Metabolomics and cytokine analysis for identification of severe drug

Subscriber access provided by UNIV AUTONOMA DE COAHUILA UADEC

Article

Metabolomics and cytokine analysis for identification of severe drug-induced liver injury Zhongyang Xie, Ermei Chen, Xiaoxi Ouyang, Xiaowei Xu, Shanshan Ma, Feiyang Ji, Daxian Wu, Sainan Zhang, Yalei Zhao, and Lanjuan Li J. Proteome Res., Just Accepted Manuscript • DOI: 10.1021/acs.jproteome.9b00047 • Publication Date (Web): 19 Apr 2019 Downloaded from http://pubs.acs.org on April 20, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 32 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

Metabolomics

and

cytokine

analysis

for

identification

of

severe

drug-induced liver injury

Zhongyang Xie1,2, † , Ermei Chen1,2, † , Xiaoxi Ouyang1,2, Xiaowei Xu2,3, Shanshan Ma1,2, Feiyang Ji1,2, Daxian Wu1,2, Sainan Zhang1,2, Yalei Zhao1,2, Lanjuan Li1,2,*

1State

Key Laboratory for Diagnosis and Treatment of Infectious Diseases, The First Affiliated Hospital, School

of Medicine, Zhejiang University, Hangzhou, China;

2

Collaborative Innovation Center for Diagnosis and

Treatment of Infectious Diseases, Zhejiang University, Hangzhou, China; 3 Department of Infectious Disease, The First Affiliated Hospital, College of Medicine, Zhejiang University, Hangzhou, China;

†These

authors contributed equally to this work

*Correspondence:

Lanjuan Li, MD, Chief of Key Laboratory of Infectious Diseases. Address: State Key

Laboratory for Diagnosis and Treatment of Infectious Diseases, The First Affiliated Hospital, School of Medicine, Zhejiang University, Hangzhou 310003, China. Tel./fax: +86 0571-87236459; E-mail: [email protected].

1

ACS Paragon Plus Environment

Journal of Proteome Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 32

1. Abstract AIM: To evaluate the levels of metabolites and cytokines in the serum of patients with severe and non-severe idiosyncratic drug-induced liver injury (DILI), and to identify biomarkers of DILI severity. METHODS:

Gas

chromatography-mass

spectrometry

(GC-MS)

and

ultra-performance

liquid

chromatography–mass spectrometry (UPLC-MS)-based metabolomic approaches were used to evaluate the metabolome of serum samples from 29 DILI patients of severity grade 3 (non-severe), 27 of severity grade 4 (severe), and 36 healthy controls (HCs). The levels of total Keratin-18 (K18), fragment K18 and 27 cytokines were determined by enzyme-linked immunosorbent assay. RESULTS: The alkaline phosphatase activity (p = 0.021) and international normalized ratio (INR) (p < 0.001) differed significantly between the severe and non-severe groups. The severe group had a higher serum fragment K18 level than the non-severe group. A multivariate analysis showed good separation between all pairs of the HC, non-severe, and severe groups. According to the orthogonal partial least squares–discriminant analysis (OPLS-DA) model, 14 metabolites were selected by GC-MS and 17 by UPLC-MS. Among these metabolites, the levels of 16 were increased and of 15 were decreased in the severe group. A pathway analysis revealed major changes in the primary bile acid biosynthesis and alpha-linolenic acid metabolic pathways. The levels of PDGF-bb, IP-10, IL-1Rα, MIP-1β, and TNF-α differed significantly between the severe and non-severe groups, and the levels of most of the metabolites were negatively correlated with those of these cytokines. An OPLS-DA model that included the detected metabolites and cytokines revealed clear separation of the severe and non-severe groups. CONCLUSION: We identified 31 metabolites and 5 cytokines related to the severity of idiosyncratic DILI. The primary bile-acid biosynthesis and alpha-linolenic acid metabolism pathways were also related to the severity of DILI. A model that incorporated the metabolites and cytokines showed clear separation between patients with severe and non-severe DILI, suggesting that these biomarkers have potential as indicators of DILI severity. Key words: Drug-induced liver injury; metabolomics; GC-MS; LC-MS; severity; keratin-18; cytokines; multivariate analysis; OPLS-DA; biomarkers


Page 3 of 32 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


2. Introduction Drug-induced liver injury (DILI) is a rare but clinically important entity; the mortality rate of DILI in the United States is approximately 10% 1. In China, the incidence of DILI has increased recently, but reliable epidemiologic data are lacking 2. A number of large multicenter studies of DILI have been conducted in the last 10 years, including the US DILI Network (DILIN) study1, 3-5. Nevertheless, the mechanisms by which drugs cause liver injury are unclear, as is the pathway by which DILI progresses to acute liver failure (ALF). On the meanwhile, the lack of specific biomarkers which is critical for the detection and prediction of DILI still remains a problem for specialists. Proteomics and metabolomics make considerable contributions to biomedical research 6-8. These approaches enable analysis of disease-related changes in the proteome and metabolome and facilitate identification of therapeutic targets and biomarkers. Metabolomics studies typically make use of gas chromatography–mass

spectrometry

(GC-MS),

liquid

chromatography–mass

spectrometry

(UPLC-MS), and nuclear magnetic resonance; of these, GC-MS and UPLC-MS are the most effective 9.

Several prior clinical

10-11,

animal

12-13,

and cell-culture

14

studies of DILI have involved omics

approaches. These studies have identified differentially abundant proteins and metabolites that are components of pathways implicated in the pathogenesis of DILI and have suggested new approaches to evaluate and predict drug-induced hepatotoxicity. However, the metabolic changes that occur during the progression of DILI and biomarkers of its severity have not been evaluated. When DILI occurs, the injured hepatocytes release damage-associated molecular patterns

(DAMPs), which activate local immune cells—kupffer cells, to release pro-inflammatory cytokines15. These released cytokines triggers the recruitment of peripheral neutrophils and monocytes to broaden the immune responses. Immune responses are reported to play vital roles in the pathogenesis of drug-induced liver injury16-17, immune dysfunction might determine the outcomes of acute liver failure patients18. Total Keratin-18(K18) and fragment K18 are potential biomarkers which reflect necrosis and apoptosis levels in vivo, and they could predict prognosis and severity of acetaminophen-induced liver injury19-20. Whereas, the relationships between these biomarkers and metabolic changes during DILI development remain mysterious and need further study. In this study, we used GC-MS and UPLC-MS to evaluate the metabolome of DILI patients with different disease severities. Furthermore, we assessed the serum levels of K18 and cytokines in 3



Page 4 of 32

patients with severe and non-severe DILI and explored the correlations among differentially abundant metabolites and cytokines. The aim was to evaluate DILI-related changes in metabolic and immune pathways and to identify biomarkers of DILI severity. To differentiate patients with severe DILI from those with moderate or mild disease, we created a multivariate model incorporating the identified metabolites and cytokines.

3. Methods 2.1 Patients A single-centered study was performed from August 1, 2015, to June 30, 2017, while the inpatients diagnosed with idiosyncratic DILI were enrolled. Hepatitis A, B, C, D, and E IgM, as well as cytomegalovirus (CMV), Epstein-Barr virus (EBV), herpes simplex virus (HSV) IgM, were tested for exclusion of viral hepatitis. Antinuclear antibody and Anti-neutrophil cytoplasmic antibodies were examined for autoimmune hepatitis. Blood samples were obtained on the day following hospitalization and were centrifuged 3500 rpm for 10 min to separate the serum, which was stored at −80°C. The clinical data of the patients were obtained from the electronic medical records system. This study was conducted in compliance with the ethical principles of the Declaration of Helsinki and was approved by the Ethics Committee of the hospital. The enrolled patients or their legal representatives provided written informed consent. 2.2 Inclusion and exclusion criteria The inclusion criteria were as follows: According to the DILIN study

21

, patients with a medication history and a hepatic biochemical

abnormality who met one or more of the following criteria: (i) Aspartate aminotransferase (AST) or alanine aminotransferase (ALT) level of >5× the upper limit of normal (ULN) or an alkaline phosphatase (ALP) level of >2 × the ULN; (ii) Serum total bilirubin (TB) level of >2.5 mg/dL and elevated AST, ALT, or ALP; (iii) International normalized ratio (INR) >1.5 with elevated AST, ALT, or ALP. The exclusion criteria were as follows: (1) Liver injury caused by acetaminophen; (2) Liver injury caused by hepatitis viruses, autoimmune liver diseases, metabolic liver diseases, or 4


Page 5 of 32 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


liver cancer; (3) Liver or bone-marrow transplantation prior to enrolment. Relevant clinical, biochemical, serological, and imaging data were obtained from the medical records. The R value was defined as the ratio of the serum ALT level (as a multiple of its ULN) to that of ALP (as a multiple of its ULN). Hepatocellular DILI was defined as R ≥5, cholestatic DILI as R ≤2, and mixed DILI as R >2 and 8), probable (6–8), possible (3–5), unlikely (1 or 2), or excluded (0). The severity of patients was defined as in the DILIN study 26 (see supplementary table 1) . 2.3 Acquisition and processing of metabolic data GC-MS sample preparation The sample preparation procedures for GC-MS were as described previously

27.

Aliquots of serum

(100 µL) were thawed to room temperature; next, heptadecanoic acid (20 µL; 1 mg/mL in methanol) and 300 µL of acetonitrile were added, and the tubes were vortex-mixed. After ultrasonic extraction and centrifugation for 10 min at 10,000 rpm and 4°C, the supernatant was dried under vacuum. Next, 50 µL of 15 mg/mL methoxyamine hydrochloride in pyridine were added to the dried samples, which were vortex-mixed and incubated at room temperature for 16 h. Subsequently, 90 µL of N-methyl-N(trimethylsily) trifluoroacetamide (MSTFA) containing 1% trimethylchlorosilane (TMCS) were added, and the samples were vortex-mixed and incubated for 2 h. After centrifugation, the supernatants were transferred to the 150-µL inner tubes of 2-mL glass autosampler vials. GC-MS data acquisition After derivatization, 1 µL of solution was injected in splitless mode into an Agilent 7890A GC system equipped with a fused silica capillary column (30 m × 0.25 mm internal diameter) chemically bonded to a 0.25-m HP-5MS stationary phase (Phenomenex). The injector temperature was set at 250°C. Helium was used as the carrier gas at a flow rate of 1 mL/min. The column temperature was maintained at 70°C for 2 min, then increased from 280°C at 15°C/min and maintained at 280°C for 2 min. The effluent was transferred into the ion source of an Agilent 5975C mass selective detector (Agilent Technologies), and masses from m/z 60 to 800 were acquired. 5



Page 6 of 32

UPLC-MS sample preparation Serum (50 µL) was mixed with 150 µL of acetonitrile, vortex-mixed, and centrifuged for 10 min at 10,000 rpm and 4°C. The supernatant was transferred to the 150-µL inner tube of a 2-mL glass autosampler vial. Aliquots (10 µL) of processed serum were pooled for quality control (QC). The QC sample was run before the sample sequence and after every eight samples. UPLC-MS data acquisition The UPLC-MS parameters were set as described previously. Briefly, a 2-µL aliquot was chromatographed on a Waters (Milford, MA, USA) Acquity UPLC system equipped with an Acquity UPLC BEH C18 analytical column (2.1 × 100 mm, 1.7 µm, 130 Å). The mobile phases were water (A) and acetonitrile (B) with 0.1% formic acid. The flow rate was set at 0.25 mL/min. The initial composition was 3% B; this was followed by a linear increase to 98% B, which was maintained for 5 min, and then increased to 100% B over 3 min. Mass spectrometry (MS) was performed on a Waters MS equipped with an electrospray ionization (ESI) source in positive-ion mode. The electrospray voltage was 3 kV, and the capillary temperature was set at 450°C. MS data were acquired in full scan mode from 50 to 1,000. Nitrogen was used as both the desolvation gas (flow rate, 600 L/h) and the cone gas (50 L/h). 2.4 Analysis of serum cytokines and total/fragment Keratin-18 (K18) levels The levels of cytokines were determined by human 27-plex assay (Bio-Plex Suspension Array System, Bio-Rad, Hercules, CA, USA). The assay contains 8 chemokines (Eotaxin, IL-7, IL-8, IP-10, MCP-1, MIP-1α, MIP-1β, RANTES), 5 growth factors (FGF basic, GCSF, GM-CSF, PDGF-BB, VEGF), 12 interleukin (IL-1β, IL-1ra, IL-2, IL-4, IL-5, IL-6, IL-9, IL-10, IL-12, IL-13, IL-15, IL-17), IFN-γ and TNF-α. These cytokines are secreted by monocytes/macrophages, neutrophils or T cells, which are almost related to all immune cells evolved in DILI progression, so we choose these 27 cytokines for further study. In brief, samples were diluted 1:1 (v:v) in sample diluent and incubated for 30 min with capture antibody-coupled magnetic beads. After washing three times, the samples were incubated for 30 min in the dark with the biotinylated detection antibody. The cytokines were detected by adding streptavidin–phycoerythrin and quantified using a BioPlex array reader. The levels of total and fragment K18 were determined by M65 (EpiDeath) and M30 (Apoptosense) enzyme-linked immunosorbent assays according to the manufacturer’s guidelines 6


Page 7 of 32 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(Peviva, Bromma, Sweden). 2.5 Data processing and metabolite identification The raw GC-MS and UPLC-MS data were processed by ChromaTOF software (version 4.34, LECO, St. Joseph, MI, USA) and MassLynx 4.1 software (Waters Co.) for peak detection, filtering, denoising, alignment, and normalization. The retention times (RTs), m/z values, and corresponding peak intensities were tabulated and imported into SIMCA-P + 12.0 (Umetrics) for multivariate analyses. Principal component analysis (PCA) and orthogonal partial least squares–discriminant analysis (OPLS-DA) were performed to investigate differences in metabolite profile among the groups. The variable of importance in the project (VIP) values and the P-values determined by two-tailed Student’s t-tests were used to identify potential biomarkers. The biomarkers detected by GC-MS were identified by comparison with the National Institute of Standards and Technology and Fiehn databases. To identify the biomarkers detected by UPLC-MS, we searched the Human Metabolome Database (http://hmdb.ca/) using the exact masses and typical MS/MS fragments. 2.6 Statistical analysis Statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS, v. 22.0; SPSS, Inc., Chicago, IL, USA). All tests were two-tailed, and P-values 3; the majority were 6–8 (probable). 7



Page 8 of 32

Table 1. Demographic, clinical, and laboratory parameters of the subjects. Healthy Control

Non-severe group

Severe Group

(n = 34)

(n = 29)

(n = 27)

51 ± 15

57 ± 14

51 ± 15

0.107

18 (52.94%)

Age Female %

P-value

13 (44.83%)

18 (66.67%)

0.104

Alcohol use

6 (20.69%)

3 (11.11%)

0.16

Hypertension

9 (31.03%)

5 (18.52%)

0.284 0.499

Latency 90 days

-

1 (3.45%)

4 (14.81%)

ALT (U/L)

18.2 ± 9.7

422.1 ± 358.5

695.7 ± 973.4

0.178

AST (U/L)

20.3 ± 6.5

327.1 ± 256.4

593.5 ± 1024.7

0.199

71.8 ± 20.3

239.2 ± 242.5

126.9 ± 35.8

0.021

12.5 ± 5.3

332.8 ± 115.4

332.8 ± 123.1

0.999

-

1.2 ± 0.2

1.9 ± 0.6

8）

Death by 90th day

0.073

Data are mean ± SD, median (p25, p75), or number (percentage). P-values for comparisons between the severe and non-severe groups. Comparisons were performed by uncoupled t-test, non-parametric Mann–Whitney U-test, or Fisher’s exact test. Abbreviations: DILI, drug-induced liver injury; SD, standard deviation; ALT, alanine aminotransferase; AST, aspartate aminotransferase; INR, international normalized ratio; K18, keratin 18; RUCAM, Roussel Uclaf causality assessment method

The serum concentrations of fragment K18 (as determined by assaying M30) and total K18 (M65) are shown in Figure 1. The faragment K18 and total K18 concentrations were significantly higher in the patient groups than in the HC group. Compared to the non-severe group, the fragment K18 concentration was higher in the severe group (p < 0.05), but the total K18 concentration and the 8


Page 9 of 32 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


fragment K18/ total K18 ratio did not differ significantly between the severe and non-severe groups.

The fragment K18 and total K18 concentrations were strongly positively correlated with the ALT and AST levels (Figure S-1). Besides, INR and bile acid also showed positive correlation with K18.

Figure 1. Serum levels of (A) fragment K18 (M30) and (B) total K18 (M65). (C) fragment K18/ total K18 (M30/M65) ratio. * p < 0.05; ** p < 0.01.

3.2 Multivariate analysis of the GC-MS and UPLC-MS data Three representative GC-MS TIC chromatograms of consecutive injections of the same sample are shown in Figure S-2A. The PCA, PLS-DA, and OPLS-DA showed good separation among the three groups (Figure 2A-D); metabolites were derived from the OPLS-DA model according to a VIP value of >1 and a significant difference (p < 0.05) between any two groups. Of the 73 metabolites differentiating between the HC and non-severe groups, 79 between the HC and severe groups, and 72 between the severe and non-severe groups, 14 overlapping metabolites were selected (Figure 2E and Table 2). Among them, the serum levels of iminodiacetic acid and glucuronic acid were higher in the severe group than the non-severe group, and the reverse was true for the remaining 12 metabolites. Clustering analysis revealed that the 14 metabolites enabled discrimination of patients with severe from those with non-severe DILI (Figure 2F).

9



Page 10 of 32

Figure 2. Metabolomic profiling by GC-MS. (A) PCA-X and (B) PLS-DA of the HC, non-severe and severe groups. (C) The 200-permutation test demonstrated no overfitting in the PLS-DA model (Q2 = (0.0, −0.201)). (D) OPLS-DA models among three groups, between the HC and non-severe groups (R2X (cum) = 0.222, R2Y (cum) = 0.988, Q2 (cum) = 0.97), HC and severe groups (R2X (cum) = 0.269, R2Y (cum) = 0.984, Q2 (cum) = 0.964), and non-severe and severe groups (R2X (cum) = 0.218, R2Y (cum) = 0.987, Q2 (cum) = 0.906). (E) Venn diagram of a selection of differentially abundant metabolites by GC-MS. (F) Clustering analysis map for the selected metabolites. Table 2. Serum levels of endogenous metabolites in the severe and non-severe groups by GC-MS. Differentially expression

Peak No.

RT

M/Z

Metabolites

1

13.23

217

Gluconic lactone

↓

0.0084

2

24.15

298

D-(glycerol 1-phosphate)

↓

0.006

3

16.70

441

uric acid

↓

0.0074

4

13.39

221

terephthalic acid

↓

0.0023

5

12.78

281

3,4-Dihydroxyphenylglycol

↓

0.0406

6

22.37

263

xanthosine

↓

Metabolomics and cytokine analysis for identification of severe drug

Recommend Documents