Comparison of sepsis rats induced by caecal ligation puncture or Staphylococcus aureus using a LC-QTOF-MS metabolomics approach

    Comparison of sepsis rats induced by caecal ligation puncture or Staphylococcus aureus using a LC-QTOF-MS metabolomics approach Zhang Lin, Xinru Liu, Lulu Sun, Jinbao Li, Zhenglin Hu, Haisheng Xie, Xianpeng Zu, Xiaoming Deng, Weidong Zhang PII: DOI: Reference:

S1567-1348(16)30182-4 doi: 10.1016/j.meegid.2016.05.010 MEEGID 2743

To appear in: Received date: Revised date: Accepted date:

23 January 2016 24 April 2016 6 May 2016

Please cite this article as: Lin, Zhang, Liu, Xinru, Sun, Lulu, Li, Jinbao, Hu, Zhenglin, Xie, Haisheng, Zu, Xianpeng, Deng, Xiaoming, Zhang, Weidong, Comparison of sepsis rats induced by caecal ligation puncture or Staphylococcus aureus using a LC-QTOF-MS metabolomics approach, (2016), doi: 10.1016/j.meegid.2016.05.010

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Comparison of sepsis rats induced by caecal ligation puncture or Staphylococcus

IP

T

aureus using a LC-QTOF-MS metabolomics approach

SC R

Zhang Lina¥, Xinru Liub¥, Lulu Sunc, d¥, Jinbao Lic, Zhenglin Hub, Haisheng Xieb, Xianpeng Zub, Xiaoming Dengc*, Weidong Zhanga, b*

School of Pharmacy, Shanghai Jiaotong University, Shanghai 200240, China;

b

School of Pharmacy, Second Military Medical University, Shanghai 200433, China;

c

Department of Anesthesiology, Changhai Hospital, Second Military Medical

MA

University, Shanghai 200433, China; d

NU

a

Department of Anesthesiology, Shanghai Ninth People Hospital Affiliated to

Shanghai Jiaotong University School of Medicine. Authors are equal to the contribution

*

Correspondence should be addressed to

CE P

TE

D

¥

Xiao-ming Deng:

Tel./fax: +86 21 25070601

AC

E-mail address: [email protected] Wei-dong Zhang:

Tel./fax: +86 21 25070386 E-mail address: [email protected]

1

ACCEPTED MANUSCRIPT

Abstract: Sepsis is a whole-body inflammatory response to infection with high mortality and is treated in intensive care units (ICUs). In the present study, to identify

IP

T

metabolic biomarkers that can differentiate sepsis models induced by caecal ligation puncture (CLP) or Staphylococcus aureus (S. aureus), small molecular metabolites in

SC R

the serum were measured by liquid chromatography quadruple time-of-flight mass spectrometry (LC-QTOF-MS) and analysed using the multivariate statistical analysis (MVA) of partial least square-discrimination analysis (PLS-DA) method. The results

NU

demonstrated that the body showed obvious metabolic disorders in the sepsis groups compared with the control group. A total of 8 potential biomarkers were identified in

MA

the CLP group, and 10 potential biomarkers were identified in the S. aureus group. These potential biomarkers primarily reflected an energy metabolism disorder,

D

inflammatory response, oxidative stress and tissue damage, which occur during sepsis,

CE P

aureus sepsis.

TE

and these markers might potentially be used to differentiate CLP from Staphylococcus

Keywords: Sepsis; Caecal ligation puncture; Staphylococcus aureus; Metabolomics;

AC

Liquid chromatography-mass spectrometry; Potential biomarker

2

ACCEPTED MANUSCRIPT

T

1. Introduction

IP

Sepsis describes the disruption of inflammatory homeostasis triggered by an

SC R

infection. Sepsis and sepsis-associated multi-organ failure might rapidly deteriorate fast severe sepsis, eventually resulting in an irrecoverable state or death if efficient treatments are not immediately administered (Buras et al., 2005). Despite intensive

NU

basic research and clinical trials, sepsis largely contributes to the morbidity and mortality of patients in ICUs, and the incidence of severe sepsis in the United States is

MA

estimated at 750, 000 cases with 210, 000 deaths per year (Angus et al., 2001). Therefore, it is obvious that efforts to develop novel therapies to treat sepsis will be of great value.

D

Animal models play an indispensable role in understanding the host response to

TE

an infection or anti-inflammatory reactions during sepsis. Sepsis models can be divided into three categories: exogenous administration of a toxin (such as

CE P

lipopolysaccharide (LPS), endotoxins or zymosan), exogenous administration of a viable pathogen (such as bacteria), or destruction of the animal’s endogenous protective barrier (such as colonic permeability) (Buras et al., 2005; Fink, 2013).

AC

However, when considering the use of animal models in studies on the development of sepsis, different models present advantages and drawbacks. The toxaemia model is less attractive because this model has short-term effects on the inflammatory cascade and shows a lack of an active nidus of infection (Buras et al., 2005; Remick and Ward, 2005). Although bacterial infection models cannot replicate some features of human sepsis, these models provide important clues about the mechanisms of host reactions to a pathogen and, more importantly, facilitate studies of a particular type of bacterial infection (such as Gram-positive versus Gram-negative bacteria). In the last decade, Gram-positive bacterial infections of Staphylococcus aureus (S. aureus) have been reported to cause more than 50% of sepsis cases (Hiramatsu et al., 1997; Opal and Cohen, 1999). When considering destroying the endogenous protective barrier, caecal 3

ACCEPTED MANUSCRIPT ligation and puncture (CLP) are typically used, and this strategy has been regarded as the gold standard for experimental sepsis. For caecal ligation and puncture (CLP), the cecum is opened to facilitate the release of faecal material into the abdominal cavity

IP

T

to generate a systemic inflammatory response that is induced by poly-bacterial infection.

SC R

Metabolomics is a representation of the actual physiological status of a biological system in response to external stimulation and perturbation (Nicholson et al., 1999). Jeremy Nicholson (Imperial College in London) initially pioneered this

NU

approach, which has currently been developed into an irreplaceable tool in fields such as biomarker discovery of diseases (Kenny et al., 2010), drug toxicity and efficacy

MA

evaluations (Nicholls et al., 2000), lipids (lipidomics) studies (German et al., 2007). Combined with the high-resolution platform of liquid chromatography and mass

D

spectrometry (LC-MS), metabolomics possesses exceptional advantages, as this

TE

method can be used to detect and quantify up to several hundred metabolites in one sample in a relative short time without any preconception or selection bias (Wilson et

CE P

al., 2005; Zelena et al., 2009). To our knowledge, there is no investigation to explore the similarities or differences of sepsis models induced by endogenous protective barrier destruction versus bacterial infection based on a metabolomics approach

AC

(Izquierdo-García et al., 2011; Lin et al., 2009; Su et al., 2014; Xu et al., 2008). Consequently, in the present study, we proposed a LC-QTOF-MS-based metabolomics approach to differentiate sepsis induced by CLP or S. aureus. The aim of the present study was to enhance the current understanding of sepsis and provide more valuable clues for future sepsis studies and treatment development. 2. Materials and methods 2.1. Chemicals Acetonitrile (LC-MS grade) and methanol (LC-MS grade) were purchased from JT Baker (NJ, USA). Formic acid (spectroscopic grade) was purchased from 4

ACCEPTED MANUSCRIPT Sigma/Aldrich (MO, USA). Distilled water was obtained using a milli-Q20 system Millipore (MA, USA). Staphylococcus aureus (ATCC 29213) was obtained from

T

American Type Culture Collection.

SC R

IP

2.2. Animals

Male specific-pathogen-free Sprague Dawley (SD) rats (250 ~ 300 g), purchased from Shanghai Experimental Animal Centre of the Chinese Academy of Sciences

NU

(Shanghai, China), were maintained under a 12 h light/dark cycle and constant temperature (20 ± 0.5 ℃) and humidity (60 % ~ 70%). Animals were provided free

MA

access to a normal rat diet and tap water. The rats were fasted overnight with water ad libitum on the night before the experiment. The experiments were conducted in accordance with the protocols reviewed and approved by the Institution Animal Care

TE

D

and Use Committee of Second Military Medical University in Shanghai, China.

CE P

2.3. Animal models of experimental sepsis rats induced by CLP and S. aureus SD rats were randomly separated into control (n = 14), CLP (n = 40) and S. aureus (n = 40) groups. The procedures for sepsis modelling were performed as

AC

previously described (Wichterman et al., 1980). For the CLP group, the rats were anaesthetized under sodium pentobarbital. A 2.5-cm incision in the lower-right quadrant of the abdomen was conducted, and the caecum was ligated and punctured using an 18-gauge needle in two layers. The abdomen was subsequently closed, and 4 mL of 0.9% saline was subcutaneously administered. After 12 hours, blood was drawn and collected in tubes to clot for 2 hours on ice. After centrifugation at 3000 g for 20 minutes, the serum was collected and stored at - 80 ℃. For the S. aureus group, the rats were intraperitoneally injected with S. aureus (DNA concentration = 1 g/kg, volume = 3 mL). After 4 hours, blood was drawn from the inferior vena cava and processed using the same procedures in CLP. Before LC-MS analysis, 100 µL of serum was extracted with 300 µL of 5

ACCEPTED MANUSCRIPT methanol and centrifuged for 10 min at 12,000 rpm. The supernatant was transferred to an auto-sampler vial for further analysis.

IP

T

2.4. Chromatography

Chromatographic separation was conducted on an Agilent Eclipse Plus C18

SC R

column, 1.8 μm, 2.1×100 mm (USA) with an Agilent 1200 HPLC system. The column was maintained at 45 ℃. The flow rate was 0.3 mL/min in positive mode and

NU

0.4 mL/min in negative mode. The gradient programme for positive mode was 0-2 min, 2% B; 2-15 min, 2%-100% B; 15-18.5 min, 100% B; 18.5-20 min, 100%-50% B;

MA

and 20-22 min, 50%-2% B. The gradient programme for negative mode was: 0-2 min, 2% B; 2-15 min, 2%-100% B; 15-17 min, 100% B; 17-20 min, 100%-50% B; and 20-23 min, 50%-2% B. The sample injection volume was 5 µl. Solvent A was water

D

mixed with 0.1% formic acid, and solvent B was acetonitrile mixed with 0.1% formic

TE

acid.

CE P

2.5. Mass spectrometry

The MS spectra were acquired on a 6520 Q-TOF MS system (Agilent, USA)

AC

with an electrospray ion (ESI) source. The capillary voltage was optimized at 4000 V for positive mode and 3900 V for the negative mode. The gas temperature was 330 ℃. The gas flow was 8 L/min. The nebulizer pressure was 30 psig. The fragment voltage was set at 130 V. Full scan spectra were performed between 100-1000 m/z. Positive and negative modes were both collected in mass detection. The MS/MS analysis was acquired in the targeted MS/MS mode with a collision energy ranging from 10 to 40 eV. 2.6. Statistical analysis and metabolite identification The obtained raw LC-QTOF-MS data were processed in Agilent Mass Hunter Qualitative Analysis (Version B.03.01, Agilent Technologies, USA) and Mass Profiler software (Agilent Technologies, USA) to detect peak and generate peak intensity lists 6

ACCEPTED MANUSCRIPT of retention time-m/z pairs. The parameters were optimized as follows: m/z ranged from 100 to 1000, peak filters were set to a centroid height exceeding 100 counts; compound filters set the base peak to more than 1000 counts. This processing step

IP

T

created MHD files that contained compound IDs (based on neutral mass and retention times), and further processing was conducted in Mass Profiler software, which aligns

SC R

mass characters across different files. The mass-clustering window was 20 ppm, the retention time-clustering window was 0.1 min. These data were handled according to the 80% rule – only variables existing in at least 80% of any group were kept for

NU

subsequent analysis. All of the peaks detected in LC-MS were exported as retention time-m/z (RT-m/z) pairs. Each ion peak area was normalized to the total peak area in

MA

every sample. The obtained RT-m/z pairs in ESI+ (94 samples×7527 RT-m/z) or ESI(94 samples×729 RT-m/z) were transferred into SIMCA-P version 11.0 (Umetrics,

D

Umea, Sweden) to perform a multivariate statistical analysis of partial least

TE

square-discriminant analysis (PLS-DA). The score plot was applied to observe the classification of different groups, and the loading plot was used to select the variables

CE P

(VIP > 1.5) contributing to that classification. T-test and fold-changes were further calculated for these variables in Matlab 7.0. Variables with p values < 0.05 and fold-changes > 1.5 (or < 0.67) were considered to be ions of potential biomarkers. The

AC

potential biomarkers were identified and interpreted based on their accurate measured masses by searching the free databases Metlin (https://metlin.scripps.edu/index.php), HMDB (http://www.hmdb.ca/) and KEGG (http://www.kegg.jp/). After database query, tandem mass spectrometry (MS/MS) was used to generate compound-specific fragmentation patterns that help facilitate structure elucidation in conjunction with reference standards and database information. 3. Results and Discussion 3.1. Statistical analysis MS-based metabolomics studies are ideal tools for the identification of perturbed pathways and characterization of the response to different modelling methods based 7

ACCEPTED MANUSCRIPT on metabolic profiles. A total of 7527 (for ESI+) and 729 (for ESI-) ions of serum samples were obtained from LC-QTOF-MS. To elucidate the metabolic characters involved in sepsis induced using different methods, supervised multivariate analysis,

IP

T

based on the PLS-DA method, was performed (Fig. 1). Matthew Barker initially proposed PLS-DA, which has recently become a prevalent vehicle for metabolomics

SC R

analysis［24］. The presentation of the classification results in visual scores and loading plots benefits learners to understand and explain the features from different types of groups. In ESI+, CLP (blue circle), S. aureus (black dot) and control (red triangle)

NU

groups were well separated across the score plot, whereas the CLP and S. aureus groups were located towards the opposite sides of the controls. This distribution

MA

indicated that the metabolic characters in the sepsis groups were different from those in the controls, and the characteristics of sepsis induced through different modelling

D

methods (CLP or S. aureus) were also different from each other. To evaluate the

TE

overall goodness of fit and predictive ability of the model, R2X and Q2 were evaluated: (1) R2X corresponding to the predictive and orthogonal variation as

CE P

explained by the model; (2) Q2 referring to the goodness of prediction calculated via cross validation. In SIMCA-P, a typical seven-fold cross-validation procedure was conducted to validate the PLS model against over-fitting. In the present study, R2X

AC

and Q2 were 0.177 and 0.766, respectively, indicating that the model has good predictive ability (Q2 > 0.5) to classify new observations into groups. The results of permutation tests showed that the models were reliable (intercepts: R2 = 0.497, Q2 = −0.0147). In ESI-, a similar distribution was observed. There was no overlapping among the three groups, and the two sepsis groups were far from each other. The PLS-DA model also generated good R2X and Q2 values (0.205 and 0.722, respectively) among the three groups. The corresponding intercepts of permutation tests were: R2 = 0.455, Q2 = −0.0717. Because there were a number of significant variables contributing to the separation among different groups in loading plots in two ion modes, we selected the most important variables with VIP values of more than 1.5. The P values and fold-changes were further considered to be objective indicators of metabolite 8

ACCEPTED MANUSCRIPT expression. In total, 234 variables (214 for ESI+, 20 for ESI-) were obtained as discriminating signals for CLP group, and 355 variables (320 for ESI+, 35 for ESI-) were chosen for S. aureus group. Moreover, some metabolites were separately

IP

T

identified and interpreted as potential biomarkers for the CLP or S. aureus group according to search metabolite databases and tandem MS fragmentation information

SC R

with reference standards (Table 1).

NU

3.2. Biomarker identification

Biomarker identification is of great importance for biological studies and clinical

MA

guidance. In total, 8 potential biomarkers were identified for the CLP group, and 10 potential biomarkers were identified for the S. aureus group compared with the control group (Table 1, Fig. 2). These potential biomarkers mainly showed that energy

D

metabolism disorders, tissue damage, oxidative damage and inflammatory responses

TE

are major characters in the development of sepsis (Fig. 3). Sepsis is typically trigged by infection or an inflammatory insult that is not

CE P

cleared by the host. The dysregulation of the inflammatory response damages the host immune system and endothelial and epithelial cells. Endothelial and epithelial cells can build barriers to protect against the infection and inflammation. Widespread

AC

immune and cellular dysfunction can subsequently disseminate, resulting in organ failure, tissue hypoxia, inflammatory factors production, reactive oxygen species (ROS) generation, and so on. In response, the small molecular metabolites in bio-fluids might significantly change to reflect the status or severity of the disease. In the present study, 8 potential biomarkers were significantly changed in the CLP group compared with the control group. Interestingly, 3 of these (glutathione, 2-oxoarginine, and 2-methylhippuric acid) were also identified as potential markers in the S. aureus group (Fig. 2a). Glutathione (GSH) is an excellent antioxidant that protects against oxidative damage. ROS is involved in the development of sepsis (Cassol-Jr et al., 2010; Kaymak et al., 2008). The negative effects of ROS include endothelial damage, cytokine release, mitochondrial impairment, lipid peroxidation and DNA damage, all 9

ACCEPTED MANUSCRIPT contributing to a free radical overload and oxidant-antioxidant imbalance (Andrades et al., 2009). As the most abundant non-protein thiol in the structure, GSH is a good antioxidant that protects against peroxidation or prevents oxygen toxicity in

IP

T

hyperbaric oxygen therapy in the clinic. In the present study, the serum GSH levels were markedly decreased in both the CLP and S. aureus groups. Previous studies

SC R

indicated that decreased GSH could protect against oxidative damage (Liu and Gaston Pravia, 2010; Lu, 2009), consistent with the results of the present study. One important thing we need to care is that GSH is sensitive to oxidation environment and

NU

it can easily convert to its oxidized form, glutathione disulfide. As a matter of fact, the accurate measurement of GSH may still be a challenge for biological analysis

MA

especially in the oxidative stress field (Asens et al., 1

ma et al., 2009). The

metabolite 2-oxoarginine is a guanidino compound, and guanidino compounds have

D

shown neurotoxicity to humans. A previous study showed that accumulated guanidino

TE

compounds, particularly 2-oxoarginine, might produce central nervous system damage via the urea cycle and other pathways in argininaemia patients (Mizutani et

CE P

al., 1987). In the present study, serum 2-oxoarginine was increased in both the CLP and S. aureus groups, likely indicating nerve injury after brain tissue failure in sepsis. The acyl glycine 2-methylhippuric acid is produced by fatty acid metabolism. The

AC

measurement of acyl glycines in bio-fluids can be used to diagnose disorders with mitochondria fatty acid beta-oxidation. Fatty acids are important sources of fuel because, when metabolized, these compounds yield large amounts of adenosine triphosphate (ATP). Many cell types, particularly from the heart and skeletal muscle, prefer fatty acids to obtain energy. Previous reports have demonstrated that critically ill patients with sepsis often experience hyper-metabolism (Chioléro et al., 1997; Wei et al., 2015; Xu et al., 2008). As the major source of energy, whole-body lipid oxidation is enhanced to satisfy energy requirements (Martinez et al., 2003; Memon et al., 1998; Samra et al., 1996). The results of the present study showed that the levels of 2-methylhippuric acid increased in the CLP group but decreased in the S. aureus group, potentially reflecting differences in the modelling methods applied, leading to many different disorders in sepsis. In this respect, 2-methylhippuric acid might be a 10

ACCEPTED MANUSCRIPT biomarker for the differentiation of CLP from S. aureus. Additionally,

5

other

potential

biomarkers

(dityrosine,

ketoleucine,

3-hydroxyphenylacetic acid, 5-hydroxyhexanoic acid and leukotriene A4) were

IP

T

identified in CLP versus the control group (Fig. 2b). Dityrosine is a fluorescent molecule that is formed from posttranslational processing, and this molecule has been

SC R

implicated as a biomarker in oxidative stress. Oxidative stress is involved in the development of many diseases, including infection. Oxidative stress reflects the increased production of ROS or a significant decrease in the effectiveness of

NU

antioxidant defences. The results of the present study showed a massive decrease in the dityrosine levels in the CLP group. Nevertheless, in mammalian tissue and urine

MA

samples, elevated dityrosine levels were observed in many pathologies, such as acute inflammation, atherosclerosis, and Alzheimer's disease (DiMarco and Giulivi, 2007).

D

This effect needs further study to determine the mechanism of decreased dityrosine in

TE

CLP. Ketoleucine has been reported as an indicator that is associated with neurological dysfunction in Maple Syrup Urine Disease (MSUD) patients. MSUD is

CE P

an aminoacidopathy resulting from the dysfunction of the branched-chain keto acid dehydrogenase. Accumulated ketoleucine compromises brain energy metabolism by blocking the respiratory chain. The results of the present study also showed elevated

AC

ketoleucine levels in CLP rats; this finding is relevant to understand the neurological dysfunction

in

sepsis

rats

(Wisniewski

et

al.,

2015).

The

antioxidant

3-hydroxyphenylacetic acid shows protective biological activity in HMDB (Wishart et al., 2009). Similar to GSH, antioxidants can protect against oxidative damage in sepsis. The results of the present study showed a significant increase in 3-hydroxyphenylacetic acid in the CLP group. This phenomenon reflects the oxidative damage

occurring

during

sepsis,

leading

to

the

production

of

more

3-hydroxyphenylacetic acid to enhance the effectiveness of antioxidant defences. The normal dicarboxylic acid 5-hydroxyhexanoic acid is a degradation product of fatty acids. Increased dicarboxylic acid is detected in human urine under conditions of abnormal fatty acid oxidation (Tserng et al., 1996). The results of the present study showed a significant increase in serum dicarboxylic acid levels in CLP rats. Thus, 11

ACCEPTED MANUSCRIPT dicarboxylic acid might also be an indicator of fatty acid metabolism disorder. Leukotriene A4 (LTA4) has been associated with inflammatory responses. The immune system is involved in host defence against an infectious source. Inflammation

IP

T

is an important component of the early immunologic response. Inappropriate or dysfunctional immune responses often underlie inflammatory diseases. LTA4 is an

SC R

intermediate in the synthesis of leukotrienes. The first step includes the oxidation of arachidonic acid to 5-hydroperoxyeicosatetraenoic acid (5-HPETE); the second step involves the dehydration of 5-HPETE to form LTA4. LTA4 is hydrolysed to

NU

leukotriene B4 or conjugated with glutathione to yield leukotriene C4/D4/E4. Leukotrienes participate in host defence reactions and pathophysiological conditions

MA

of inflammation (Mendes and Silveira, 2014). Therefore, the decreased LTA4 observed in the present study might reflect the severity of inflammation in

D

CLP-induced sepsis.

TE

In the S. aureus group, 10 metabolites were identified as potential biomarkers compared with the control group. The remaining 7 potential biomarkers included indoxylsulfuric

acid,

cholic

acid,

chenodeoxyglycocholate,

CE P

L-carnitine,

sphinganine-1-phosphate, 3-hydroxydodecanoic acid and alpha-linolenic acid (Fig. 2c). L-carnitine and 3-hydroxydodecanoic acid are both associated with fatty acid

AC

metabolism disorder in S. aureus. Similar to CLP, energy metabolism dysfunction is considered to be a key feature in sepsis (Hecker et al., 2014). The medium-chain fatty acid 3-hydroxydodecanedioic acid participates in fatty acid metabolism. The measurement of 3-hydroxydodecanedioic acid might reflect long-chain hydroxyacyl CoA dehydrogenase deficiency (LCHAD) and fatty acid oxidation disorders (Chickos et al., 2002). The signs and symptoms of LCHAD occur in many diseases, particularly in metabolic errors of fatty acid metabolism and L-carnitine metabolism. L-carnitine is a transporter for the delivery of long chain fatty acids to the mitochondria for oxidation (Panchal et al., 2015). In the S. aureus group, 3-hydroxydodecanoic acid was significantly decreased and L-carnitine was significantly increased. Thus, the appearance of these two markers further confirmed the energy metabolism dysfunction in the S. aureus group. Indoxylsulfuric acid is a metabolite of the 12

ACCEPTED MANUSCRIPT common amino acid tryptophan. Plasma concentrations of indoxylsulfuric acid have been reported as tissue injury factors and markers of renal function (Stanfel et al., 1986). Other studies have proposed that indoxylsulfuric acid is also involved in

IP

T

oxidative stress (Kato et al., 2003), potentially reducing the superoxide scavenging ability in the kidneys (Owada et al., 2008), increasing ROS production in tubular cells,

SC R

and increasing NAD(P)H oxidase activity in endothelial cells (Dou et al., 2007). The results of the present study showed a significant decrease in indoxylsulfuric acid in S. aureus rats. This effect might be an indication of oxidative stress and renal damage

NU

observed in sepsis.

Cholic acid and chenodeoxyglycocholate are both associated with bile acid

MA

metabolism disorders. Bile acids are steroid acids are predominantly detected in the bile of mammals and other vertebrates. Different molecular forms of bile acids can be

D

synthesized by different species. Cholic acid is the most abundant bile acid produced

TE

in the liver. Chenodeoxyglycocholate, the glycine conjugation of chenodeoxycholic acid, is one of the major bile salts in bile. The main function of bile acids is to

CE P

facilitate the formation of micelles, which promote the digestion and absorption of fat. Bile acid secretion is the major determinant of bile formation and an important indicator of overall hepatic function. Increasing evidence has shown that affected bile

AC

acid secretion or cholestasis frequently occurs in inflammatory diseases (Bhogal and Sanyal, 2013; Green et al., 1996; Reynolds et al., 1995). Staphylococcus aureus infection induced the release of systemic and intrahepatic pro-inflammatory cytokines, subsequently resulting in profound liver failure or hepatic dysfunction. The results of the present study showed that cholic acid and chenodeoxyglycocholate were both decreased in the S. aureus group. Therefore, these observations carry important implications for the diagnosis of hepatic dysfunction. Sphinganine-1-phosphate is a ceramide in sphingolipid metabolism. Ceramide is a bioactive lipid that regulates immune cell function. The ceramide concentration in sepsis patients has been identified as a candidate tool to predict multi-organ dysfunctions (Delogu et al., 1999; Drobnik et al., 2003). Therefore, the decreased sphinganine-1-phosphate in serum reflects organ dysfunction in S. aureus. Moreover, alpha-linolenic acid (ALA) is a 13

ACCEPTED MANUSCRIPT typical omega-3 polyunsaturated fatty acid (PUFA). Omega-3 fatty acids have anti-inflammatory effects because these molecules inhibit increases in the protein and mRNA expression of iNOS and COX-2 during inflammation (Calder, 2003; Ren and

IP

T

Chung, 2007; Ren et al., 2007). The results of the present study showed a significant decrease in ALA in the S. aureus group, likely associated with the excessive release of

SC R

inflammatory mediators during sepsis. Therefore, ALA might be a good candidate for the severity of inflammation.

NU

4. Conclusion

MA

Metabolomics studies are of great value for early diagnosis and treatment in intensive care units. In the present study, we reported for the first time LC-QTOF-MS analysis to compare the metabolic profiles in rats with sepsis induced by CLP or S.

D

aureus. Combined with multivariate statistical analysis of PLS-DA, 8 potential

TE

biomarkers were identified in the CLP group, and 10 were identified in the S. aureus group. These potential biomarkers were primarily involved in energy metabolism

CE P

disorders, inflammatory response, oxidative stress and tissue damage. Metabolomics studies are of great value for understanding the complex nature of sepsis on a global scale. More work is needed to validate and extend these potential biomarkers into

AC

practice in the future.

Author contributions All of the authors contributed equally to this work. Acknowledgements The work was supported by Professor of Chang Jiang Scholars Program, NSFC (81520108030, 81573318, 81373301, 1302658, 81171788 and 81270128). Conflict of interest All authors declared no conflicts of interest. 14

ACCEPTED MANUSCRIPT

References

T

Andrades, M., Ritter, C., Oliveira, M.R.D., Streck, E.L., Moreira, J.C.F., Dal-Pizzol, F., 2009. Antioxidant

IP

treatment reverses organ failure in rat model of sepsis: role of antioxidant enzymes imbalance, neutrophil infiltration, and oxidative stress. J Surg Res. 167, e307–e313.

SC R

Angus, D.C., Linde-Zwirble, W.T., Lidicker, J., Clermont, G., Carcillo, J., Pinsky, M.R., 2001. Epidemiology of severe sepsis in the United States: analysis of incidence, outcome, and associated costs of care. Crit Care Med. 29, 1303-1310.

Asensi, M., Sastre, J., Pallardo, F.V., Garcia de la Asuncion, J., Estrela, J.M., Vina, J., 1994. A biological samples. Anal Biochem. 217, 323-328.

NU

high-performance liquid chromatography method for measurement of oxidized glutathione in Bhogal, H.K., Sanyal, A.J., 2013. The molecular pathogenesis of cholestasis in sepsis. Front Biosci. 5,

MA

87-96.

Buras, J.A., Holzmann, B., Sitkovsky, M., 2005. Animal Models of sepsis: setting the stage. Nat Rev Drug Discov. 4, 854-865.

Calder, P.C., 2003. N-3 polyunsaturated fatty acids and inflammation: from molecular biology to the

D

clinic. Lipids. 38, 343-352.

Cassol-Jr, O.J., Comim, C.M., Silva, B.R., Hermani, F.V., Constantino, L.S., Felisberto, F., Petronilho, F.,

TE

Hallak, J.E.C., Martinis, B.S.D., Zuardi, A.W., 2010. Treatment with cannabidiol reverses oxidative stress parameters, cognitive impairment and mortality in rats submitted to sepsis by cecal ligation and

CE P

puncture. Brain Res. 1348, 128–138.

Chickos, J.S., Way, B.A., Wilson, J., Shaharuzzaman, M., Laird, J., Landt, M., 2002. Analysis of 3-hydroxydodecanedioic acid for studies of fatty acid metabolic disorders: preparation of stable isotope standards. J Clin Lab Anal. 16, 115-120. Chioléro, R., Revelly, J.-P., Tappy, L., 1997. Energy metabolism in sepsis and injury. Nutrition. 13, 45-51.

AC

Delogu, G., Famularo, G., Amati, F., Signore, L., Antonucci, A., Trinchieri, V., Di Marzio, L., Cifone, M.G., 1999. Ceramide concentrations in septic patients: a possible marker of multiple organ dysfunction syndrome. Crit Care Med. 27, 2413-2417. DiMarco, T., Giulivi, C., 2007. Current analytical methods for the detection of dityrosine, a biomarker of oxidative stress, in biological samples. Mass Spectrom Rev. 26, 108-120. Dou, L., Jourde-Chiche, N., Faure, V., Cerini, C., Berland, Y., Dignat-George, F., Brunet, P., 2007. The uremic solute indoxyl sulfate induces oxidative stress in endothelial cells. J Thromb Haemost. 5, 1302-1308. Drobnik, W., Liebisch, G., Audebert, F.X., Frohlich, D., Gluck, T., Vogel, P., Rothe, G., Schmitz, G., 2003. Plasma ceramide and lysophosphatidylcholine inversely correlate with mortality in sepsis patients. J Lipid Res. 44, 754-761. Fink, M.P., 2013. Animal models of sepsis. Virulence. 5, 143-153. German, J.B., Gillies, L.A., Smilowitz, J.T., Zivkovic, A.M., Watkins, S.M., 2007. Lipidomics and lipid profiling in metabolomics. Curr Opin Lipidol. 18, 66-71. Green, R.M., Beier, D., Gollan, J.L., 1996. Regulation of hepatocyte bile salt transporters by endotoxin and inflammatory cytokines in rodents. Gastroenterology. 111, 193-198. Hecker, M., Sommer, N., Voigtmann, H., Pak, O., Mohr, A., Wolf, M., Vadasz, I., Herold, S., Weissmann, 15

ACCEPTED MANUSCRIPT N., Morty, R.E., Seeger, W., Mayer, K., 2014. Impact of short- and medium-chain fatty acids on mitochondrial function in severe inflammation. JPEN J Parenter Enteral Nutr. 38, 587-594. Hiramatsu, K., Hanaki, H., Ino, T., Yabuta, K., Oguri, T., Tenover, F.C., 1997. Methicillin-resistant Staphylococcus aureus clinical strain with reduced vancomycin susceptibility. J Antimicrob Chemoth.

T

40, 135-136. Izquierdo-García, J., Nin, N., Ruíz-Cabello, J., Rojas, Y., de Paula, M., López-Cuenca, S., Morales, L., diagnosis of experimental sepsis. Intensive Care Med. 37, 2023-2032.

IP

Martínez-Caro, L., Fernández-Segoviano, P., Esteban, A., Lorente, J., 2011. A metabolomic approach for

SC R

Kato, A., Odamaki, M., Hishida, A., 2003. Association between blood indoxyl sulfate and carbonyl stress marker in hemodialysis patients. Clin Nephrol. 60, 161-167.

Kaymak, C., Kadioglu, E., Ozcagli, E., Osmanoglu, G., Izdes, S., Agalar, C., Basar, H., Sardas, S., 2008. Oxidative DNA damage and total antioxidant status in rats during experimental gram-negative sepsis.

NU

Hum Exp Toxicol. 27, 485-491.

Kenny, L.C., Broadhurst, D.I., Dunn, W., Brown, M., North, R.A., McCowan, L., Roberts, C., Cooper, G.J., Kell, D.B., Baker, P.N., 2010. Robust early pregnancy prediction of later preeclampsia using

MA

metabolomic biomarkers. Hypertension. 56, 741-749.

Lin, Z.Y., Xu, P.B., Yan, S.K., Meng, H.B., Yang, G.J., Dai, W.X., Liu, X.R., Li, J.B., Deng, X.M., Zhang, W.D., 2009. A metabonomic approach to early prognostic evaluation of experimental sepsis by 1H NMR and pattern recognition. NMR Biomed. 22, 601-608.

D

Liu, R.M., Gaston Pravia, K.A., 2010. Oxidative stress and glutathione in TGF-beta-mediated

TE

fibrogenesis. Free Radic Biol Med. 48, 1-15.

Lu, S.C., 2009. Regulation of glutathione synthesis Molecular aspects of medicine. 30, 42-59. Martinez, A., Chiolero, R., Bollman, M., Revelly, J.P., Berger, M., Cayeux, C., Tappy, L., 2003. Assessment

CE P

of adipose tissue metabolism by means of subcutaneous microdialysis in patients with sepsis or circulatory failure. Clin Physiol Funct I 23, 286-292. Memon, R.A., Feingold, K.R., Moser, A.H., Fuller, J., Grunfeld, C., 1998. Regulation of fatty acid transport protein and fatty acid translocase mRNA levels by endotoxin and cytokines. Am J Physiol.

AC

274, E210-217. Mendes,

M.T.,

Silveira,

P.F.,

2014.

The

interrelationship

between

leukotriene

B4

and

leukotriene-A4-hydrolase in collagen/adjuvant-induced arthritis in rats. Biomed Res Int. 730421, 20. Mizutani, N., Hayakawa, C., Ohya, Y., Watanabe, K., Watanabe, Y., Mori, A., 1987. Guanidino compounds in hyperargininemia. Tohoku J Exp Med. 153, 197-205. Nicholls, A.W., Nicholson, J.K., Haselden, J.N., Waterfield, C.J., 2000. A metabonomic approach to the investigation of drug-induced phospholipidosis: an NMR spectroscopy and pattern recognition study. Biomarkers. 5, 410-423. Nicholson, J.K., Lindon, J.C., Holmes, E., 1999. 'Metabonomics': understanding the metabolic responses of living systems to pathophysiological stimuli via multivariate statistical analysis of biological NMR spectroscopic data. Xenobiotica. 29, 1181-1189. Opal, S.M., Cohen, J., 1999. Clinical gram-positive sepsis: does it fundamentally differ from gram-negative bacterial sepsis? Crit Care Med. 27, 1608-1616. Owada, S., Goto, S., Bannai, K., Hayashi, H., Nishijima, F., Niwa, T., 2008. Indoxyl sulfate reduces superoxide scavenging activity in the kidneys of normal and uremic rats. Am J Nephrol. 28, 446-454. Panchal, S.K., Poudyal, H., Ward, L.C., Waanders, J., Brown, L., 2015. Modulation of tissue fatty acids by L-carnitine attenuates metabolic syndrome in diet-induced obese rats. Food Funct. 6, 2496-2506. 16

ACCEPTED MANUSCRIPT Remick, D.G., Ward, P.A., 2005. Evaluation of endotoxin models for the study of sepsis. Shock. 1, 7-11. Ren, J., Chung, S.H., 2007. Anti-inflammatory effect of alpha-linolenic acid and its mode of action through the inhibition of nitric oxide production and inducible nitric oxide synthase gene expression via NF-kappaB and mitogen-activated protein kinase pathways. J Agric Food Chem. 55, 5073-5080. acid isolated from Actinidia polygama fruits. Arch Pharm Res. 30, 708-714.

T

Ren, J., Han, E.J., Chung, S.H., 2007. In vivo and in vitro anti-inflammatory activities of alpha-linolenic

IP

Reynolds, M.A., Siow, Y., Collet, D., Klar, E., Richardson, J.D., Vitale, G.C., 1995. The role of bile acid secretion in the hepatic response to operation and infection. Am Surg. 61, 469-474.

SC R

Samra, J.S., Summers, L.K., Frayn, K.N., 1996. Sepsis and fat metabolism. Br J Surg. 83, 1186-1196. Stanfel, L.A., Gulyassy, P.F., Jarrard, E.A., 1986. Determination of indoxyl sulfate in plasma of patients with renal failure by use of ion-pairing liquid chromatography. Clin Chem. 32, 938-942. Su, L., Huang, Y., Zhu, Y., Xia, L., Wang, R., Xiao, K., Wang, H., Yan, P., Wen, B., Cao, L., Meng, N., Luan,

NU

H., Liu, C., Li, X., Xie, L., 2014. Discrimination of sepsis stage metabolic profiles with an LC/MS-MS-based metabolomics approach. BMJ Open Respir Res. 1, e000056. Tserng, K.Y., Griffin, R.L., Kerr, D.S., 1996. Distinction of dicarboxylic aciduria due to medium-chain

MA

triglyceride feeding from that due to abnormal fatty acid oxidation and fasting in children. Metabolism. 45, 162-167.

Wei, Y., Gao, N., Zhang, Z., Zu, X., Hu, Z., Zhang, W., Yin, J., Liu, X., 2015. Metabolic changes at the early stage of sepsis induced by cecal ligation and puncture in rats and the interventional effects of

D

Huang-Lian-Jie-Du-Tang. J Chromatogr B Analyt Technol Biomed Life Sci.

TE

Wichterman, K.A., Baue, A.E., Chaudry, I.H., 1980. Sepsis and septic shock--a review of laboratory models and a proposal. J Surg Res. 29, 189-201. Wilson, I.D., Plumb, R., Granger, J., Major, H., Williams, R., Lenz, E.M., 2005. HPLC-MS-based methods

CE P

for the study of metabonomics. J Chromatogr B. 817, 67-76. Wishart, D.S., Knox, C., Guo, A.C., Eisner, R., Young, N., Gautam, B., Hau, D.D., Psychogios, N., Dong, E., Bouatra, S., Mandal, R., Sinelnikov, I., Xia, J., Jia, L., Cruz, J.A., Lim, E., Sobsey, C.A., Shrivastava, S., Huang, P., Liu, P., Fang, L., Peng, J., Fradette, R., Cheng, D., Tzur, D., Clements, M., Lewis, A., De Souza,

AC

A., Zuniga, A., Dawe, M., Xiong, Y., Clive, D., Greiner, R., Nazyrova, A., Shaykhutdinov, R., Li, L., Vogel, H.J., Forsythe, I., 2009. HMDB: a knowledgebase for the human metabolome. Nucleic Acids Res. 37, D603-D610.

Wisniewski, M.S., Carvalho-Silva, M., Gomes, L.M., Zapelini, H.G., Schuck, P.F., Ferreira, G.C., Scaini, G., Streck, E.L., 2015. Intracerebroventricular administration of alpha-ketoisocaproic acid decreases brain-derived neurotrophic factor and nerve growth factor levels in brain of young rats. Metab Brain Dis. 20, 20. Xu, P., Lin, Z., Meng, H., Yan, S., Yang, Y., Liu, X., Li, J., Deng, X., Zhang, W., 2008. A metabonomic approach to early prognostic evaluation of experimental sepsis. J Infect. 56, 474-481. Y

, K., 2009. A

practical HPLC method to measure reduced(GSH) and oxidized (GSSG) glutathione concentrations in animal tissues. Journal of Animal & Veterinary Advances. 8, 343-347. Zelena, E., Dunn, W.B., Broadhurst, D., Francis-McIntyre, S., Carroll, K.M., Begley, P., O'Hagan, S., Knowles, J.D., Halsall, A., Wilson, I.D., Kell, D.B., 2009. Development of a robust and repeatable UPLC-MS method for the long-term metabolomic study of human serum. Anal Chem. 81, 1357-1364.

17

ACCEPTED MANUSCRIPT

Figure legends: Figure 1 PLS-DA score plots of rat serum from the control (red triangle), CLP (blue

Figure 2 The significantly changed metabolites among (

) control (n=14), (

) CLP

) S. aureus (n=40) groups: (a) glutathione; 2-oxoarginine;

SC R

(n=40) and (

IP

T

circle) and S. aureus (black dot) groups in ESI+ and ESI- modes.

2-methylhippuric acid (b) dityrosine; ketoleucine; 3-hydroxyphenylacetic acid; 5-hydroxyhexanoic acid; leukotriene A4 (c) L-carnitine; indoxylsulfuric acid; cholic

NU

acid; chenodeoxyglycocholate; sphinganine-1-phosphate; 3-hydroxydodecanoic acid; and alpha-linolenic acid. (PA = peak area, *P < 0.05, **P < 0.01).

MA

Figure 3 The identified potential biomarkers and associated biological functions: (■) identified in both sepsis groups; (▲) identified in the CLP group; and (▼) identified

AC

CE P

TE

D

in the S. aureus group.

18

ACCEPTED MANUSCRIPT

Figures

SC R

IP

T

Fig. 1

ESI+ 50 40

NU

30 20

0

MA

t[2]

10

-10 -20

-40 -50

-40

-30

-20

TE

-60

D

-30

t[2]

10

0

0

10

20

30

40

50

t[1]

R2X[2] = 0.0409665

Ellipse: Hotelling T2 (0.95) SIMCA-P+ 11 - 2015/12/22 10:15:03

ESI-

AC

20

CE P

R2X[1] = 0.0785777

-10

-10

-10

0

10

t[1] R2X[1] = 0.0659571

R2X[2] = 0.0699472

Ellipse: Hotelling T2 (0.95) SIMCA-P+ 11 - 2015/12/22 10:26:04

19

60

ACCEPTED MANUSCRIPT Fig. 2 (a)

*

**

600000

IP PA

PA

2000

**

800000

3000

3000

PA

*

4000

T

**

**

4000

2 - m e t h y lh ip p u r ic a c id

2 - o x o a r g in in e

g lu t a th io n e 5000

2000

400000

200000

D

TE

s u

P

re

L

u a . S s u re

C

u a . S

u a . S

le u k o tr ie n e A 4

**

3000

*

L

o tr n co

s u re

L

C

n

co

u a . S

5 -h y d r o x y h e x a n o ic a c id

P

l

0

P

l

o

tr

s

C

re

L

u

P

l o tr

C

20000

0

0

n

l

40000

20000

5000

**

60000

40000

2000

co

o

80000

PA

MA

4000

3 -h y d r o x y p h e n y la c e tic a c id

**

60000

PA

PA

o

au S.

NU

80000

**

6000

2000

PA

3000

PA

C

re

L C

S.

k e t o le u c in e

d it y r o s in e

4000

tr

SC R

P

l nt co

au

C

re

L

ro

us

P

l ro nt co

(b) 8000

0

us

0

0

n

1000

1000

2000

1000

s re u

n

S

.

a

co

in d o x y ls u lfu r ic

**

a c id

c h o lic

4000

a c id

15000

**

*

2000000

3000

10000

1000000

PA

1500000

PA

PA

u

o tr

u

a

.

S

AC

L -c a r n itin e

2500000

P

l

re

C

n co

(c)

L

s u

0

CE P

tr

L

o

P

l

0

C

1000

2000

5000 500000

1000

0

s

0

s re .

c

a

o

u

n

C

tr

L

u

P

l o

s u re

c h e n o d e o x y g ly c o c h o la te

s p h in g a n in e -1 -p h o s p h a te

*

3000

S

S

.

c

a

o

u

n

C

tr

L

o

P

l

u re S

.

c

a

o

u

n

C

tr

L

o

P

l

0

3 -h y d r o x y d o d e c a n o ic

**

100000

80000

3000

60000

PA

PA

PA

2000

**

4000

2000

40000

1000

1000 20000

0

0

40000

20000

s u

P

re

L

a

o

u

C

tr

o

l

0

n

. S

20

s u

P

re

L

u

tr

S

.

c

a

o

n

C

o

s . S

a c id

**

c

PA

re a

c

a lp h a -lin o le n ic 80000

60000

u

P L

o

u

n

C

tr

s u re S

.

c

a

o

u

n

C

tr

L

o

P

l

o

l

l

0

a c id

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

Fig. 3

21

ACCEPTED MANUSCRIPT Table

307.09

Glutathione

1.15

173.07

2-oxoarginine

6.92

193.07

2-methylhippuric acid

1.27

382.11

Dityrosine

5.67

130.06

Ketoleucine

6.14

174.03

3-hydroxyphenylacetic acid

7.62

154.06

5-hydroxyhexanoic acid

10.48

318.22

Leukotriene A4

0.90

161.11

L-carnitine

6.27

213.01

9.38

408.29

10.95

449.32

Chenodeoxyglycocholate

11.77

381.27

Sphinganine-1-phosphate

13.43

238.16

3-hydroxydodecanoic acid

15.31

278.22

Alpha-linolenic acid

S. aureus

** -58.6 * 18.4 ** 6.6 ** -5070434.7

** -5.0 * 7.5 ** -2.0

TE

D

MA

** 12.7

** 5.0 * 3.5 ** -18.6

AC

Cholic acid

22

Pathways

Glutathione metabolism Arginine and proline metabolism

Phenylalanine metabolism Unknown Valine, leucine and isoleucine degradation Phenylalanine metabolism Fatty acid metabolism Arachidonic acid metabolism

** 1.7 * -10.0 ** -8124504.8 * -5.2 ** -2.9 ** -3.8 ** -4.8

Indoxylsulfuric acid

T

0.92

CLP

IP

Name

SC R

Mass

NU

RT

CE P

Table 1. The potential biomarkers for sepsis rats induced by CLP or S. aureus.

Fatty acid metabolism Unknown Bile acid metabolism Bile acid metabolism Sphingolipid metabolism Fatty acid metabolism Alpha-linolenic acid metabolism

ACCEPTED MANUSCRIPT Highlights

T

A metabolomics method was proposed to separate sepsis induced by CLP or S. aureus.

IP

8 potential biomarkers were identified in CLP whereas 10 were found in S. aureus.

AC

CE P

TE

D

MA

NU

SC R

These potential biomarkers can be used to differentiate CLP from S. aureus sepsis.

23