Association Between Circulating Bile Acid Alterations and Nonalcoholic Steatohepatitis Independent of Obesity and Diabetes Mellitus

Youngae Jung; Bo Kyung Koo; Seo Young Jang; Dain Kim; Heeyeon Lee; Dong Hyeon Lee; Sae Kyung Joo; Yong Jin Jung; Jeong Hwan Park; Taekyeong Yoo; Murim Choi; Min Kyung Lee; Sang Won Kang; Mee Soo Chang; Won Kim; Geum-Sook Hwang


Liver International. 2021;41(12):2892-2902. 

In This Article


Study Subjects

We analysed clinical and histological data from the 'Boramae NAFLD cohort' (NCT 02206841)[17,18] of Seoul Metropolitan Government Seoul National University Boramae Medical Center. The eligibility criteria of this study were as follows: (i) ≥18 years old, (ii) bright liver echogenicity on ultrasound scan,[19] and (iii) unexplained high alanine aminotransferase (ALT) levels above the upper reference value within the past 6 months.[20] The following conditions were used as the exclusion criteria for this study: (i) viral hepatitis (hepatitis B or C), (ii) autoimmune hepatitis, primary biliary cholangitis or primary sclerosing cholangitis, (iii) drug-induced liver injury or steatosis, (iv) Wilson disease or hemochromatosis, (v) excessive alcohol consumption (male >30 g/d, female >20 g/d),[21] and (vi) diagnosis of malignancy within the past year. Among the eligible subjects, those with at least two of the following risk factors underwent liver biopsy:[22] diabetes mellitus, central obesity (waist circumference ≥90 cm for males or ≥80 cm for females), high triglyceride levels (≥150 mg/dL), low high-density lipoprotein (HDL)-cholesterol levels (<40 mg/dL for males or <50 mg/dL for females), insulin resistance, hypertension and clinically suspected nonalcoholic steatohepatitis (NASH) or fibrosis. The entire study cohort was divided into the discovery (n = 241) and validation (n = 136) cohorts according to the enrolment period. The discovery cohort consisted of individuals enrolled from Jan 2013 to Jul 2016; those enrolled thereafter formed the validation cohort, which included more recently enrolled subjects (between Jul 2016 and May 2019) at the same institution.[23,24] BMI was classified based on the World Health Organization (WHO) Asia–Pacific criteria as follows: BMI ≥25 kg/m2 indicated obesity.[25] This study was conducted in accordance with the provisions of the Declaration of Helsinki for the participation of human subjects in research and was approved by the Institutional Review Board of Boramae Medical Center (IRB No. 16-2013-45). Written informed consent was obtained from each subject in the study cohort.

Clinical Parameters and Assessments

The methods used for measuring visceral adipose tissue (VAT) area have been described elsewhere.[26] Systemic insulin resistance (HOMA-IR) was evaluated using the homeostasis model assessment (HOMA).[27] This study was conducted in accordance with the ethical guidelines of the 1975 Declaration of Helsinki for the participation of human subjects and was approved by the Institutional Review Board of Boramae Medical Center. Written informed consent was obtained from each study participant.

Liver Histology

NAFLD was defined as the presence of ≥5% macrovesicular steatosis identified via histological examination. NASH was diagnosed based on Brunt et al's criteria: an overall pattern of histological hepatic injury consisting of steatosis, lobular inflammation and ballooning.[28,29] We graded steatosis, lobular inflammation and ballooning according to the NAFLD activity scoring system. Fibrosis was staged according to criteria described by Kleiner et al.[30] Significant fibrosis was defined as F2-F4. No-NAFLD controls were not suspected of having NAFLD, according to liver biopsy performed during an evaluation for living donor liver transplantation or the characterization of solid liver masses that were suspected to be hepatic adenoma or focal nodular hyperplasia based on an abdominal imaging study.

Global Profiling of Polar Metabolites

Global profiling analysis was performed using ultra-performance liquid chromatography (UPLC; ACQUITY™ UPLC system; Waters) coupled with quadrupole time-of-flight mass spectrometry (Q TOF-MS; Triple TOF 5600; SCIEX). Fifty microliters of each serum sample were extracted using 500 μL of chloroform/methanol (2:1, v/v) and separation was performed on an Acquity UPLC HSS T3 C18 column (2.1 × 100 mm with 1.7-μm particles; Waters). Mobile phases A and B comprised water and acetonitrile with 0.1% (v/v) formic acid respectively. Samples were eluted at 450 μL/min for 18 minutes. All samples were pooled in equal amounts to generate a quality control (QC) sample. Solvent blanks and QC sample injections were carried out between every 12 samples to assess analytical reproducibility. The mass spectrometer was analysed in the electrospray ionization positive and negative ion modes, and the mass range was 50–1000 m/z. Spectral data were analysed with MarkerView software (SCIEX) and normalized to total spectral area and quality control (QC) samples. Metabolites were putatively identified according to available information (accurate mass, fragment ions and/or retention time) that matched data from online databases (Human Metabolome Database [HMDB] and METLIN).

BA Analysis

BA analysis was performed using UPLC (Agilent 1290 Infinity) coupled with triple quadrupole mass spectrometry (TQ-MS; Agilent 6495). Fifty microliters of each serum sample were extracted using 200 μL of methanol dried using SpeedVac Concentrators. Separation was performed on an Acquity UPLC HSS T3 C18 column (2.1 × 50 mm with 1.7-μm particles; Waters). Mobile phases A and B comprised of water with 5 mM ammonium acetate and methanol-acetonitrile (1:1, v/v) respectively. Samples were eluted at 500 μL/min for 9 minutes. The linear gradient elution was as follows: 0–2.0 minutes, 35%-40% B; 2.0–2.5 minutes, 40%-45% B; 2.5–3.5 minutes, 45%-50% B; 3.5–4.6 minutes, 50%-55% B; 4.6–5.7 minutes, 55%-80% B; 5.7–5.9 minutes, 80%-85% B; 5.9–6.5 minutes, isocratic 85% B; 6.5–6.51 minutes, 85%-35% B; 6.51–9.0 minutes, isocratic 35% B. MS/MS experiments were conducted in negative ion mode with the following parameters: capillary voltage of 3.0 kV, nitrogen nebulizer gas at 30 psi, drying gas temperature of 200°C, drying gas flow rate of 16 L/min, sheath gas temperature of 300°C, sheath gas flow rate of 11 L/min and nozzle voltage of 500 V. Quantification was performed in multiple reaction monitoring (MRM) mode. Individual BA standards were used for the optimization of the conditions of each metabolite and the preparation of a series of calibration solutions to generate calibration curves. The MRM transition, collision energy, and retention time are provided in Table S1. Measured BA concentrations below the limit of quantification (LOQ) were imputed as the half value of the LOQ of the respective BA. In the case of BA concentrations in which 30% or more are below the LOQ, those BAs were excluded from further analysis.

Transcriptomic Analysis

Using an ordinary TRIzol (Invitrogen) RNA extraction protocol, total RNA was isolated from 245 liver biopsy samples (29 NAFL and 21 NASH nonobese subjects; 103 NAFL and 92 NASH obese subjects). Total RNA was quantified and assessed by a BioAnalyzer (Agilent Technologies, Inc.), and the mean RIN value was 8.35 (4.5–9.5). For RNA-seq, cDNA library preparation was performed using a TruSeq Stranded Total RNA Sample Prep Kit (Illumina, Inc.). The libraries were sequenced on Illumina HiSeq platform to generate 100 bp reads. Read alignment, mapping and quantification were performed using Tophat and HTseq.[31] Mean number of reads per sample is 74.8 million (53.8–125.6). Mean ratio of reads that were mapped to human transcriptome is 97.9% (90.4–99.0). For the DEG (differentially expressed gene) analysis, DESeq2 packages[32] in R was used. All count values are normalized with their size factors. P-values are attained by the Wald test and corrected for multiple testing using the Benjamini and Hochberg method implemented in DESeq2 package.

FGF19 Analysis

The serum FGF19 levels were quantified using a human FGF19 ELISA kit (Abcam) according to the manufacturer's instructions. Fifty microliters of standard and serum samples diluted to 12.5% with 50 μL of antibody mix were pipetted into appropriate wells of an anti-tag-coated microplate followed by incubation for 1 hour at room temperature. After washing away unbound materials, a 100 μL aliquot of 3,3',5,5'-tetramethylbenzidine (TMB) development solution was added to the wells followed by incubation for 10 mins at room temperature. The development of the blue colour from catalysis by horseradish peroxidase (HRP) was stopped by adding stop solution, and the optical density (OD) at 450 nm was measured using an Infinite 200 PRO microplate reader (Tecan Group Ltd., Männedorf, Switzerland). The standard stock solution was diluted with sample diluent, and a standard dilution series from 3.9 to 250 pg/mL was prepared to generate a standard curve. The concentrations of each sample were generated using the standard curve and calculated.

Statistical Analysis

Multivariate statistical analysis was performed using SIMCA-P+ software, version 12.0 (Umetrics) to identify significant metabolites. To identify significantly altered metabolites between NASH and NAFL patients in global metabolic profiling, features with variable importance in projection (VIP) >1.0 in the partial least squares-discriminant analysis (PLS-DA) model with a P-value <.05 in the Mann-Whitney test were selected among all features. Statistical analyses were performed using MedCalc statistical software (MedCalc Software Ltd, Ostend, Belgium) and R version 3.6.0 ( Data for clinical parameters are expressed as the mean ± standard deviation, median (interquartile range) and number (%), and for metabolite comparisons, data are expressed as the median (interquartile range). P-values were obtained from the independent t test, Mann-Whitney U-test, chi-square test, and the ranked analysis of covariance (ANCOVA) test with adjustment for age, gender, HOMA-IR and PNPLA3 genotypes. For two-way comparisons between no-NAFLD vs NAFL and NAFL vs NASH data, P-values were corrected for multiple comparisons using the Bonferroni test and significance level was readjusted to 0.05/2 (0.025). Spearman's correlation was used to determine the relationships between variables.