- Email: [email protected]
Metabonomics screening of serum identifies pyroglutamate as a diagnostic biomarker for nonalcoholic steatohepatitis
Accepted Manuscript Metabonomics screening of serum identifies pyroglutamate as a diagnostic biomarker for nonalcoholic steatohepatitis
Suwen Qi, Depeng Xu, Qiaoliang Li, Ni Xie, Jun Xia, Qin Huo, Pu Li, Qiwen Chen, Si Huang PII: DOI: Reference:
S0009-8981(17)30324-8 doi: 10.1016/j.cca.2017.08.022 CCA 14842
To appear in:
Clinica Chimica Acta
Received date: Revised date: Accepted date:
30 January 2017 11 August 2017 20 August 2017
Please cite this article as: Suwen Qi, Depeng Xu, Qiaoliang Li, Ni Xie, Jun Xia, Qin Huo, Pu Li, Qiwen Chen, Si Huang , Metabonomics screening of serum identifies pyroglutamate as a diagnostic biomarker for nonalcoholic steatohepatitis, Clinica Chimica Acta (2017), doi: 10.1016/j.cca.2017.08.022
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 Title page Article title Metabonomics screening of serum identifies pyroglutamate as a diagnostic biomarker for nonalcoholic steatohepatitis
PT
Running title
RI
Metabolic biomarker of NASH
SC
Authors
Suwen Qia, Depeng Xua, Qiaoliang Li*, Ni Xie*b, Jun Xia b, Qin Huo b, Pu Lic, Qiwen
NU
Chena, Si Huanga
a
MA
Affiliations and addresses
Department of Biomedical Engineering, School of Medicine, Shenzhen University,
Shenzhen Second Hospital, The First Affiliated hospital of Shenzhen University,
PT E
b
D
Shenzhen 518060, China
Shenzhen 518060, P.R.China
Department of Laboratory Medicine, The Second Hospital Affiliated to Chongqing
CE
c
AC
Medical University, 410006, Chongqing, China *Corresponding author Name: Qiaoliang Li, PhD Email: [email protected] Phone & fax: +86 755 2562 6750 Address: Shenzhen University, Shenzhen, Guangdong Province, 518020, P.R. China
1
ACCEPTED MANUSCRIPT Abstract Objective A key step in managing non-alcoholic fatty liver disease (NAFLD) is to differentiate nonalcoholic steatohepatitis (NASH) from simple steatosis (SS). Method Serum samples were collected from three groups: NASH patients (N = 21),
PT
SS patients (N = 38) and healthy controls (N = 31). High performance liquid
RI
chromatography-mass spectrometry (HPLC-MS) was used to analyse the metabolic
SC
profile of the serum samples. The acquired data were processed by multivariate principal component analysis (PCA) and orthogonal partial least-squares-discriminant
NU
analysis (OPLS-DA) to identify novel metabolites. The potential biomarkers were
MA
quantitatively determined and their diagnostic power was further validated. Results A total of 56 metabolites were capable of distinguishing NASH from SS
D
samples based on the OPLS-DA model. Pyroglutamate was found to be the most
PT E
promising factor in distinguishing the NASH from SS groups. With an optimal cut-off value of 4.82 mmol/L, the sensitivity and specificity of the diagnosis of NASH were
CE
72% and 85%, respectively. The area under the receiver operating characteristic
AC
(AUROC) of the pyroglutamate levels of NASH versus SS patients was more than those of tumor necrosis factor-α, adiponectin and interleukin-8. Conclusion These data suggest that pyroglutamate may be a new and useful biomarker for the diagnosis of NASH. Key words: Non-alcoholic fatty liver disease; nonalcoholic steatohepatitis; simple steatosis; metabolic profiling; pyroglutamate
2
ACCEPTED MANUSCRIPT 1. Instruction Non-alcoholic fatty liver disease (NAFLD) is a spectrum of diseases that covers simple steatosis (SS) and the more severe progressive form of NAFLD—non-alcoholic steatohepatitis (NASH) [1]. NASH is a hallmark of NAFLD,
PT
which shows typical pathologic characteristics, such as lobular inflammation,
RI
hepatocyte clarification or ballooning and fibrosis progression. Hence, early diagnosis
SC
of NASH is extremely important for the treatment and prognosis of NAFLD patients [1-3].
NU
The actual gold standard for diagnosing NASH is to conduct a histologic evaluation
MA
of an invasive liver biopsy [4]. However, it is not always logistically feasible due to the high risk of complications. Imaging techniques and serum biomarkers are two
D
non-invasive approaches for the diagnosis of NASH. Ultrasound, computed
PT E
tomography (CT) and magnetic resonance imaging (MRI) are commonly employed imaging techniques for qualitative or quantitative assessments of hepatic steatosis [5,
CE
6]. Controlled attenuation parameter (CAP), which is a high accuracy ultrasound
AC
based technique, is the most promising method for NASH diagnosis, but it needs to be implemented with a probe in patients [7]. CT is inappropriate for assessing hepatic steatosis as it suffers from poor accuracy in detecting mild hepatic steatosis. MRI-proton density fat fraction (PDFF) is more accurate and sensitive, though the high costs associated with this technique limit its availability [8]. The biomarker-based liver fat score of NAFLD is the preferred method for quantification identification of steatosis. Serum biomarkers often are analyzed in 3
ACCEPTED MANUSCRIPT panels to establish live fat score models of NAFLD. SteatoTest TM (Biopredictive, Paris, France) is a score algorithm for the quantification of steatosis. The score ranges from 0 to 1 based on the values of twelve parameters that can be determined via biochemical assays, such as alanine transaminase [9], haptoglobin, total bilirubin,
PT
gamma-glutamyl transferase (GGT) and body mass index (BMI) [10]. Meta-analysis
RI
has shown that the SteatoTest TM has an area under receiver operating characteristic
SC
curve (AUROC) of 0.80 for diagnosing individuals with more than 33% steatosis [11]. Some biomarkers that correlate with NAFLD, such as tumor necrosis factor-α
NU
(TNF-α), adiponectin and Interleukin-8 (IL-8), had the same diagnostic score.
MA
However, no panel or single biomarker could effectively discriminate NASH from SS [12-14]. The aim of the proposed metabonomics study was to identify highly sensitive
PT E
diagnosis of NASH.
D
and specific biomarkers that are convenient and applicable for the differential
2. Materials and Methods
CE
2.1 Patients and controls
AC
A total of 59 participants (38 SS and 21 NASH) were enrolled from the Department of Gastroenterology and Infectious Diseases of the Second Hospital Affiliated to Chongqing Medical University from November 2015 to March 2016. NASH and SS were diagnosed based on the results of histological examinations or the combined results of clinical, laboratory and imaging examinations. Patients with known causes of chronic liver disease, including viral hepatitis (HBV and HCV), Wilson disease, haemochromatosis, or autoimmune hepatitis, were excluded. Patients with a history of 4
ACCEPTED MANUSCRIPT alcoholic consumption > 20 g/d for a period of > 5 years or > 50 g/d for a period of > 2 years were excluded [15]. Thirty-one healthy volunteers served as the controls. The baseline clinical characteristics of the NAFLD patients and controls are summarized in Table 1. Approximately 5 mL of fasted peripheral venous blood from patients and
PT
controls was collected, and then gently mixed by inverting the tube 8 to 10 times.
RI
After 30-40 min at room temperature for a clot to form, the whole blood were
SC
centrifuged 15 min at 3000 rpm at room temperature. One thousand microliters of serum were pipetted carefully into labeled tubes and stored in -80°C freezer. The
NU
procedure was conducted by the same personnel and the whole operation should be
MA
completed within 1 hr. If hemolysis or lipidemia was observed, the sample should be excluded.
D
This study was performed according to the guidelines of Ethics Committee of
PT E
Chongqing Medical University (Chongqing, China; approval No. CMU-2016-07) , which abided by the Declaration of Helsinki on ethical principles for medical
CE
research involving human subjects.
AC
2.2 HPLC-MS analysis
Approximately 500 µL of serum was mixed with 800 µL of cold methanol and precipitated at 12,000 RPM at 4°C for 15 min. Two-hundred microliters of supernatant was analyzed using high performance liquid chromatograph-mass spectrometer (HPLC-MS, Ultimate 3000 LC, Orbitrap Elite; Thermo Fisher Scientific, Waltham, MA, USA) analysis. The mass spectrum (MS) parameters in positive ion mode (ESI+ MODE) were as follows: Heater Temp, 300°C; Sheath Gas Flow rate, 45 5
ACCEPTED MANUSCRIPT arb; Aux Gas Flow Rate, 15 arb; Sweep Gas Flow Rate, 1arb; spray voltage, 3.0 KV; Capillary Temp, 350°C; and S-Lens RF Level, 30%. Standard controls for pyroglutamate were purchased from Sigma (St. Louis, MO, USA) and were used to prepare calibration curves.
PT
HPLC-MS data were preprocessed using SIEVE software (Thermo Fisher Scientific)
RI
and then normalized and edited in EXCEL 2010 software (Microsoft Corporation,
SC
Redmond, WA, USA). The results were visualized as a 2D data matrix, including metabolites’ retention time/mass-to-charge ratio (rt/mz), molecular weights
NU
(compMW) and peak intensities.
MA
2.3 Pattern recognition analysis
The metabolic profiles of the serum were first examined by unsupervised principal
D
component analysis (PCA) to obtain the score plots of NASH, SS and healthy controls
PT E
[16]. After an initial overview of the PCA analysis, orthogonal partial least-squares-discriminant analysis (OPLS-DA) was performed to develop a more
CE
sophisticated model with the specific discriminant information between the different
AC
groups [17, 18]. Variable importance in the projection (VIP) values of the OPLS-DA model (threshold value > 1) combined with t-test p values (p < 0.05) were adopted to find differences in the expression of metabolites. The metabolites were assigned according to the mass-to-charge ratio found in an online database (https://metlin.scripps.edu). 2.4 Measurements of biomarkers The quantitative determination of pyroglutamate was tested by HPLC-MS. 6
ACCEPTED MANUSCRIPT Calibration curves for pyroglutamate were prepared at different concentrations from 0.1 mmol/L to 20 mmol/L. TNF-α and IL-8 were quantified using a Luminex 200 (Austin, Texas, USA). The adiponectin levels were measured using an ELISA assays kit from Phoenix Pharmaceuticals, Inc. (Belmont, CA, USA).
PT
2.5 Data analysis
RI
Statistical analysis was performed using SPSS11.5 software (SPSS Company,
SC
Chicago, Illinois, USA). The data were compared using the t-test, and the threshold P value was set at 0.05 throughout the study.
NU
3. Results
MA
The HPLC-MS signals of all common metabolites were identified, and the extracted data were subsequently analyzed using multivariate statistics, including PCA and
D
OPLS-DA. From all samples, we obtained 1761 features in ESI+ MODE. Mass
PT E
chromatograms of serum obtained from control, simple steatosis nonalcoholic steatohepatitis were displayed in Figure 1. The edited data matrix was imported into
CE
SIMCA-P software (v13.0) and subsequently analyzed using multivariate statistics.
AC
3.1 Principal component analysis PCA was performed on samples to determine the overall metabolic differences between groups. In the PCA score plot, a single point represented a serum metabolic profile. Figure 2A displays the two-dimensional PCA score plots of controls and SS. The first two principal components (PC1 and PC2) exhibit the most efficient features contained in the data set. Twelve principal components were obtained with a R2X of 0.734 and a Q2 Y of 0.235. In general, for the unsupervised model, an R2X value 7
ACCEPTED MANUSCRIPT greater than 0.4 means that it is reliable [19]. Similarly, Figure 2B shows the PCA score plots of SS and NASH with an R2X of 0.694 and a Q2Y of 0.258. 3.2 OPLS-DA OPLS-DA is a supervised pattern recognition algorithm that is more focused on
PT
discriminatory variations. Healthy control and SS samples were discriminated with an
RI
R2 X of 0.306, R2 Y of 0.844, and Q2 Y of 0.717(Figure 3A). SS and NASH were
SC
discriminated with an R2 X of 0.324, R2 Y of 0.972, and Q2 Y of 0.956 (Figure 3B). According to a threshold value of VIP greater than 1, a total of 108 metabolites were
NU
capable of distinguishing NAFLD from healthy controls and 56 metabolites were
MA
capable of distinguishing NASH from SS. The metabolic profiles showed that the main altered metabolites were glutamate, kynurenic acid, among others. To screen
D
remarkable metabolites, we applied a stricter criterion that stated that the VIP of
PT E
metabolites should be greater than 1.3. Five metabolites, including uracil, α-linolenic acid, glutamate, L-glutamine and pyroglutamate, were identified as prominent factors
CE
in discriminating NASH from SS. The fold changes of pyroglutamate between SS
AC
versus control and NASH versus SS were 1.56 and 2.26, respectively, which were higher than those of the other metabolites (Table 2). 3.3 Biomarkers levels in serum A standard curve of y= 3366.8x + 5826.7 (peak area vs concentration) for pyroglutamate was obtained with a correlation coefficient of 0.9779. Two different levels of the sera (3.57 mmol/L, 9.61 mmol/L) were measured twenty times to evaluate the imprecision. The variable coefficient (CV) were 4.53% and 3.27%, 8
ACCEPTED MANUSCRIPT respectively. A known amount of pyroglutamate (15.0 mmol/L) was added to two human sera at different levels as spike-and-recovery controls. The average recoveries were 97.56% and 102.35%. We selected 25 control, 25 SS and 25 NASH samples to quantitatively determine the
PT
levels of pyroglutamate, TNF-α, IL-8 and adiponectin in serum. The results showed
RI
that the level of pyroglutamate was highest in the NASH, followed by the SS and
SC
control samples.
The median pyroglutamate level in the NAFLD (including NASH and SS) group was
NU
5.19 mmol/L (95% CI 3.89–8.19 mmol/L), which was dramatically higher than that of
MA
the control group, 1.48 mmol/L (95% CI 0.48–3.97 mmol/L) (Figure 4A). ROC analysis of the pyroglutamate levels of NAFLD samples versus controls yielded
D
AUROC of 0.880 (Figure 4B). With an optimal cut-off value of 3.20 mmol/L, the
PT E
sensitivity and specificity for the diagnosis of NAFLD were 73% and 87%, respectively.
CE
The pyroglutamate levels in the NASH group were 7.89 mmol/L (95% CI 5.35–10.41
AC
mmol/L), which was higher than that of the SS group at 3.57 mmol/L (95% CI 0.98– 7.34 mmol/L). With a cut-off value of 4.82 mmol/L, ROC analysis yielded an AUROC of 0.846; the sensitivity and specificity for the differential diagnosis of NASH from SS were 72% and 85%, respectively. The level of adiponectin, IL-8 and TNF-α in patients with SS and NASH are shown in Table 3. Groupwise comparisons showed decreased adiponectin (Se 64%, Sep 56%, AUROC 0.723), decreased IL-8 (Se 65%, Sep 68%, AUROC 0.757) and increased 9
ACCEPTED MANUSCRIPT TNF-α (Se 72%, Sep 76%, AUROC 0.810), which were associated with NASH. The ROC curves of pyroglutamate, adiponectin, IL-8 and TNF-α of NASH versus SS samples are shown in Figure 5. In summary, pyroglutamate performed better than the others as a single diagnostic biomarker of NASH. A combination of 2-3 biomarkers is
PT
usually used in clinical lab to improve diagnosis sensitivity. The diagnostic
RI
performances of pyroglutamate combination with adiponectin, IL-8, TNF-α are
SC
summarized in Table 4. In addition, the combination of pyroglutamate with standard laboratory parameters such as GGT and triglycerides also are displayed in Table 4.
NU
4. Discussion
MA
Distinguishing NASH from SS is a key challenge in the management of NAFLD patients. Serum biomarkers are the most available noninvasive metrics for the
D
diagnosis of NASH [20]. In this study, we studied the serum metabolic profiles of
PT E
NAFLD based on HPLC-MS to identify NASH biomarkers. PCA score plots clearly distinguished the control, SS and NASH groups. As a highly sensitive analytical
CE
method, OPLS-DA, identified the altered metabolites in the NASH group compared
AC
with the SS group. Five metabolites, including uracil, α-Linolenic acid, glutamate, L-glutamine and pyroglutamate, were identified as the most prominent biomarkers. The fold changes of pyroglutamate between SS and control and NASH and SS were 1.56 and 2.26, respectively, which were higher than those of the other biomarkers. Pyroglutamate is also known as 5-oxoproline. The metabolism of pyroglutamate is related to the gamma-glutamyl cycle, which is the primary pathway for glutathione synthesis and degradation [21, 22]. The level of pyroglutamate is often used to screen 10
ACCEPTED MANUSCRIPT for glutathione synthetase deficiency in pediatric patients [23]. Children with deficiencies of either glutathione synthase or 5-oxoprolinase will present with metabolic acidosis or other abnormalities due to the accumulation of pyroglutamic acid [24-26]. When glycine is insufficiently available to support glutathione recovery
PT
via the renal gamma-glutamyl cycle, the rising level of pyroglutamate is regarded as
RI
an index of glycine insufficiency. In addition, the components of gamma-glutamyl
SC
cycle such as glutathione (CompMW: 307.08, retention time: 0.94) and cysteine (CompMW: 283.28, retention time: 7.23) were identified in the LC-MS data.
NU
Although both were capable of distinguishing between SS or NASH group and
MA
controls, glutathione and cysteine did not show significant changing between NASH and SS.
D
Furthermore, pyroglutamate has been proposed as a potential biomarker for enhanced
PT E
oxidative stress in liver cells [27, 28]. Studies have shown enhanced levels of pyroglutamate in various biofluids or tissues following administration of hepatic
CE
toxicants, such as acetaminophen [29] and bromobenzene [30]. A growing number of
AC
reports have documented increased pyroglutamate in the blood, urine, or both in cases of metabolic acidosis complicating paracetamol poisoning [31]. However, there is no study focusing on the role of pyroglutamate in the progress of NAFLD. The metabolic profiling conducted in this study showed that pyroglutamate is a remarkable metabolite that is correlated with NAFLD. Its sensitivity and specificity are equal to those of SteatoTestTM, which is a steatosis score algorithm based on 12 clinical parameters. The AUROC value (0.880) for the diagnosis of NAFLD was more 11
ACCEPTED MANUSCRIPT than 0.80 of that of SteatoTestTM [11]. Pyroglutamate profiling has the capacity to identify NAFLD from healthy controls. Nevertheless, more should be done to discriminate patients with SS and NASH because both are very similar. The performance of pyroglutamate in diagnosing
PT
NASH was carefully compared with that of adiponectin, IL-8 and TNF-α, which are
RI
correlated with NASH. Low adiponectin is a characteristic of steatosis progression
SC
and leads to an increase of intra-cellular fatty acids [12, 32]. In this study, the level of adiponectin in the SS group was higher than that in controls, whereas there was no
NU
significant difference between the NASH and SS groups. IL-8 is an independent
MA
factor that is associated with NASH and was noticeably decreased in NASH compared to SS samples. The levels of TNF-α increased gradually from healthy
D
controls to SS, then to NASH. Although adiponectin, IL-8 and TNF-α showed some
PT E
diagnostic significance, the results showed that their sensitivity and specificity were inferior to those of pyroglutamate.
CE
In addition to using single serum biomarkers related to NASH, some non-invasive
AC
models based on several biomarkers have been used to diagnosis NASH. The fatty liver index (FLI) is a routinely used tool based on four components, including BMI, waist circumference, triglycerides, and GGT [33, 34]. However, the unreliability of the components limited the interpretation of FLI [35]. Munkhzul et al. [33] established an index of NASH (ION) model based on anthropometric, clinical, ultrasound, laboratory, and mortality data, and its sensitivity (81%) was better, but its specificity (82%) was less than the pyroglutamate biomarker. In consideration of the 12
ACCEPTED MANUSCRIPT cost and operating performance, pyroglutamate still has an advantage in diagnosing NASH. The panel of pyroglutamate with other biomarkers contributed to enhance the diagnostic performance. The sensitivity of combination of pyroglutamate, adiponectin and TNF-α reached 92%.
PT
The difference in expression of pyroglutamate between the SS and NASH groups may
RI
be a reflection of their role in disease pathogenesis. It was reported that increased
SC
pyroglutamate excretion is an indication of abnormal response to oxidative stress [36]. NAFLD is accompanied by broad effects on mitochondria, which is the main location
NU
for fatty acid oxidation. When SS progresses to NASH, mitochondria have impaired
MA
respiratory coupling, which may increase the reductive load on respiration and form reactive oxygen species (ROS);ROS damage hepatocytes, trigger inflammation, and
D
elevate pyroglutamate levels [37, 38] .In conclusion, we identified pyroglutamate to
PT E
be a novel biomarker for the diagnosis of NASH. Future studies with larger sample sizes in clinical settings are worth pursuing.
CE
Acknowledgements
AC
This work was supported by the Project of the National Science Foundation of China (61401285
and
81401750),
Natural
Science
Foundation
of
Guangdong
(2016A030313029), Scientific Plan Program of Guangdong (2014A020212038), Innovation Program of Shenzhen (JCYJ20150330102720122), and International Cooperation Foundation of Shenzhen (GJHZ20160301163138685), Shenzhen Science Plan
(JCYJ2016030711492524,
JCYJ20150330102720122,
CXZZ201418182638764) and The Fund of Shenzhen University (2015006). 13
and
ACCEPTED MANUSCRIPT References [1]
V. Ajmera, E.R. Perito, N.M. Bass, N.A. Terrault, K.P. Yates, R. Gill, et al.,Novel plasma biomarkers associated with liver disease severity in adults with nonalcoholic fatty liver disease, Hepatology. 65(2017) 65-77. E. Bugianesi, C. Rosso, and H. Cortez-Pinto,How to diagnose NAFLD in
PT
[2]
M.E. Rinella and A.J. Sanyal,Management of NAFLD: a stage-based approach,
SC
[3]
RI
2016, J Hepatol. 65(2016) 643-4.
Nat Rev Gastroenterol Hepatol. 13(2016) 196-205.
C. Stern and L. Castera,Non-invasive diagnosis of hepatic steatosis, Hepatol
NU
[4]
[5]
MA
Int. 2016)
S. Kinner, S.B. Reeder, and T. Yokoo,Quantitative Imaging Biomarkers of
J. Pirhonen, J. Arola, S. Sadevirta, P. Luukkonen, S.M. Karppinen, T.
PT E
[6]
D
NAFLD, Dig Dis Sci. 61(2016) 1337-47.
Pihlajaniemi, et al.,Continuous Grading of Early Fibrosis in NAFLD Using
CE
Label-Free Imaging: A Proof-of-Concept Study, PLoS One. 11(2016)
[7]
AC
e0147804.
R.P. Myers, A. Pollett, R. Kirsch, G. Pomier-Layrargues, M. Beaton, M. Levstik, et al.,Controlled Attenuation Parameter (CAP): a noninvasive method for the detection of hepatic steatosis based on transient elastography, Liver Int. 32(2012) 902-10.
[8]
E. Hashimoto, M. Taniai, and K. Tokushige,Characteristics and diagnosis of NAFLD/NASH, J Gastroenterol Hepatol. 28 Suppl 4(2013) 64-70. 14
ACCEPTED MANUSCRIPT [9]
J. Vulevic, A. Juric, G.E. Walton, S.P. Claus, G. Tzortzis, R.E. Toward, et al.,Influence of galacto-oligosaccharide mixture (B-GOS) on gut microbiota, immune parameters and metabonomics in elderly persons, Br J Nutr. 114(2015) 586-95. T. Poynard, V. Ratziu, S. Naveau, D. Thabut, F. Charlotte, D. Messous, et
PT
[10]
RI
al.,The diagnostic value of biomarkers (SteatoTest) for the prediction of liver
[11]
SC
steatosis, Comp Hepatol. 4(2005) 10.
T. Poynard, G. Lassailly, E. Diaz, K. Clement, R. Caiazzo, J. Tordjman, et
NU
al.,Performance of biomarkers FibroTest, ActiTest, SteatoTest, and NashTest
MA
in patients with severe obesity: meta analysis of individual patient data, PLoS One. 7(2012) e30325.
L. Abenavoli and V. Peta,Role of adipokines and cytokines in non-alcoholic
D
[12]
[13]
PT E
fatty liver disease, Rev Recent Clin Trials. 9(2014) 134-40. M.H. Jarrar, A. Baranova, R. Collantes, B. Ranard, M. Stepanova, C. Bennett,
CE
et al.,Adipokines and cytokines in non-alcoholic fatty liver disease, Aliment
[14]
AC
Pharmacol Ther. 27(2008) 412-21. S. Abiru, K. Migita, Y. Maeda, M. Daikoku, M. Ito, K. Ohata, et al.,Serum cytokine and soluble cytokine receptor levels in patients with non-alcoholic steatohepatitis, Liver Int. 26(2006) 39-45. [15]
V.R. Mas, R. Fassnacht, K.J. Archer, and D. Maluf,Molecular mechanisms involved in the interaction effects of alcohol and hepatitis C virus in liver cirrhosis, Mol Med. 16(2010) 287-97. 15
ACCEPTED MANUSCRIPT [16]
B. Worley and R. Powers,PCA as a practical indicator of OPLS-DA model reliability, Curr Metabolomics. 4(2016) 97-103.
[17]
S. Wagner, K. Scholz, M. Donegan, L. Burton, J. Wingate, and W. Volkel,Metabonomics and biomarker discovery: LC-MS metabolic profiling
PT
and constant neutral loss scanning combined with multivariate data analysis
Y. Wang, H. Man, J. Gao, X. Liu, X. Ren, J. Chen, et al.,Plasma metabonomics
SC
[18]
RI
for mercapturic acid analysis, Anal Chem. 78(2006) 1296-305.
study on toxicity biomarker in rats treated with Euphorbia fischeriana based
S. Qi, X. Ouyang, L. Wang, W. Peng, J. Wen, and Y. Dai,A pilot metabolic
MA
[19]
NU
on LC-MS, Biomed Chromatogr. 30(2016) 1386-96.
profiling study in serum of patients with chronic kidney disease based on (1)
H. Yki-Jarvinen,Diagnosis of non-alcoholic fatty liver disease (NAFLD),
PT E
[20]
D
H-NMR-spectroscopy, Clin Transl Sci. 5(2012) 379-85.
Diabetologia. 59(2016) 1104-11. D.B. Liss, M.S. Paden, E.S. Schwarz, and M.E. Mullins,What is the clinical
CE
[21]
AC
significance of 5-oxoproline (pyroglutamic acid) in high anion gap metabolic acidosis following paracetamol (acetaminophen) exposure?, Clin Toxicol (Phila). 51(2013) 817-27. [22]
Y.M. Yu, C.M. Ryan, Z.W. Fei, X.M. Lu, L. Castillo, J.T. Schultz, et al.,Plasma L-5-oxoproline kinetics and whole blood glutathione synthesis rates in severely burned adult humans, Am J Physiol Endocrinol Metab. 282(2002) E247-58. 16
ACCEPTED MANUSCRIPT [23]
C.L. Foot, J.F. Fraser, and D.V. Mullany,Pyroglutamic acidosis in a renal transplant patient, Nephrol Dial Transplant. 20(2005) 2836-8.
[24]
I.A. Almaghlouth, J.Y. Mohamed, M. Al-Amoudi, L. Al-Ahaidib, A. Al-Odaib, and F.S. Alkuraya,5-Oxoprolinase deficiency: report of the first human
L. Zand, A. Muriithi, E. Nelsen, P.M. Franco, E.L. Greene, Q. Qian, et
RI
[25]
PT
OPLAH mutation, Clin Genet. 82(2012) 193-6.
SC
al.,Severe anion gap metabolic acidosis from acetaminophen use secondary to 5-oxoproline (pyroglutamic acid) accumulation, Am J Med Sci. 344(2012)
P. Tailor, T. Raman, C.L. Garganta, R. Njalsson, K. Carlsson, E. Ristoff, et
MA
[26]
NU
501-4.
al.,Recurrent high anion gap metabolic acidosis secondary to 5-oxoproline
E.J. Rolleman, E.J. Hoorn, P. Didden, and R. Zietse,Guilty as charged:
PT E
[27]
D
(pyroglutamic acid), Am J Kidney Dis. 46(2005) e4-10.
unmeasured urinary anions in a case of pyroglutamic acidosis, Neth J Med.
T. Soga, R. Baran, M. Suematsu, Y. Ueno, S. Ikeda, T. Sakurakawa, et
AC
[28]
CE
66(2008) 351-3.
al.,Differential metabolomics reveals ophthalmic acid as an oxidative stress biomarker indicating hepatic glutathione consumption, J Biol Chem. 281(2006) 16768-76. [29]
D.T.
Lawrence,
L.K.
Bechtel,
N.P.
Charlton,
and
C.P.
Holstege,5-oxoproline-induced anion gap metabolic acidosis after an acute acetaminophen overdose, J Am Osteopath Assoc. 110(2010) 545-51. 17
ACCEPTED MANUSCRIPT [30]
N.J. Waters, C.J. Waterfield, R.D. Farrant, E. Holmes, and J.K. Nicholson,Integrated
metabonomic
analysis
of
bromobenzene-induced
hepatotoxicity: novel induction of 5-oxoprolinosis, J Proteome Res. 5(2006) 1448-59. S. Geenen, C. Guallar-Hoyas, F. Michopoulos, J.G. Kenna, K.L. Kolaja, H.V.
PT
[31]
RI
Westerhoff, et al.,HPLC-MS/MS methods for the quantitative analysis of
SC
5-oxoproline (pyroglutamate) in rat plasma and hepatic cell line culture medium, J Pharm Biomed Anal. 56(2011) 655-63.
R. Kumar, S. Prakash, S. Chhabra, V. Singla, K. Madan, S.D. Gupta, et
NU
[32]
MA
al.,Association of pro-inflammatory cytokines, adipokines & oxidative stress with insulin resistance & non-alcoholic fatty liver disease, Indian J Med Res.
M. Otgonsuren, M.J. Estep, N. Hossain, E. Younossi, S. Frost, L. Henry, et
PT E
[33]
D
136(2012) 229-36.
al.,Single non-invasive model to diagnose non-alcoholic fatty liver disease
CE
(NAFLD) and non-alcoholic steatohepatitis (NASH), J Gastroenterol Hepatol.
[34]
AC
29(2014) 2006-13. G. Calori, G. Lattuada, F. Ragogna, M.P. Garancini, P. Crosignani, M. Villa, et al.,Fatty liver index and mortality: the Cremona study in the 15th year of follow-up, Hepatology. 54(2011) 145-52. [35]
M.A. Borman, F. Ladak, P. Crotty, A. Pollett, R. Kirsch, G. Pomier-Layrargues, et al.,The Fatty Liver Index has limited utility for the detection and quantification of hepatic steatosis in obese patients, Hepatol Int. 7(2013) 18
ACCEPTED MANUSCRIPT 592-9. [36]
G. Brooker, J. Jeffery, T. Nataraj, M. Sair, and R. Ayling,High anion gap metabolic acidosis secondary to pyroglutamic aciduria (5-oxoprolinuria): association with prescription drugs and malnutrition, Ann Clin Biochem.
W.B. Ashworth, N.A. Davies, and I.D. Bogle,A Computational Model of
RI
[37]
PT
44(2007) 406-9.
SC
Hepatic Energy Metabolism: Understanding Zonated Damage and Steatosis in NAFLD, PLoS Comput Biol. 12(2016) e1005105.
NU
M. Ono, N. Okamoto, and T. Saibara,The latest idea in NAFLD/NASH
CE
PT E
D
MA
pathogenesis, Clin J Gastroenterol. 3(2010) 263-70.
AC
[38]
19
ACCEPTED MANUSCRIPT Table 1. Clinical and biochemical characteristics of NAFLD and controls Controls (n=31) AGE (yrs)
(n=38)
52.8±14.6 25:7
31:9
(n=21) 56.3±12.9 24:8
Body mass index
47.3±6.8
47.5±8.1
Alanine aminotransferase (U/L)
35.6±7.2
65.7±20.8*
Aspartate aminotransferase (U/L)
21.3±6.7
38.7±12.8
50.7±23.5*
gamma -glutamyl transferase (U/L)
36.8±8.4
78.8±14.3
152±15.9*†
Triglyceride (mg/dL)
136±66.4
150±84.1
198±95.2*†
180±30.7
197±31.8
192±32.5
104±20.6
110±25.9
134±40.5*
D
Glucose (mg/dL)
41.4±23.5
SC
NU
MA
Cholesterol (mg/dL)
PT E
*p < 0.05, compared to controls. †
58.7±14.5
NASH
RI
Gender (M:F)
SS
PT
Parameter
p < 0.05, compared to SS.
AC
CE
Data refer to the number of patients or the mean ± SD.
20
47.5±8.4
ACCEPTED MANUSCRIPT Table 2. Fold changes of identified metabolites between the SS versus control groups or NASH versus SS groups based on the OPLS-DA. SS vs Control
NASH vs SS
Uracil
1.34
1.63
α-Linolenic Acid
1.33
1.52
Glutamate
1.56
1.80
L-Glutamine
1.56
Pyroglutamate
1.56
RI
PT
Metabolite
AC
CE
PT E
D
MA
NU
SC
1.81
21
2.26
ACCEPTED MANUSCRIPT Table 3. Groupwise comparisons of serum biomarkers for healthy controls and patients with SS and NASH Control (n = 25)
SS (n = 25)
NASH (n = 25)
(mmol/L)
1.48±1.21
3.18±1.53*
7.85±3.22†
Adiponectin (ug/mL)
13.25±8.36
11.69±6.25
7.08±4.40†
TNF-α (pg/mL)
3.89±3.21
4.26±2.96
IL-8 (pg/ml)
25.15±13.87
26.02±10.69
†
p < 0.05, compared to SS.
RI
SC
NU MA
*p < 0.05, compared to controls.
PT
Pyroglutamate
AC
CE
PT E
D
Data refer to the number of patients or the mean ± SD.
22
8.99±4.63† 17.73±8.70†
ACCEPTED MANUSCRIPT Table 4. The diagnostic performance of biomarkers for NASH
Biomarker
Sensitivity
Specificity
Positive
Negative
predictive value
predictive value
Accuracy
64%
56%
59%
60%
60%
TNF-α
72%
76%
75%
73%
74%
IL-8
65%
68%
86%
triglyceride
54%
75%
68%
GGT
57%
80%
pyroglutamate
72%
85%
pyroglutamate +adiponectin
82%
80%
pyroglutamate+TNF-α
83%
77%
pyroglutamate+IL-8
85%
82%
66%
62%
65%
74%
65%
69%
62%
90%
82%
58%
93%
80%
70%
86%
79%
91%
69%
82%
86%
81%
78%
88%
85%
86%
93%
89%
91%
87%
83%
84%
86%
pyroglutamate+ triglyceride
83%
86%
85%
83%
85%
pyroglutamate +GGT
85%
80%
82%
88%
85%
91%
76%
79%
90%
84%
+adiponectin+IL-8
90%
AC
pyroglutamate+
CE
+TNF-α
92%
PT E
+adiponectin+TNF-α pyroglutamate+IL-8
triglyceride+GGT
NU
D
pyroglutamate
MA
pyroglutamate
RI
40%
SC
PT
adiponectin
23
ACCEPTED MANUSCRIPT Figure legends: Figure 1. Mass chromatograms of serum obtained from control (A), simple steatosis (B) nonalcoholic steatohepatitis (C).
PT
Figure 2. (A) The PCA score plots of healthy controls (blue square) and SS patients
RI
(red triangle). (B) The PCA score plots of the SS (red triangle) and NASH
SC
(inverse yellow triangle) groups.
NU
Figure 3. (A) The OPLS-DA score plots of healthy controls (blue square) and SS
D
(inverse yellow triangle) groups.
MA
patients (red triangle). (B) The OPLS-DA score plots of SS (red triangle) and NASH
PT E
Figure 4. (A) The serum pyroglutamate levels in controls and patients with SS and NASH. The short transverse line represents the median of diff erent groups. (B) The
CE
ROC analysis of pyroglutamate in serum of NAFLD versus control with AUROC of
AC
0.880 and SS versus controls with AUROC of 0.846, respectively.
Figure 5. The ROC analysis of pyroglutamate, adiponectin, TNF-α and IL-8 in serum of NASH versus SS.
24
ACCEPTED MANUSCRIPT
AC
CE
PT E
D
MA
NU
SC
RI
PT
Fig. 1
25
ACCEPTED MANUSCRIPT
AC
CE
PT E
D
MA
NU
SC
RI
PT
Fig. 2
26
ACCEPTED MANUSCRIPT
AC
CE
PT E
D
MA
NU
SC
RI
PT
Fig. 3
27
ACCEPTED MANUSCRIPT
AC
CE
PT E
D
MA
NU
SC
RI
PT
Fig. 4
28
ACCEPTED MANUSCRIPT
AC
CE
PT E
D
MA
NU
SC
RI
PT
Fig. 5
29
ACCEPTED MANUSCRIPT Highlights 1. High performance liquid chromatography-mass spectrometry (HPLC-MS) was
used to analyse the metabolic profile of the serum samples from patients with nonalcoholic steatohepatitis (NASH) from simple steatosis (SS), and controls.
PT
2. Five metabolites, including uracil, α-linolenic acid, glutamate, L-glutamine and
RI
pyroglutamate, were identified as prominent factors in discriminating NASH from
SC
SS.
3. Pyroglutamate was found to be the most promising factor in distinguishing the
NU
NASH from SS groups with an optimal cut-off value of 4.82 mmol/L, the
MA
sensitivity and specificity of the diagnosis of NASH were 72% and 85%, respectively.
D
4. The potential diagnostic power of pyroglutamate was superior to other biomarkers
AC
CE
PT E
such as factor-α (TNF-α), adiponectin and interleukin-8 (IL-8).
30
Suwen Qi, Depeng Xu, Qiaoliang Li, Ni Xie, Jun Xia, Qin Huo, Pu Li, Qiwen Chen, Si Huang PII: DOI: Reference:
S0009-8981(17)30324-8 doi: 10.1016/j.cca.2017.08.022 CCA 14842
To appear in:
Clinica Chimica Acta
Received date: Revised date: Accepted date:
30 January 2017 11 August 2017 20 August 2017
Please cite this article as: Suwen Qi, Depeng Xu, Qiaoliang Li, Ni Xie, Jun Xia, Qin Huo, Pu Li, Qiwen Chen, Si Huang , Metabonomics screening of serum identifies pyroglutamate as a diagnostic biomarker for nonalcoholic steatohepatitis, Clinica Chimica Acta (2017), doi: 10.1016/j.cca.2017.08.022
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 Title page Article title Metabonomics screening of serum identifies pyroglutamate as a diagnostic biomarker for nonalcoholic steatohepatitis
PT
Running title
RI
Metabolic biomarker of NASH
SC
Authors
Suwen Qia, Depeng Xua, Qiaoliang Li*, Ni Xie*b, Jun Xia b, Qin Huo b, Pu Lic, Qiwen
NU
Chena, Si Huanga
a
MA
Affiliations and addresses
Department of Biomedical Engineering, School of Medicine, Shenzhen University,
Shenzhen Second Hospital, The First Affiliated hospital of Shenzhen University,
PT E
b
D
Shenzhen 518060, China
Shenzhen 518060, P.R.China
Department of Laboratory Medicine, The Second Hospital Affiliated to Chongqing
CE
c
AC
Medical University, 410006, Chongqing, China *Corresponding author Name: Qiaoliang Li, PhD Email: [email protected] Phone & fax: +86 755 2562 6750 Address: Shenzhen University, Shenzhen, Guangdong Province, 518020, P.R. China
1
ACCEPTED MANUSCRIPT Abstract Objective A key step in managing non-alcoholic fatty liver disease (NAFLD) is to differentiate nonalcoholic steatohepatitis (NASH) from simple steatosis (SS). Method Serum samples were collected from three groups: NASH patients (N = 21),
PT
SS patients (N = 38) and healthy controls (N = 31). High performance liquid
RI
chromatography-mass spectrometry (HPLC-MS) was used to analyse the metabolic
SC
profile of the serum samples. The acquired data were processed by multivariate principal component analysis (PCA) and orthogonal partial least-squares-discriminant
NU
analysis (OPLS-DA) to identify novel metabolites. The potential biomarkers were
MA
quantitatively determined and their diagnostic power was further validated. Results A total of 56 metabolites were capable of distinguishing NASH from SS
D
samples based on the OPLS-DA model. Pyroglutamate was found to be the most
PT E
promising factor in distinguishing the NASH from SS groups. With an optimal cut-off value of 4.82 mmol/L, the sensitivity and specificity of the diagnosis of NASH were
CE
72% and 85%, respectively. The area under the receiver operating characteristic
AC
(AUROC) of the pyroglutamate levels of NASH versus SS patients was more than those of tumor necrosis factor-α, adiponectin and interleukin-8. Conclusion These data suggest that pyroglutamate may be a new and useful biomarker for the diagnosis of NASH. Key words: Non-alcoholic fatty liver disease; nonalcoholic steatohepatitis; simple steatosis; metabolic profiling; pyroglutamate
2
ACCEPTED MANUSCRIPT 1. Instruction Non-alcoholic fatty liver disease (NAFLD) is a spectrum of diseases that covers simple steatosis (SS) and the more severe progressive form of NAFLD—non-alcoholic steatohepatitis (NASH) [1]. NASH is a hallmark of NAFLD,
PT
which shows typical pathologic characteristics, such as lobular inflammation,
RI
hepatocyte clarification or ballooning and fibrosis progression. Hence, early diagnosis
SC
of NASH is extremely important for the treatment and prognosis of NAFLD patients [1-3].
NU
The actual gold standard for diagnosing NASH is to conduct a histologic evaluation
MA
of an invasive liver biopsy [4]. However, it is not always logistically feasible due to the high risk of complications. Imaging techniques and serum biomarkers are two
D
non-invasive approaches for the diagnosis of NASH. Ultrasound, computed
PT E
tomography (CT) and magnetic resonance imaging (MRI) are commonly employed imaging techniques for qualitative or quantitative assessments of hepatic steatosis [5,
CE
6]. Controlled attenuation parameter (CAP), which is a high accuracy ultrasound
AC
based technique, is the most promising method for NASH diagnosis, but it needs to be implemented with a probe in patients [7]. CT is inappropriate for assessing hepatic steatosis as it suffers from poor accuracy in detecting mild hepatic steatosis. MRI-proton density fat fraction (PDFF) is more accurate and sensitive, though the high costs associated with this technique limit its availability [8]. The biomarker-based liver fat score of NAFLD is the preferred method for quantification identification of steatosis. Serum biomarkers often are analyzed in 3
ACCEPTED MANUSCRIPT panels to establish live fat score models of NAFLD. SteatoTest TM (Biopredictive, Paris, France) is a score algorithm for the quantification of steatosis. The score ranges from 0 to 1 based on the values of twelve parameters that can be determined via biochemical assays, such as alanine transaminase [9], haptoglobin, total bilirubin,
PT
gamma-glutamyl transferase (GGT) and body mass index (BMI) [10]. Meta-analysis
RI
has shown that the SteatoTest TM has an area under receiver operating characteristic
SC
curve (AUROC) of 0.80 for diagnosing individuals with more than 33% steatosis [11]. Some biomarkers that correlate with NAFLD, such as tumor necrosis factor-α
NU
(TNF-α), adiponectin and Interleukin-8 (IL-8), had the same diagnostic score.
MA
However, no panel or single biomarker could effectively discriminate NASH from SS [12-14]. The aim of the proposed metabonomics study was to identify highly sensitive
PT E
diagnosis of NASH.
D
and specific biomarkers that are convenient and applicable for the differential
2. Materials and Methods
CE
2.1 Patients and controls
AC
A total of 59 participants (38 SS and 21 NASH) were enrolled from the Department of Gastroenterology and Infectious Diseases of the Second Hospital Affiliated to Chongqing Medical University from November 2015 to March 2016. NASH and SS were diagnosed based on the results of histological examinations or the combined results of clinical, laboratory and imaging examinations. Patients with known causes of chronic liver disease, including viral hepatitis (HBV and HCV), Wilson disease, haemochromatosis, or autoimmune hepatitis, were excluded. Patients with a history of 4
ACCEPTED MANUSCRIPT alcoholic consumption > 20 g/d for a period of > 5 years or > 50 g/d for a period of > 2 years were excluded [15]. Thirty-one healthy volunteers served as the controls. The baseline clinical characteristics of the NAFLD patients and controls are summarized in Table 1. Approximately 5 mL of fasted peripheral venous blood from patients and
PT
controls was collected, and then gently mixed by inverting the tube 8 to 10 times.
RI
After 30-40 min at room temperature for a clot to form, the whole blood were
SC
centrifuged 15 min at 3000 rpm at room temperature. One thousand microliters of serum were pipetted carefully into labeled tubes and stored in -80°C freezer. The
NU
procedure was conducted by the same personnel and the whole operation should be
MA
completed within 1 hr. If hemolysis or lipidemia was observed, the sample should be excluded.
D
This study was performed according to the guidelines of Ethics Committee of
PT E
Chongqing Medical University (Chongqing, China; approval No. CMU-2016-07) , which abided by the Declaration of Helsinki on ethical principles for medical
CE
research involving human subjects.
AC
2.2 HPLC-MS analysis
Approximately 500 µL of serum was mixed with 800 µL of cold methanol and precipitated at 12,000 RPM at 4°C for 15 min. Two-hundred microliters of supernatant was analyzed using high performance liquid chromatograph-mass spectrometer (HPLC-MS, Ultimate 3000 LC, Orbitrap Elite; Thermo Fisher Scientific, Waltham, MA, USA) analysis. The mass spectrum (MS) parameters in positive ion mode (ESI+ MODE) were as follows: Heater Temp, 300°C; Sheath Gas Flow rate, 45 5
ACCEPTED MANUSCRIPT arb; Aux Gas Flow Rate, 15 arb; Sweep Gas Flow Rate, 1arb; spray voltage, 3.0 KV; Capillary Temp, 350°C; and S-Lens RF Level, 30%. Standard controls for pyroglutamate were purchased from Sigma (St. Louis, MO, USA) and were used to prepare calibration curves.
PT
HPLC-MS data were preprocessed using SIEVE software (Thermo Fisher Scientific)
RI
and then normalized and edited in EXCEL 2010 software (Microsoft Corporation,
SC
Redmond, WA, USA). The results were visualized as a 2D data matrix, including metabolites’ retention time/mass-to-charge ratio (rt/mz), molecular weights
NU
(compMW) and peak intensities.
MA
2.3 Pattern recognition analysis
The metabolic profiles of the serum were first examined by unsupervised principal
D
component analysis (PCA) to obtain the score plots of NASH, SS and healthy controls
PT E
[16]. After an initial overview of the PCA analysis, orthogonal partial least-squares-discriminant analysis (OPLS-DA) was performed to develop a more
CE
sophisticated model with the specific discriminant information between the different
AC
groups [17, 18]. Variable importance in the projection (VIP) values of the OPLS-DA model (threshold value > 1) combined with t-test p values (p < 0.05) were adopted to find differences in the expression of metabolites. The metabolites were assigned according to the mass-to-charge ratio found in an online database (https://metlin.scripps.edu). 2.4 Measurements of biomarkers The quantitative determination of pyroglutamate was tested by HPLC-MS. 6
ACCEPTED MANUSCRIPT Calibration curves for pyroglutamate were prepared at different concentrations from 0.1 mmol/L to 20 mmol/L. TNF-α and IL-8 were quantified using a Luminex 200 (Austin, Texas, USA). The adiponectin levels were measured using an ELISA assays kit from Phoenix Pharmaceuticals, Inc. (Belmont, CA, USA).
PT
2.5 Data analysis
RI
Statistical analysis was performed using SPSS11.5 software (SPSS Company,
SC
Chicago, Illinois, USA). The data were compared using the t-test, and the threshold P value was set at 0.05 throughout the study.
NU
3. Results
MA
The HPLC-MS signals of all common metabolites were identified, and the extracted data were subsequently analyzed using multivariate statistics, including PCA and
D
OPLS-DA. From all samples, we obtained 1761 features in ESI+ MODE. Mass
PT E
chromatograms of serum obtained from control, simple steatosis nonalcoholic steatohepatitis were displayed in Figure 1. The edited data matrix was imported into
CE
SIMCA-P software (v13.0) and subsequently analyzed using multivariate statistics.
AC
3.1 Principal component analysis PCA was performed on samples to determine the overall metabolic differences between groups. In the PCA score plot, a single point represented a serum metabolic profile. Figure 2A displays the two-dimensional PCA score plots of controls and SS. The first two principal components (PC1 and PC2) exhibit the most efficient features contained in the data set. Twelve principal components were obtained with a R2X of 0.734 and a Q2 Y of 0.235. In general, for the unsupervised model, an R2X value 7
ACCEPTED MANUSCRIPT greater than 0.4 means that it is reliable [19]. Similarly, Figure 2B shows the PCA score plots of SS and NASH with an R2X of 0.694 and a Q2Y of 0.258. 3.2 OPLS-DA OPLS-DA is a supervised pattern recognition algorithm that is more focused on
PT
discriminatory variations. Healthy control and SS samples were discriminated with an
RI
R2 X of 0.306, R2 Y of 0.844, and Q2 Y of 0.717(Figure 3A). SS and NASH were
SC
discriminated with an R2 X of 0.324, R2 Y of 0.972, and Q2 Y of 0.956 (Figure 3B). According to a threshold value of VIP greater than 1, a total of 108 metabolites were
NU
capable of distinguishing NAFLD from healthy controls and 56 metabolites were
MA
capable of distinguishing NASH from SS. The metabolic profiles showed that the main altered metabolites were glutamate, kynurenic acid, among others. To screen
D
remarkable metabolites, we applied a stricter criterion that stated that the VIP of
PT E
metabolites should be greater than 1.3. Five metabolites, including uracil, α-linolenic acid, glutamate, L-glutamine and pyroglutamate, were identified as prominent factors
CE
in discriminating NASH from SS. The fold changes of pyroglutamate between SS
AC
versus control and NASH versus SS were 1.56 and 2.26, respectively, which were higher than those of the other metabolites (Table 2). 3.3 Biomarkers levels in serum A standard curve of y= 3366.8x + 5826.7 (peak area vs concentration) for pyroglutamate was obtained with a correlation coefficient of 0.9779. Two different levels of the sera (3.57 mmol/L, 9.61 mmol/L) were measured twenty times to evaluate the imprecision. The variable coefficient (CV) were 4.53% and 3.27%, 8
ACCEPTED MANUSCRIPT respectively. A known amount of pyroglutamate (15.0 mmol/L) was added to two human sera at different levels as spike-and-recovery controls. The average recoveries were 97.56% and 102.35%. We selected 25 control, 25 SS and 25 NASH samples to quantitatively determine the
PT
levels of pyroglutamate, TNF-α, IL-8 and adiponectin in serum. The results showed
RI
that the level of pyroglutamate was highest in the NASH, followed by the SS and
SC
control samples.
The median pyroglutamate level in the NAFLD (including NASH and SS) group was
NU
5.19 mmol/L (95% CI 3.89–8.19 mmol/L), which was dramatically higher than that of
MA
the control group, 1.48 mmol/L (95% CI 0.48–3.97 mmol/L) (Figure 4A). ROC analysis of the pyroglutamate levels of NAFLD samples versus controls yielded
D
AUROC of 0.880 (Figure 4B). With an optimal cut-off value of 3.20 mmol/L, the
PT E
sensitivity and specificity for the diagnosis of NAFLD were 73% and 87%, respectively.
CE
The pyroglutamate levels in the NASH group were 7.89 mmol/L (95% CI 5.35–10.41
AC
mmol/L), which was higher than that of the SS group at 3.57 mmol/L (95% CI 0.98– 7.34 mmol/L). With a cut-off value of 4.82 mmol/L, ROC analysis yielded an AUROC of 0.846; the sensitivity and specificity for the differential diagnosis of NASH from SS were 72% and 85%, respectively. The level of adiponectin, IL-8 and TNF-α in patients with SS and NASH are shown in Table 3. Groupwise comparisons showed decreased adiponectin (Se 64%, Sep 56%, AUROC 0.723), decreased IL-8 (Se 65%, Sep 68%, AUROC 0.757) and increased 9
ACCEPTED MANUSCRIPT TNF-α (Se 72%, Sep 76%, AUROC 0.810), which were associated with NASH. The ROC curves of pyroglutamate, adiponectin, IL-8 and TNF-α of NASH versus SS samples are shown in Figure 5. In summary, pyroglutamate performed better than the others as a single diagnostic biomarker of NASH. A combination of 2-3 biomarkers is
PT
usually used in clinical lab to improve diagnosis sensitivity. The diagnostic
RI
performances of pyroglutamate combination with adiponectin, IL-8, TNF-α are
SC
summarized in Table 4. In addition, the combination of pyroglutamate with standard laboratory parameters such as GGT and triglycerides also are displayed in Table 4.
NU
4. Discussion
MA
Distinguishing NASH from SS is a key challenge in the management of NAFLD patients. Serum biomarkers are the most available noninvasive metrics for the
D
diagnosis of NASH [20]. In this study, we studied the serum metabolic profiles of
PT E
NAFLD based on HPLC-MS to identify NASH biomarkers. PCA score plots clearly distinguished the control, SS and NASH groups. As a highly sensitive analytical
CE
method, OPLS-DA, identified the altered metabolites in the NASH group compared
AC
with the SS group. Five metabolites, including uracil, α-Linolenic acid, glutamate, L-glutamine and pyroglutamate, were identified as the most prominent biomarkers. The fold changes of pyroglutamate between SS and control and NASH and SS were 1.56 and 2.26, respectively, which were higher than those of the other biomarkers. Pyroglutamate is also known as 5-oxoproline. The metabolism of pyroglutamate is related to the gamma-glutamyl cycle, which is the primary pathway for glutathione synthesis and degradation [21, 22]. The level of pyroglutamate is often used to screen 10
ACCEPTED MANUSCRIPT for glutathione synthetase deficiency in pediatric patients [23]. Children with deficiencies of either glutathione synthase or 5-oxoprolinase will present with metabolic acidosis or other abnormalities due to the accumulation of pyroglutamic acid [24-26]. When glycine is insufficiently available to support glutathione recovery
PT
via the renal gamma-glutamyl cycle, the rising level of pyroglutamate is regarded as
RI
an index of glycine insufficiency. In addition, the components of gamma-glutamyl
SC
cycle such as glutathione (CompMW: 307.08, retention time: 0.94) and cysteine (CompMW: 283.28, retention time: 7.23) were identified in the LC-MS data.
NU
Although both were capable of distinguishing between SS or NASH group and
MA
controls, glutathione and cysteine did not show significant changing between NASH and SS.
D
Furthermore, pyroglutamate has been proposed as a potential biomarker for enhanced
PT E
oxidative stress in liver cells [27, 28]. Studies have shown enhanced levels of pyroglutamate in various biofluids or tissues following administration of hepatic
CE
toxicants, such as acetaminophen [29] and bromobenzene [30]. A growing number of
AC
reports have documented increased pyroglutamate in the blood, urine, or both in cases of metabolic acidosis complicating paracetamol poisoning [31]. However, there is no study focusing on the role of pyroglutamate in the progress of NAFLD. The metabolic profiling conducted in this study showed that pyroglutamate is a remarkable metabolite that is correlated with NAFLD. Its sensitivity and specificity are equal to those of SteatoTestTM, which is a steatosis score algorithm based on 12 clinical parameters. The AUROC value (0.880) for the diagnosis of NAFLD was more 11
ACCEPTED MANUSCRIPT than 0.80 of that of SteatoTestTM [11]. Pyroglutamate profiling has the capacity to identify NAFLD from healthy controls. Nevertheless, more should be done to discriminate patients with SS and NASH because both are very similar. The performance of pyroglutamate in diagnosing
PT
NASH was carefully compared with that of adiponectin, IL-8 and TNF-α, which are
RI
correlated with NASH. Low adiponectin is a characteristic of steatosis progression
SC
and leads to an increase of intra-cellular fatty acids [12, 32]. In this study, the level of adiponectin in the SS group was higher than that in controls, whereas there was no
NU
significant difference between the NASH and SS groups. IL-8 is an independent
MA
factor that is associated with NASH and was noticeably decreased in NASH compared to SS samples. The levels of TNF-α increased gradually from healthy
D
controls to SS, then to NASH. Although adiponectin, IL-8 and TNF-α showed some
PT E
diagnostic significance, the results showed that their sensitivity and specificity were inferior to those of pyroglutamate.
CE
In addition to using single serum biomarkers related to NASH, some non-invasive
AC
models based on several biomarkers have been used to diagnosis NASH. The fatty liver index (FLI) is a routinely used tool based on four components, including BMI, waist circumference, triglycerides, and GGT [33, 34]. However, the unreliability of the components limited the interpretation of FLI [35]. Munkhzul et al. [33] established an index of NASH (ION) model based on anthropometric, clinical, ultrasound, laboratory, and mortality data, and its sensitivity (81%) was better, but its specificity (82%) was less than the pyroglutamate biomarker. In consideration of the 12
ACCEPTED MANUSCRIPT cost and operating performance, pyroglutamate still has an advantage in diagnosing NASH. The panel of pyroglutamate with other biomarkers contributed to enhance the diagnostic performance. The sensitivity of combination of pyroglutamate, adiponectin and TNF-α reached 92%.
PT
The difference in expression of pyroglutamate between the SS and NASH groups may
RI
be a reflection of their role in disease pathogenesis. It was reported that increased
SC
pyroglutamate excretion is an indication of abnormal response to oxidative stress [36]. NAFLD is accompanied by broad effects on mitochondria, which is the main location
NU
for fatty acid oxidation. When SS progresses to NASH, mitochondria have impaired
MA
respiratory coupling, which may increase the reductive load on respiration and form reactive oxygen species (ROS);ROS damage hepatocytes, trigger inflammation, and
D
elevate pyroglutamate levels [37, 38] .In conclusion, we identified pyroglutamate to
PT E
be a novel biomarker for the diagnosis of NASH. Future studies with larger sample sizes in clinical settings are worth pursuing.
CE
Acknowledgements
AC
This work was supported by the Project of the National Science Foundation of China (61401285
and
81401750),
Natural
Science
Foundation
of
Guangdong
(2016A030313029), Scientific Plan Program of Guangdong (2014A020212038), Innovation Program of Shenzhen (JCYJ20150330102720122), and International Cooperation Foundation of Shenzhen (GJHZ20160301163138685), Shenzhen Science Plan
(JCYJ2016030711492524,
JCYJ20150330102720122,
CXZZ201418182638764) and The Fund of Shenzhen University (2015006). 13
and
ACCEPTED MANUSCRIPT References [1]
V. Ajmera, E.R. Perito, N.M. Bass, N.A. Terrault, K.P. Yates, R. Gill, et al.,Novel plasma biomarkers associated with liver disease severity in adults with nonalcoholic fatty liver disease, Hepatology. 65(2017) 65-77. E. Bugianesi, C. Rosso, and H. Cortez-Pinto,How to diagnose NAFLD in
PT
[2]
M.E. Rinella and A.J. Sanyal,Management of NAFLD: a stage-based approach,
SC
[3]
RI
2016, J Hepatol. 65(2016) 643-4.
Nat Rev Gastroenterol Hepatol. 13(2016) 196-205.
C. Stern and L. Castera,Non-invasive diagnosis of hepatic steatosis, Hepatol
NU
[4]
[5]
MA
Int. 2016)
S. Kinner, S.B. Reeder, and T. Yokoo,Quantitative Imaging Biomarkers of
J. Pirhonen, J. Arola, S. Sadevirta, P. Luukkonen, S.M. Karppinen, T.
PT E
[6]
D
NAFLD, Dig Dis Sci. 61(2016) 1337-47.
Pihlajaniemi, et al.,Continuous Grading of Early Fibrosis in NAFLD Using
CE
Label-Free Imaging: A Proof-of-Concept Study, PLoS One. 11(2016)
[7]
AC
e0147804.
R.P. Myers, A. Pollett, R. Kirsch, G. Pomier-Layrargues, M. Beaton, M. Levstik, et al.,Controlled Attenuation Parameter (CAP): a noninvasive method for the detection of hepatic steatosis based on transient elastography, Liver Int. 32(2012) 902-10.
[8]
E. Hashimoto, M. Taniai, and K. Tokushige,Characteristics and diagnosis of NAFLD/NASH, J Gastroenterol Hepatol. 28 Suppl 4(2013) 64-70. 14
ACCEPTED MANUSCRIPT [9]
J. Vulevic, A. Juric, G.E. Walton, S.P. Claus, G. Tzortzis, R.E. Toward, et al.,Influence of galacto-oligosaccharide mixture (B-GOS) on gut microbiota, immune parameters and metabonomics in elderly persons, Br J Nutr. 114(2015) 586-95. T. Poynard, V. Ratziu, S. Naveau, D. Thabut, F. Charlotte, D. Messous, et
PT
[10]
RI
al.,The diagnostic value of biomarkers (SteatoTest) for the prediction of liver
[11]
SC
steatosis, Comp Hepatol. 4(2005) 10.
T. Poynard, G. Lassailly, E. Diaz, K. Clement, R. Caiazzo, J. Tordjman, et
NU
al.,Performance of biomarkers FibroTest, ActiTest, SteatoTest, and NashTest
MA
in patients with severe obesity: meta analysis of individual patient data, PLoS One. 7(2012) e30325.
L. Abenavoli and V. Peta,Role of adipokines and cytokines in non-alcoholic
D
[12]
[13]
PT E
fatty liver disease, Rev Recent Clin Trials. 9(2014) 134-40. M.H. Jarrar, A. Baranova, R. Collantes, B. Ranard, M. Stepanova, C. Bennett,
CE
et al.,Adipokines and cytokines in non-alcoholic fatty liver disease, Aliment
[14]
AC
Pharmacol Ther. 27(2008) 412-21. S. Abiru, K. Migita, Y. Maeda, M. Daikoku, M. Ito, K. Ohata, et al.,Serum cytokine and soluble cytokine receptor levels in patients with non-alcoholic steatohepatitis, Liver Int. 26(2006) 39-45. [15]
V.R. Mas, R. Fassnacht, K.J. Archer, and D. Maluf,Molecular mechanisms involved in the interaction effects of alcohol and hepatitis C virus in liver cirrhosis, Mol Med. 16(2010) 287-97. 15
ACCEPTED MANUSCRIPT [16]
B. Worley and R. Powers,PCA as a practical indicator of OPLS-DA model reliability, Curr Metabolomics. 4(2016) 97-103.
[17]
S. Wagner, K. Scholz, M. Donegan, L. Burton, J. Wingate, and W. Volkel,Metabonomics and biomarker discovery: LC-MS metabolic profiling
PT
and constant neutral loss scanning combined with multivariate data analysis
Y. Wang, H. Man, J. Gao, X. Liu, X. Ren, J. Chen, et al.,Plasma metabonomics
SC
[18]
RI
for mercapturic acid analysis, Anal Chem. 78(2006) 1296-305.
study on toxicity biomarker in rats treated with Euphorbia fischeriana based
S. Qi, X. Ouyang, L. Wang, W. Peng, J. Wen, and Y. Dai,A pilot metabolic
MA
[19]
NU
on LC-MS, Biomed Chromatogr. 30(2016) 1386-96.
profiling study in serum of patients with chronic kidney disease based on (1)
H. Yki-Jarvinen,Diagnosis of non-alcoholic fatty liver disease (NAFLD),
PT E
[20]
D
H-NMR-spectroscopy, Clin Transl Sci. 5(2012) 379-85.
Diabetologia. 59(2016) 1104-11. D.B. Liss, M.S. Paden, E.S. Schwarz, and M.E. Mullins,What is the clinical
CE
[21]
AC
significance of 5-oxoproline (pyroglutamic acid) in high anion gap metabolic acidosis following paracetamol (acetaminophen) exposure?, Clin Toxicol (Phila). 51(2013) 817-27. [22]
Y.M. Yu, C.M. Ryan, Z.W. Fei, X.M. Lu, L. Castillo, J.T. Schultz, et al.,Plasma L-5-oxoproline kinetics and whole blood glutathione synthesis rates in severely burned adult humans, Am J Physiol Endocrinol Metab. 282(2002) E247-58. 16
ACCEPTED MANUSCRIPT [23]
C.L. Foot, J.F. Fraser, and D.V. Mullany,Pyroglutamic acidosis in a renal transplant patient, Nephrol Dial Transplant. 20(2005) 2836-8.
[24]
I.A. Almaghlouth, J.Y. Mohamed, M. Al-Amoudi, L. Al-Ahaidib, A. Al-Odaib, and F.S. Alkuraya,5-Oxoprolinase deficiency: report of the first human
L. Zand, A. Muriithi, E. Nelsen, P.M. Franco, E.L. Greene, Q. Qian, et
RI
[25]
PT
OPLAH mutation, Clin Genet. 82(2012) 193-6.
SC
al.,Severe anion gap metabolic acidosis from acetaminophen use secondary to 5-oxoproline (pyroglutamic acid) accumulation, Am J Med Sci. 344(2012)
P. Tailor, T. Raman, C.L. Garganta, R. Njalsson, K. Carlsson, E. Ristoff, et
MA
[26]
NU
501-4.
al.,Recurrent high anion gap metabolic acidosis secondary to 5-oxoproline
E.J. Rolleman, E.J. Hoorn, P. Didden, and R. Zietse,Guilty as charged:
PT E
[27]
D
(pyroglutamic acid), Am J Kidney Dis. 46(2005) e4-10.
unmeasured urinary anions in a case of pyroglutamic acidosis, Neth J Med.
T. Soga, R. Baran, M. Suematsu, Y. Ueno, S. Ikeda, T. Sakurakawa, et
AC
[28]
CE
66(2008) 351-3.
al.,Differential metabolomics reveals ophthalmic acid as an oxidative stress biomarker indicating hepatic glutathione consumption, J Biol Chem. 281(2006) 16768-76. [29]
D.T.
Lawrence,
L.K.
Bechtel,
N.P.
Charlton,
and
C.P.
Holstege,5-oxoproline-induced anion gap metabolic acidosis after an acute acetaminophen overdose, J Am Osteopath Assoc. 110(2010) 545-51. 17
ACCEPTED MANUSCRIPT [30]
N.J. Waters, C.J. Waterfield, R.D. Farrant, E. Holmes, and J.K. Nicholson,Integrated
metabonomic
analysis
of
bromobenzene-induced
hepatotoxicity: novel induction of 5-oxoprolinosis, J Proteome Res. 5(2006) 1448-59. S. Geenen, C. Guallar-Hoyas, F. Michopoulos, J.G. Kenna, K.L. Kolaja, H.V.
PT
[31]
RI
Westerhoff, et al.,HPLC-MS/MS methods for the quantitative analysis of
SC
5-oxoproline (pyroglutamate) in rat plasma and hepatic cell line culture medium, J Pharm Biomed Anal. 56(2011) 655-63.
R. Kumar, S. Prakash, S. Chhabra, V. Singla, K. Madan, S.D. Gupta, et
NU
[32]
MA
al.,Association of pro-inflammatory cytokines, adipokines & oxidative stress with insulin resistance & non-alcoholic fatty liver disease, Indian J Med Res.
M. Otgonsuren, M.J. Estep, N. Hossain, E. Younossi, S. Frost, L. Henry, et
PT E
[33]
D
136(2012) 229-36.
al.,Single non-invasive model to diagnose non-alcoholic fatty liver disease
CE
(NAFLD) and non-alcoholic steatohepatitis (NASH), J Gastroenterol Hepatol.
[34]
AC
29(2014) 2006-13. G. Calori, G. Lattuada, F. Ragogna, M.P. Garancini, P. Crosignani, M. Villa, et al.,Fatty liver index and mortality: the Cremona study in the 15th year of follow-up, Hepatology. 54(2011) 145-52. [35]
M.A. Borman, F. Ladak, P. Crotty, A. Pollett, R. Kirsch, G. Pomier-Layrargues, et al.,The Fatty Liver Index has limited utility for the detection and quantification of hepatic steatosis in obese patients, Hepatol Int. 7(2013) 18
ACCEPTED MANUSCRIPT 592-9. [36]
G. Brooker, J. Jeffery, T. Nataraj, M. Sair, and R. Ayling,High anion gap metabolic acidosis secondary to pyroglutamic aciduria (5-oxoprolinuria): association with prescription drugs and malnutrition, Ann Clin Biochem.
W.B. Ashworth, N.A. Davies, and I.D. Bogle,A Computational Model of
RI
[37]
PT
44(2007) 406-9.
SC
Hepatic Energy Metabolism: Understanding Zonated Damage and Steatosis in NAFLD, PLoS Comput Biol. 12(2016) e1005105.
NU
M. Ono, N. Okamoto, and T. Saibara,The latest idea in NAFLD/NASH
CE
PT E
D
MA
pathogenesis, Clin J Gastroenterol. 3(2010) 263-70.
AC
[38]
19
ACCEPTED MANUSCRIPT Table 1. Clinical and biochemical characteristics of NAFLD and controls Controls (n=31) AGE (yrs)
(n=38)
52.8±14.6 25:7
31:9
(n=21) 56.3±12.9 24:8
Body mass index
47.3±6.8
47.5±8.1
Alanine aminotransferase (U/L)
35.6±7.2
65.7±20.8*
Aspartate aminotransferase (U/L)
21.3±6.7
38.7±12.8
50.7±23.5*
gamma -glutamyl transferase (U/L)
36.8±8.4
78.8±14.3
152±15.9*†
Triglyceride (mg/dL)
136±66.4
150±84.1
198±95.2*†
180±30.7
197±31.8
192±32.5
104±20.6
110±25.9
134±40.5*
D
Glucose (mg/dL)
41.4±23.5
SC
NU
MA
Cholesterol (mg/dL)
PT E
*p < 0.05, compared to controls. †
58.7±14.5
NASH
RI
Gender (M:F)
SS
PT
Parameter
p < 0.05, compared to SS.
AC
CE
Data refer to the number of patients or the mean ± SD.
20
47.5±8.4
ACCEPTED MANUSCRIPT Table 2. Fold changes of identified metabolites between the SS versus control groups or NASH versus SS groups based on the OPLS-DA. SS vs Control
NASH vs SS
Uracil
1.34
1.63
α-Linolenic Acid
1.33
1.52
Glutamate
1.56
1.80
L-Glutamine
1.56
Pyroglutamate
1.56
RI
PT
Metabolite
AC
CE
PT E
D
MA
NU
SC
1.81
21
2.26
ACCEPTED MANUSCRIPT Table 3. Groupwise comparisons of serum biomarkers for healthy controls and patients with SS and NASH Control (n = 25)
SS (n = 25)
NASH (n = 25)
(mmol/L)
1.48±1.21
3.18±1.53*
7.85±3.22†
Adiponectin (ug/mL)
13.25±8.36
11.69±6.25
7.08±4.40†
TNF-α (pg/mL)
3.89±3.21
4.26±2.96
IL-8 (pg/ml)
25.15±13.87
26.02±10.69
†
p < 0.05, compared to SS.
RI
SC
NU MA
*p < 0.05, compared to controls.
PT
Pyroglutamate
AC
CE
PT E
D
Data refer to the number of patients or the mean ± SD.
22
8.99±4.63† 17.73±8.70†
ACCEPTED MANUSCRIPT Table 4. The diagnostic performance of biomarkers for NASH
Biomarker
Sensitivity
Specificity
Positive
Negative
predictive value
predictive value
Accuracy
64%
56%
59%
60%
60%
TNF-α
72%
76%
75%
73%
74%
IL-8
65%
68%
86%
triglyceride
54%
75%
68%
GGT
57%
80%
pyroglutamate
72%
85%
pyroglutamate +adiponectin
82%
80%
pyroglutamate+TNF-α
83%
77%
pyroglutamate+IL-8
85%
82%
66%
62%
65%
74%
65%
69%
62%
90%
82%
58%
93%
80%
70%
86%
79%
91%
69%
82%
86%
81%
78%
88%
85%
86%
93%
89%
91%
87%
83%
84%
86%
pyroglutamate+ triglyceride
83%
86%
85%
83%
85%
pyroglutamate +GGT
85%
80%
82%
88%
85%
91%
76%
79%
90%
84%
+adiponectin+IL-8
90%
AC
pyroglutamate+
CE
+TNF-α
92%
PT E
+adiponectin+TNF-α pyroglutamate+IL-8
triglyceride+GGT
NU
D
pyroglutamate
MA
pyroglutamate
RI
40%
SC
PT
adiponectin
23
ACCEPTED MANUSCRIPT Figure legends: Figure 1. Mass chromatograms of serum obtained from control (A), simple steatosis (B) nonalcoholic steatohepatitis (C).
PT
Figure 2. (A) The PCA score plots of healthy controls (blue square) and SS patients
RI
(red triangle). (B) The PCA score plots of the SS (red triangle) and NASH
SC
(inverse yellow triangle) groups.
NU
Figure 3. (A) The OPLS-DA score plots of healthy controls (blue square) and SS
D
(inverse yellow triangle) groups.
MA
patients (red triangle). (B) The OPLS-DA score plots of SS (red triangle) and NASH
PT E
Figure 4. (A) The serum pyroglutamate levels in controls and patients with SS and NASH. The short transverse line represents the median of diff erent groups. (B) The
CE
ROC analysis of pyroglutamate in serum of NAFLD versus control with AUROC of
AC
0.880 and SS versus controls with AUROC of 0.846, respectively.
Figure 5. The ROC analysis of pyroglutamate, adiponectin, TNF-α and IL-8 in serum of NASH versus SS.
24
ACCEPTED MANUSCRIPT
AC
CE
PT E
D
MA
NU
SC
RI
PT
Fig. 1
25
ACCEPTED MANUSCRIPT
AC
CE
PT E
D
MA
NU
SC
RI
PT
Fig. 2
26
ACCEPTED MANUSCRIPT
AC
CE
PT E
D
MA
NU
SC
RI
PT
Fig. 3
27
ACCEPTED MANUSCRIPT
AC
CE
PT E
D
MA
NU
SC
RI
PT
Fig. 4
28
ACCEPTED MANUSCRIPT
AC
CE
PT E
D
MA
NU
SC
RI
PT
Fig. 5
29
ACCEPTED MANUSCRIPT Highlights 1. High performance liquid chromatography-mass spectrometry (HPLC-MS) was
used to analyse the metabolic profile of the serum samples from patients with nonalcoholic steatohepatitis (NASH) from simple steatosis (SS), and controls.
PT
2. Five metabolites, including uracil, α-linolenic acid, glutamate, L-glutamine and
RI
pyroglutamate, were identified as prominent factors in discriminating NASH from
SC
SS.
3. Pyroglutamate was found to be the most promising factor in distinguishing the
NU
NASH from SS groups with an optimal cut-off value of 4.82 mmol/L, the
MA
sensitivity and specificity of the diagnosis of NASH were 72% and 85%, respectively.
D
4. The potential diagnostic power of pyroglutamate was superior to other biomarkers
AC
CE
PT E
such as factor-α (TNF-α), adiponectin and interleukin-8 (IL-8).
30