Hepatitis C virus (HCV) is estimated to infect over 185 million people worldwide, often leading to severe liver damage (Thrift et al., 2017; Mohd Hanafiah et al., 2013; https://www.cdc.gov/hepatitis/hcv/cfaq.htm). The virus can cause both acute and chronic hepatitis, ranging from asymptomatic to serious, lifelong illnesses including liver cirrhosis and cancer. For some, the virus is cleared spontaneously without any intervention, but for the majority of individuals, chronic HCV infection develops over time (Thrift et al., 2017). Globally, an estimated 58 million people have chronic HCV infections, but this is likely to be an underestimate as studies tend to exclude high-risk groups such as incarcerated persons (Thrift et al., 2017; Blach et al., 2022). Moreover, while HCV infections have recently decreased in Asia, they are currently on the increase in North America primarily due to an aging population, illustrating that the disease burden from HCV infection will be with us for some time to come (Yang et al., 2023).

Women account for ∼35% of HCV cases, and an estimated 15 million women of childbearing age (15–49 years old) have HCV worldwide according to cohort studies (Pott et al., 2018; Dugan et al., 2021). The prevalence of HCV in pregnant women is estimated at between 1% and 8% of the global population, with prevalence dependent upon demographics (Arshad et al., 2011; Le Campion et al., 2012). The pathogenesis of HCV infection during pregnancy remains poorly understood, as does the effect it poses on pregnancy outcomes (Yeung et al., 2014; He et al., 2023). However, some studies report an increased rate and/or risk of low birth weight, small for gestational age, preterm birth, low Apgar scores, congenital malformations, use of intensive care, assisted ventilation, and overall perinatal mortality relative to HCV-negative women (Yeung et al., 2014; He et al., 2023).

It is widely accepted that there is a significant risk of mother-to-infant transmission, occurring in about 5% of cases, and the risk of vertical transmission increases ∼10-fold if the mother is coinfected with HIV (Le Campion et al., 2012; He et al., 2023). Over the past two decades, a significant effort has been made to combat HCV with the development of novel antiviral therapies. However, there are currently no guidelines implemented for their use in pregnant women. Thus, the effect of off-label use of these medications on pregnancy outcomes remains undefined.

Recent studies of HIV antiviral use during pregnancy have been correlated with adverse effects on the developing fetus, including increased rate of preterm delivery, adrenal dysfunction, and elevated plasma concentrations of adrenal steroid hormones such as dehydroepiandrosterone 3-sulfate (DHEA-S) (Kariyawasam et al., 2014, 2020; https://clinicalinfo.hiv.gov/en/guidelines/perinatal/safety-toxicity-arv-agents-protease-inhibitors-lopinavir-ritonavir-kaletra). It has been hypothesized that some of these effects may stem from inhibition of cytochrome P450 3A7 (CYP3A7), the predominant enzyme expressed in fetal livers (Kandel and Lampe 2021). CYP3A7 accounts for ∼50% of the total hepatic CYP content in fetal livers and is thus the major metabolizing enzyme for xenobiotics (Li and Lampe, 2019). It also regulates an important function during development; CYP3A7 metabolizes DHEA-S to 16α-hydroxy-DHEA-S (Kitada et al., 1987; Torimoto et al., 2007; Kandel and Lampe 2021; Li and Lampe, 2019), and this metabolite is a precursor molecule to the production of estriol (Li and Lampe, 2019; Kandel and Lampe, 2021), a critically important hormone in bringing the pregnancy to full term (Chatuphonprasert et al., 2018). Low estriol levels during the last trimester of pregnancy have been linked to both premature birth and low birth weight, two leading causes of infant mortality (Deng et al., 2022; Yilmaz Gulec et al., 2022).

While HCV antivirals have demonstrated clinical effectiveness, significant questions remain regarding their metabolism and their relationship to reported hepatotoxicity. Whereas some drugs in this class have been reported to be substrates of CYP3A4, no reports have studied the effect they have on fetal CYP3A7. Although CYP3A4 and CYP3A7 share ∼88% amino acid sequence similarity and overlapping substrate specificity, the Kcats of these enzymes differ significantly; CYP3A7 has, on average, a hundredfold lower catalytic rate compared with CYP3A4 (Li and Lampe, 2019). CYP3A7’s lower catalytic activity, along with its broad range of substrates, makes it vulnerable to drug–drug or drug–hormone interactions that may lead to adverse drug reactions.

Herein, we investigated the inhibitory effects of 13 Food and Drug Administration (FDA)-approved HCV antiviral drugs on CYP3A7 activity. This was initially assessed with a high-throughput fluorescent assay previously developed by our laboratory that utilizes the probe dibenzylfluorescein (DBF) (Work et al., 2021). Drugs that reduced CYP3A7 activity by 50% or more in this assay were further characterized by determining their IC50 values using an liquid Chromatography with tandem mass spectrometry (LC-MS/MS)-based method and measuring 16α-hydroxy-DHEA-S formation. We also investigated if any of these compounds were time-dependent independent (TDI) or mechanism-based inactivators (MBI) of CYP3A7 using a single-point fluorescent assay, and any positive hits were again followed up by further characterization using LC-MS/MS and DHEA-S as an activity probe. Based on our data, HCV antivirals could pose a serious threat to the developing fetus and neonate by blocking CYP3A7 metabolism of DHEA-S, thus affecting the fetal–maternal communication axis.

Although the World Health Organization has proposed a strategy to reduce and eliminate hepatitis C by 2030, HCV remains a global issue today and still affects roughly 100 million people around the world (Dore and Bajis, 2021). While there is already a significant burden on the general population, this is exacerbated in pregnant women who have HCV, as there are few to no treatment guidelines available, and it is unclear what threat current treatments pose to the developing fetus. The little information we do have on the effect of HCV antivirals on fetal development comes from either 1) animal studies, which historically have a poor correlation with human outcomes, or 2) previously reported exposure during pregnancy, which is a very limited dataset.

When it comes to determining fetal and neonatal exposure, it must be taken into account that the drug must first pass the placenta of the pregnant individual to get to the fetus or be transferred by breastmilk to the neonate. To date, placental transfer of some HCV antivirals has been observed in rabbits (grazoprevir) and in rats (sofosbuvir, ledipasvir, velpatasvir, grazoprevir, and glecaprevir) (Freriksen et al., 2019). If the HCV antiviral is passed to the fetus or neonate, one of its main modes of clearance is by cytochrome P450 CYP3A7. This enzyme is highly expressed in fetal and neonatal livers, accounting for up to 50% of the total CYP content and 87% to 100% of the total CYP3A content (Kitada et al., 1987; Li and Lampe, 2019). CYP3A7 also plays a vital role in fetus and neonatal development, as it metabolizes DHEA-S into 16α-hydroxy-DHEA-S, a precursor to estriol production. Thus, in this study we sought to investigate the risk HCV antiviral therapy may pose to developing fetuses and neonates by characterizing CYP3A7 inhibition by these drugs.

The 13 HCV antiviral drugs investigated in this study are presented in Table 1. The FDA labels most of these compounds as Pregnancy Category B due to their lack of embryotoxicity during animal testing, but human and rodent interspecies differences can be significant, particularly in regard to metabolism. Some of the other drugs examined do not yet have a pregnancy category label since they are investigational drugs in clinical trials, and others are labeled X due to their required coadministration with ribavirin and interferon α-2b, as this medication shows significant teratogenic and embryocidal effects at as low as 1/20th of the recommended human dose (https://www.accessdata.fda.gov/drugsatfda_docs/label/2002/20903s23lbl.pdf). Of the 13 HCV antivirals, 11 were reported to be CYP3A4 substrates or inhibitors. However, of those 11, only 5 mention their effect on CYP3A4 on their respective FDA label (these are boceprevir, telaprevir, grazoprevir, velpatasvir, and glecaprevir) (Brennan et al., 2015; Eley et al., 2015; Isakov et al., 2016; Ahmed et al., 2017; Smolders et al., 2017; Freriksen et al., 2019; Miao et al., 2020; https://www.accessdata.fda.gov/drugsatfda_docs/label/2021/214187s000lbl.pdf; https://www.accessdata.fda.gov/drugsatfda_docs/label/2011/201917lbl.pdf; https://www.accessdata.fda.gov/drugsatfda_docs/label/2011/202379lbl.pdf; https://www.accessdata.fda.gov/drugsatfda_docs/label/2011/202258lbl.pdf). Despite the fact that none of the aforementioned drugs are deemed safe for use during pregnancy and no clinical trials have been performed in this patient population group, we did find reports of their use during pregnancy. Velpatasvir, ledipasvir, paritaprevir, and sofosbuvir all have been previously used during pregnancy (Chappell et al., 2020; AbdAllah et al., 2021; Kislovskiy et al., 2021; Zeng et al., 2022; Curtis and Chappell, 2023).

HCV antiviral drugs, their corresponding structures, reported use in pregnant persons, and CYP3A4 inhibition characteristics

We began with an initial screen of these antivirals to determine their propensity for inhibition of CYP3A7 metabolism of DBF. Although this reaction is not unique or specific to CYP3A7, previous findings determined that CYP3A7 demonstrated maximal activity with this fluorescent substrate, producing the lowest Km and highest signal-to-noise ratio (Work et al., 2021). Additionally, these assays use only recombinant CYP3A7 and NADPH reductase supplemented with cytochrome b5, without interference from other CYP enzymes or alternative clearance pathways. Using this technique, we discovered that 8 of the 13 tested HCV antivirals inhibited CYP3A7 by 50% or more at a concentration of 20 µM (Fig. 1). Of the eight aforementioned drugs, ledipasvir was the only antiviral not reported to be a CYP3A4 inhibitor or substrate (Table 1). However, we were unable to further investigate ledipasvir-mediated inhibition, as well as the other NS5a protease inhibitor velpatasvir, due to solubility issues with each compound in reaction mixtures; at higher concentrations each drug was found to precipitate out of solution. In contrast, of the five drugs that did not significantly inhibit CYP3A7, four of them are reported to be CYP3A4 substrates or inhibitors (boceprevir, narlaprevir, telaprevir, and grazoprevir) (Table 1). While these two enzymes are ∼88% identical in amino acid sequence, these results demonstrate the slight, but important, variations in substrate specificity that may alter toxicological outcomes.

We further characterized five of the drugs that showed > 50% inhibition of CYP3A7 by determining their half-maximal inhibitory concentrations using DHEA-S as an activity probe, as this is a more physiologically relevant marker. Paritaprevir, glecaprevir, asunaprevir, danoprevir, and simeprevir all had IC50 values ranging between 10 µM and 21 µM. These five drugs, along with dasabuvir and grazoprevir, all contain a nitrogen-sulfone bond. Sulfate sidechains are typically favorable to CYP3A7 as the naturally occurring endogenous substrate, DHEA-S, contains a sulfate moiety that anchors it in the active site; CYP3A7 is known to be a more efficient metabolizer of DHEA-S than DHEA (Li and Lampe, 2019). This may help explain these drugs’ higher affinity for CYP3A7 binding and thus inhibition of DHEA-S metabolism. It is probable that dasabuvir is also an inhibitor of DHEA-S metabolism by CYP3A7, but we were unable to test high enough concentrations of dasabuvir due to solubility issues and DMSO constraints in reactions. Similarly, grazoprevir most likely inhibits DHEA-S metabolism but less potently (i.e., higher IC50 value) than the five drugs we followed up on based on our criteria from the DBF assay results.

Upon further investigation, paritaprevir showed an indication of being a mechanism-based inactivator of CYP3A7, causing a ∼twofold shift in the CYP3A7 IC50 value from 11 µM to 5.5 µM (Table 2, Fig. 3). This drug inactivated CYP3A7 DHEA-S metabolism, with KI and kinact values of 4.66 µM and 0.00954 minute−1, respectively. Although the KI value is rather moderate compared with the other known CYP3A7 inactivator ritonavir [KI = 0.392 µM, (Kandel and Lampe, 2021)], it is still within the range to pose a potential threat to the neonate. The Km of DHEA-S for CYP3A7 has been reported to be between 5 µM and 20 µM, so paritaprevir could compete with DHEA-S for binding to CYP3A7 and inactivate the enzyme, blocking downstream production of estriol and leading to disruption of the fetal–maternal communication axis.

IC50 shift of paritaprevir and ritonavir due to 30-minute preincubation in presence of NADPH

While paritaprevir has been reported to be a CYP3A4 substrate (Shen et al., 2016; Ahmed et al., 2017; Shebley et al., 2017; Smolders et al., 2017), there are no reports regarding paritaprevir exhibiting mechanism-based inactivation of CYP3A4. This could be due to the selection of test systems used, i.e., human liver microsomes (HLMs) instead of isolated/recombinant CYP3A4. In more complicated testing systems, such as HLMs, other CYPs and Phase II enzymes could be contributing to paritaprevir binding and metabolism, masking the TDI or MBI effect it may have on CYP3A4 alone (Shebley et al., 2017). This suggests that the TDI may not be significant in regards to CYP3A4. However, in a developing fetus who may be exposed to the drug, it could prove very significant due to the fact that CYP3A7 is the only CYP3A enzyme in the fetal liver and is known to have an average Kcat 1/100 that of CYP3A4 for most drug substrates (Li and Lampe, 2019), hence increasing the risk for drug–drug interactions and drug toxicity in the developing fetus and neonate. Post-marketing surveillance revealed instances of sudden alanine aminotransferase elevations and drug-induced hepatotoxicity with paritaprevir in combination with other HCV antiviral inhibitors (Kumar et al., 2019; http://www.ncbi.nlm.nih.gov/books/NBK548747/). While the mechanism of liver injury is currently unknown, it is possible that reactive metabolites produced from one or more of the drugs in the combination treatment may play a role. Additionally, reactive metabolites could function to inhibit the clearance pathways for other CYP3A drug substrates, leading to drug–drug interactions and adverse side effects.

In a study performed by Shen et al. (2016), the group identified 18 different oxidized metabolites of paritaprevir in plasma, urine, and feces after healthy male subjects received a 200 mg dose of [14C]paritaprevir coadministered with 100 mg of ritonavir. The group further determined that the two most abundant oxidized metabolites (M2 and M24) were formed by hepatic enzymes, most likely CYP3A4. Both the M2 and M24 metabolites are proposed to be oxidized at or near the C9 carbon of the substrate. The M24 metabolite was postulated to form from an unstable epoxide intermediate produced by CYP3A4 in vivo that readily reacts with GSH to form a glutathione adduct (Shen et al., 2016; Supplemental Fig. 1). While the complete characterization of the CYP3A7 metabolite profile of paritaprevir is beyond the scope of our study reported here, we have attempted to provide a plausible mechanism for protein adduction and inactivation based on the reactivity of this known CYP3A metabolite (Supplemental Fig. 2). While this potential inactivation pathway still needs to be confirmed, it may go some way to explain the TDI observed with paritaprevir and CYP3A7.

To gain additional insight into the structural details of the paritaprevir-CYP3A7 interaction, we conducted an in silico docking study. The most stable binding interaction between paritaprevir and CYP3A7 is represented in Fig. 5. This binding pose places the C9 carbon of paritaprevir in the closest position to the heme iron (∼4.5 Å; Fig. 5B), suggesting a possible site of oxidation that corresponds well with what has been previously reported for CYP3A4 (Shen et al., 2016). However, this needs to be validated with future metabolite identification studies. Despite this, the number of interactions between paritaprevir and residues within the CYP3A7 active site suggests a high degree of stability between the ligand and the protein (Fig. 5C).

Although there are likely differences between the metabolism of paritaprevir by CYP3A4 and CYP3A7, it is plausible that oxidation at the C9 position could produce the reactive epoxide intermediate identified previously, leading to CYP3A7 alkylation at nucleophilic residues, such as cysteine thiols, e.g., Cys58 (Supplemental Fig. 2A), causing CYP3A7 inactivation. The effects of paritaprevir on CYP3A inhibition may have initially been missed in the clinical studies due to the overwhelming inhibitory effect of ritonavir on CYP3A4 (Shebley et al., 2017). Experiments are currently underway to determine if the TDI effects of paritaprevir differ between CYP3A7 and CYP3A4 in recombinant systems and HLMs.

While the IC50s of CYP3A7 by the hepatitis C antivirals tested were all comparatively high (5 µM–20 µM), they may be in a clinically relevant concentration range. The recommendations for HCV treatment range between 8 and 24 weeks of daily antiviral administration depending on the stage of infection/disease, with concomitant treatment of two to four antivirals at doses between 100 mg and 500 mg daily (Talal et al., 2018; Smolders et al., 2019). The Cmax recorded in nonpregnant, HCV-free adults also has a wide range depending on treatment plan, drug, and usage of pharmacokinetic enhancers like ritonavir. The results vary widely between 100 ng/mL and 15,000 ng/mL, or roughly 0.1 to 20 µM, as seen in the case of paritaprevir dosed with or without ritonavir (Menon et al., 2016). As both drugs inhibit and inactivate CYP3A7, the high circulating plasma levels in adults makes drug exposure to the fetus highly probable, and this could lead to significant adverse effects due to their CYP3A7 interactions. Additionally, the high Cmax values could be further exacerbated if patients have chronic HCV infection, affecting normal liver function and metabolism of these drugs. While no studies exist on fetal exposure to maternal HCV treatment (e.g., umbilical cord concentration, placental concentrations, etc.) due to a plethora of ethical and logistical reasons, it can be speculated that, due to the high plasma levels and evidence of placenta and breastmilk transfer of antivirals, these drugs could pose a serious threat to fetuses and neonates.