Drug Disposition during Pregnancy: Transport, Metabolism, and the Gut Microbiome

Abstract

Medication use during pregnancy is increasingly prevalent and medically necessary to maintain maternal health. Yet, a quarter of medications used by pregnant women have unknown safety for the pregnant population. It is therefore imperative to better understand how drug disposition is altered by pregnancy for each drug. Bupropion (BUP) is an antidepressant and smoking cessation aid that is prescribed to pregnant women. It is extensively metabolized to three pharmacologically active metabolites erythrohydrobupropion (EB), hydroxybupropion (OHB) and threohydrobupropion (TB). However, at the time of study, the overall disposition mechanisms remained poorly understood, and most studies have focused their efforts on elucidating metabolism. Hence, we screened major hepatic uptake and efflux transporters for the transport of BUP and the three active metabolites using cell lines that overexpressed organic anion transporting polypeptide (OATP) 1B1, OATP1B3, OATP2B1, OATP4A1, organic cation transport (OCT)1, breast cancer resistance protein (BCRP), multidrug resistance-associated protein 2 (MRP2), and P-glycoprotein (P-gp). We found that although there was observably significant net active uptake by OATP overexpressing cell lines, the active uptake could not be inhibited by prototypical inhibitors of OATPs. We also reported that none of the compounds were substrates of OCT1, BCRP, MRP2, or P-gp. Hence, the overall contribution of hepatic transporters in BUP disposition should be minor. Metabolism is another key drug processing event that can be altered during pregnancy. To extend efforts in this regard, we explored N-acetyltransferase 2 (NAT2) dependent pharmacokinetics on oral hydralazine in pregnant women. Hydralazine is used to treat hypertension during pregnancy because of its low risk to the fetus. Only the parent compound is reported to be pharmacologically active, and its major route of elimination is by metabolism via NAT2 to its major inactive metabolite 3-methyl-1,2,4-triazolo[3,4-a]phthalazine (MTP). NAT2 is highly polymorphic and hydralazine plasma concentrations in non-pregnant subjects could vary up to 15-fold, and adverse events have been reported at high doses in slow acetylators (SA). In this study, 12 pregnant women in their second or third trimester who were already receiving oral hydralazine (5-25mg QID) as a part of their antihypertension treatments were recruited. During one steady state interval, serial blood samples and buccal swabs were collected for NAT2 genotyping. In total, we classified 6 subjects to be rapid acetylators (RA) and 6 to be SAs. We found that SAs had significantly slower weight-adjusted apparent oral hydralazine clearance (20 vs 70 L/h, P < 0.05), higher dose-normalized area under time-concentration curve (5.9 vs 1.5 ng*h/ml, P < 0.05), lower dose-normalized peak concentrations (4.04 vs 0.77 ng/mL, P < 0.05), and larger weight-adjusted apparent oral volume of distribution (116 vs 302, P < 0.05) compared to RAs. We observed no gestational age effects between subjects in their second or third trimester. And dose effects were also not observed. Taken together, this was the first study to show NAT2 dependent oral hydralazine pharmacokinetics in pregnancy. Next, we explored the effect of the gut microbiome on hepatic drug processing genes during pregnancy using C57BL/6 conventional (CV) and germ-free (GF) mice. Four groups of female mice were used, namely CV non-pregnant (CVNP), GF non-pregnant (GFNP), CV pregnant (CVP), and GF pregnant (GFP) mice. Pregnant mice were sacrificed on gestation day 15. Hepatic mRNA was quantified by RNA-sequencing; hepatic proteomics and plasma metabolomics were performed by LC-MS/MS. We observed that pregnancy-induced effects were similar in both CV and GF mice, but the magnitudes of changes were significantly different for some hepatic enzymes. Most remarkable was the Cyp3a isoform Cyp3a11, with a 2-fold greater downregulation by pregnancy at the mRNA level in GF mice compared to CV mice, but at protein level both were elevated around 2-fold by pregnancy. Microsomal incubation revealed induction of Cyp3a activity by pregnancy in both CV and GF mice, but the induction was 5 times greater in CV mice compared to GF mice. We also analyzed the effect of the gut microbiome on overall metabolic pathways. We performed untargeted hepatic transcriptomics and untargeted plasma metabolomics on the same mice. We found that the same pathways were significantly enriched by pregnancy in CV and GF mice. However, four pathways were further modulated by germ-free in pregnancy, namely retinol metabolism, linoleic acid metabolism, arachidonic acid metabolism, and steroid hormone biosynthesis. Taken together, our studies provide clear evidence that the gut microbiome can alter host hepatic enzymes and metabolic pathways in pregnancy.