measurements.
Arterial systolic and diastolic BP (SBP and DBP; mmHg) of the rats were monitored once a week at the same time each day during the modelling and drug intervention periods using the BP-2000 Blood Pressure Analysis System (Visitech Systems, Inc., North Carolina, USA) with the noninvasive tail-cuf method. Briefy, the rats were allowed to move inside a holder for 10 min to ensure good blood fow to the tails, and fve preliminary measurements were taken in each session. Te room was kept free of noise or disturbance to increase the reliability and reproducibility of BP measurements. All rats were examined six times in parallel.

In Vitro Vascular Ring
Model. Rats were anesthetized with pentobarbital (50 mg/kg, i.p.) and placed supine on the operating surface.
Te thoracic aorta was quickly dissected and placed in an ice-cold Krebs solution that had been prepared 20 min in advance under 95% O 2 and 5% CO 2 . Te surrounding fat and connective tissue were carefully removed from the aortas under a dissecting microscope while avoiding traction and injury to blood vessels, and 3 mm long vascular rings with intact endothelium were mounted in an Multiwire Myograph System (DMT 620M, Myo Technology A/S, Danish), bathed in 5 ml Krebs solution at 37°C, and blistered continuously with mixed gas. Te vessels were then stretched in a stepwise manner to their optimal resting tension of 20 mN and then balanced for 1 h before the experiment. Te thoracic aortic segment was excited twice with a 5 ml, 60 mM potassium (K + ) solution to test the functional integrity of the vascular smooth muscle. 1 μM phenylephrine (PE) was added to preconstrict the vascular ring, followed by 10 μM acetylcholine (ACh). Te vascular endothelial function was considered intact if blood vessels relaxed to more than 80%. Te chambers were then washed until the vascular ring tension returned to the baseline level. In addition, the aortas rings were incubated with L-NAME for 20 min and then treated with PE and RS. Te tension change was recorded using the PowerLab 8 signal analysis system. Te efect of the diferent interventions on the vascular ring was calculated in terms of the maximum diastolic rate, that is, the percentage of contraction after and before administration.

Sample Acquisition and Histological
Assay. Te animals were deeply anesthetized with pentobarbital, and blood was collected by abdominal aorta puncture and centrifuged at 3500 rpm for 15 min to obtain serum. Te thoracic aorta was immediately dissected, and one part was immersed in 2.5% glutaraldehyde (Lot: CR2101167, Cusabio, China) and one in 4% paraformaldehyde (Lot: CR2101179, Cusabio, China) for scanning electron microscopy (SEM) and hematoxylin and eosin (HE) staining, respectively. In addition, a small portion was snap frozen for molecular analyses.
2.6. Measurement of Endothelin-1 (ET-1) and NO. Te levels of ET-1 and NO in the rat sera were measured using a specifc Rat Endothelin-1 Enzyme-Linked Immunosorbent Assay Kit (Lot: G14012112, Wuhan Cusabio Co., Ltd., China) and Nitric Oxide Assay Kit (Lot: 20210604, Nanjing Jiancheng Bioengineering Institute, China) as per the manufacturer's instructions.

RNA Sequencing.
Total RNA was extracted from the frozen aorta using TRIzol reagent (Lot: 252612, Termo Fisher Scientifc Inc., USA) according to the manufacturer's instructions. Samples were subjected to agarose gel electrophoresis to detect any contamination or degradation, and the purity and concentration of RNA were determined using the NanoPhotometer ® N120 spectrophotometer (IMPLEN Corp., Munich, Germany). Te integrity and quantity of RNA were assessed by the Agilent Bioanalyzer 2100 System, and three biological replicates per group were sequenced by Novogene Co., Ltd. (Beijing, China). Te DEGs between groups were screened using FC > 2 and P adj < 0.05 as thresholds and functionally annotated by GO analysis. containing 0.05% formic acid. Te sequence was as follows: 0-3 min, 0-2% B; 3-9 min, 2-40% B; 9-18 min, 40-98% B; and 18-21 min, 98% B. Te sample loading amount was 2 μl, the column temperature was set at 45°C, and the fow rate was set at 0.30 ml/min. MS analysis was performed in both positive and negative ionization modes equipped with a heated electrospray ionization source and Xcalibur 3.0 software. Optimal analysis conditions of the MS were set as follows: sheath gas 45 arb and auxiliary gas 10 arb; capillary voltage in positive and negative ion modes at 3500 V and 3000 V; capillary temperature 350°C; mass range 80-1000 m/ z with a resolution of 70000; and S-lens RF level 55.

Data Processing.
To ensure data stability and reproducibility, 10 μl of each serum sample was evenly mixed to make quality control (QC) samples. Te acquired mass spectrometry data (.raw) were frst converted to the mzXML format, and the peak identifcation, peak alignment, retention time, and peak area were extracted using R software. Peaks with missing values greater than 50% were fltered. HMDB (https://www.hmdb.ca/), METLIN (https://www. metlin.scripps.edu/), and KEGG (https://www.genome.jp/ kegg/) databases and Xcalibur 3.0 software were used to identify metabolites. Te absolute value of the mass error was less than 10 ppm. Principal component analysis (PCA) and orthogonal partial least squares discriminant analysis (OPLS-DA) were used to fnd METs. Te MetaboAnalyst 5.0 (https://www.metaboanalyst.ca/) online platform was used for metabolic pathway analysis.

Integrated Transcriptomics and Metabolomics. Te
MetScape module of Cytoscape software, based on the KEGG database, was used to build the network of DEGs and potential metabolites. Te target genes corresponding to the METs were searched in the HMDB database. Te proteinprotein interaction (PPI) network of DEGs and the gene targets of potential METs was constructed using the STRING database with a confdence score of 0.7.  Table 1.
2.11. Statistical Analysis. All data were analyzed using SPSS 22.0 software (SPSS Inc., Chicago, IL, USA) and visualized using GraphPad Prism 8.0 software (GraphPad Software, San Diego, CA, United States). Multiple groups were compared using one-way analysis of variance and Tukey's post hoc test and P < 0.05 was considered statistically signifcant.

RS Administration Decreased BP in Prehypertensive Rats
by Improving Endothelial Function. Te initial BP of the Evidence-Based Complementary and Alternative Medicine prehypertensive rats was 160/85 mmHg compared to 140/ 70 mmHg in normal controls, indicating a successful model. Te intervention with RS signifcantly decreased both SBP and DBP compared to those in the untreated PHT rats. Te therapeutic efects of valsartan were manifested earlier and the SBP showed an obvious decrease after four weeks ( b P < 0.01) (Figure 1(a)). Given the crucial role of the vascular endothelium in prehypertension, we next evaluated the efect of RS on the aortic endothelium. As shown in Figure 1(b), the endothelium of the thoracic aorta of PHT animals was partially disintegrated (marked by black arrows) and showed signifcant degeneration of the endothelial cells (ECs) and vascular wall thickening. RS intervention decreased the exfoliation of the ECs and restored the structural integrity of the vascular endothelium. SEM examination of the aortas (Figure 1(c)) further revealed a well-ordered, rope-shaped vascular endothelium in the control group, with interconnected ECs and minimal damage (marked by red arrows). In contrast, ECs in the PHT group showed extensive shedding and loss of intercellular junctions, resulting in cavities. However, the intervention with RS signifcantly reduced endothelial abscission. Taken together, RS alleviated the prehypertensive symptoms in rats by restoring the aortic vascular endothelium.

RS-Induced Vasodilation by Modulating ET-1 and NO
Levels. Endothelins are potent vasoconstrictors, whereas endothelium-derived NO is the most powerful vasodilator. Lowered NO bioavailability and increased production of vasoconstrictor substances eventually lead to vascular ED [28]. Te PHTrats had signifcantly higher serum ET-1 levels compared to controls, which decreased after treatment with RS and valsartan. On the contrary, NO levels were markedly reduced in PHT animals and restored by RS and valsartan interventions. Furthermore, RS elevated serum NO more efectively compared to valsartan (Figure 1(d)). Tis strongly indicated that RS can reverse vascular ED in prehypertensive rats by inducing vasodilation. To validate this hypothesis, we tested the efect of RS on the thoracic aorta in vitro using the vascular ring tension method. A cumulative gradient of RS had no signifcant efect on the tension of the rat aortic ring at rest. However, RS signifcantly relaxed the aorta precontracted with PE in a dose-dependent manner, and the maximum relaxation rate was approximately 80%. L-NAME is a nonselective inhibitor of endothelial nitric oxide synthase (eNOS), which can block the synthesis of NO and lead to vasoconstriction. As shown in Figure 1(e), pretreatment with L-NAME (100 mM) signifcantly inhibited the relaxation curve of the vascular rings in the presence of RS. Tis suggests that hypertension-related ED may involve an altered basal release of NO and that the antihypertensive efects of RS rely on NO-mediated vasodilation.

Transcriptomic Alterations with RS Treatment.
To further dissect the molecular mechanisms underlying the antihypertensive efects of RS, we compared the transcriptomes of the thoracic aorta of diferent groups. Using P adj < 0.05 and FC > 2, we identifed 496 DEGs between the PHT and the control group (PHT vs. C), of which 272 genes were upregulated and 224 genes were downregulated in the former. RS signifcantly upregulated 55 genes and downregulated 71 genes in the PHT group (Figure 2(a)). Furthermore, the gene expression profle of the RS-treated animals was closer to that of the controls than that of the untreated PHT animals (Figure 2(b)), indicating that RS also reversed the molecular changes associated with prehypertension. A Venn diagram of the DEGs between the two datasets ( Figure 2(c)) revealed 65 DEGs in response to RS, of   Table 2). Gene Ontology (GO) analysis further indicated that the 496 DEGs were enriched for signifcant pathways, including "response to the hormone," "antigen processing and presentation of peptide," "response to a toxic substance," "circadian rhythm," "response to mechanical stimulus," etc. (Figure 3(a)). Compared to the PHT group, the signifcantly enriched GO terms in the R group included "vasoconstriction," "regulation of protein acetylation," "antigen processing and presentation of peptide," "circadian rhythm," and "negative regulation of muscle cell apoptotic process" (Figure 3(b)). Taken together, circadian rhythm, antigen processing, and presentation of the peptide are crucial determinants of the antihypertensive efects of RS.

RS Afects Multiple Metabolic Pathways in Hypertensive
Rats. Te metabolic changes induced by RS were determined by comparing the serum metabolite profles of the diferent groups. Total ion chromatograms (TIC) of sera from the control, PHT, and RS-treated groups are shown in Figures 4(a) and 4(b), which indicate a clear separation of metabolites along with diferences in some peak intensities. PCA further reiterated the diferences between the groups through data dimensionality reduction (Figures 4(c)-4(f )). Furthermore, analysis of the QC samples showed that the detection method had good reproducibility and stability and that our data were reliable. Good separation was achieved between the control and PHT groups, indicating that prehypertension altered the metabolic profle, resulting in signifcant diferences in the content of endogenous metabolites. On the other hand, the metabolites of the R group were closer to those of the control group, suggesting that the RS intervention improved metabolism. OPLS-DA was performed to identify metabolites that can discriminate between samples (Figures 4(g)-4(j)) on the basis of variable importance in projection (VIP). METs were screened using VIP > 1, P value < 0.05, and FC ≥ 2 or ≤0.5 as the thresholds. As shown in Table 3, there were 52 potential METs after RS intervention, which were functionally annotated using the MetaboAnalyst 5.0 platform. Te METs were enriched in 17 metabolic pathways (impact > 0.01), including linoleic acid metabolism, selenocompound metabolism, D-glutamine, and D-glutamate metabolism, arginine biosynthesis, histidine metabolism, arginine and proline metabolism, purine metabolism, pentose and glucuronate interconversions, sphingolipid metabolism, and so on (Table 4). Taken together, RS exerts its antihypertensive efects by targeting multiple pathways.

Integrated Transcriptomics and Metabolomics.
Te network of DEGs and METs was then constructed to explore their causal relationship in response to RS ( Figure 5(a)). Ugt1a1 interacted with androstanedione and androsterone sulfate through androgen and estrogen biosynthesis and metabolism pathways under the action of glucuronosyltransferase, and was associated with 7alpha-hydroxycholest-4-en-3-one through bile acid biosynthesis. Additionally, prostaglandin E2 and lipoxin A4 were formed from arachidonate by arachidonic acid (AA) metabolism with nicotinamide adenine dinucleotide (NAD) (+) or nicotinamide adenine dinucleotide phosphate (NADP) (+) as acceptors. Aldh1a3 interacted with 2-phenylacetamide and carnosine through tyrosine metabolism and histidine metabolism pathways, respectively, and interacted with Lglutamate and L-cystine through the urea cycle and metabolism. Due to the limited nature of MetScape, some metabolites and genes were not recognized. PPI network analysis using the STRING database revealed 448 nodes and  Evidence-Based Complementary and Alternative Medicine 7 136 interacting pairs, with an average node degree of 5.83 and an average local clustering coefcient of 0.46. As shown in Figure 5(b), eight DEGs, including Pik3c2a, Hspa8, Dnaja1, Ugt1a1, Aldh1a3, Dbp, Rasd1, and Arntl, were associated with potential metabolites.

Verifcation of the Key Target Genes of RS in Prehypertensive Rats.
Te target genes of RS identifed above were validated in the thoracic aorta by qRT-PCR. As shown in Figure 6, Pik3c2a, Akt1, eNOS, Hspa8, Dnajal, and Arntl were downregulated, and Dbp, Rasd1, Ugt1a1, and Aldh1a3 were upregulated in the PHT group compared to the controls. Pik3c2a and Akt1 did not show a signifcant diference, but they showed a downward trend in the PHT group. RS treatment increased the transcript levels of Pik3c2a, Akt1, eNOS, Hspa8, Dnajal, and Arntl and decreased those of Dbp, Rasd1, Ugtlal, and Aldh1a3. Interestingly, Akt1 showed a trend similar to that of Pik3c2a, while eNOS mRNA was signifcantly downregulated in the PHT group and increased with RS treatment. Tese fndings further support our hypothesis that RS alleviates hypertension through NOmediated vasodilation.

Discussion
Prehypertensive patients are at a higher risk of developing hypertension and CVD compared to those with normal BP [29], but are often overlooked. ED is the pathological basis of hypertension and prehypertension, which was also established in the prehypertensive rat model in terms of elevated ET-1 and lower NO levels in sera. Te RS intervention reversed these changes and also repaired the damaged aortic endothelium. Te vascular ring assay showed that RS reduced BP by improving relaxation of the vascular endothelium. Furthermore, the vasodilatory efect of RS on the thoracic aorta of a prehypertensive rat was confrmed by the increased vasoconstriction seen after preincubating the blood vessels with the eNOS inhibitor L-NAME.
Early metabolomics studies have shown that the serum metabolite profle of individuals with borderline hypertension is distinct from that of hypertensive patients [30]. Tis strongly indicated that the prehypertensive stage is associated with changes in specifc metabolites that can potentially be used for early diagnosis. We identifed 52 METs in the sera of RS-treated hypertensive rats, most of which have been reported previously to be related to prehypertension or hypertension. To compensate for the limitations of untargeted metabolomics, we integrated the transcriptomes and metabolomes to construct a network and identify the key targets of drug intervention. Eight putative RS targets were related to pathways involved in BP regulation and may be promising biomarkers for early diagnosis and risk assessment.
RS may exert its efects through the phosphatidylinositol signaling pathway, which plays an important role in BP regulation. Pik3c2a, a member of the phosphoinositide 3kinases (PI3K) family, phosphorylates the 3-hydroxyl of the phosphatidylinositol (PI) ring, producing a second messenger that relays signals via multiple pathways. In addition, Pik3c2a is critical for EC junctions [31] and survival [32,33], and its absence can alter platelet structure and viscosity through shear stress [34]. Laminar shear stress increases the interaction between Sirt1 and eNOS, as well as eNOS deacetylation, to enhance NO production and reverse ED [35]. Phosphoinositide-dependentkinase-1 (PDK1) interacts with phosphatidylserine (PS) and phosphatidylinositol-3,4,5-trisphosphate (PIP 3 ) through diferent kinase domains and activates Akt [36]. Akt1 is the main subtype of Akt, which is expressed in ECs and preferentially phosphorylates eNOS and promotes NO release in vitro [37]. Te levels of PS, Pik3c2a, Akt1, and eNOS were signifcantly increased after RS intervention, indicating that it improves endothelial function and reduces BP by regulating phosphoinositol signaling.
Elevated ET-1 during hypertension increases vascular superoxide levels, which further aggravates ED. Te main sources for oxidative excess in the vasculature are NADPH   Evidence-Based Complementary and Alternative Medicine   [28,38]. XOD directly oxidizes xanthine to produce superoxide anion and uric acid (UA), which is the fnal product of purine degradation and is depleted in the absence of purine-nucleoside phosphorylase (PNP). Enhanced levels of superoxide anions also decrease NO bioavailability and further promote ED. Studies show that UA independently increases the risk of prehypertension [39]. Terefore, lowering plasma UA levels may prevent prehypertension [40]. Xanthine and adenosine levels decreased, and those of xanthosine increased signifcantly after RS treatment. Terefore, the therapeutic efect of RS can also be attributed to the regulation of the purine metabolism pathway.
Most physiological functions in humans, including BP regulation, follow circadian rhythms. Disturbance of endogenous circadian rhythms increases the risk of       14 Evidence-Based Complementary and Alternative Medicine hypertension [41]. Te transcription factors BMAL1 and CLOCK form a heterodimer that drives the transcription of circadian genes (Per and Cry), tissue-specifc genes like Edn1, and clock-controlled genes such as Dbp by binding to E-Box response elements. BMAL1, also known as aryl hydrocarbon receptor nuclear translocator-like (Arntl) protein, is downregulated in hypertensive women [42]. Studies show that PPARc activator reduces BP by inducing the aortic expression of BMAL1 mRNA [43]. Compared to hypertensive rats, the expression levels of BMAL1 and CLOCK are higher in WKY rats [44]. Arntl levels increased after RS intervention, indicating that it may be critical for its antihypertensive efects. Circulating ET-1, the product of the Edn1 gene, also exhibits a circadian rhythm [45]. Trough a negative feedback loop, Per and Cry inhibit their own transcription by forming a complex with the BMAL1-CLOCK heterodimer. In addition to the negative autoregulatory feedback loop of Per and Cry, the Dbp-mediated loop also amplifes the circadian oscillation. Sirt1 regulates the circadian rhythm by histone deacetylation [46] and increases the expression of endothelial NO [47], which is negatively correlated with Dbp [44]. RS signifcantly reduced the level of Dbp, which may reduce BP by upregulating eNOS. Rasd1, also known as Dexras1, regulates circadian rhythm in response to external signals and can also activate physiological NO signaling [48]. Furthermore, Rasd1 can inhibit the activity of cyclic adenosine monophosphate (cAMP) [49]. Tus, RS-mediated downregulation of Rasd1 may reverse the inhibition of cAMP and relax the blood vessels [50]. Linoleic acid (LA), a precursor of AA biosynthesis, is an essential fatty acid that maintains physiological levels of prostaglandins and thromboxanes and regulates vascular tone [51]. LA is metabolized by various enzymes, including cytochrome P450 (CYP), lipoxygenase (LO), and cyclooxygenase (COX), of which CYP1A2 has a strong catalytic activity. Previous studies have shown that hypertensive rats have lower levels of LA and a higher level of AA compared to normotensive controls due to inhibition of COX activity [52]. Terefore, restoring the levels of both compounds can prevent hypertension [53] through improved NO bioavailability and amelioration of ED [54]. Tus, an RSmediated increase in NO levels can be due to the increase in LA, which may be one of the active substances regulating BP.
Eicosanoid metabolites are produced by AA metabolism in blood vessels, and their levels in CVD are the major determinants of ED [51]. AA is also metabolized through the CYP, LO, and COX pathways. Te products of the CYPdependent metabolism of AA are associated with increased renal vascular resistance in prehypertensive SHR [55]. Te COX-mediated synthesis of prostaglandin E2 (PGE2) is diminished in prehypertensive rats [56]. Te efect of PGE2 on blood vessels depends on its receptor [57]. For instance, after binding to the receptors EP2 and EP4, PGE2 promotes cAMP production and induces vascular smooth muscle relaxation [50]. RS probably exerts its antihypertensive efect by increasing PGE2 levels. Prostaglandin E synthase 3 (PTGES3) catalyzes the oxidoreduction of prostaglandin H2 (PGH2) to PGE2, which functions as a cochaperone with heat shock protein 90. Dnaja1 encodes Hsp40 proteins, which act as cochaperones for Hspa8. Overexpression of Hspa8 inhibited EC apoptosis and promoted angiogenesis and vascular remodeling following vascular injury [58]. Genetic variations in Hspa8 correlate inversely with the risk of hypertension [59]. Furthermore, Hspa8 is also upregulated in patients with arterial hypertension, which may be a protective response to EC injury [60]. Te expression of negative (e, f ) ion modes. OPLS-DA score plots based on data from the diferent groups obtained in the positive (g, h) and negative (i, j) ion modes. C represents the control group, PHT represents the model group, R represents RS-treated group, and QC represents the quality control group.    Hspa8 and Dnaja1 increased in the RS-treated group, which probably promoted the transformation of PGH2 to PGE2 via PTGES3, eventually lowering BP. We also observed an upregulation of 12(R)-HETE and a downregulation of lipoxin A4(LXA4), which are AA metabolites produced via the LO pathway in vascular ECs. A previous study showed that 12(R)-HETE acts as an agonist of the vasodilator in preconstricted mouse artery [61]. LXA4 has anti-infammatory, vasodilatory, and antioxidant efects [62], and low LXA4 levels may be an indicator of the   Figure 6: qRT-PCR for target genes using divergent primers. Te mRNA levels of Pik3c2a (a), Akt1 (b), eNOS (c), Hspa8 (d), Dnaja1 (e), Arntl (f ), Dbp (g), Rasd1 (h), Ugt1a1 (i), and Aldh1a3 (j) were detected. * P < 0.05, * * P < 0.01, * * * P < 0.001, and * * * * P < 0.0001 vs. the PHT group. development of hypertension [63]. On the contrary, increased LXA4 in the plasma of preeclampsia women plays a central role in disease development [64].
UDP glucuronosyltransferase family 1 member A1 (Ugt1a1) is the only enzyme that can metabolize bilirubin and control serum bilirubin levels. Bilirubin is the end product of heme catabolism and a risk factor for CVD [65]. Genetic variations in Ugt1a1 are associated with elevated bilirubin and an increased risk of hypertension in individuals with African ancestry [66]. Mice treated with anti-Ugt1a1 antibodies show a reduction in BP, which can be attributed to lower oxidative stress and increased NO levels [67]. Ugt1a1 catalyzes glucuronidation at the 3-OH of dihydrocafeic acid (DHCA), which scavenges reactive oxygen species in ECs and increases the level of eNOS to protect the ECs from oxidative damage [68]. We found that RS downregulated Ugt1a1 and increased DHCA levels, indicating that RS can exert an antihypertensive function by increasing the level of eNOS.
Aldehyde dehydrogenase family 1 member 3 (Aldh1a3) regulates proliferation in the neointima and the medial layer of vascular smooth muscle [69]. Inhibition of excessive Aldh1a3 activity can alleviate intimal hyperplasia. Additionally, a variation in the Aldh1a3 gene has been linked to the hydrochlorothiazide response in hypertensive individuals [70], and the increased Aldh1a3 led to intimal hyperplasia and lumen stenosis in a rat model of carotid artery injury [71]. Te expression of Aldh1a3 decreased after administration of RS, indicating that it may lower BP by reducing Aldh1a3 expression.

Conclusions
RS can efectively reduce the BP of prehypertensive rats and improve vascular ED by regulating purine, linoleic acid, AA metabolism, circadian rhythm, and the phosphatidylinositol signaling pathway. Pik3c2a, Hspa8, Dnaja1, Arntl, Ugt1a1, Dbp, Rasd1, and Aldh1a3 may be the key genes targeted by RS intervention in prehypertension (Figure 7). More research is needed to validate the clinical potential of our fndings.

Data Availability
Te original data of the RNA-seq analysis has been uploaded to NCBI, https://www.ncbi.nlm.nih.gov/sra/PRJNA801895, which can be downloaded. Te raw data used to support the fndings of this study are available from the corresponding authors upon request.