Metabolic Proling in Mice Infected with Echinococcosis

Background: Echinococcosis is a severe zoonotic parasitic disease, which seriously affects the health of the hosts. The diagnosis of echinococcosis depends on objective detection of clinical symptoms. However, the patient is often in the late stage of the disease when the symptoms appear, limiting the early diagnosis of echinococcosis. The treatment and prognosis of the patients are seriously hampered due to long-term asymptomatic latency. Metabolomics is a new discipline developed in the late 1990s. It reects a series of biological responses in a pathophysiological process by demonstrating the changes in metabolism under the inuence of internal and external factors. When the organism is invaded by pathogens, the alteration in characteristics of metabolites in cells becomes exceedingly sensitive. Here, we used a liquid chromatography with tandem mass spectrometry (LC-MS/MS) system-based metabolomics approach to determine the molecular mechanism of cystic echinococcosis (CE) and to develop an effective method for CE diagnosis. Methods: Eight-weeks-old female Balb/c mice were divided into normal and Echinococcus granulosus infection groups. To develop the Echinococcus granulosus infection model, mice were infected with protoscolex. Six weeks later, the abdomen of mice was obviously bulged. An LC-MS/MS system-based metabolomics approach was used for the analysis of the liver and feces to reveal the metabolic proles of mice with echinococcosis. Results: We found that the metabolism of nucleotides, alkaloids, amino acids, amides, and organic acids in mice is closely interrelated with Echinococcus granulosus infection Conclusion: The metabolomics approach used in this study provides a reference for a highly sensitive and specic diagnostic and screening method for echinococcosis.


Background
Cystic echinococcosis (CE) or hydatidosis is a type of near-cosmopolitan zoonosis caused by Echinococcus granulosus sensu lato [1,2]. The life cycle of Echinococcus granulosus involves two mammalian hosts. For the adult tapeworms, carnivores (canines and cats) are the de nitive hosts, while ungulates and odontoids are intermediate hosts [3]. CE is prevalent in Western China, South America, Central Asia, Mediterranean, and East Africa [4]. The main risk factors are dogs and livestock [5][6][7][8].
Humans are not usually directly involved in the spread of CE. However, humans or intermediate hosts accidentally ingest eggs, which incubate in the intestines and release oncospheres [9]. The oncospheres are transported to the liver through the portal vein and lymphatic vessels where they settle and develop into larvae (hydatid cysts), rarely reaching the lungs, brains, bones or any other organ of a human or intermediate host [2,10]. The growth of CE cysts is very slow [11]. There are no obvious symptoms in the early stage of CE, with more than half of cysts expressing no change in size in 10-15 years [12]. The clinical symptoms, such as epigastric discomfort or loss of appetite, commonly appear when the cyst becomes more than 10 cm long, and gradually develop causing damage and dysfunction in the parasitic organs (mainly liver) [13,14]. Clinically, the CE patients are in the late stage of echinococcosis when they present in the hospital. The cyst is frequently misdiagnosed as a tumor [15]. Therefore, determination of a positive and effective approach to discriminate CE patients from healthy people is conducive to the early diagnosis and treatment of echinococcosis, which can signi cantly improve the survival rate of patients.
Metabolomics, a new high-throughput sequencing technology introduced in recent years, is an important branch of omics and it has great potential in drug toxicity or safety evaluation [16]. Metabolomics re ects many factors, such as gene changes, nutritional status, pathogenesis, natural environment changes, drug treatment, physiological response, and pathological characteristics [17][18][19][20][21]. It is a dynamic observation of the occurrence and development of diseases and the transformation process [22]. In metabolomic studies, small molecule metabolites are used as the research object. It can be used to study drug metabolism, nd differential metabolites, and explore the target and mechanism of action with the corresponding statistical analysis software using high sensitivity instruments (gas chromatographymass spectrometry, nuclear magnetic resonance, liquid chromatography-mass spectrometry) [23][24][25].
Metabolomics has been widely used in a variety of hepatopathy to identify potential early biomarkers and metabolic pathways [26],which is a feasible measure to illustrate host-parasite interactions [27][28][29]. The metabolic state and biochemical activity of cells or tissues can be directly re ected by metabolites such as amino acids, lipids, or sugars [30,31]. metabolic pro ling approaches have been widely used in various studies on diseases caused by atworms, Fasciola hepatica or Onchocerca volvulus [32][33][34][35]. Nevertheless, reports on the application of the LC-MS/MS-based metabolomics approach in echinococcosis is limited. In this study, we used a LC-MS/MS system-based metabolomics approach and multivariate statistical analyses to investigate the molecular mechanism of CE and to provide a potential valuable reference for the diagnosis of CE.

Establishment of hydatidosis model
Twenty 8-weeks-old female Balb/c mice, weighting about 18-22 g, were randomly divided into two groups: normal control group (n = 10) and Echinococcus granulosus infection group (n = 10). The mice were intraperitoneally injected with 2 ml PBS diluted protoscolex(containing 2000 protoscolex) to develop the Echinococcus granulosus infection model, The normal control group was intraperitoneally injected with the same 2 ml PBS. The liver and feces of both groups were obtained by dissection after 6 weeks of feeding. The mice models infected with Echinococcus granulosus were con rmed through routine pathological section staining.
The animal experiments were implemented in strict accordance with the laboratory animal management, with the approval from the Ningxia Medical University animal ethics committee and in accordance with the National Institute of Health guidelines for the Care and Use of Laboratory Animals.

Sample preparation
25mg liver and fecal Samples were weighed and transferred to a new clean EP tube. Add 500 µL extract solution (acetonitrile: methanol: water = 2:2:1) with isotopically labeled internal standard mixture to the sample. After 30 s of gentle vortex, the samples were homogenized at 35 Hz for 4 min and ultrasonicated in ice-water bath for 5 min. The procedure of homogenization and ultrasonication cycle was repeated three times. Then, the samples were incubated at -40°C for 1 hour and centrifuged at 4°C for 15 min at 12000 rpm. The ultimate resulting supernatant was transferred to a fresh and clean glass bottle for analysis. The quality control sample was prepared by mixing the supernatants of all samples of liver or feces in equal aliquot amounts. The MS/MS spectra were collected by QE HFX mass spectrometer in the control of the acquisition software (Xcalibur, Thermo). The acquisition software continuously evaluates the full scan MS spectrum in this mode. The conditions of ESI source were set as following: aux gas ow rate -10 Arb, capillary temperature -320°C,sheath gas ow rate -50 Arb, full MS resolution -60000, collision energy -10/30/60 in NCE mode, MS/MS resolution -7500, spray voltage -3.5 kV (positive) or -3.2 kV (negative), respectively.

Metabolomics statistical analysis
The original data were converted to mzXML format by using ProteoWizard, and processed by an in-house program, which was based on XCMS and was developed using R. Metabolites were annotated using an in-house MS2 database (BiotreeDB). The cutoff value for annotation was set to 0.3.

Statistical analysis
SPSS software (version 20.0 for Windows) was used to analyzed the tata.and have been expressed as mean ± SD. In all cases, P < 0.05 was considered to be signi cant.

Multivariate statistical analysis
In order to determine the metabolic changes in the liver and feces due to echinococcosis initiated following Echinococcus granulosus infection, two groups of samples were analyzed through principal component analysis (PCA). The metabolites of the liver were found to be moderate groupings between the infection and normal control groups under positive ion mode (POS) and negative ion mode (NEG) based on PCA ( Fig. 1A and 1B). The interpretation rates of the rst principal component (PC1) and second principal component (PC2) under POS were 33.5% and 17.6%, respectively, while under NEG, the respective interpretation rates were 30.9% and 15.6%. To better highlight the metabolic changes between the two groups, Orthogonal Projections to Latent Structures-Discriminant Analysis (OPLS-DA) was performed. The difference between the two groups was evident in the OPLS-DA score plots ( Fig. 1C and 1D). A similar tendency was observed in the metabolic changes in feces ( g. S1A, S1B, S1C, S1D). The group separation revealed that hydatid disease could cause obvious metabolic changes in the liver and feces of mice. All the samples were within the 95% con dence interval. In addition, the OPLS-DA models were supposed to be a good explanation for predictive powers Q 2 and variations R 2 Y and with permutation tests (Fig. 1E, 1F, S1E, and S1F).

Screening of differential metabolites
In order to distinguish metabolic markers between the infection and control groups, the data are presented in the form of the volcano plots. The results of the metabolic changes were determined based on variable importance in the projection (VIP) > 1 and P < 0.05. The metabolites with enormous changes have been represented with blue or red dots and larger circle shapes, located in the upper left or right corner of the volcano plots. From the metabolites of the liver, 248 and 131 distinct metabolic molecules were screened under POS and NEG ( Fig. 2A and 2B). In the feces, 201 and 58 various metabolites were screened under POS and NEG ( g. S2A and S2B). The top 15 metabolites with the most signi cantly different metabolism under POS and NEG in the liver and feces are shown in Table 1, Table 2, Table S1, and Table S2. In the livers of the infection group, the representative substances such as bile acid (deoxyviolaceinic acid), glycerides (2-O-(alpha-d-mannosyl)-d-glycerate), amino acids (glutaminylhistidine), and nucleotides (cytidine 2',3'-cyclic phosphate) were considerably upregulated, while nucleotides (inosinic acid and 5-uorodeoxyuridine monophosphate) were remarkably downregulated under POS and in NEG. Furthermore, in the feces of the mice with echinococcosis, metabolites such as nucleotides (dTMP), alkaloids (piperine), amino acids (d-pantethine, d-aspartic acid, and gamma-glutamylleucine), and amides (N-acetylhistamine) were upregulated, while nucleosides (5′methylthioadenosine), amino acids (3-methylcrotonylglycine, selenocystemic acid), organic acids (hexadecanedioic acid, dimethylmalonic acid, and ascorbic acid), and alkaloids (cytokinin b) were downregulated under POS and NEG (Table 1, 2, 3, and 4). The results suggest that the metabolism of nucleotides, alkaloids, amino acids, amides, and organic acids in mice is closely interrelated with Echinococcus granulosus infection.

Hierarchical clustering analysis of different metabolites
The discrepant metabolites by ltrating usually have similarity or complementarity of functions in biology, displaying similar or opposite expression characteristics among different experimental groups and the same metabolic pathway. The hierarchical clustering analysis of these characteristics of metabolites is helpful to classify the metabolites with the same characteristics into one category, as well as the variation characteristics of modern metabolites among the experimental groups.
For every sample, we calculated the Euclidean distance matrix to determine the quantitative value of the differential metabolites. The differential metabolites of the liver were clustered through the complete linkage method, and displayed in the thermodynamic diagram ( Fig. 3A and 3B). Through further evaluation, we selected 91 and 47 differential metabolic molecules under POS and NEG. Liver metabolites mainly included esters, amino acids, carbohydrates, and lipids. Compared with the normal control group, the expression of glyceric acid, acetylcysteine, myricetin, L-arabitol, L-asparagine, tryptophan, L-phenylalaine, racemethionine, D-alanine, D-proline, L-valine, and D-mannose were all increased in the infection group. Conversely, in the infection groups the levels of cholesterol sulfate, ribothymidine, ascorbic acid, alpha-linolenic acid, inosinic acid, adenine, thiamine, maltotetraose, creatinine, and imidazoleacetic acid were evidently decreased. Additionally, the levels of other metabolites were changed by various degrees. Fecal metabolic molecules were analyzed similarly; 107 and 23 differential molecules were selected under POS and NEG ( g. S3A and S3B). The expression of metabolic molecules such as medical acid, anandamide, piperine, N-carbamoyl putrescine, isoleucyl phenylalanine, D-aspartic acid, cytidine, and L-valine were sharply increased in the infection group. Besides, metabolic molecules such as trimethylaminoacetone, histidine, ophthalmic acid, hexadecanedioic acid, methylthioadenosine, traumatic acid, and phenylacetic acid showed a downward trend in the infection group. Coincidentally, L-valine was upregulated in the liver and feces in the infection group. The heat map ( Fig. 3A and 3B) shows that the different metabolites had good classi cation results.

Correlation analysis of different metabolites
The complex metabolic reactions and their regulation do not occur in isolation after the mice are infected with Echinococcus granulosus. When different molecules work together, their interaction and regulation eventually lead to systematic changes in metabolomics. We calculated the correlation coe cient of the quantitative values of different metabolites and expressed it as r. In this study, 91 and 47 differential metabolic molecules in the liver under POS and NEG were analyzed through Pearson method. There was a clear correlation between amino acid metabolism ( Fig. 4A and 4B). We also found that 107 and 23 differential molecules in the feces under POS and NEG were related ( g. S4A and S4B).

KEGG annotation of different metabolites
The analysis of the metabolic and regulatory pathways and systematic understanding of the changes in biological processes caused by changes in experimental conditions, pathogenesis of traits or diseases, and mechanism of drug action can be more comprehensive through KEGG annotation of different molecules. From the KEGG pathway map, we found that the differential metabolic molecules between the infection and control groups involved multiple pathways, including energy metabolism, material transport, signal transduction, and cell cycle regulation (Fig. 5A and 5B). In all, 46 pathways in the liver were involved with the differential metabolic molecules. In the feces, these differentially metabolized molecules involved 57 pathways ( g. S5A and S5B). For example, in the liver of the infected group, there were four differentially metabolized molecules that were enriched in purine metabolic pathways and three differentially metabolized molecules that were enriched in amino acid metabolic pathways. Similarly, in the feces of the infected group, three differential metabolic molecules were involved in pyrimidine metabolism, and two differential molecules were involved in arginine and proline metabolism. The more differential metabolic molecules involved in certain pathway, the more obvious the difference in the metabolic pathway between the infection and control groups.
3.6 Analysis of metabolic pathways of different metabolites KEGG annotation analysis revealed all the pathways involved in the differential metabolites. In order to determine whether these metabolic pathways were closely related to the experimental conditions, it was necessary to further analyze the metabolic pathways of the differential metabolites. Through analysis by synthesis of the pathways (including topological analysis and enrichment analysis), we could further explore the pathways and nd the key pathways with the highest correlation with metabolic differences. The results of metabolic pathway analysis were presented in bubble chart. In the liver, there were seven metabolic pathways under POS and NEG, three of which were in common. The metabolic pathways of tyrosine and tryptophan biosynthesis, phenylalanine, valine, leucine and isoleucine biosynthesis and phenylalanine metabolism were notably associated with the occurrence and development of hydatid disease (Fig. 6A and 6B). However, there were 15 metabolic pathways under POS and NEG, one of which is in common ( g.S6A and S6B), which revealed that pantothenate and CoA biosynthesis is also considerably related with echinococcosis.

Discussion
In this study, we analyzed the metabolic footprint and signi cant changes in metabolism, which revealed the substantial changes in the mice metabolome due to CE with the multivariate statistical analysis of liver and feces. In all, 138 metabolic molecules were selected to discriminate CE mice from healthy mice.
Seven metabolic pathways appeared under POS and NEG, which can be considered to be related with echinococcosis. The three common metabolic pathways were tyrosine and tryptophan biosynthesis, phenylalanine, valine, leucine and isoleucine biosynthesis and phenylalanine metabolism, which indicated that amino acid metabolism is considerably associated with the occurrence and development of hydatid disease. These metabolic differences may provide a novel viewpoint into the biological mechanisms that occur during CE infection. Interestingly, 130 metabolic molecules in feces were considered to distinguish between the infection and control groups. These molecules were enriched in 15 pathways, one of which is in common under POS and NEG. The pantothenate and CoA biosynthesis pathway indicated that the metabolism of fatty acid and pyruvate in mice with echinococcosis would be disturbed.
The main organ involved in echinococcosis is the liver. With the cyst parasitization and growth, it continuously obtains nutrition from the liver. When it grows large enough, it oppresses the liver to change its structure and function [1,36]. The results revealed that CE can lead to signi cant changes in amino acid metabolism. The level of circulating of amino acids indicate the equilibrium between muscle and liver metabolism [37]. It was con rmed by increased phenylalanine and tyrosine levels in the infection group. Tyrosine is the foremost product of Phenylalanine. It has been reported that the concentrations of these two metabolites are the same. The conversion of phenylalanine to tyrosine occurs uniquely in the liver [38,39]. The catabolism of amino acids occurs mainly in the liver [40]. Liver injury can lead to changes in amino acid metabolism, mainly manifested as the decrease in free branched chain amino acids and increase in free amino acids (phenylalanine, tryptophan and tyrosine) [41]. In contrast with the normal mice, the molecular metabolites of the liver and feces were related in infected mice. Among the rst 15 most differentially metabolized molecules, tolclofos methyl was increased in the liver and reduced in the feces, chlorfenson was downregulated in the liver and upregulated in the feces while aminosalicylate sodium anhydrous and Disul were both decreased in the liver and feces. In conclusion, the reason why amino acids (phenylalanine, tyrosine, tryptophan, valine, leucine, and isoleucine) were increased in the infection group may be related to liver dysfunction. We noticed that purine metabolism and alpha-linolenic acid metabolism were abnormal in CE mice, which suggests that other pathways may participate in the metabolism after hydatid infection as well.
The limitations of this study were as follows. First, the current results need more samples to verify and further improve its reliability and accuracy. Secondly, the comprehensive application of LC-MS with GC-MS or NMR will expand the coverage of metabolomics. Moreover, a variety of omics technologies (such as genomics and proteomics) can be cross-validated and may better support the experimental results. Finally, no effective method was determined to distinguish echinococcosis from other types of liver disease in this study. If the method is developed, it is essential to prevent clinical misdiagnosis because echinococcosis is similar to other hepatic diseases. The feasibility of using Fischer ratio and unique metabolic characteristics of echinococcosis to identify different liver diseases warrants further study.
The LC-MS method is suitable for metabolomics of hydatid disease due to its repeatability, multiple metabolite coverage in one measurement and short detection duration. The systematic research of multiple metabolites of small molecules is helpful in revealing the overall metabolic changes induced by Echinococcus granulosus. The early diagnosis of echinococcosis may be feasible by combining both imaging techniques and metabolomics. For example, chemical-exchange weighted magnetic resonance imaging techniques are used to yield maps weighted by metabolites of interest [42][43][44]. These maps are obtained by exchange with water and are characterized by signi cant enhancement of the signal and the observation of small metabolic changes that cannot be noticed by other imaging methods. Unfortunately, the common problem with these technologies is the lack of speci city. LC-MS-based metabolomics with speci city can be combined to enhance the analysis and better explain the data.

Conclusion
In our study, LC-MS-based metabolomics was used for investigating metabolic diversi cation of CE. It was con rmed that amino acid metabolism (phenylalanine, tyrosine, tryptophan, valine, leucine, and isoleucine) and pantothenate and CoA biosynthesis changed signi cantly in hydatid infection. The particular metabolic changes found in this study may provide new understanding of the molecular mechanism of CE and provide some meaningful clues for the early diagnosis and therapeutic intervention of CE.

Availability of data and materials
The data of the results of this study are included in this paper and its attached les.

Ethics approval and consent to participate
All the experiments were approved by the ethics committee of Ningxia Medical University and complied with the guidelines of all institutions and countries on experimental animals. The data obtained met the appropriate ethical requirements.

Consent for publication
Not applicable.

Competing interests
Page 10/13 The author states that they have no competing interests.
coe cient represents relevance. The Q2 and R2Y values re ect the model predictability and the fraction of explained variance, respectively. S2. volcano plots in feces (A and B) A and B. volcano plots, each dot in the volcano map represents a metabolite, the abscissa shows the Fold change value (take the logarithm of cardinal number 2), the ordinate represents the P-value of student's t-test (take the negative number of base logarithm of 10), and the size of the scatter represents the VIP value of OPLS -DA model, the larger the scatter, the greater the VIP value. S3. heat map in feces (A and B)