Metabolic Biomarkers of Liver Failure in Cell Models and Patient Sera: Toward Liver Damage Evaluation In Vitro

Abstract

Recent research has concentrated on the development of suitable in vitro cell models for the early identification of hepatotoxicity during drug development in order to reduce the number of animal models and to obtain a better predictability for hepatotoxic reactions in humans. The aim of the presented study was to identify translational biomarkers for acute liver injury in human patients that can serve as biomarkers for hepatocellular injury in vivo and in vitro in simple cell models. Therefore, 188 different metabolites from patients with acute-on-chronic liver failure before and after liver transplantation were analyzed with mass spectrometry. The identified potential metabolic biomarker set, including acylcarnitines, phosphatidylcholines and sphingomyelins, was used to screen primary and permanent hepatocyte culture models for their ability to model hepatotoxic responses caused by different drugs with known and unknown hepatotoxic potential. The results obtained suggest that simple in vitro cell models have the capability to display metabolic responses in biomarkers for liver cell damage in course of the treatment with different drugs and therefore can serve as a basis for in vitro models for metabolic analysis in drug toxicity testing. The identified metabolites should further be evaluated for their potential to serve as a metabolic biomarker set indicating hepatocellular injury in vitro as well as in vivo.

Keywords: drug toxicity testing; hepatocellular injury; in vitro cell models; in vivo–in vitro translation; metabolic biomarker.

Figure 1

Metabolites changed in patients with an acute-on-chronic liver failure before and after liver transplantation and identified metabolic biomarker set for hepatocellular injury. (A) Mean log2 fold changes of metabolite concentrations in patients with an acute-on-chronic liver failure before (LT 0 h) and at several time points after liver transplantation (LT 2 h, LT 12 h, LT 24 h) as opposed to post OP control patients (post-OP 2 h, post-OP 72 h, post-OP 168 h) each in comparison to healthy controls (for a detailed heatmap of log2 fold changes for each patient, see Supplementary Figure S1). (B) Detailed view of mean values of metabolites found with significant (p < 0.05) log2 fold changed concentrations in patients with hepatocellular injury as opposed to post-OP patients, each in comparison to healthy control patients potentially serving as a translational metabolic biomarker set for hepatocellular injury. Metabolites that were not significantly changed are highlighted in gray.

Figure 2

Log2 fold changes of metabolites significantly changed in patients with an acute-on-chronic liver failure before liver transplantation (LT 0 h) as opposed to log2 fold changes of metabolites after liver transplantations (LT 2 h, LT 12 h, LT 24 h). Significant changes are marked (* p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001).

Figure 3

Metabolites changed in primary human hepatocytes in response to the exposure with several drugs (A) Mean log2 fold changes of metabolite concentrations in primary human hepatocytes from two donors in comparison to vehicle control (for a detailed heatmap of log2 fold changes for each sample, see Supplementary Figure S2). (B) Detailed view of log2 fold changes of metabolites found with significant (p < 0.05) log2 fold changed concentrations in PHH from two donors in comparison to vehicle control opposed to log2 fold changes in patients with an acute-on-chronic liver failure before liver transplantation. Metabolites that were not significantly changed are highlighted in gray.

Figure 4

Metabolites changed in HepG2/C3A hepatocytes in response to the exposure with several drugs. (A) Mean log2 fold changes of metabolite concentrations HepG2/C3A in comparison to vehicle control opposed to log2 fold changes in patients with an acute-on-chronic liver failure before liver transplantation (for a detailed heatmap of log2 fold changes for each sample, see Supplementary Figure S3). (B) Detailed view of log2 fold changes of metabolites found with significant (p < 0.05) log2 fold changed concentrations in HepG2/C3A in comparison to vehicle control as opposed to log2 fold changes in patients with an acute-on-chronic liver failure before liver transplantation. Metabolites that were not significantly changed are highlighted in gray.

Figure 5

Flowchart of patient cohort. The patient cohort includes samples of patients with liver insufficiency before and at several time points (2 h, 12 h, 24 h) after liver transplantation (LT), samples of patients with different surgical interventions at different time points (2 h, 72 h, 168 h) after the intervention (post OP) and samples of healthy controls.