Regulation on BCAAs catabolism pathway plays the key role in cyclophosphamide-induced leucopenia BALB/c mice after the treatment of a typical Traditional Chinese Medicine of Lvjiao Buxue Granules

Background: Cyclophosphamide is a common tumor chemotherapy drug used to treat various cancers, but the resulting immunosuppression leads to leukopenia, which is a serious limiting factor in clinical application. Therefore, the introduction of immunomodulators as adjuvant therapy may help to reduce the hematological side effects of cyclophosphamide. Lvjiao Buxue Granules has been widely used in clinical treatment of gynecological diseases such as anemia and irregular menstruation, and recently, it has been found to increase the role of white blood cells, but its mechanism of action is still unclear. In this research, we applied the 1 H-NMR metabolomics approach to characterize metabolites in cyclophosphamide-induced leucopenia mice spleen, so as to fully understand the metabolic processes of leucopenia and improve the leukocyte function of Lvjiao Buxue Granules. Methods: Cyclophosphamide was used to establish the leucopenia mice with cancer chemotherapy and the content of white blood cells, red blood cells, hemoglobin, platelets, and other routine blood indexes were measured. The changes of endogenous metabolites in spleen analyzed by 1 H-NMR metabolomics technique were investigated the regulation effect of LBG in mice with leukopenia. Afterward, the chemical components-targets-differential metabolites network of Lvjiao Buxue Granules was constructed by the use of biological targets network, thus leucopenia-relevant metabolism pathways were dissected. Results: The blood routine parameters and organ indexes levels of leucopenia mice with cancer chemotherapy were improved by Lvjiao Buxue Granules. The metabolomic study revealed that 15 endogenous metabolites in mice spleen were considered as potential biomarkers of Lvjiao Buxue Granules for its protective effect. Integrated analysis of metabolomics and biological targets network indicated that Lvjiao Buxue Granules exerted the leukocyte elevation activity by inhibiting the branched-chain amino acids (BCAAs) degradation pathway and increasing the levels of valine, leucine and isoleucine. high dose of Lvjiao Buxue Granules group (LBG-H), and Diyu Shengbai tablets group (DST), each group was consists of 8 mouse. The mice in LBG-L, LBG-M, or LBG-H groups were administered Lvjiao Buxue Granules (3 g/kg, 6 g/kg, 12 g/kg) suspension daily, and the mice in the DST group were administered Diyu Shengbai tablets (0.14 g/kg) suspension daily. Mice in the control and model groups received an equal volume of vehicle orally. The 4T1 breast cancer model was established in which the tumor grew to approximately 5*5 cm 2 . The ve cyclophosphamide-induced groups would be injected with cyclophosphamide in the dose of 80 mg/kg by intraperitoneal to 4T1 breast cancer model on the 1st day, 3rd day, 5th day and 7th day individually. The treatment lasted for 7 days, and the state of the mice was observed and the weight was measured daily.


Background
Cyclophosphamide is the most common cancer chemotherapy agent and used in anticancer for various types of cancer, especially breast cancer [1,2]. Unfortunately, immunosuppression induced by cyclophosphamide causes the occurrence of leukopenia, which is a seriously limiting factor in clinical application [3,4]. Therefore, introducing immunomodulatory agents as supportive therapy might be useful in alleviating hematotoxicity side effects of cyclophosphamide.
After constant efforts, the treatment against side effects of chemotherapy still leaves much to be desired. In recent years, Prescriptions made from a combination of multiple drugs have been considered as a promising therapeutic strategy for improving anti-tumor effects and reducing the side effects of chemotherapy drugs [5]. Traditional Chinese medicine (TCM) prescriptions are usually made up of some different kinds of herbs, with the advantages of low toxicity and multiple targets [6]. Through overall and multi-target therapies, it has a comprehensive therapeutic effect in multi-factorial diseases.
In the past few decades, as good medicine for de ciency of Qi and blood, fatigue and weakness, Lvjiao Buxue Granules has been widely used in the clinical treatment of anemia irregular menstruation and other gynecological diseases, and recently it has been discovered elevating leukocytes effect [7]. However, the research on drug e cacy and mechanism of action is relatively rare. Lvjiao Buxue Granules generally comprised of 6 herbs: Asini Corii Colla, Astragali Radix, Codonopsis Radix, Rehmanniae Radix Praeparata, Atractylodis Macrocephalae Rhizoma and Angelicae Sinensis Radix at a ratio of 36:30:30:20:15:10.
Speci c immune function can be stimulated by Asini Corii Colla in cyclophosphamide-induced mice [8].
The compatibility of Astragali Radix and Angelicae Sinensis Radix (Such as Danggui Buxue Tang is the famous prescription and is formed of Astragali Radix and Angelicae Sinensis Radix.) could increase the quantity of bone marrow mononuclear cells and peripheral blood leukocyte, enhance immunity and improve microcirculation [9]. Some other previous studies have also demonstrated that Codonopsis Radix, Rehmanniae Radix Praeparata and Atractylodis Macrocephalae Rhizoma exhibited effects of leukocyte elevation and lmmuno-enhancement [10,11]. Moreover, all the above herbs in the Lvjiao Buxue Granules are commonly and safely used in the formula of TCM, and rare adverse reactions were reported in clinical applications. However, up to now, the action mechanism of Lvjiao Buxue Granules in the treatment of leucopenia remained poorly understood.
In recent years, metabolomics is used to clarify the scienti c effects related to the effectiveness mechanism, material basis and compatibility of TCM, and it will provide the technical support for the evaluation of the effectiveness of TCM, the basis of prescription substances and the understanding essence of TCM syndromes [12][13][14]. With the continuous development of metabolomics technology, it has been increasingly concentrated on metabolomics to reveal the pharmacodynamics and mechanisms of TCM prescriptions, such as Xiao Yao San and Baihe Dihuang Tang [15,16]. In this research, we applied the 1 H-NMR metabolomics approach to characterize metabolites in cyclophosphamide-induced leucopenia mice spleen, so as to fully understand the metabolic processes of leucopenia and improve the leukocyte function of Lvjiao Buxue Granules.
In this study, we rstly constructed a leucopenia mice model with cancer chemotherapy, in which mice manifest similar syndromes to those of patients with leukopenia in the clinic. We then tested metabolites in mice spleen and compared the levels of endogenous metabolites between healthy mice, leucopenia mice and leucopenia mice treated with Lvjiao Buxue Granules, and revealed the related metabolic pathways as well. Next, this study was applied to Biological targets network analysis methods to generate a leucopenia "chemical components-targets-differential metabolites" regulatory network centered. Finally, through a comprehensive analysis of the biological target network and metabolomics results, a regulatory network for the "herb-chemical-constituent-targets-pathway-metabolites" was constructed to characterize the pharmacological effect of Lvjiao Buxue Granules in the treatment of leukopenia. The strategies on characterizing the main constituents of Lvjiao Buxue Granules and its therapeutic effects and mechanisms of leucopenia will provide a theoretical basis for the clinical application of this traditional Chinese medicine prescription.  Shengbai tablets group (DST), each group was consists of 8 mouse. The mice in LBG-L, LBG-M, or LBG-H groups were administered Lvjiao Buxue Granules (3 g/kg, 6 g/kg, 12 g/kg) suspension daily, and the mice in the DST group were administered Diyu Shengbai tablets (0.14 g/kg) suspension daily. Mice in the control and model groups received an equal volume of vehicle orally. The 4T1 breast cancer model was established in which the tumor grew to approximately 5*5 cm 2 . The ve cyclophosphamide-induced groups would be injected with cyclophosphamide in the dose of 80 mg/kg by intraperitoneal to 4T1 breast cancer model on the 1st day, 3rd day, 5th day and 7th day individually. The treatment lasted for 7 days, and the state of the mice was observed and the weight was measured daily.

Sample collection and determination
After 1h of the last treatment on the 7th day, 0.4 mL blood samples were collected into 1.0 mL tube with EDTA within via the orbital blood. Animal blood analyzer (HEMAVET950) was applied to evaluate peripheral blood routine parameters of 400 μL whole blood: white blood cell count (WBC), neutrophil count (NE), lymphocyte count (LY), monocytes count (MO), red blood cell count (RBC), hemoglobin (HGB), red blood cell volume (HCT), platelet count (PLT), mean red cell volume (MCV) and mean corpuscular hemoglobin (MCH). On the 8th day, mice were sacri ced, and spleen, thymus and liver tissues were immediately weighed and collected. The calculation formula of the organ index is as follows: organ index = organ weight (mg)/bodyweight (g). Quickly transfer the spleen to a refrigerator at -80 ℃ and use it for metabolomics analysis.

Sample preparation for NMR measurements
After thawing the spleen tissue, take about 40 mg, cut it (on ice), add 650 μL of MeOH and H 2 O (v/v, 2:1) to a 2 mL centrifuge tube, and homogenize and extract twice on an ice bath. The homogenate was centrifuged at 4 ℃, 13 000 r·min -1 for 15 min. The supernatants were combined, transferred to a 2 mL centrifuge tube and blown with nitrogen. The dried sample was dissolved in 700 μL of phosphate buffer (pH 7.40, containing D 2 O, 0.1 mo/L, Na 2 HPO 4 /Na H 2 PO 4 , 0.01% TSP), centrifuge at 13 ℃, 13,000 r·min -1 for 20 min, 600 μL supernatant was transferred into a 5 mm NMR tube for 1 H NMR analysis.

Metabolomics analysis
The 1 H NMR spectral data were collected on a Bruker 600 MHz AVANCE III NMR spectrometer(Bruker, Germany). The sample was the Noesygppr1d sequence to suppress the water peak. The number of scans was 64 scans, and each scan required an acquisition time of 2.654s. The speci c parameters were as follows: spectral width was 12 345.7 Hz; spectrum size was 65 536 data points; pulse width (PW) was 30° (12.7 μs); fourier transform LB was 0.3 Hz and relaxation delay time was 1.0 s. The 1 H NMR spectrum of the spleen was corrected for chemical shifts using TSP (δ0.00) as the standard. The spectrum in the region of δ0.60 to 9.49 was divided into 0.01 equal widths and integrated. All resulting integration data are "mass" normalized to eliminate weight differences of spleen tissue.
Simca-P 14.1 (Umetrics, Sweden) was used to perform multivariate data analysis. Firstly, by principal component analysis (PCA) of the normalized data, to identify the degree of dispersion between the control group and the model group, and the outliers were eliminated. Next, partial least-squares discriminant analysis (PLS-DA) was used to distinguish the differences in metabolic pro les between the control, model and drug groups. Orthogonal-projection to latent structure-discriminant analysis (OPLS-DA) was used to nd differential metabolites between the control group and the model group. Finally, ANOVA analysis was performed on the metabolites in SPSS 16.0 software, with VIP 1 and P 0.05 as differential metabolites.

Biological targets network analysis
All components of Lvjiao Buxue Granules were collected from the TCMSP database. For all ingredients, the initial structure formats (e.g., mol2 and SDF) were transformed into a uni ed SDF format using the Open Babel toolkit (version 2.4.1). The ingredients with suitable OB≥30% and DL≥0.18 were chosen as candidate ingredients for further research, which is used as a selection criterion for the ingredients in the traditional Chinese herbs. After ADME screening, some ingredients that did not meet the three screening criteria were also selected because of their high content and high biological activity. All databases and software mentioned above are public.
The PharmMapper server was used for potential target prediction analysis. Metabolite data were imported in Metascape, a plugin of Cytoscape 3.7.1 and subjected to metabolic enzyme analysis. Finally, Cytoscape 3.7.1 software was used to construct the "herb-chemical constituent-targets-pathwaymetabolite" regulatory network of Lvjiao Buxue Granules for leucopenia treatment.

Determination of the levels of BCKDHA and ACADS
In order to further verify the results of biological targets network, the content of the key rate-limiting enzymes BCKDHA (branched chain keto acid dehydrogenase E1, alpha polypeptide) and ACADS (Acyl-CoA Dehydrogenase Short Chain) on the BCAAs degradation pathway were determined. The levels of BCAAs and the degradation rate of them could be affected by the content of these enzymes.
The levels of BCKDHA and ACADS in the liver tissue lysates were determined by ELISA kits according to manufacturer instructions.

Effect of LBG on blood routine parameters and organ indexes in cyclophosphamide-treated mice
The biochemical indexes of the peripheral blood were observed by evaluating the blood toxicity of cyclophosphamide. The parameters of WBC, NE, LY, MO, RBC, HGB, HCT and MCH in the model group were signi cantly lower than those in the control group(P 0.01; Table 1), and the PLT parameters were signi cantly higher than those of the control group(P 0.05; Table 1). The increase and decrease of these peripheral blood routine parameters are the main diagnostic criteria of leukopenia, indicating that the model of leukopenia was successfully replicated. Compared to the model group, the expressions of WBC, NE, LY, MO, RBC, HGB and MCH in the LBG-L, LBG-M and LBG-H were signi cantly increased (P 0.05; Table 1), and the e cacy was better than the DST group. The results suggested that Lvjiao Buxue Granules exerted well effect on leukocyte elevation activity. Table 1 Changes in indexes of blood routine examination on leucopenia mice with LBG treatment (n = 8)

LBG regulates metabolic disorders in leukopenia mice
To excavate the speci c marker metabolites that resulted in differentially expressed metabolite changes in mice caused by cyclophosphamide-induced leukopenia, the supervised multivariate methods PLS-DA and OPLS-DA were used for processing. (Fig. 3). The PLS-DA pattern recognition analysis of all the group trends was shown in Fig. 3A, in which the control group was completely separated from the model group, and the LBG-L, LGB-M, LGB-H groups were separated from the model group, with a tendency closer to the control. It was indicated that the leucopenia model was successfully established and the Lvjiao Buxue Granules exhibited an excellent leukocyte elevation effect.
The validity of the analysis was performed by using 200 permutation tests, in which all R 2 and Q 2 values were lower than the original ones. (Intercepts: R 2 =0.854, Q 2 =0.649) (Fig. 3B). Furthermore, differential metabolites between control and the model groups were discovered by OPLS-DA, which is a supervised pattern recognition method that could improve the discovery effect of differential metabolites (Fig. 3C). The corresponding loading (S+V)-plot with color-coded was illustrated in Fig. 3D, and the metabolites contributed to the leukocyte elevation effect were identi ed by corresponding (S+V)-plots and statistical analysis.
The disturbed metabolite variances of the different groups could be related to the metabolic changes associated with leucopenia and the increase of leukocytes in Lvjiao Buxue Granules. The changes of the differential metabolites between control and model groups in mice spleen were shown in Fig. 4. Compared with the control group, the elevated levels of TMAO, pyruvate, GPC, taurine and glutamate in the model group were evident in the spleen samples from. Additionally, lower levels of choline, myoinositol, tyrosine, valine, iso-leucine, PC, α-glucose, leucine and phenylalanine in the spleen of the model group compared with the control group we observed. The changes in these endogenous metabolites are considered to be a direct result of the leucopenia resulting from DST. Meaningfully, the levels of metabolites were regulated by LBG treatment (Fig. 4), suggesting that LBG may play a leukocyte elevation effect by leveling off the divergences of the metabolites.
Correlation analysis of differential metabolites, blood routine parameters and organ indexes The blood routine parameters and organ indexes are well-known for the evaluation of leucopenia. An analysis of the correlation between the blood routine parameters, organ indexes and differential metabolites can be used to screen for speci c biomarkers. Pearson's correlation analysis method was used for the investigation of the relationship among the differential metabolites, blood routine parameters and organ indexes of all groups. The correlation map was shown in Fig. 5A, where the color re ects the correlation strength and sign, the WBC, RBC, HGB, HCT, NE, LY, MO and MCH presented negative correlations with glutamate, pyruvate, aspartate, GPC, taurine and positive correlations with valine, iso-leucine, leucine, choline, PC, myo-inositol, α-glucose, tyrosine and phenylalanine. Besides, the uctuation of the spleen index showed correlations with differential metabolites (P 0.01) (Fig. 5A). The correlation network of blood routine parameters, organ indexes and differential metabolites based on Pearson's was shown in Fig. 5B, which could be served as differential metabolites for assessing the leucopenia and the effect of LBG.

Establishment of "different metabolites-enzymes/genes" metabolic network
The metabolic networks involved in a slice of enzymes and genes were constructed by using the Netscape plug-in running on Cytoscape 3.7.1, that the internal correlation of the differential metabolites based on enzyme or gene levels could be better understood (Fig. 6). As a result, 154 candidate genes or enzymes related to differential metabolites were tentatively found out, and they will be served as targets for subsequent biological targets network analysis for the construction of targets-metabolites interactions network (Fig. 6). Finally, to map the metabolic pathway of identi ed differential metabolites from leucopenia-associated researches, the enrichment analysis was performed and metabolic pathways were obtained for further investigation.

Biological targets network and metabolomics integration analysis Screening active compounds of Lvjiao Buxue Granules
In the current study, two ADME-related models, including OB and DL, were employed to screen for active ingredients in Lvjiao Buxue Granules. After ADME screening, some ingredients that did not meet the screening criteria were selected because of their high content and high biological activity. First of all, a gross of 87, 134, 76, 55 and 125 candidate ingredients were obtained from Astragali Radix, Codonopsis Radix, Rehmanniae Radix Praeparata, Atractylodis Macrocephalae Rhizoma, Angelicae Sinensis Radix, respectively. The ingredients were retrieved from these ingredients via the ADME parameters and literature con rmation. Consequently, a total of 35 chemical ingredients of Lvjiao Buxue Granules were ltered out for further analysis (Table S1).
Construction of "chemical components-targets-differential metabolites" regulatory network and correlative Pathways 490 targets of the 35 components in Lvjiao Buxue Granules were predicted using the Pharm Mapper server, and 28 of them were closely related to differential metabolites of leukopenia with a total frequency of 496 (Table S2). Cytoscape software was used to establish the "Chemical Components-Targets-Differential Metabolites" regulatory network of Lvjiao Buxue Granules, which was an indication that the correlations of 35 compounds, 28 metabolite-associated target proteins, and 11 differential metabolites were presented in Fig. 7. Analysis of the "Chemical Compositions-Targets" correlation revealed that multiple compounds could act on the same target, and multiple targets could be affected by the same compound. For instance, the L-amino-acid oxidase could be the target of ferulic acid, caffeic acid, biatractylolide simultaneously, while the compound of catalpol could act on Aldo-keto reductase family 1 member B1, Eukaryotic translation initiation factor 4E-binding protein 3, and Glucose-6-phosphatase at the same time (Fig. 7).
To assess the molecular mechanisms of Lvjiao Buxue Granules effects on leucopenia, the Cytoscape with its plugin ClueGO was utilized for KEGG pathway analysis, followed by the analysis of target-related pathway. A total of 9 signi cant pathways are predicted (in the term of P-Value 0.05), among which the Valine, leucine and isoleucine degradation was the most closely related one (Table 3). According to the results of biological targets network and metabolomics, a preliminary prediction of the mechanism by which Lvjiao Buxue Granules plays leukocyte elevation activity could be as follows: Lvjiao Buxue Granules can inhibit the Valine, leucine and isoleucine degradation and add the levels of Valine, leucine and isoleucine (Fig. 4A).

Molecular docking veri cation
The chemical-components-targets-differential metabolites regulatory network was constructed and the valine, leucine and isoleucine degradation pathway were focused that the correlations of 35 compounds, 7 metabolite-associated target proteins, and 3 differential metabolites were presented in Fig. S3.
Furthermore, the Systems Dock (http://systemsdock.unit.oist.jp/iddp/home/index) method was used to evaluate the binding potential between selected Valine, leucine and isoleucine degradation pathway targets and the chemical components with top 7 degrees (Degree≥5) (Fig. S2). The docking score of SystemsDock can directly indicate the protein-ligand binding potential. The 3D structures and PDB ID of the above 7 selected targets were gathered from the PDB database (https://www.rcsb.org/) ( Table  S3).The results showed that 6 targets (ACAA2, BCAT1, BCAT2, DLD, HADHA and IL4I1) had 3D structures, while the 3D structure of ACAA1 did not exist. As shown in Fig. 8, the docking scores of 95 pairs of targetcompound combinations were mostly greater than native ligand, which showed that they possessed great binding activity.
Experimental validation of BCAAs catabolism pathway by ELISA To further con rm the results of the biological targets network, we experiment to verify the BCAAs catabolism pathway. Two key and irreversible reactions of BCAAs catabolism require the participation of branched-chain keto acid dehydrogenase (BCKDH) complex and acyl-CoA dehydrogenase (ACAD). The levels of BCKDHA and ACADS were signi cantly increased (P 0.01) in leukopenia model mice (Fig. 8).
After oral administration of Lvjiao Buxue Granules, the levels of BCKDHA and ACADS were signi cantly Cyclophosphamide can degrade aldehydes, phosphoramides, nitrogen mustard and acrolein.

Organ indexes
As the most important immune organs in mammals, spleen and thymus are the places where immune cells grow and proliferate [27]. The developmental status of the spleen and thymus directly affects immune function and disease resistance [28][29][30]. The results of this experiment showed that the spleen and thymus indexes of the high, medium and low dose Lvjiao Buxue Granules groups were signi cantly higher than those of the cyclophosphamide-induced group. It is suggested that Lvjiao Buxue Granules could resist the toxic effects of cyclophosphamide on the development of spleen and thymus. Since the cyclophosphamide is metabolized in the liver which is the tissue represents a major target for cyclophosphamide-induced tissue damage [31,32]. The liver is the main drug metabolism organ in mammals. Metabolized cyclophosphamide causes damage to mitochondria and damage to cellular respiration, which affects the onset of lipid peroxidation and an increase in reactive oxygen species [33,34]. In this work, signi cant increases in Liver index were observed through the impact of cyclophosphamide, and this altered were regulated by Lvjiao Buxue Granules' treatment. Hence, Lvjiao Buxue Granules plays a role in increases in the Liver index through reduced lipid peroxidation and reactive oxygen species production.

BCAAs catabolism
The branched-chain amino acids (BCAAs) catabolism have been focused on a great deal of diseases, especially liver cirrhosis, renal failure, sepsis and cancer (Fig. 9)  In the current study, we used 19 active ingredients of Lvjiao Buxue Granules as a probe to conduct molecular docking with potential targets of BCAAs catabolism, and the molecular docking showed that they possessed great binding activity. A study revealed that astragaloside , formononetin, calycosin and ferulic acid showed promoting hematopoiesis through regulating cyclin-related proteins, promoting cell cycle transformation, and promoting HSC proliferation [41]. Another study showed that caffeic acid can down-regulated the expression of TLR-2 and HLA-DR, and inhibited the production of cytokines, then exerted an immunomodulatory action on human monocytes [42]. Studies of receiver biases suggest that the chemical composition of the Lvjiao Buxue Granules has a better activity of immunity regulation and hematopoiesis. The results of molecular docking showed that there were complex interactions between them, which showed the characteristics of multi-components and multi-targets. In this study, the levels of valine, leucine and isoleucine decreased signi cantly in the cyclophosphamide-induced mice, which was a suggestion that the leukopenia was associated with the branched-chain amino acids catabolism. Moreover, the branched-chain amino acids were positively correlated with WBC, NE, LY and MO in the leucopenia mice, which were an indication that they played key roles in the progression of leucopenia.
Compared with cyclophosphamide-induced mice, the levels of valine, leucine and isoleucine can be improved signi cantly in the Lvjiao Buxue Granules groups, which was a suggestion that the Lvjiao Buxue Granules could play a key role of the leukocyte elevation effect by inhibiting the BCAAs catabolism. In this study, two key enzymes, the BCKDHA and ACADS enzyme levels in a leukopenia model mice, signi cantly increased in the BCAAs catabolism pathway. Abnormal expression of BCAAs degrading enzymes causes BCAAs overly decomposed in leukopenia model mice. Lvjiao Buxue Granules can reverse the levels of two enzymes, thereby improving the abnormal metabolism of BCAAs.
In addition, it should be noted that whether Lvjiao Buxue Granules affects the activity of some enzymes, such as rate-limiting enzyme in the BCAAs catabolism, there is a need for further investigations. In the future, molecular biology research, isotope-labeled tracking experiments or targeted metabolomics technology could be used to further explore the biological connotation of BCAAs catabolism and their biological roles in leukopenia and the leukocyte elevation effect of Lvjiao Buxue Granules. At the same time, the key information of a biological system will be obtained with the integration of metabolites, enzymes and genes altered under a situation of leukopenia.

Conclusion
In this research, the metabolomics and the biological targets network approach were integrated to analyze and to evaluate the leukocyte elevation effects of Lvjiao Buxue Granules on mice with cyclophosphamide-induced leucopenia. It is a demonstration that the regulation on the BCAAs catabolism pathway could play a key role in cyclophosphamide-induced leucopenia BALB/c mice after the treatment of a typical Traditional Chinese Medicine of Lvjiao Buxue Granules. Phosphocholine.

Declarations
Availability of data and materials Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.

Consent for publication
Not applicable.
Ethics approval and consent to participate Experiments were approved by the Ethics Committee on Animal Experiments of the Shanxi University. The 1H NMR spectra of mice spleen tissue in C, M and LBG-M groups.  The relative content of the heatmap of differential metabolites in mice spleens. The ribbon -3~3: represents the relative content of the differential metabolites from low to high.

Figure 5
Correlation analysis of differential metabolites, blood routine parameters and organ indexes. (A)The correlation map of differential metabolites, blood routine parameters and organ indexes. Red and blue represent positive and negative correlations respectively (*P<0.05, **P<0.01). (B) Correlation network of differential metabolites, blood routine parameters and organ indexes based on Pearson's correlation coe cients. Purple, yellow and green node represent differential metabolites, blood routine parameters and organ indexes. Red and green line represent positive and negative correlations. Line thickness re ects the magnitude of the correlation coe cients.

Figure 6
The metabolic networks involved in enzymes, genes and related-compound were established based on the differential metabolites. Red, light-blue and pink node represent differential metabolites, enzymes/genes and related-compound. "chemical components-targets-differential metabolites" regulatory network of LBG. Yellow nodes represent the chemical components, green nodes represent the targets, and pink node represent differential metabolites.