Methods to study mechanisms underlying altered hepatic drug elimination during pregnancy

Abstract

Hepatic drug metabolism is a major route of drug elimination, mediated by multiple drug-metabolizing enzymes. Any changes in the rate and extent of hepatic drug metabolism can lead to altered drug efficacy or toxicity. Accumulating clinical evidence indicates that pregnancy is accompanied by changes in hepatic drug metabolism. In this article, we discuss in vitro and in vivo tools used to study the mechanisms underlying the altered drug metabolism during pregnancy, focusing on primary hepatocyte culture, transgenic animal models, and use of probe drugs to assess change in enzymatic activity. The information obtained from these studies has enabled prediction of clinical PK changes for a given drug in pregnant women.

Hepatic drug metabolism

Hepatic metabolism is the major route of elimination for about 75% of marketed drugs.¹ Hepatic drug metabolism is mediated by multiple drug-metabolizing enzyme systems that belong to phase I and phase II metabolic pathways. Phase I enzymes mediate functionalization reactions while phase II enzymes catalyze conjugations. Cytochrome P450 (CYP) and UDP-glucuronosyltransferase (UGT) are major phase I and II enzymes, respectively. Although there are 18 families of these heme-containing CYP enzymes in humans, enzymes in the families CYP1, 2 and 3 are known to catalyze Phase I oxidative reactions for the majority of medications. Among the multiple enzymes in each family, CYP1A2, CYP2A6, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1, CYP3A4, and CYP3A5 are the major CYPs responsible for drug metabolism.¹ Phase II UGTs catalyze the conjugation of a glucuronide group to xenobiotics and endogenous molecules, producing products that are more water soluble and readily excreted into the urine or bile.² Isoforms in UGT1 and UGT2 families are involved in drug elimination.² Other Phase II enzyme catalyze the addition of different compounds, such as amino acids, sulfur moiety, acetyl and methyl groups.

Tools to study prediction of changes in hepatic drug metabolism during pregnancy

Due to the availability of large chemical libraries for identification of potential new drug candidates and the need to prioritize compounds for drug development in pharmaceutical industry, multiple in vitro drug metabolism models have been developed to predict in vivo hepatic clearance of drugs. Of these models, the two most commonly used are primary hepatocyte culture and transgenic mouse models. Based on easy manipulation of environments (e.g., culture media), such models have also enabled identification of different factors that modulate hepatic drug metabolism (e.g., drug-drug interactions) as well as detailed understanding of the underlying molecular mechanisms.^{3, 4}

(1). Primary human hepatocyte culture

Primary human hepatocytes are isolated from liver tissue by the two-step collagenase perfusion technique that involves sequential perfusion of the liver with a calcium chelator and collagenase. These agents disrupt the intercellular junctions and the supporting extracellular matrix, leading to a suspension of parenchymal cells (i.e., hepatocytes).⁵ For isolation of human hepatocytes, whole livers (not used for orthotopic transplantation) or liver tissues resected from patients from various medical conditions are typically used,⁶ and thus availability of hepatocyte is often limited due to scarcity of the tissue. Freshly isolated human hepatocytes are the closest in vitro model to human liver that can produce a metabolic profile of a given drug that is similar to that found in vivo.^{3, 4} As a result, human hepatocyte suspension is commonly used to estimate the metabolic capacity of the liver for a given drug, which is then used to predict in vivo drug disposition. For example, hepatic clearance of a drug can be estimated based on hepatic blood flow rate (a physiological value) and intrinsic metabolic clearance (experimentally obtained from hepatocyte suspension) using a predictive model. Despite the merits, primary human hepatocytes are limited by high variability (part of which originates from the isolation steps) and short life spans. Hepatocytes in suspension lose their differentiation status, including the expression of drug-metabolizing enzymes, within hours after isolation unless cryopreserved.⁷

Culturing of primary hepatocytes is one of the methods used to maintain cell viability for days after isolation. The freshly isolated hepatocytes are plated onto matrix-coated wells and cultured for days using appropriate media.⁷ While this extends the life span of hepatocytes, it does not prevent ongoing de-differentiation. For example, mRNA levels of most drug metabolizing enzymes decrease by >90% within days of culturing.⁸ Interestingly, however, hepatocytes in culture maintain the capability to respond to the inducers of drug-metabolizing enzymes, such as rifampin.^{9, 10} Rifampin binds to a hepatic transcription factor celled pregnane X receptor (PXR), which is then activated to enhance the promoter activity of target genes, including CYP3A4.¹¹ Rifampin added to the culture media increases CYP3A4 mRNA/protein/activity levels in the hepatocytes after 24–72 h treatment (with daily media change). Other chemicals that exhibit clinical drug-drug interactions also modulate CYP expression in cultured human hepatocytes,⁹ providing mechanistic understanding of the clinically observed drug interactions.¹² Of note, marked interindividual variability in the magnitude of CYP induction has been observed (e.g., 195 ± 263-fold induction in CYP3A4 mRNA level by 10 μM rifampin in hepatocytes from 18 different donors),¹³ likely reflecting batch-to-batch differences in donor liver conditions, hepatocyte isolation procedures, and subsequently, the rate of de-differentiation. As a result, the extent of CYP induction triggered by a test compound is often interpreted with caution and compared against that by a reference compound, such as rifampin. Despite these limitations, human hepatocyte culture has been accepted as a gold standard model to study regulation of drug-metabolizing enzymes and to predict potential drug-drug or drug-disease interactions.¹⁴

Pregnancy is accompanied by rising concentrations of multiple reproductive hormones. For example, the plasma concentrations of estrogen and progesterone increase at term pregnancy by up to 500-fold and 4 to 5-fold, respectively, as compared to pre-pregnancy.¹⁵ The effects of these hormones on the expression of hepatic drug-metabolizing enzymes have been previously investigated in human hepatocyte culture. In hepatocytes from female donors (hepatocytes from male donors were not tested), estradiol and progesterone increased mRNA transcript levels of many CYP enzymes (e.g., CYP2A6, CYP2B6, CYP2C8, CYP3A4, and CYP3A5), leading to subsequent increases in enzyme activities.¹² The most prominent effects were observed with CYP2A6, CYP2B6, and CYP3A4.¹² The directional change for these enzymes was consistent with that reported in pregnant women, suggesting the potential involvement of female hormones in CYP induction during pregnancy. In addition to female hormones, plasma concentration of cortisol also increases at term pregnancy compared to postpartum, although the magnitude of increase in cortisol concentration (by ~3-fold) is much smaller than that for female hormones.¹⁶ Cortisol increases CYP3A4 activity by 5 to15-fold in primary human hepatocytes when compared to vehicle-treated cells.^{17, 18} Notably, culture media free of corticosteroids was used in the latter studies. Corticosteroids are typically added in the media to enhance and maintain differentiation of primary hepatocytes. Thus, CYP3A4 induction by cortisol may in part reflect the greater differentiation status of hepatocytes caused by the addition of corticosteroids. Regardless, the magnitudes of CYP3A4 induction with cortisol treatment were comparable to the observed increases in elimination of indinavir (a CYP3A4 substrate) for the different gestational time points of pregnancy.¹⁸

Other hormones whose concentrations rise during pregnancy were also tested for their effects on CYP expression in human hepatocytes. Plasma concentrations of prolactin increase ~30-fold during pregnancy,¹⁹ and certain growth hormones, including growth hormone variant and placental lactogen, are produced only during pregnancy. In primary human hepatocytes, these growth hormones, excluding placental lactogen, did not alter CYP expression/activity,^{18, 20} and placental lactogen increased only CYP2E1 expression. Whether pregnancy is indeed associated with increased rate of CYP2E1-mediated drug metabolism remains to be determined. Together, these results support multiple hormones whose concentrations increase during pregnancy as potential factors responsible for altered drug metabolism during pregnancy, especially for CYP2A6, CYP2B6, and CYP3A4 substrates. This information may provide a basis to predict the temporal changes in hepatic drug metabolism during pregnancy that can be later verified clinically.

(2). Transgenic mouse models

Over decades, mice have been used as the in vivo animal model for various physiological studies. Indeed, about 80% of mouse proteins appear to have strict 1:1 orthologues in the human genome,²¹ supporting the usefulness of the mouse model in physiological studies. However, genes encoding drug-metabolizing enzymes belong to the remaining 20% of mouse proteins that lack the strict 1:1 relationship,²¹ indicating that the CYP gene families have expanded and changed since the last common ancestor. Not surprisingly, marked interspecies differences in the expression and regulation of drug-metabolizing enzymes have been reported between mice and humans,²² limiting the use of mouse models for drug metabolism studies. For example, while numerous clinical studies indicate enhanced CYP2D6-mediated drug metabolism in pregnant women (reviewed in ^{23, 24}), hepatic expression of endogenous CYP2D homologs in rodents (e.g., mouse Cyp2d22) was found decreased during pregnancy.²⁵ This likely reflects the interspecies differences in upstream regulatory sequences for the respective genes. To overcome the limitations, various transgenic mouse models have been developed where human genes along with their upstream regulatory sequences were injected into mouse embryo and the genetic materials were integrated into the mouse genome.²² The resulting transgenic mice have provided a rare opportunity to examine how human CYP promoters respond to various physiological and pathological conditions in vivo.

CYP2D6 is one of the major drug-metabolizing enzymes and is responsible for eliminating ~25% of marketed drugs.²⁶ CYP2D6-humanized transgenic (tg-CYP2D6) mice were established about 20 years ago to characterize CYP2D6-mediated drug metabolism in an in vivo system.²⁷ Notably, the genome of tg-CYP2D6 mice harbors both the structural CYP2D6 gene (wild-type) and the 2.5-kb upstream regulatory region of CYP2D6; thus, regulation of CYP2D6 gene expression during the complex physiological changes accompanying pregnancy can be examined. When CYP2D6 levels were measured at different gestational time points of pregnancy in tg-CYP2D6 mice, CYP2D6 expression/activity was found to be 2–3-fold higher at term pregnancy, which returned to the pre-pregnancy level after delivery.²⁸ The magnitude of increase in CYP2D6-mediated drug metabolism was similar to the clinically observed changes in pregnant women, validating tg-CYP2D6 mice as a potential in vivo model to study mechanisms underlying CYP2D6 induction during pregnancy. Subsequent mechanisms studies revealed that CYP2D6 induction during pregnancy is mediated by transcriptional up-regulation of the CYP2D6 promoter, which is potentially triggered by changes in hepatic retinoid homeostasis during pregnancy.²⁸ Retinoids are a class of compounds belonging to vitamin A that are absorbed from food, stored in the liver, and converted to bioactive retinoid all-trans retinoic acid when needed. The hepatic contents of all-trans retinoic acid were found to be lower at term pregnancy (vs. pre-pregnancy levels) in tg-CYP2D6 mice.²⁸ These results suggest a role for altered retinoid homeostasis in CYP2D6 induction during pregnancy. How pregnancy impacts retinoid homeostasis remains to be determined.

CYP3A4 is a drug metabolizing enzyme responsible for metabolizing ~50% of marketed drugs. CYP3A4-promoter-humanized (tg-CYP3A4) mouse line was established in early 2000’s to study transcriptional regulation of the human CYP3A4 gene in vivo.²⁹ The genome harbors a 13-kb human CYP3A4 promoter driving the firefly luciferase gene expression such that CYP3A4 expression levels are reflected in the levels of luciferase activities in tissues.²⁹ In these mice, pregnancy led to ~4-fold increases in CYP3A4 promoter activity as compared to pre-pregnancy,³⁰ indicating that enhanced CYP3A4-mediated drug metabolism observed in pregnant women can be attributable to transcriptional up-regulation of CYP3A4 gene. This result is also consistent with the enhanced CYP3A4 transcripts seen in hormone-treated human hepatocytes.

(3). Probe Drugs

Probe drugs are medications that are metabolized almost exclusively by a single enzyme, so that clearance of that drug reflects the metabolic capacity of the responsible enzyme in that environment. Although less ideal, occasionally a selected probe drug (e.g. dextromethorphan) may have more than one primary metabolite, with each pathway being uniquely catalyzed by a single enzyme (CYP2D6 and CYP3A4). Calculating the clearance of the parent drug to each of the different metabolites provides an estimate of the activity of the respective enzymes. Additional probe molecules have been identified that are transported by a single transporter and can be used to evaluate activity of that specific transporter. The FDA recommends the use of probe drugs to assess changes in enzyme or transporter activity under various conditions, including pre-marketing evaluation of new medications for potential drug-drug interactions and investigating changes in enzyme activity in different disease states and special populations, including pregnancy.³¹ Probe drugs can be used to evaluate enzyme or transport activity both in human subjects as well as in transgenic mouse models and in in vitro, hepatic cell culture studies.³² Probe medications can also be administered in combination, as a “cocktail”, allowing simultaneous assessment of multiple enzymatic pathways. As various combinations of probe drugs have been evaluated, current recommendations advise using regimens that favor oral over intravenous drug administration, minimize invasiveness of the assessments, and utilize reduced or micro dosing whenever possible.³³ Studies utilizing probe drugs in pregnant women have elucidated changes in several important Phase I enzymes. Tracey et al in 2005³⁴ gave 35 pregnant woman an oral cocktail of caffeine 100 mg (probe for CYP1A2 activity) and dextromethorphan 30 mg (CYP2D6 O-demethylation and CYP3A N-demethylation) in early second trimester, late second trimester, late third trimester and 6 to 8 weeks postpartum. Their results demonstrated that compared to postpartum, CYP1A2 activity progressively decreased with advancing gestational age (−32.8% +/− 22.8%, −48.1% +/− 27%, −65.2% +/− 15.3%), CYP2D6 activity (poor metabolizers were excluded) sequentially increased across pregnancy (25.6% +/− 58.3%, 34.8% +/− 41.4%, 47.8% +/− 24.7%) and CYP3A activity went up in early pregnancy and stayed elevated at 35–38%. Using a 2 mg dose of midazolam in 13 pregnant women, Hebert, et al in 2008³⁵ documented a significant increase in CYP3A activity in the third trimester compared to 6 to 10 weeks postpartum: area under the concentration time curve decreased by 46 26%, maximum concentration decreased by 28 ± 32% and apparent oral clearance increased by 108 ± 62%. In 1993, CYP2C19 activity was shown to decrease in pregnancy compared to postpartum using the conversion of probe drug proguanil to cycloguanil.³⁶

Prospective.

Primary hepatocytes and transgenic mouse models have served as models that complement each other to identify potential factors responsible for altered drug metabolism during pregnancy, namely female hormones, retinoids, and corticosteroids. While the extent of their relative contribution to altered drug metabolism in pregnant women remains to be determined clinically, mechanistic understanding of different factors that govern drug disposition has provided a basis to predict the PK changes of a given drug in an individual. For example, a systems pharmacology approach has allowed prediction of PK changes of a drug at different gestational time points of pregnancy by integrating physiological data as well as results from preclinical mechanistic studies.^37–39 Studies utilizing probe drugs to evaluate changes in specific enzymatic activity during pregnancy have corroborated some of these pre-clinical findings. The information from these various types of studies should guide the design of focused clinical PK studies in pregnant women to further test the predictions based on our mechanistic studies.