Metabolomic Analysis of the Urine from Rats with Collagen-Induced Arthritis with the Effective Part of Caulophyllum robustum Maxim

Rheumatoid arthritis (RA) is a chronic autoimmune disease with high incidence and high disability and recurrence rates. Caulophyllum robustum Maxim (C. robustum) is a traditional Chinese medicine (TCM) with main effective parts (CRME) commonly used for RA treatment. To explore the mechanism of CRME in RA, we used metabolomics to investigate the effect of CRME intervention on urine metabolism in rats with collagen-induced arthritis (CIA). CIA rats were randomly divided into normal control, CIA model, and CRME groups. A metabolomics approach, using Ultra-Performance Liquid Chromatography-Quadrupole-Time-of-Flight/Mass Spectrometry, was developed to perform urinary metabolic profiling. Differential metabolites were identified by comparing the CIA model and CRME groups. Preliminarily, 56 significant differential metabolites were identified in urine, and 20 metabolic pathways were disturbed by the CIA. The amount of 16 different metabolites changed in urine after CRME intervention. The production of these metabolites involves tryptophan, tyrosine, energy, cholesterol, and vitamin metabolism. CRME has anti-inflammatory and immunosuppressive effects in CIA model rats. By examining the endogenous metabolite levels, we identified potential CRME targets and pathways involved in the treatment of RA. The results of our metabolic studies indicate that CRME regulates amino acid, vitamin, energy, and lipid metabolism pathways to treat RA and may provide a new explanation for the anti-RA mechanism of CRME.


Introduction
Rheumatoid arthritis (RA) is a chronic autoimmune and systemic inflammatory disease that results in joint injury, disability, and decreased quality of life and has distinct signatures of persistent synovitis and hyperplasia [1]. Worldwide, approximately 0.24% of the population suffers from RA. Although RA has been studied for many years, the pathogenesis of this disease is still incompletely understood. Currently, there are no ideal methods or medicines to cure RA. However, it is known that endogenous metabolism is changed in RA patients [2]. Moreover, studies and clinical test results indicate that traditional Chinese medicine (TCM) can play important roles in alleviating the symptoms of RA [3].
Caulophyllum robustum Maxim (C. robustum) is also known as "Hong Mao Qi". C. robustum is widely distributed in northeast, northwest, and southwest China, and the main effective parts of this plant are its roots and rhizomes. C. robustum can dispel wind, remove obstruction in the meridians, promote blood circulation, and stop the pain, and it has been used to treat rheumatic arthralgia, irregular menstruation, stomachache, and abdominal pain [4][5][6]. C. robustum Maxim extract (CRME) mainly contains saponins and alkaloids [7], and spectrum-activity relationship results indicate that multiple CRME components have antiinflammatory and immunosuppressive effects [8]. CRME has obvious dose-dependent therapeutic effects in RA animal models where it can regulate the levels of tumor necrosis factor and interleukin family inflammatory cytokines [9] and regulate cellular and humoral immunity [10,11]. CRME also can protect cardiocytes [12], slow the formation and proliferation of synovial pannus [13], and reduce articular cartilage and bone erosion and destruction [13]. Together, these results indicate that CRME is effective in the prevention and treatment of RA and immune hyperactivity. However, the mechanism of TCM is complex, and it is difficult to elucidate the effects of TCM at the molecular level.
Metabonomics is a new omics technology that combines systems biology, genomics, and proteomics [14]. Metabolomics is mainly used to determine endogenous metabolites in biological samples using qualitative and quantitative modern analytical techniques (including LC-MS, GC-MS, and NMR) [15]. Changes in the endogenous metabolites produced before and after TCM administration were investigated using chemical informatics (PCA, PLS-DA, and OPLS-DA) [14,16,17]. Using this approach, differentially produced metabolites can be used as biomarkers of the pathologic and physiologic body state, can illustrate the efficacy of TCM pharmacodynamic material, and uncover the mechanism of TCM action [18].
We used a combination of metabonomics and modern analytical techniques to explore the effects of CRME on the urine metabolic profiles in collagen-induced arthritis (CIA) rats. en, we explored the role of CRME in regulating small molecular metabolites in the body and metabolic pathways by examining the changes in biomarkers. Research of the related metabolic pathways was used to explore the mechanism of CRME action, its anti-inflammatory effects, and its targets in the body. is molecular-level understanding can guide the clinical application of CRME. e original medicinal materials for CRM (100 g) were extracted twice with 70% ethanol (v/v) at reflux, and the extract was placed in an AB-8 macroporous adsorption resin column and collected. e 70% eluate obtained was dried by evaporation in a rotary evaporator, and the solvent was recovered to obtain the effective part of C. robustum Maxim (CRME). e six main CRME components, CaulosideH, LeonticinD, CaulosideG, CaulosideD, CaulosideB, and CaulosideC, were determined by high-performance liquid chromatography (HPLC). e results showed that the CRME was composed of 6.21%, 5.14%, 28.88%, 10.19%, 10.19%, and 2.26% for CaulosideH, LeonticinD, CaulosideG, CaulosideD, CaulosideB, and CaulosideC, respectively.

Model Construction and Drug Administration.
Wistar rats were selected as subjects. e appropriate amount of CII was precision weighed, and a 2 mg/ml solution was generated by dissolving in glacial acetic acid. e solution was stored at 4°C overnight. e collagen solution was emulsified with CFA in a 1 : 1 ratio on the first day of immunization and fully emulsified in the ice bath environment, and eventually, the emulsion droplet dripped into the water is not dispersed as appropriate. Rats were intradermally injected with the CII emulsion (0.2 mL) at the base of the tail at three equidistant locations, and the first basic immunization was recorded as d0. Collagen solution was emulsified with IFA in a 1 : 1 ratio on the second day of booster immunization. A secondary immunization, or boost, using 0.2 mL of the CII emulsion was given in the same way on day 7. Rats in the normal group were not treated and had normal feeding.
Rats were divided into normal control group, CIA model group, and CRME group (n � 8 in each group). Gavage was started on the day of the second immunization. e doses were CRME 69.58 mg/kg for 30 consecutive days. e blank group and the CIA model group were given equal volumes of distilled water.

2.5.
Handling of Animal Samples. After successful modeling of CIA rats, all experimental rats were continuously administered CRME for 30 days. Rats had free access to water, without a standard diet, for a 12 h period on days 10, 20, and 30 during which time they were placed into metabolic cages for urine sample collection. e fresh urine samples with sodium azide were stored at −80°C immediately before use. All urine samples were centrifuged at 5000 rpm for 10 min, and the supernatant was stored at −80°C until required for analysis. Prior to analysis, 150 μL of urine from each sample was vortex-mixed with 150 μL cold methanol acetonitrile (2 : 1, v/v) solution and 10 μL internal standard (2-chloro-L-phenylalanine, 0.3 mg/mL, methanol configuration), subjected to ultrasonic extraction for 5 minutes and left to stand for 20 minutes at −20°C. Urine samples were collected after centrifugation at 14,000 rpm for 10 min at 4°C and weighed. Clear supernatant (200 μL) was transferred to a sampling vial for UPLC-MS analysis. e flow rate was set at 0.40 mL/min, and the injection volume was set as 3 μL for all analyses.
Mass spectrometric data were collected using a Mass Spectrometer equipped with an ESI source operating in positive or negative ion mode. e sample injection and the collision voltage were 1.0 kV and 40 V, respectively. e ionic source and dissolving temperatures were 120°C and 500°C, respectively, and the desolvation gas flow was 900 L/ h. Centroid data was collected from 50 to 1,000 m/z with a scan time of 0.1 s and an interscan delay of 0.02 s over a 13 min analysis time.

Urine
Metabolic Profiling Analysis. Baseline filter, peak identification, integration, retention time correction, peak alignment, and normalization were performed by the metabolomics processing software of the instrument Progenesis QI (Waters Corporation) using raw data to obtain a data matrix with retention time, mass-to-charge ratio, and peak intensity. e normalized urine sample data matrix was imported into the SIMCA-P + 14.0 software package (Umetrics, Umea, Sweden) for unsupervised principal components analysis (PCA) to observe the overall sample distribution and the stability of the analysis process. en, the supervised (orthogonal) partial least square method, (O) PLS-DA, was used to distinguish overall metabolic profile differences and identify metabolite differences between groups. Variables with variable important in projection (VIP) scores greater than 1 were considered differential variables. To prevent the model overfitting, the quality of the model was investigated using seven interactive verification cycles and 200 response sequencing tests.

Differential Metabolite Screening and Identification.
Screening the metabolite differences between groups (VIP > 1, p < 0.05, Fold Change (FC) > 1.2) was achieved using multidimensional analysis of PLS-DA and single-dimensional analysis (Student's t-test) methods. Differential metabolites were selected for qualitative analysis using http://www.hmdb.ca/and self-built databases.
According to the variable projection importance (VIP) parameter in the PLS-DA model, the potential biomarkers were evaluated, and the different variables of VIP > 1 were selected according to the order of VIP, and then, the compounds with significant differences among the three groups were screened by t-test. Only the different variables with p < 0.05 were considered as potential biomarkers. According to the matching results and HMDB metabolite database search, as shown in Figure 1 and Table 1, 16 metabolites were identified, which were preliminarily considered as potential biomarkers induced by the effective part of Caulophyllum robustum Maxim.

Effect of CRME on Urine Metabolism Profiles and Differential Metabolites Identified in CIA Rats.
e changes in differential metabolites in the last urine collection of experimental rats administered CRME were analyzed to analyze the substances with callback during the treatment for analyzing the effect of effective parts of CRME on the differential metabolites in the urine of rats. e metabolic pathways of differential metabolites with callbacks were analyzed and determined.

Assessment of the CIA Model and the Anti-Inflammatory Effects of CRME.
e rats of the normal control group had normal feeding and behavior, and their weight gradually increased. Rats of the CIA group showed a trend of decreasing weight, and their activity decreased after 12 days (p < 0.01). Compared with the model group, the bodyweight of the rats in each group increased over the period of CRME administration (p < 0.05) (Figure 2(a)). e arthritis index (AI) in CIA rats that are administered CRME was examined.
e AI value of rats in each group reached the maximum on day 16. Compared with the CIA model group, the AI values of rats in each group showed a decreasing trend from the 10 th day after administration, and the AI values at different Evidence-Based Complementary and Alternative Medicine time points in each administration group, including the CRME group (p < 0.05), significantly differed from those in the model group (p < 0.01) (Figure 2(b)). Compared with the blank group, the swelling of the right hind limb in the CIA model group was most evident on day 10 (p < 0.01). Compared with the model group, the foot volume of rats in each group began to significantly decrease after 7 days of CRME administration (p < 0.05 or p < 0.01) (Figure 2(c)).

Spectral Data Analysis.
A visual inspection of the base peak ion flow chromatogram (BPI) was performed for all samples.
is revealed that all samples had strong signal strength, large peak capacity, and good retention time. Urine samples from rats in the CRME group were analyzed after 30 days of CRME administration. BPI chromatograms in positive and negative ion modes are shown in Figure S1.

Metabolic Profiling Analysis of Urine.
A mathematical model was established for urine samples in blank and model groups using PCA and PLS-DA. e score chart ( Figure 3(a)) and PCA analysis parameter values (R 2 X � 0.655) reveal a total metabolic difference between blank and CIA model groups, and the variation of the sample within the group is greater. ese results show that the analysis process was stable, and the physiological state of the model group obviously differed from that of the blank group. e PLS-DA score plots (R 2 X � 0.537, R 2 Y � 0.871, Q 2 � 0.649) of urine samples indicated that the mathematical model is reliable and stable and can be used to find potential differential metabolites. Compared with the normal group, the samples from model rats metabolically differed based on the PLS-DA score plot (Figure 3(b)), indicating that the metabolism of endogenous substances differs in rats after injecting collagen and leads to significant changes in the metabolic fingerprint of urine samples in rats.
e closer the slope of the normal R 2 Y and Q 2 Y lines is to the horizontal line, the more likely the model is to be overfitted. Additionally, Q 2 is generally required to be less than zero when using the RPT test ( Figure 3(c)). In this experiment, Q 2 � −0.631 of urine samples, indicating that the model is well-fitted. e points at both ends of the s-type are potential biomarkers and are determined by VIP values greater than 1.0 in the S-plot diagram (Figure 3(d)). e PLS-DA method was used to evaluate the overall metabolism of the normal control group, CIA model group, and the CRME administration group. Urine sample data from rats in each group were analyzed by PLS-DA on days 10, 20, and 30 ( Figure 4). PLS-DA score (cum) � 0.791) revealed dynamic metabolite changes in the drug groups on the selected days. e metabolic state of the CRME group differed more from that of the model group with prolonged treatment time (Figure 4), indicating that the metabolic state of CIA rats gradually improved after CRME treatment and that CRME has an obvious therapeutic effect on CIA rats.

Identification of Biomarkers and Analysis of Metabolic
Pathways.
e PLS-DA analysis value of variable importance in projection (VIP), and the t-test method was combined to select compounds with differences (p < 0.05). For the urine samples of model rats, 56 differential metabolites (34 in positive, 22 in negative) were confirmed in the final experiment (Table S1). CII significantly disturbs metabolites in rats, leading to a large number of differential metabolites in urine.
We mapped 56 differential metabolites using the KEGG database and obtained 20 associated metabolic pathways, including phenylalanine, tyrosine, and tryptophan biosynthesis (Impact 1), taurine and hypotaurine metabolism (Impact 0.43), phenylalanine metabolism (Impact 0.41), glycine, serine, and threonine metabolism (Impact 0.29), pentose and glucose conversion (Impact 0.27), histidine metabolism (Impact 0.24), tyrosine metabolism (Impact 0.17), riboflavin metabolism (Impact 0.17), and tryptophan metabolism (Impact 0.09). ese pathway analysis results are presented graphically ( Figure 5) and in a detailed table (Table 2). e link between these pathways and the identified metabolites is illustrated in Figure 6. e metabolic profile data of the urine of rats at 30 days was analyzed to identify the differential changes in metabolites. e effective parts of CRME had different degrees of differential metabolite callbacks using SPSS18.0. Among them, 16 biomarkers have callback effects in urine samples ( Figure 1 and Table 1) and are involved in eight metabolic pathways, including tryptophan metabolism, nicotinic acid and nicotinamide metabolism, tyrosine metabolism, phenylalanine metabolism, tricarboxylic acid (TCA cycle), pentose, glucuronic acid conversion, metabolism of cholesterol, and bile acid synthesis.
e pathogenesis of CIA is closely related to the disordered amino acid, glucose, vitamin, and lipid metabolism, which may also be potential metabolic pathway targets for drug therapy.

Discussion
RA is a disease of unknown etiology that is related to metabolism. According to the theory of TCM, the occurrence of RA is affected by the three major factors of cold (han), damp (shi), and wind (feng), which together affect the stability of the human body environment [3].
ere are many drugs clinically used to treat RA, but there is no treatment that is completely effective, nor can RA be completely cured. erefore, it is important to focus on the mechanism of action of potential drugs for RA treatment. TCM is characterized by multiple components, multiple targets, and high levels of safety. TCM has unique therapeutic advantages for RA, but the mechanism of action is complex and usually involves a wide range of metabolic processes in vivo. Clarification of the material basis of TCM effects at the molecular level is a difficult and pressing topic in TCM research. e overall characteristics of metabonomics have led many scholars to apply this approach to the study of TCM metabolites and their pharmacodynamics. Yun et al. studied the interference effect of tetrandrine, in tetrandrine powder, on endogenous metabolites in the urine of CIA rats using the 1H-NMR metabolomics method. ey found 23 potential biomarkers of CIA models, which were mainly involved in energy, amino acid, lipid, and intestinal microbial metabolism. e protective effect of pirenoxine on CIA rats may be regulated by these metabolic pathways and potential biomarkers [19]. Li et al. used the UPLC-Q-TOF-MS metabonomic analysis method to study the anti-RA mechanism of clematis triterpene saponins. ey found that clematis triterpene saponins can remove citric acid and p-cresol glucose in the serum and urine of CIA rats. Twentyfour biomarkers, including aldehyde and creatinine, return to normal levels following clematis triterpene saponin treatment. is then regulates the overall metabolic pathways of glycerophospholipid metabolism, arachidonic acid metabolism, steroid hormone synthesis, amino acid metabolism, and pantothenic acid and coenzyme A biosynthesis [10]. e effective part of Caulophyllum robustum Maxim mainly contains saponins and alkaloids, which are the main material basis of anti-RA of C. robustum [8]. Saponin is an important chemical component of traditional Chinese medicine, which has extensive pharmacological effects. Most of the natural saponins exist in the form of sapogenin-linked glycosides. In the process of intestinal metabolism and liver metabolism, most pentacyclic triterpenoid saponins are easily metabolized and converted into aglycones by various enzymes and humoral environments, resulting in biological activity. e bioavailability of pentacyclic triterpenoid saponins is generally low due to poor oral absorption and metabolism [20]. Alkaloids, as an important natural product, are numerous and complex in structure. In recent years, it has been found that alkaloids Evidence-Based Complementary and Alternative Medicine

Evidence-Based Complementary and Alternative Medicine
can be metabolized not only in the liver but also in the intestine, and intestinal flora has a certain metabolic transformation effect on some alkaloids. For the diagnosis of RA, a complex disease, looking for related biomarkers by metabonomics, can improve the accuracy of diagnosis and provide new ideas for the treatment of traditional Chinese medicine [21]. Biomarkers are signal indicators that can reflect the physiological and pathological conditions of the human body and are widely used in disease prediction, diagnosis, and treatment [22]. In a complex disease like RA, finding relevant biomarkers can improve the accuracy of diagnosis and provide new insights into TCM treatment [21]. C. robostum Maxim (CRM) has a long history in the treatment of RA in China. However, to date, there have been no studies examining CRM-anti-RA metabolomics. Moreover, there is also a lack of studies on the changes related to metabolic pathways and the identification of biomarkers in RA following treatment with CRME. erefore, in this study, we established a CII-induced CIA rat model and performed metabolomics analysis of rat urine samples to explore the mechanism of CRME in ameliorating the effects of RA. Key biomarkers were screened and differential metabolites with significant changes in urine samples and their related metabolic pathways were identified.

Amino Acid Metabolism.
Phenylalanine hydroxylase converts phenylalanine to tyrosine. Tyrosine is converted to tyramine by amino acid decarboxylase. Tyramine is further metabolized into dopamine, norepinephrine, and other substances with physiological activities. Tyrosine can also be hydrolyzed to fumaric acid, acetoacetic acid, and gentisic acid, into the tricarboxylic acid cyclic. In CIA rats, the levels of hydroquinone and m-cresol decreased and levels of tyrosine, acyl-tyrosine, tyramine, and benzaldehyde increased. e level of m-cresol in the urine of AA rats was increased [23]. erefore, abnormal tyrosine metabolism will lead to phenylalanine metabolism and energy metabolism disorders, which will affect the overall condition of the body. e level of tyrosine returned to normal in CIA rats after CRME intervention, indicating that CRME could treat RA by regulating tyrosine metabolism.
Tryptophan is the second essential amino acid in animals and plays a key role in the synthesis and metabolism of proteins, enzymes, vitamins, and neurotransmitters [24]. e levels of tryptophan, 5-methoxyamine, and indole acetaldehyde metabolites are increased, and the levels of 2indolecarboxylic acid, 3-methylindole, indoxyl, kynurenine, and indole-3-methanol are decreased. Tryptophan is deaminated by intestinal flora to produce indoles, which are absorbed into the blood by the intestinal tract and metabolized into indole sulfate by the liver. Indole sulfate has high protein affinity and various cytotoxicity in vivo. Tryptophan plays an important role in regulating T cell-mediated autoimmune disease [25], and it can affect T cell proliferation in patients with RA by regulating the IDO/TTS tryptophan metabolism pathway. It is also possible to clear autoreactive T cell-mediated chronic immune inflammation. Kynurenine is an important metabolite in the tryptophan metabolism pathway. e ratio of kynurenine to tryptophan is an immune status index in vivo and is significantly higher in the blood of patients with RA than in those without RA [26]. e  Evidence-Based Complementary and Alternative Medicine levels of 2-indolecarboxylic acid and indole phenol returned to normal after the intervention with CRME, suggesting that CRME can alleviate the inflammatory symptoms of RA by regulating the response of tryptophan in the intestinal flora. e histidine content in the urine of CIA rats is high, and the leucine to histidine ratio can be a potential biomarker for clinical osteoarthritis. Early studies have found that histidine metabolism is elevated in patients with RA, but the specific pathological mechanism remains unclear. is is consistent with our findings that histamine metabolism disorder is also a pathological feature of RA.

Glutathione Metabolism.
Glutathione is a tripeptide consisting of glutamic acid, cysteine, and glycine and contains c-amide bonds and mercapto groups. Glutathione is found in almost every cell of the body and helps the immune system function [27]. Glycine receptors are distributed in the cell membranes of monocytes, macrophages, neutrophils, T-lymphocytes, and vascular endothelial cells and participate in immune responses and inflammation. Liu et al. [28] also found that acute gouty arthritis can cause disorders of the glutathione metabolism pathway. In the field of medicine, it has been reported that glutathione can treat eye diseases such as cataract, alleviate hepatitis, and protect the liver [29] and has antioxidant [27] and anti-HIV effects [30]. Glutathione is involved in apoptosis, signal transduction, and regulation of nitric oxide signaling pathways and induces oxidative stress responses in cells [31]. e levels of methionine, glycine, and phenylacetyl glycine in urine samples from CIA rats were decreased, indicating that glutathione metabolism was disordered in CIA rats.

Energy Metabolism.
e tricarboxylic acid cycle is the final energy-producing carbohydrate, lipid, and amino acid metabolic pathway and is the hub of carbohydrate, lipid, and amino acid metabolism [32]. e tricarboxylic acid cycle also provides small molecular precursors for other synthetic metabolic processes. For example, oxaloacetic acid is the precursor of aspartic acid, α-ketoglutaric acid is the precursor for glutamic acid synthesis, and some amino acids can be converted to sugar through this pathway. Creatine and its phosphorylated form are intermediates of energy metabolism, and a decrease in these molecules may be related to energy demand. Since phosphocreatine has a direct role in cellular energy transport, creatine can supply energy to the muscles in the form of stored creatine phosphate. As energy requirements increase, muscles can rapidly resynthesize ADP into ATP through creatine and creatine kinase systems. e levels of glutaric acid and α -ketoglutaric acid in urine samples from CIA rats decreased, indicating disordered energy metabolism in the model rats. Abnormal energy metabolism was found in AA rats, and α-ketoglutaric acid was used as a biomarker of RA.

Vitamin Metabolism.
e nicotinic acid, retinoic acid, methyl nicotinamide, coenzyme Q9, and α-CEHC levels in urine samples from CIA rats increased, while N1-methyl-4pyridone-3-formamide levels decreased. Retinoic acid and coenzyme Q9 are vitamin A and vitamin H molecules, respectively. Nicotinic acid, methyl nicotinamide, and N1-methyl-4-pyridone-3-formamide are nicotinamide components and belong to the B group vitamins. B vitamins are water-soluble and can promote the process of biological oxidation and metabolism. Tryptophan provides pyridine nucleotides to the liver through the nicotinamide adenine dinucleotide pathway to produce nicotinamide, which is converted to N-methyl nicotinamide by N-methyltransferase and then to N1-methyl-4pyridinone-3-formamide by aldehyde oxidase. a-CEHC can reflect vitamin E metabolism status in the body. Vitamin E is a fat-soluble antioxidant and is an important component of both cell membranes and lipoproteins. When vitamin E is lacking, free radicals are produced in the body that not only damage the structure and function of the cell membrane but also cause the immune system to decline and lead to abnormal metabolism. erefore, vitamins are important for maintaining and regulating normal metabolism. e change in vitamins in the urine of CIA rats indicates that RA will cause the immune system to decline and affect normal metabolic dysfunction. It can be inferred that RA may lead to decreased immune system activity and disordered metabolic function according. In CIA rats, the levels of α-CEHC and N1-methyl-4-pyridine-3-formamide returned to normal after CRME intervention, indicating that CRME can regulate vitamin E and vitamin B3 metabolic pathways and regulate the disease state of RA through vitamins.
Cholesterol, also known as dihydrocholesterol, is a sterol. Trihydroxycholestane induces apoptosis in malignant glioma cells by activating internal and external apoptotic pathways, and Bcl-2 family proteins play an important regulatory role in this process. Stigmasterol has strong anti-inflammatory effects and directly used as an antiinflammatory drug. Cholesterol is an important component of cell membranes and plasma lipoprotein and is the precursor of many important steroids including bile acid, adrenocortical hormone, estrogen, androgen, and vitamin D3. Most cholesterol can be converted into bile acid, and the imbalance of cholesterol metabolism can have adverse effects on the body.
Leukotriene A4 (LTA4) and leukotriene B4 (LTB4) are bioactive lipid mediators produced by 5-lipoxygenase metabolizing arachidonic acid. Leukotrienes are chemical mediators in some allergic, inflammatory, and cardiovascular diseases and play an important role in RA and gout. While the leukotriene content in the body should be significantly increased with inflammation, we found that it was decreased in CIA rats. e specific reasons for this require further study. After CRME intervention, the levels of 17β-estradiol-3-glucuronoside, cholestanol, trihydroxycholestane, stigmasterol, and 7-hydroxy-3-oxycholic acid returned to normal, indicating that CRME can treat RA by regulating steroid hormone and bile acid metabolism.

Other
Metabolism. Pyridinol has emerged as a biomarker of bone metabolism in recent years [33]. After an osteoporotic fracture, deoxypyridinoline shows a trend to increase in the urine. e observed decrease of the urine pyridinol content in CIA rats represents abnormal bone metabolism in vivo [34].
In summary, amino acids play an important role in the pathogenesis of RA. e amino acid metabolism changes observed in CIA rats may be the main cause of immune activation in the pathogenesis of RA. After intervention with CRME, 16 metabolites including benzoic acid, hippuric acid, hydroquinone, tyrosine, and indophenol returned to normal levels, indicating that CRME can treat RA by regulating amino acid, energy, and lipid metabolism.
Fifty-six potential biomarkers related to CIA and 20 metabolic pathways disturbed by CIA were preliminarily identified. Based on these results, the pathogenesis of RA and metabolic disorders was proposed. We describe that phenylalanine, tyrosine, and tryptophan biosynthesis (Impact 1), taurine and subtaurine metabolism (Impact 0.43), phenylalanine metabolism (Impact 0.41), glycine, serine, and threonine metabolism (Impact 0.29), pentose and glucose conversion (Impact 0.27), histidine metabolism (Impact 0.24), tyrosine metabolism (Impact 0.17), riboflavin metabolism (Impact 0.17), and tryptophan metabolism (Impact 0.09) were significantly disturbed. After the CRME intervention CIA model rats, 16 kinds of differential metabolites were regulated to the normal state, and the disordered metabolic spectrum returned to normal levels. e levels of 2-indolecarboxylic acid, hydroquinone, benzoic acid, and hippuric acid were significantly upregulated, while the levels of 17-β-estradiol-3-glucuronoside, cholestanol, 3hydroxycholestane, stigmasterol, tyrosine, D-ribose, N1methyl-4-pyridone-3-formamide, 2-diaminobutyric acid, monoisobutylphthalic acid, α-CEHC, and 7-hydroxyoxycholic acid were significantly downregulated. Together, these results show the urine metabolic spectrum reveals that the pathogenesis of CIA in rats is closely related to the amino acid, energy, and lipid metabolism. CRME has a certain regulatory effect on disorders of amino acid, vitamin, and lipid metabolism caused by CIA. e metabonomics results suggest that regulating amino acid, vitamin, energy, and lipid metabolism may be a potential target pathway for CRME in the treatment of RA.  Figure 6: Correlation networks of the perturbed metabolic pathways in collagen-induced arthritis urine models. Red means that the biomarker levels were increased in the model group. Green means that the biomarker levels were decreased in the model group.

Conclusions
is study employed an innovative approach to look for biomarkers and metabolic pathways related to the pathogenesis of RA by analyzing urine samples. On the basis of traditional metabonomics research, the potential biomarkers found were quantified using LC-MS, and the quantitative indexes were given. Commonly used clinical therapeutic drugs were used for methodological verification, and a new metabonomics method was developed to study the overall action mechanism of RA therapeutic drugs. e established metabolomics model was used to study the mechanism of CRME action against RA. e 56 metabolites in the urines might be associated with the treatment response of C. robustum, of which, 16 metabolites were found to be obviously changed after C. robustum treatment. ese endogenous metabolites such as tyrosine, alpha-CEHC, and 2-indolecarboxylic acid were assumed as potential biomarkers of RA, which are connected with various metabolic pathways such as amino acid, vitamin, energy, and lipid metabolism. e anti-RA effect of Caulophyllum robustum Maxim is exerted by regulating pathways of amino acid, vitamin, energy, and lipid metabolism. is is the first time that a metabolomic approach has been adopted to account for the anti-RA action of Caulophyllum robustum Maxim. e method of this study reflects the "holistic" characteristics of TCM and is expected to provide strong technical support for the study of the mechanism of CRME in the treatment of RA.

Data Availability
Baseline filter, peak identification, integration, retention time correction, peak alignment, and normalization were performed by the metabolomics processing software of the instrument Progenesis QI (Waters Corporation) using raw data to obtain a data matrix with retention time, mass-tocharge ratio, and peak intensity. e normalized urine sample data matrix was imported into the SIMCA-P + 14.0 software package (Umetrics, Umea, Sweden) for unsupervised principal components analysis (PCA) to observe the overall sample distribution and the stability of the analysis process. en, the supervised (orthogonal) partial least square method, (O)PLS-DA, was used to distinguish overall metabolic profile differences and identify metabolite differences between groups. Variables with variable important in projection (VIP) scores greater than 1 were considered differential variables. To prevent the model overfitting, the quality of the model was investigated using seven interactive verification cycles and 200 response sequencing tests. Screening the metabolite differences between groups (VIP > 1, p < 0.05, Fold Change (FC) > 1.2) was achieved using multidimensional analysis of PLS-DA and single-dimensional analysis (Student's t-test) methods. Differential metabolites were selected for qualitative analysis using http:// www.hmdb.ca/and self-built databases.

Conflicts of Interest
e authors declare no financial/commercial conflicts of interest.

Authors' Contributions
Shaowa Lü conceived, designed, and conducted experiments and revised the manuscript. Xu Dan performed the experiments, analyzed data, and interpreted results. Zhu Mingtao wrote the manuscript. Wang Qiuhong and Kuang Haixue designed the study and supervised research. Guo Yuyan and Li Guoyu provided suggestions. All the authors approved the final version of the manuscript.