(T1430-10-64) Development of a Quantitative Systems Toxicology Model to Predict Drug-Induced Liver Injury in Pediatrics

Purpose: Drug-induced liver injury (DILI) is an underrecognized cause of pediatric liver disease which accounts for almost 20% of pediatric acute liver failure cases, and is a major reason for liver transplantation in the USA [1]. However, challenges such as inadequate numbers of available subjects, the need for special infrastructure and expertise, and ethical considerations often preclude extensive clinical studies in pediatric populations. In this study, our primary goal was to develop a pediatric representation in a quantitative systems toxicity (QST) modeling platform, DILIsym^®, to explore children's relative susceptibility to DILI mediated by bile acid transport inhibition.

Methods: DILIsym is a multi-scale, mathematical model of DILI, which includes key liver cell populations, intracellular biochemical systems (e.g., mitochondrial function, bile acid synthesis and transport), drug exposure, and drug-mediated toxicological mechanisms [2]. Healthy and diseased adult populations were previously represented within DILIsym. In the current study, we introduced a novel representation of pediatrics that includes pediatric-specific physiology. As a proof of concept, we represented individuals from four age groups: toddler, preschool, school age, and adolescent (1-, 4-, 10-, and 14-year-old, respectively). The PEAR-Physiology^TM in GastroPlus® was used to accurately calculate the pediatrics physiology such as body weight, organ weights, organ volumes, and organ blood flow rates. Parameters representing age-specific expression/activity of transporters and enzymes involved in bile acid homeostasis such as sodium/taurocholate cotransporting polypeptide (NTCP), multidrug resistance-associated protein (MRP) 3, MRP4, bile salt export pump (BSEP) and cholesterol 7 alpha-hydroxylase (CYP7A1) were optimized based on available clinical ontogeny data [3]–[6]. Parameters without available ontogeny data were further optimized to recapitulate the clinically observed range of bile acid concentrations in respective age groups [7] including chenodeoxycholic acid (CDCA), CDCA-amide, lithocholic acid (LCA), LCA-amide, LCA-sulfate, bulk bile acid, and total bile acid which is the sum of all the types of bile acids aforementioned. Next, a two-week simulation of a hypothetical inhibitor of bile acid transporters was performed in four pediatric individuals and a representative healthy adult subject to explore the potential impact of age on bile acid-mediated DILI. The same liver drug concentration-time profile (Figure 2A) with NTCP, BSEP, and basolateral inhibition constants were used for simulations in all age groups. Simulated outputs such as liver adenosine triphosphate (ATP), the fraction of viable hepatocytes, and plasma alanine transaminase (ALT, a biomarker of liver injury) were evaluated in the pediatric individuals and compared with the representative adult subject.

Results: The serum bile acid levels simulated in our pediatric and adult subjects recapitulated the reported, decreasing trend from young to older ages [8]–[10] (Figure 1A). While large variability has been reported in observed bile acid levels in pediatric subjects, potentially due to the small pool of participants [7], simulated bile acid levels in our representative pediatric individuals aligned reasonably well with the observed data ranges. Specifically, the total serum bile acid level in simulated pediatric individuals were within 90%-110% of the average reported values for respective age groups. The serum CDCA levels in simulated 1-, 4-, and 10-year-old subjects were higher than the average values, but still stayed in the reported ranges (Figure 1B). The serum LCA levels in the simulated 10- and 14-year-old subjects were between 65%-82% compared with the average reported values for the respective age groups (Figure 1C). In the 1- and 4-year-old subjects, simulated serum LCA levels were slightly overpredicted compared with the average reported values for the respective age groups, but were still within the reported range. Due to the lack of data on hepatic bile acid concentrations in pediatrics, we assumed that liver bile acid levels could slightly deviate from the adult subject. The simulated liver bile acid concentrations in pediatrics were within 90%-135% compared with the representative healthy adult [11]. Simulations with the hypothetical inhibitor of bile acid transporters showed similar hepatotoxicity responses in pediatric and adult subjects. In all age groups, the hypothetical inhibitor of bile acid transporters led to similar levels of hepatic bile acid accumulation, reduction of liver ATP, decline of fraction viable hepatocytes, and increase in plasma ALT; these toxicity responses occurred slightly quicker in pediatrics than in adults, but the magnitude of responses was comparable between pediatrics and the healthy adult (Figure 2).

Conclusion: Four pediatric individuals (1- ,4- ,10-, and 14-year-old subjects) represented in DILIsym recapitulated observed serum bile acid profiles reasonably well [7]. The QST modeling approach leveraging known physiology and available clinical data paves the path towards predictions of DILI in pediatrics.

References: [1] M. D. Amin, S. Harpavat, and D. H. Leung, “Drug-induced liver injury in children,” Curr Opin Pediatr, vol. 27, no. 5, pp. 625–633, Oct. 2015, doi: 10.1097/MOP.0000000000000264.
[2] P. B. Watkins, “Quantitative Systems Toxicology Approaches to Understand and Predict Drug-Induced Liver Injury,” Clin Liver Dis, vol. 24, no. 1, pp. 49–60, Feb. 2020, doi: 10.1016/j.cld.2019.09.003.
[3] S. Sharma et al., “Hepatobiliary Disposition of 17-OHPC and Taurocholate in Fetal Human Hepatocytes: A Comparison with Adult Human Hepatocytes,” Drug Metab Dispos, vol. 41, no. 2, pp. 296–304, Feb. 2013, doi: 10.1124/dmd.112.044891.
[4] B. Prasad et al., “Ontogeny of Hepatic Drug Transporters as Quantified by LC-MS/MS Proteomics,” Clin Pharmacol Ther, vol. 100, no. 4, pp. 362–370, Oct. 2016, doi: 10.1002/cpt.409.
[5] F. Freudenberg, F. Gothe, F. Beigel, C. Rust, and S. Koletzko, “Serum 7-alpha-hydroxy-4-cholesten-3-one as a marker for bile acid loss in children,” J Pediatr, vol. 163, no. 5, pp. 1367-1371.e1, Nov. 2013, doi: 10.1016/j.jpeds.2013.06.083.
[6] B. D. van Groen et al., “Proteomics of human liver membrane transporters: a focus on fetuses and newborn infants,” Eur J Pharm Sci, vol. 124, pp. 217–227, Nov. 2018, doi: 10.1016/j.ejps.2018.08.042.
[7] J. Jahnel, E. Zöhrer, H. Scharnagl, W. Erwa, G. Fauler, and T. Stojakovic, “Reference ranges of serum bile acids in children and adolescents,” Clin Chem Lab Med, vol. 53, no. 11, pp. 1807–1813, Oct. 2015, doi: 10.1515/cclm-2014-1273.
[8] S. Niijima, “Studies on the conjugating activity of bile acids in children,” Pediatr Res, vol. 19, no. 3, pp. 302–307, Mar. 1985, doi: 10.1203/00006450-198503000-00010.
[9] W. F. Balistreri, “Fetal and neonatal bile acid synthesis and metabolism — Clinical implications,” J Inherit Metab Dis, vol. 14, no. 4, pp. 459–477, Jul. 1991, doi: 10.1007/BF01797918.
[10] A. F. Hofmann, “The Continuing Importance of Bile Acids in Liver and Intestinal Disease,” Archives of Internal Medicine, vol. 159, no. 22, pp. 2647–2658, Dec. 1999, doi: 10.1001/archinte.159.22.2647.
[11] K. D. Setchell et al., “Bile acid concentrations in human and rat liver tissue and in hepatocyte nuclei,” Gastroenterology, vol. 112, no. 1, pp. 226–235, Jan. 1997, doi: 10.1016/s0016-5085(97)70239-7.

Figure 1. Comparison of simulated vs. observed bile acid concentrations in respective age groups [7]. (A) Simulated serum total bile acids in four pediatric individuals. (B) Simulated serum CDCA. (C) Simulated serum LCA. Error bar represents the standard deviation.

Figure 2. Comparison of simulated hepatotoxicity responses for adult and four pediatric individuals (1-, 4-, 10-, 14-year-old) using a hypothetical inhibitor of bile acid transporters. Simulations using a hypothetical inhibitor of bile acid transporters with competitive NTCP inhibition (Ki=126.5 umol/L), mixed BSEP inhibition (BSEP Ki = 2.4 umol/L, BSEP alpha Ki=2.4), and mixed basolateral efflux transporter inhibition (basolateral Ki = 12.9 umol/L, basolateral alpha Ki = 2.1) combined with a dynamic liver concentration-time profile (A) lead to the accumulation of bile acids in the liver (B), reduction of liver ATP (C), decrease in the viable fraction of all hepatocytes (D), and increase in plasma ALT (E) in four pediatric individuals and a representative healthy adult.