Metabolic Profiles in Madin–Darby Canine Kidney Cell Lines Infected with H3N2 Canine Influenza Viruses

Abstract

Virus replication and host cell growth require host cell metabolic networks to provide energy and precursors for the synthesis of macromolecules. The aim of this study was to investigate the most direct changes in energy metabolism and small-molecule metabolism of Madin–Darby canine kidney (MDCK) cells infected with H3N2 canine influenza virus (CIV) and to determine whether small metabolites contribute to the pathogenesis of CIV. To study the metabolomics of MDCK cells infected with H3N2 CIV, we used liquid chromatography–tandem mass spectrometry combined with multivariate statistical analysis. The results showed that 798 positive ions were detected, among which 33 were upregulated and 11 were downregulated, and 406 negative ions were detected, among which 33 were upregulated and 9 were downregulated. Through Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis, we found that these differentially expressed molecules were mainly concentrated in the steroid hormone biosynthesis, amino sugar and nucleotide sugar metabolism, sphingolipid metabolism, vitamin B6 metabolism, cysteine and methionine metabolism, vitamin digestion and absorption, arginine and proline metabolism, biosynthesis of amino acids, and folate biosynthesis metabolic pathways. These pathways are involved in energy metabolism and nucleic acid and protein synthesis, which are essential for virus replication. Our experimental data suggest that H3N2 CIV infection reconstitutes/influences cellular metabolic processes, which in turn may contribute to viral replication. These findings are important for the development of enzyme inhibitors or metabolites for the identification of antiviral drugs. In addition, understanding the metabolic interaction between CIV and host cells is also very important for the complex pathogenicity of CIV, providing certain guidance for the treatment of canine influenza.

Introduction

Canine influenza virus (CIV), an influenza A virus, which is the influenza A virus, can cause cough, runny nose, fever, and other clinical symptoms in dogs. There are two main subtypes of CIVs: equine H3N8 and avian H3N2 (5,21,32). The influenza A virus genome consists of eight single-stranded RNA segments of negative polarity that encode ribonucleoproteins (PA, PB1, PB2, NP), hemagglutinin (HA), neuraminidase (NA), nonstructural proteins (NS1 and NS2), and matrix proteins (M1 and M2). In addition, the nonstructural proteins PA-X, PA-N182, PA-N155 (18), PB1-F2 (41), PB1-N40 (42), M42 (17), NS3 (30), and PB2-S1 have also been discovered (2).

The metabolome is defined as the total of all small-molecular-weight metabolites (<1,000 Da) that an organism or cell produces during a specific physiological period. Metabolomics, which is an omics technology that emerged after genomics, transcriptomics, and proteomics, is an important part of systems biology. Metabolomics aims to investigate the changes in metabolites in biological systems (cells, tissues, etc.) after stimulation or disturbance or with change over time. By determining the different metabolites between the experimental group and the control group, metabolomics studies biological processes. Metabolomics is a more phenotypic omics compared with transcriptomics or proteomics; it is an extension of transcriptomics and proteomics, and can reflect the physiological state of organism more directly and accurately. Inferring biochemical mechanisms from metabolomics data is of great significance, and it can provide new insights into the pathophysiological mechanisms of diseases and enhance diagnostic and prognostic tools (10,13,27). For example, people have successfully found that fatty acids are an important marker of abnormal lipid metabolism in diabetic patients by metabolomic analysis (39,44); nucleosides are important tumor markers as a class of metabolites (26,45). Metabolomics also has important research and therapeutic value in cardiovascular disease (19), post-transplant detection (31) and inflammatory bowel disease (23).

Previous metabolomics reports have shown great promise in assessing the mechanisms of viral infection by human cytomegalovirus, dengue virus, herpes simplex virus-1, classical swine fever virus, Zika virus, and respiratory syncytial virus (1,6,8,15,20,37,38). Metabolomics research has been instrumental to diagnose and characterize certain metabolic diseases, such as cancer, diabetes, obesity, and metabolic syndrome (4,7,14,16,22,34,48). Currently, common metabolic pathways that have been well studied include the tricarboxylic acid cycle, glycolysis, and arachidonic acid/inflammatory pathways.

To explore the complex interactions between CIV and host cells, transcriptomics and proteomics methods have been used to analyze the relevant cellular mechanisms (11,33,35,46,47). However, the systemic changes of metabolites in CIV-infected cells remain unknown. In this study, we used liquid chromatography–tandem mass spectrometry (LC-MS/MS) (9,40) and characterized the metabolomics of Madin–Darby canine kidney (MDCK) cells infected with H3N2 CIV with multivariate statistical analysis.

Materials and Methods

Viruses and cells

This study did not involve animal experiments and IRB is not applicable. The CIV H3N2 (A/canine/Guangdong/B/2013) was isolated from the nasal swab of a pet dog with severe respiratory symptoms in Guangdong Province. MDCK cells were obtained from the American Type Culture Collection (ATCC) and propagated in Dulbecco's modified Eagle's medium (DMEM) with 10% fetal bovine serum.

Titration of viral infectivity and HA activity

To measure viral titers, cells cultivated in 96-well plates were infected with a series semilogarithmic dilution (10⁻³, 10^−3.5…10⁻⁸) of CIV. The HA titer was measured after the cytopathic lesion stopped developing (usually 72–96 h). The TCID₅₀ titer was calculated by the Reed-Muench method. A specific reference (No. CNICSOP07013) was made to the titration of tissue cell infection volume by the National Influenza Center of China.

Virus growth curve

MDCK cells with 80% confluence in a 25-cm² cell culture flask were infected with CIV (multiplicity of infection [MOI] = 1 and 0.1). Subsequently, 100 μL of viral supernatant was harvested at 6, 12, 18, 24, 30, 36, and 42 hpi (repeated three times). The virus growth curve was drafted according to viral titers.

Sample preparation and LS-MS/MS analysis

MDCK cells were cultured in a 10-cm Petri dish. When the cells reached 80% confluence, the cells were inoculated with H3N2 CIV (MOI = 1) for 24 h (the time point of 24 h was chosen as the fastest reproduction rate of the strain was from 18 to 24 h, and it had almost completed the proliferation at about 24 h. The control group received the same treatment without virus, six duplicate samples in each group), the cells were washed repeatedly three times with precooled PBS, then the cells were removed from the Petri dish, and 1 mL of precooling 60% methanol solution (Thermo Fisher) was added to quench intracellular metabolism. The cells were then centrifuged at 4°C for 1 min at 1,000 g low speed. The supernatant was removed, and the cells were collected in 2-mL cryopreserved tubes. The samples were placed in frozen liquid nitrogen for 15 min and then stored at −80°C until analysis. The samples were then sent to Novogene (Beijing, China) for untargeted metabolomics analysis.

Cells (100 mg) were individually grounded with liquid nitrogen, and 100 μL of the homogenates was resuspended with precooled 100% methanol (−20°C) followed by well vortexing. The samples were incubated at −20°C for 60 min and then centrifuged at 14,000 g and 4°C for 15 min. Then, the supernatants were transferred to a fresh microcentrifuge tube and dried under vacuum in a centrifugal evaporator. The dried metabolite pellets were redissolved in 80% methanol and analyzed by LC-MS/MS.

LC-MS/MS (43) analyses were performed using a Vanquish UHPLC system (Thermo Fisher) coupled with an Orbitrap Q Exactive HF-X mass spectrometer (Thermo Fisher) operating in the data-dependent acquisition mode. Samples were injected onto an Accucore HILIC column (100 × 2.1 mm, 2.6 μm) using a 20-min linear gradient at a flow rate of 0.3 mL/min. The eluants of the positive polarity mode were eluant A (0.1% formic acid, FA, in 95% acetonitrile, ACN, 10 mM ammonium acetate) and eluant B (0.1% FA in 50% ACN, 10 mM ammonium acetate). The eluants of the negative polarity mode were eluant A (95% ACN, 10 mM ammonium acetate, pH 9.0) and eluant B (50% ACN, 10 mM ammonium acetate, pH 9.0). The solvent gradient was set as follows: 2% B, 1 min; 2–50% B, 16.5 min; 50–2% B, 2.5 min. The Q Exactive HF-X mass spectrometer was operated in positive/negative polarity mode with a spray voltage of 3.2 kV, capillary temperature of 320°C, sheath gas flow rate of 35 arb, and aux gas flow rate of 10 arb.

The quality control in the LC-MS/MS analyses

Metabolomes are susceptible to external factors and change rapidly. Therefore, data quality control (QC) is an essential step to obtain reproducible and accurate metabolome results. The Pearson correlation coefficient between QC samples was calculated based on the peak area value. The higher the correlation of QC samples (R2 is closer to 1), the better the stability of the whole detection process and the higher the data quality.

Screening differential metabolites

Because of the characteristics of high dimensionality and high correlation between variables, traditional univariate analysis cannot quickly and accurately explore the potential information in the data. Therefore, multivariate statistical methods, such as principal component analysis (PCA) and partial least squares discrimination analysis (PLS-DA), should be used in the metabolomics data analysis. Based on the maximum retention of the original information, downscaling and regression analyses were carried out on the collected multidimensional data, followed by screening of differential metabolites and subsequent analysis.

Differential metabolite analysis

In this experiment, cluster analysis of differential metabolites, correlation analysis of differential metabolites, and enrichment analysis of differential metabolite Kyoto Encyclopedia of Genes and Genomes (KEGG) (24) were used to conduct in-depth analysis of the data.

Results

Growth kinetics of H3N2 in MDCK cells

To determine the fastest growth time of CIV (GD2014) in MDCK cells, MDCK cells were infected by CIV (MOI = 1 and MOI = 0.1), and the TCID50 of CIV was detected at the time points of 6, 12, 18, 24, 30, 36, and 42 h, and the infectivity (TCID50/mL) of CIV was plotted. The results showed that the infection measures differ by a factor of 10, with almost no significant difference at the late stage of infection, and the growth curves after cells are infected with different doses are similar, possibly because MDCK cells are very suitable for the growth of influenza virus and the fastest growth time of CIV was 18–24 h after the infection of cells (Fig. 1).

FIG. 1.

Replication kinetics of H3N2 in MDCK cells. The data were analyzed using one-way ANOVA (p < 0.05). Data are presented as the mean ± SD of independent triplicate experiments. MDCK, Madin–Darby canine kidney.

Screening differential metabolites

PCA was used to observe the overall distribution trend between the two groups of samples. The figure clearly shows that the control group and H3N2 group are divided into two parts with obvious differences (Fig. 2).

FIG. 2.

PCA (A is the positive ion mode, and B is the negative ion mode). In the figure, abscissa PC1 and ordinate PC2 represent the scores of principal components ranked first and second, respectively, and scatter points of different colors represent samples of different experimental groups. The ellipse is a 95% confidence interval (when the number of biological repeats is less than 4, a 95% confidence ellipse cannot be shown). PCA, principal component analysis. Color images are available online.

PLS-DA is a supervised discriminant statistical method. In this method, partial least squares regression (3) was used to establish the relationship model between metabolite expression and sample category to predict the sample category. The PLS-DA model of each comparison group was established, and the model evaluation parameter (R2, Q2) was obtained by sevenfold cross-validation. The closer R2 and Q2 are to 1, the more stable and reliable the model is. Through PLS-DA, the positive ion mode had R2 = 0.99 and Q2 = 0.73, and the negative ion mode had R2 = 0.97 and Q2 = 0.7, indicating that our data model with differential expression was well established (Fig. 3).

FIG. 3.

PLS-DA scatter point diagram and sorting validation diagram (A is the positive ion mode, and B is the negative ion mode). The scatter point diagram is obtained, and the abscissa is the score of the sample on the first principal component; the ordinate is the score of the sample on the second principal component; R2Y represents the interpretation rate of the model, Q2Y was used to evaluate the prediction ability of the PLS-DA model, and when R2Y is greater than Q2Y, the model is well established. PLS-DA, partial least squares discrimination analysis. Color images are available online.

Differential expression metabolite analysis of MDCK cells infected with H3N2

In the PLS-DA model, the variable importance in the projection (VIP) value of the first principal component was adopted. The VIP value represents the contribution rate of different metabolites in different groups. Fold change (FC) is the ratio of the mean of repeated quantitative values of each metabolite in the comparison group. In addition, combined with the p-value of the t-test to find the differential expression metabolites, the threshold value was set as VIP >1.0, difference multiple FC >1.5 or FC <0.667, and p-value <0.05. There were 798 positive metabolites, with 44 significant differences in H3N2 versus control, and 33 upregulated and 11 downregulated metabolites. Four hundred six negative metabolites were identified, with 42 significant differences compared with the control group, among which 33 were upregulated and 9 were downregulated (Supplementary Data S1).

Differential metabolite volcano diagram

The volcanic figure shows differences in metabolites of the overall distribution of horizontal expressions of multiple changes in metabolites in different groups (log₂ FC). The ordinate shows a significant difference level (-log₁₀ p–value). Each dot on the volcanic figure represents a metabolite, and a significant increase in metabolites in the red dots and a significant decrease in metabolites in the green dots. The size of the dots represents the VIP value. It can be seen from the figure that several metabolites increase most significantly, including N(6)-methyladenosine, 4-o-(beta-l-araf)-cis-l-hyp, 2′-o-methyladenosine, n-acetylneuraminic acid (Fig. 4, Supplementary Data S1).

FIG. 4.

Volcano diagram of differential metabolites (A is the positive ion mode, and B is the negative ion mode). Red dots represent significant upregulation, green dots represent significant downregulation, and gray dots represent no significant difference. Color images are available online.

Cluster analysis of differential metabolites

Cluster analysis was used to determine metabolic patterns of metabolites under different experimental conditions. Metabolites with similar metabolic patterns may have similar functions or participate in the same metabolic process or cell pathway. Therefore, the function of some metabolites can be inferred by clustering metabolites with the same or similar metabolic patterns. Hierarchical cluster analysis was carried out on the obtained metabolites of each group, and the differences in metabolic expression patterns between and within the two groups were obtained. As can be seen from the figure, the control group and the experimental group are obviously divided into two categories, with significant differences and very high experimental repeatability (Fig. 5).

FIG. 5.

Heat diagram of differential metabolite clustering (A is the positive ion mode, and B is the negative ion mode). The longitudinal is the clustering of samples and the transverse is the clustering of metabolites. The shorter the cluster branches, the higher the similarity (the corresponding names of metabolite numbers in the figure are shown in Table 1). Color images are available online.

Table 1.

The Corresponding Names of Metabolite Numbers

Metabolite numbers	Compound name or ChemSpider ID	Molecular formula
Com_2968_pos	(1S,2S)-5-Bromo-3-phenoxy-3,5-cyclohexadiene-1,2-diol	C12 H11 Br O3
Com_4123_pos	Coumestrol	C15 H8 O5
Com_716_pos	Diphenyl sulfoxide	C12 H10 O S
Com_540_pos	Hexose	C6 H12 O6
Com_781_pos	2-Amino-1,3,4-octadecanetriol	C18 H39 N O3
Com_4378_pos	(3beta)-7,11-Dioxolanostan-3-yl acetate	C32 H52 O4
Com_1157_pos	1-stearoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine zwitterion	C41 H80 N O8 P
Com_4174_pos	Androstanolone	C19 H30 O2
Com_784_pos	(+/−)-Camphor	C10 H16 O
Com_1582_pos	2-Decylfuran	C14 H24 O
Com_216_pos	1,3-Dihydroxy-2-propanyl (9Z)-9-tetradecenoate	C17 H32 O4
Com_2466_pos	Eglumetad	C8 H11 N O4
Com_2131_pos	Pyridoxal	C8 H9 N O3
Com_2857_pos	9,12-Hexadecadienoylcarnitine	C23 H41 N O4
Com_2530_pos	Lys-phe	C15 H23 N3 O3
Com_1891_pos	N-Palmitoyl-L-tyrosine	C25 H41 N O4
Com_3406_pos	(2R)-2,3-Dihydroxypropyl nonadecanoate	C22 H44 O4
Com_520_pos	Palmitoyl carnitine	C23 H45 N O4
Com_1706_pos	Stearoyl carnitine	C25 H49 N O4
Com_773_pos	Myristoyl glycine	C16 H31 N O3
Com_574_pos	D-(−)-Glutamine	C5 H10 N2 O3
Com_3374_pos	2-Acetamido-2-deoxyglucose	C8 H15 N O6
Com_109_pos	2,3-Dehydro-2-deoxy-N-acetylneuraminic acid	C11 H17 N O8
Com_160_pos	8-Methyl-8-azabicyclo[3.2.1]oct-3-yl (3S)-1,2-dithiolane-3-carboxylate	C12 H19 N O2 S2
Com_686_pos	N-Acetylneuraminic acid	C11 H19 N O9
Com_1780_pos	4-O-(beta-L-Araf)-cis-L-Hyp	C10 H17 N O7
Com_683_pos	N(6)-Methyladenosine	C11 H15 N5 O4
Com_3209_pos	(2E)-2,3-Diphenyl-3-(2-pyridinyl)acrylonitrile	C20 H14 N2
Com_3978_pos	S-Adenosyl-L-methionine	C15 H22 N6 O5 S
Com_1418_pos	3-(phosphoacetamido)-L-alanine	C5 H11 N2 O7 P
Com_3628_pos	Marinobufagenin	C24 H32 O5
Com_2986_pos	Benzoic acid, 4-[4-[4-(1,1-dimethylethyl)phenyl]-2-hydroxybutoxy]	C21 H26 O4
Com_1063_pos	(9Z)-9-Cycloheptadecen-1-ol	C17 H32 O
Com_1239_pos	2-(p-Isopropylphenyl)propene	C12 H16
Com_2603_pos	10-[(5E)-3,4-Dihydroxy-5-isopropyl-5-hepten-2-yl]-5,6-dihydroxy-7a,9a-dimethylhexadecahydro-3H-benzo[c]indeno[5,4-e]oxepin-3-one	C29 H48 O6
Com_865_pos	PALGLY	C18 H35 N O3
Com_1571_pos	Mallotochromene	C24 H26 O8
Com_1558_pos	Psychosine	C24 H47 N O7
Com_1996_pos	b,e-Carotene-3,3′-diol	C40 H56 O2
Com_4309_pos	(14R,26R)-17,20,23-Trihydroxy-2,40-dimethyl-26-[(10-methylundecanoyl)oxy]-17,23-dioxido-11,29-dioxo-12,16,18,22,24,28-hexaoxa-17lambda ∼5∼,23lambda ∼5∼-diphosphahentetracontan-14-yl 17-methyloctadecanoate	C66 H128 O17 P2
Com_3140_pos	(14R,26R)-17,20,23-Trihydroxy-2,39-dimethyl-14-[(10-methylundecanoyl)oxy]-17,23-dioxido-11,29-dioxo-12,16,18,22,24,28-hexaoxa-17lambda ∼5∼,23lambda ∼5∼-diphosphatetracontan-26-yl 20-methylhenicosanoate	C68 H132 O17 P2
Com_2997_pos	Roxane	C19 H28 N2 O4
Com_2643_pos	all-trans-4,4′-diapophytofluene	C30 H46
Com_1220_pos	1,26-Hexacosanediyl (2E,2′E)bis[3-(4-hydroxy-3-methoxyphenyl)acrylate]	C46 H70 O8
Com_781_neg	(2R)-1-{[(2-Aminoethoxy)(hydroxy)phosphoryl]oxy}-3-hydroxy-2-propanyl (5Z,8Z,11Z,14Z,17Z)-5,8,11,14,17-icosapentaenoate	C25 H42 N O7 P
Com_170_neg	Lysophosphatidylethanolamine (0:0/22:6(4Z,7Z,10Z,13Z,16Z,19Z))	C27 H44 N O7 P
Com_416_neg	3,6-Anhydro-1-O-[2-(dodecanoyloxy)ethyl]-2,4,5-tris-O-(2-hydroxyethyl)hexitol	C26 H50 O10
Com_403_neg	3alpha,7alpha,12alpha-trihydroxy-5beta-cholestan-26-al	C27 H46 O4
Com_486_neg	1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine	C39 H76 N O8 P
Com_256_neg	Dppe	C37 H74 N O8 P
Com_550_neg	1-[(9Z)-hexadecenyl]-2-[(9Z)-hexadecenoyl]-sn-glycero-3-phosphocholine	C40 H78 N O7 P
Com_771_neg	(2R)-3-{[(2-Aminoethoxy)(hydroxy)phosphoryl]oxy}-2-[(1Z)-1-hexadecen-1-yloxy]propyl (6Z,9Z,12Z)-6,9,12-octadecatrienoate	C39 H72 N O7 P
Com_470_neg	1-stearoyl-sn-glycero-3-phosphoethanolamine	C23 H48 N O7 P
Com_464_neg	2,3,4,5-tetrachloro-4′-biphenylol	C12 H6 Cl4 O
Com_228_neg	Fludrocortisone acetate	C23 H31 F O6
Com_433_neg	2-Heptadecylfuran	C21 H38 O
Com_258_neg	DOA	C22 H42 O4
Com_174_neg	Lisofylline	C13 H20 N4 O3
Com_29_neg	Taurohyocholic acid	C26 H45 N O7 S
Com_22_neg	Thiolutin	C8 H8 N2 O2 S2
Com_766_neg	(2R)-3-{[(2-Aminoethoxy)(hydroxy)phosphoryl]oxy}-2-hydroxypropyl (9Z)-9-hexadecenoate	C21 H42 N O7 P
Com_456_neg	2-[(11Z,14Z)-icosadienoyl]-sn-glycero-3-phosphoethanolamine	C25 H48 N O7 P
Com_122_neg	Neopterin	C9 H11 N5 O4
Com_412_neg	3{[(2,3Dihydroxypropoxy)(hydroxy)phosphoryl]oxy}-2-hydroxypropyl stearate	C24 H49 O9 P
Com_25_neg	Tetrofosmin	C18 H40 O4 P2
Com_411_neg	25-hydroxy-cholesterol	C27 H46 O2
Com_410_neg	Difurfuryl monosulfide	C10 H10 O2 S
Com_135_neg	2-(n-Tetradecyl)cyclobutanone	C18 H34 O
Com_397_neg	3-Hydroxy-1-phenyl-1-icosanone	C26 H44 O2
Com_310_neg	Batilol	C21 H44 O3
Com_515_neg	1-hexadecanal	C16 H32 O
Com_211_neg	g-Tokoferol	C28 H48 O2
Com_202_neg	Homaline	C30 H42 N4 O2
Com_94_neg	Pazoxide	C12 H10 Cl2 N2 O2 S
Com_802_neg	(25R)-3-Oxocholest-4-en-26-al	C27 H42 O2
Com_421_neg	2′-O-Methyladenosine	C11 H15 N5 O4
Com_124_neg	N-Acetylneuraminic acid	C11 H19 N O9
Com_322_neg	Arabinosylhypoxanthine	C10 H12 N4 O5
Com_318_neg	Avagacestat	C20 H17 Cl F4 N4 O4 S
Com_402_neg	3b-Hydroxy-5-cholenoic acid	C24 H38 O3
Com_386_neg	3β-Androstanediol	C19 H32 O2
Com_204_neg	Hexose	C6 H12 O6
Com_1_neg	ZMP	C9 H15 N4 O8 P
Com_521_neg	1-alpha,24R,25-Trihydroxyvitamin D2	C28 H44 O4
Com_497_neg	1-oleoyl-sn-glycero-3-phospho-D-myo-inositol	C27 H51 O12 P
Com_40_neg	Sarmentoloside	C29 H44 O11

Correlation analysis of differential metabolites

There is a synergistic or mutually exclusive relationship between different metabolites. For example, if the variation trend of certain metabolites is the same, it is a positive correlation. Contrary to the change trend of some metabolites, it was negatively correlated. The purpose of differential metabolite correlation analysis is to check the consistency of metabolite and metabolite change trends and to analyze the correlation between metabolites by calculating the Pearson correlation coefficient between all metabolites.

As the linear relationship between the two metabolites increases, the positive correlation tends to 1, and the negative correlation tends to −1. In addition, the significance of correlation analysis of metabolites was statistically tested. A significance level p-value <0.05 was selected as the threshold of significance correlation. The correlation analysis of metabolites in the difference table showed that Cationic: Com_109_pos (2,3-Dehydro-2-deoxy-N-acetylneuraminic acid), Com_160_pos (8-methyl −8-azabicyclo[3.2.1]oct-3-yl (3S) −1,2-dithiolane-3-carboxylate), Com_683_pos (N(6)-methyladenosine), Com_1780_pos (4-o-(beta-l-araf)-cis-l-hyp), Com_686_pos (n-acetylneuraminic acid), Com_3209_pos ((2E)-2,3-diphenyl-3-(2-pyridinyl)acrylonitrile) have a high correlation; Anionic: Com_310_neg (batilol), Com_411_neg (25-hydroxy-cholesterol), Com_515_neg (1-hexadecanal), Com_135_neg (2-(n-tetradecyl)cyclobutanone), Com_211_neg (g-Tokoferol) is medium high correlation and Com_433_neg (2-Heptadecylfuran) with Com_258_neg (DOA) is medium high correlation, which may be related to virus replication (Fig. 6).

FIG. 6.

Correlation of differentially expressed metabolites (A is the positive ion mode, and B is the negative ion mode). The highest correlation was 1; the full positive correlation is indicated in red; the lowest correlation was −1; and the full negative correlation is indicated in blue. The part without color represents a p-value >0.05. The figure shows the correlation of differential metabolites of the top 20 ranked by p-value from smallest to largest (the corresponding names of metabolite numbers in the figure are shown in Table 2). Color images are available online.

Table 2.

The Corresponding Names of Metabolite Numbers

Metabolite numbers	Compound name or ChemSpider ID	Molecular formula
Com_109_pos	2,3-Dehydro-2-deoxy-N-acetylneuraminic acid	C11 H17 N O8
Com_160_pos	8-Methyl-8-azabicyclo[3.2.1]oct-3-yl (3S)-1,2-dithiolane-3-carboxylate	C12 H19 N O2 S2
Com_683_pos	N(6)-Methyladenosine	C11 H15 N5 O4
Com_1780_pos	4-O-(beta-L-Araf)-cis-L-Hyp	C10 H17 N O7
Com_686_pos	N-Acetylneuraminic acid	C11 H19 N O9
Com_3209_pos	(2E)-2,3-Diphenyl-3-(2-pyridinyl)acrylonitrile	C20 H14 N2
Com_1706_pos	Stearoyl carnitine	C25 H49 N O4
Com_3374_pos	2-Acetamido-2-deoxyglucose	C8 H15 N O6
Com_2466_pos	Eglumetad	C8 H11 N O4
Com_3406_pos	(2R)-2,3-Dihydroxypropyl nonadecanoate	C22 H44 O4
Com_1571_pos	Mallotochromene	C24 H26 O8
Com_1558_pos	Psychosine	C24 H47 N O7
Com_773_pos	Myristoyl glycine	C16 H31 N O3
Com_781_pos	2-Amino-1,3,4-octadecanetriol	C18 H39 N O3
Com_865_pos	PALGLY	C18 H35 N O3
Com_2968_pos	(1S,2S)-5-Bromo-3-phenoxy-3,5-cyclohexadiene-1,2-diol	C12 H11 Br O3
Com_716_pos	Diphenyl sulfoxide	C12 H10 O S
Com_1996_pos	b,e-Carotene-3,3′-diol	C40 H56 O2
Com_4378_pos	(3beta)-7,11-Dioxolanostan-3-yl acetate	C32 H52 O4
Com_4123_pos	Coumestrol	C15 H8 O5
Com_421_neg	2′-O-Methyladenosine	C11 H15 N5 O4
Com_124_neg	N-Acetylneuraminic acid	C11 H19 N O9
Com_416_neg	3,6-Anhydro-1-O-[2-(dodecanoyloxy)ethyl]-2,4,5-tris-O-(2-hydroxyethyl)hexitol	C26 H50 O10
Com_403_neg	3alpha,7alpha,12alpha-trihydroxy-5beta-cholestan-26-al	C27 H46 O4
Com_1_neg	ZMP	C9 H15 N4 O8 P
Com_310_neg	Batilol	C21 H44 O3
Com_411_neg	25-hydroxy-cholesterol	C27 H46 O2
Com_515_neg	1-hexadecanal	C16 H32 O
Com_802_neg	(25R)-3-Oxocholest-4-en-26-al	C27 H42 O2
Com_433_neg	2-Heptadecylfuran	C21 H38 O
Com_322_neg	Arabinosylhypoxanthine	C10 H12 N4 O5
Com_135_neg	2-(n-Tetradecyl)cyclobutanone	C18 H34 O
Com_464_neg	2,3,4,5-tetrachloro-4′-biphenylol	C12 H6 Cl4 O
Com_771_neg	(2R)-3-{[(2-Aminoethoxy)(hydroxy)phosphoryl]oxy}-2-[(1Z)-1-hexadecen-1-yloxy]propyl (6Z,9Z,12Z)-6,9,12-octadecatrienoate	C39 H72 N O7 P
Com_497_neg	1-oleoyl-sn-glycero-3-phospho-D-myo-inositol	C27 H51 O12 P
Com_781_neg	(2R)-1-{[(2-Aminoethoxy)(hydroxy)phosphoryl]oxy}-3-hydroxy-2-propanyl (5Z,8Z,11Z,14Z,17Z)-5,8,11,14,17-icosapentaenoate	C25 H42 N O7 P
Com_410_neg	Difurfuryl monosulfide	C10 H10 O2 S
Com_258_neg	DOA	C22 H42 O4
Com_486_neg	1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine	C39 H76 N O8 P
Com_211_neg	g-Tokoferol	C28 H48 O2

Enrichment analysis of differential metabolite KEGG

The bubble map of the KEGG enrichment pathway was drawn according to the enrichment results (only the top 20 results were shown). The results show that different metabolites are mainly concentrated in metabolic pathways such as energy metabolism, and nucleic acid and protein synthesis. Including are hormone biosynthesis (hormones can be regulated by the metabolic activity of various tissues and cells to influence the physiological activities of the body), metabolism of amino sugars and sugar nucleotides (which can directly influence the nucleic acid and protein synthesis), metabolism of vitamin B6 (vitamin B6 group can be quickly converted into auxiliary enzymes pyridoxal phosphate and pyridoxamine phosphate, the relationship between these two kinds of enzymes and protein metabolism is very close), cysteine and methionine metabolism (methionine is one of the essential amino acids to constitute the human body, and participate in protein synthesis. Because it cannot be produced in the body itself, it must be obtained from the outside. Methionine metabolism into cysteine, maintaining their balance, is very important for protein synthesis), vitamin absorb (vitamins are a kind of regulating substances and play an important role in the metabolism), arginine and proline metabolism (arginine is an essential amino acid in the body, the main end products of metabolism are proline and ornithine, and maintaining the balance of arginine and proline is very important for protein synthesis), amino acid biosynthesis (amino acids are the basic unit of protein synthesis and very important for protein synthesis), and folic acid biosynthesis and metabolic pathways (in organisms, the synthesis of proteins, nucleotides, pantothenic acids, and the methylation of molecules all require the participation of a carbon unit. Folic acid is an extremely important cofactor that mediates the transfer of a carbon unit. Folic acid is mainly involved in the de novo synthesis of purines and pyrimidines) (Fig. 7 and Supplementary Data S2).

FIG. 7.

Bubble diagram of KEGG enrichment (A is the positive ion mode, and B is the negative ion mode). The x-coordinate in the figure is x/y (the number of differential metabolites in the corresponding metabolic pathway/the total number of identified metabolites in this pathway). The higher the value, the higher the concentration of differential metabolites in this pathway. The color of the points represents the p-value value of the hypergeometric test. The smaller the value is, the more reliable and statistically significant the test is. The size of the dots represents the number of differential metabolites in the corresponding pathway, and the larger the number, the more the differential metabolites in the pathway. KEGG, Kyoto Encyclopedia of Genes and Genomes. Color images are available online.

Discussion

Material metabolism and energy metabolism are the most direct ways to affect life. Virus replication is dependent on the energy and biosynthesis precursors provided by host cell metabolism. Studying the biological processes involved in differential metabolites can more directly and accurately reflect the physiological state of an organism and reveal its biological activity mechanism.

Through untargeted metabolomics analysis, this experiment found a large number of differentially expressed metabolites, which were mainly enriched by folate (vitamin B9) biosynthesis—a coenzyme of the carbon unit transferase system in the biochemical reaction in vivo, acting as a carbon unit transporter, which is involved in the synthesis of purine and thymine, which leads to further synthesis of DNA and RNA. Folate is involved in amino acid metabolism and acts as a carbon unit carrier in the mutual conversion between glycine and serine, histidine and glutamic acid, and homocysteine and methionine (36). Steroid hormone biosynthesis—steroid hormone-induced gene activation plays a definite role in maintaining life and regulating sexual function and in organism development and immune regulation (25). Vitamin B6 metabolism—vitamin B6 serves as a cofactor for many proteins and enzymes. Vitamin B6 is essential for serine de novo biosynthesis (28,29). Sphingolipid metabolism—sphingolipids are part of the cellular structure and are important signaling molecules that are closely related to cell growth, aging, meiosis, maturation, and death (12). Amino sugar and nucleotide sugar metabolism—biosynthesis of amino acid pathways, cysteine and methionine metabolism, and arginine and proline metabolism—is essential for energy metabolism and nucleic acid and protein synthesis. Interestingly, in cancer, there are significantly upregulated molecules in this pathway.

In this experiment, through the growth curve of CIV, we found that the fastest growth period of CIV in MDCK cells was between 18 and 24 h, and the metabolite change was the largest at 24 h. Therefore, this experiment analyzed the metabolite change at 24 h when MDCK cells were infected with CIV.

In this study, MDCK cells were infected with H3N2 CIV, and changes in cell metabolism were detected. Dogs were not used because metabolomics requires consistency in sample handling, and dogs have individual differences that do not reflect the metabolic differences caused by H3N2 CIV in all dogs. However, consistency is not found in the analysis of multiple dogs. Moreover, more than 10 animal samples are required, which does not conform to animal welfare. With the continuous improvement and optimization of its methodology, metabolomics research will become a more efficient and accurate way to diagnose diseases.

Footnotes

Authors' Contributions

P.T. and P.Z. carried out the data analysis and drafted the article. W.X. prepared materials for the experiments. G.L. and S.L. conceived the study. All authors read and approved the final article.

Author Disclosure Statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Funding Information

This work was supported by the Guangdong Provincial Natural Science Foundation under Grant (Number 2017A030310367).

Supplementary Material

Supplementary Data S1

Supplementary Data S2

References

Birungi

, Chen

, Loy

, et al. Metabolomics approach for investigation of effects of dengue virus infection using the EA.hy926 cell line. J Proteome Res, 2010; 9:6523–6534.

Boergeling

, Rozhdestvensky

, Schmolke

, et al. Evidence for a novel mechanism of influenza virus-induced type I interferon expression by a defective RNA-encoded protein. PLOS Pathog, 2015; 11:e1004924.

Boulesteix

, and Strimmer

. Partial least squares: a versatile tool for the analysis of high-dimensional genomic data. Brief Bioinform, 2006; 8:32–44.

Cirulli

, Guo

, Leon Swisher

, et al. Profound perturbation of the metabolome in obesity is associated with health risk. Cell Metab, 2019; 29:488–500.e2.

Crawford

, Dubovi

, Castleman

, et al. Transmission of equine influenza virus to dogs. 2005; 310:482–485.

Cui

, Pang

, Lee

, et al. Serum metabolome changes in adult patients with severe dengue in the critical and recovery phases of dengue infection. PLoS Negl Trop Dis, 2018; 12:e0006217.

Dai

, Arceo

, Arnold

, et al. Metabolomics of oncogene-specific metabolic reprogramming during breast cancer. Cancer Metab, 2018; 6:5.

Diop

, Vial

, Ferraris

, et al. Zika virus infection modulates the metabolomic profile of microglial cells. PLoS One, 2018; 13:e0206093.

Dunn

, Broadhurst

, Begley

, et al. Procedures for large-scale metabolic profiling of serum and plasma using gas chromatography and liquid chromatography coupled to mass spectrometry. Nat Protoc, 2011; 6:1060–1083.

10.

Fiehn

. Metabolomics—the link between genotypes and phenotypes. Plant Mol Biol, 2002; 48:155–171.

11.

, Luo

, Ye

, et al. Integrated lung and tracheal mRNA-Seq and miRNA-Seq analysis of dogs with an avian-like H5N1 canine influenza virus infection. Front Microbiol, 2018; 9:303.

12.

Gault

, Obeid

, and Hannun

. An overview of sphingolipid metabolism: from synthesis to breakdown. Adv Exp Med Biol, 2010; 688:1–23.

13.

German

, Watkins

, and Fay

. Metabolomics in practice: emerging knowledge to guide future dietetic advice toward individualized health. J Am Diet Assoc, 2005; 105:1425–1432.

14.

Gorad

, Ellingsen

, Bathen

, et al. Identification of metastasis-associated metabolic profiles of tumors by 1H-HR-MAS-MRS. Neoplasia, 2015; 17:767–775.

15.

Gou

, Zhao

, Yuan

, et al. Metabolic profiles in cell lines infected with classical swine fever virus. Front Microbiol, 2017; 8:691.

16.

, Zhao

, Huang

, et al. Very low carbohydrate diet significantly alters the serum metabolic profiles in obese subjects. J Proteome Res, 2013; 12:5801–5811.

17.

Helen

. Wise ECHB. Identification of a novel splice variant form of the influenza A virus M2 ion channel with an antigenically distinct ectodomain. PLoS Pathog, 2012; 8:e1002998.

18.

Jagger

, Wise

, Kash

, et al. An overlapping protein-coding region in influenza A virus segment 3 modulates the host response. Science, 2012; 337:199–204.

19.

Joanne

, Henrik

, Elaine

, et al. Rapid and noninvasive diagnosis of the presence and severity of coronary heart disease using 1H-NMR-based metabonomics. Nat Med, 2002; 8:1439–1444.

20.

Joshua Munger

SUBH

. Dynamics of the cellular metabolome during human cytomegalovirus infection. PLoS Pathog, 2006; 2:e132.

21.

, Shi

, Jiao

, et al. Avian-origin H3N2 canine influenza A viruses in Southern China. Infect Genet Evol, 2010; 10:1286–1288.

22.

, Li

, Wang

, et al. Progress in metabonomics of type 2 diabetes mellitus. Molecules, 2018; 23:1834.

23.

Marchesi

, Holmes

, Khan

, et al. Rapid and noninvasive metabonomic characterization of inflammatory bowel disease. J Proteome Res, 2007; 6:546–551.

24.

Minoru

, and Susumu

. KEGG: Kyoto Encyclopedia of Genes and Genomes. Nucleic Acids Res, 2000; 28:27–30.

25.

Nancy

. Steroid hormones regulate gene expression posttranscriptionally by altering the stabilities of messenger RNAs. Biol Reprod, 2005; 6:1290–1296.

26.

Odunsi

, Wollman

, Ambrosone

, et al. Detection of epithelial ovarian cancer using1H-NMR-based metabonomics. Int J Cancer, 2005; 113:782–788.

27.

, and Voit

. Strategies for comparing metabolic profiles: implications for the inference of biochemical mechanisms from metabolomics data. IEEE/ACM Trans Comput Biol Bioinfor, 2017; 14:1434–1445.

28.

Ramos

, Pras-Raves

, Gerrits

, et al. Vitamin B6 is essential for serine de novo biosynthesis. J Inherit Metab Dis, 2017; 40:883–891.

29.

Rosenberg

, Ischebeck

, and Commichau

. Vitamin B6 metabolism in microbes and approaches for fermentative production. Biotechnol Adv, 2017; 35:31–40.

30.

Selman

, Dankar

, Forbes

, et al. Adaptive mutation in influenza A virus non-structural gene is linked to host switching and induces a novel protein by alternative splicing. Emerg Microbes Infect, 2012; 1:1–10.

31.

Serkova

, Zhang

, Coatney

, et al. Early detection of graft failure using the blood metabolic profile of a liver recipient. Transplantation, 2007; 83:517–521.

32.

Song

, Lee

, Kang

, et al. Experimental infection of dogs with avian-origin canine influenza A virus (H3N2). Emerg Infect Dis, 2009; 15:56–58.

33.

, Tian

, Hong

, et al. Global and quantitative proteomic analysis of dogs infected by avian-like H3N2 canine influenza virus. Front Microbiol, 2015; 6:228.

34.

Tam

, Ng

, Tan

, et al. Metabolite profiling in identifying metabolic biomarkers in older people with late-onset type 2 diabetes mellitus. Sci Rep, 2017; 7:4392.

35.

Tao

, Ning

, Hao

, et al. Comparative analysis of whole-transcriptome RNA expression in MDCK cells infected with the H3N2 and H5N1 canine influenza viruses. Front Cell Infect Microbiol, 2019; 9:76.

36.

Tapiero

, Tew

, Gaté

, et al. Prevention of pathologies associated with oxidative stress and dietary intake deficiencies: folate deficiency and requirements. In. France: Elsevier SAS, 2001:381–390.

37.

Turi

, Romick-Rosendale

, Gebretsadik

, et al. Using urine metabolomics to understand the pathogenesis of infant respiratory syncytial virus (RSV) infection and its role in childhood wheezing. Metabolomics, 2018; 14:135.

38.

Vastag

, Koyuncu

, Grady

, et al. Divergent effects of human cytomegalovirus and herpes simplex virus-1 on cellular metabolism. PLoS Pathog, 2011; 7:e1002124.

39.

Wang

, Kong

, Guan

, et al. Plasma phospholipid metabolic profiling and biomarkers of type 2 diabetes mellitus based on high-performance liquid chromatography/electrospray mass spectrometry and multivariate statistical analysis. Anal Chem, 2005; 77:4108–4116.

40.

Want

, Wilson

, Gika

, et al. Global metabolic profiling procedures for urine using UPLC-MS. Nat Protoc, 2010; 5:1005–1018.

41.

Weisan Chen

PACD

. A novel influenza A virus mitochondrial protein that induces cell death. Nat Med, 2001; 7:1306–1312.

42.

Wise

, Foeglein

, Sun

, et al. A complicated message: Identification of a novel PB1-related protein translated from influenza A virus segment 2 mRNA. J Virol, 2009; 83:8021–8031.

43.

, Fan

, Rieser

, et al. Recent advances in high-throughput quantitative bioanalysis by LC–MS/MS. J Pharm Biomed Anal, 2007; 44:342–355.

44.

Yang

, Xu

, Hong

, et al. Discrimination of type 2 diabetic patients from healthy controls by using metabonomics method based on their serum fatty acid profiles. J Chromatogr B, 2004; 813:53–58.

45.

Yang

, Xu

, Zheng

, et al. Diagnosis of liver cancer using HPLC-based metabonomics avoiding false-positive result from hepatitis and hepatocirrhosis diseases. J Chromatogr B, 2004; 813:59–65.

46.

Zhao

, Su

, Zhou

, et al. Comparative analysis of microRNAs from the lungs and trachea of dogs (Canis familiaris) infected with canine influenza virus. Infect Genet Evol, 2014; 21:367–374.

47.

Zheng

, Fu

, Wang

, et al. Comparative analysis of microRNA expression in dog lungs infected with the H3N2 and H5N1 canine influenza viruses. Microb Pathog, 2018; 121:252–261.

48.

Zhong

, Xu

, Bruno

, et al. Targeted high performance liquid chromatography tandem mass spectrometry-based metabolomics differentiates metabolic syndrome from obesity. Exp Biol Med (Maywood), 2017; 242:773–780.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.04 MB

0.01 MB

0.00 MB