The ABC transporters BCRP/ABCG2, MDR1/ABCB1, MRP2/ABCC2, MRP3/ABCC3, MRP4/ABCC4, the solute carriers OATP4A1/SLCO4A1 and ASBT/SLC10A2, and the P-type ATPase FIC1/ATP8B1 are reported to be localized at the apical side of the placental syncytium (Ontsouka et al., 2021) which is consistent with their role in the retrograde transport towards the maternal compartment (Figure 2). Many of the apical transporters mentioned above have partially overlapping substrate specificities. Substrates include cholesterol, nutritional compounds, xenobiotics, different BA species (e.g., taurocholate and cholate), estrogens and prostaglandins and their sulfated derivates (Allikmets et al., 1998; Zelcer et al., 2003; Lam et al., 2005). For some of these transporters consistent expression throughout gestation has been found in the human placenta. Depending on the study, ABCG2 gene expression has been shown to both increase and remain unchanged across gestation (Mathias et al., 2005; Meyer Zu Schwabedissen et al., 2005; Yeboah et al., 2006; Petrovic et al., 2015; Sieppi et al., 2016). The role of placental ABCG2 and MRP3 in the increase of maternal serum BAs was recently confirmed (Liu et al., 2021) implying a protective function of these transporters for the fetus. ABCB1 (both mRNA and protein) has been reported to be highly expressed in the first trimester compared to term placentas (Mathias et al., 2005). In rats, placental mRNA expression of Oatp1b2 decreased from gestational day 13 to day 20, then increased again immediately before birth, while the opposite pattern was reported for placental Oatp-4 mRNA levels (St-Pierre et al., 2004). In contrast, ATP8B1 mRNA expression was significantly reduced (33-fold) in the third compared to the first trimester in humans (Patel et al., 2003). In general, these data suggest that the investigated transporters are involved in the maintenance of maternal-fetal BA homeostasis at different gestational stages.

The ABC transporter MDR3/ABCB4 as well as the solute carriers OATP2B1/SLCO2B1, OATP3A1/SLCO3A1, NTCP/SLC10A1, OST-α/SLC51A1/, OST-β/SLC51B1, and the small protein mEH/EPHX1 were reported to be located at the basal side of the syncytiotrophoblast layer (Ontsouka et al., 2021). Given their transport directionality and specificity, these proteins modulate fetal exposure to a broad range of substrates, including nutritional compounds, sulfated steroids, and BA species (Dippe et al., 1996; Tamai et al., 2000; Huber et al., 2006; Roth et al., 2012). ABCB4 mRNA expression in the human placenta was increased 4-fold in the third trimester compared to the first trimester (Patel et al., 2003). Considering the unidirectionality of the substrate transport by ABCB4 towards the fetal circulation (Figure 2), its gestation-dependent increase may suggest an incremental role of ABCB4 in sustaining fetal development. ABCB4 is potentially also important in transporting BAs to their sites of receptor-mediated actions in the developing fetus as well as in the placenta itself. Moreover, given the BA transport directionality suggested for NTCP, OATP2B1, OATP3A1, OST-α/-β, and mEH in an earlier study (Ontsouka et al., 2021), these transporters may represent active players in the retrograde BA transport system towards the placenta/mother, thus protecting the fetus from overexposure to BA.

The most common BA-related disorder specific to pregnancy is intrahepatic cholestasis of pregnancy (ICP), a disease that can occur as early as the seventh week of gestation but appears more frequently in the third trimester. It is characterized by high levels of maternal BA concentrations with a cut-off value of 10 μmol/L (Manzotti et al., 2019) and up to a maximum value of 430 μmol/L (Lammert et al., 2000). ICP is clinically characterized by pruritus, abnormal liver function tests (liver transaminases), and supra-physiological levels of circulating BAs (Lammert et al., 2000). There is a well-documented correlation between increased maternal blood BA levels and increased rates of fetal complications. In cases of severe ICP, where maternal serum BA concentrations reach 100 μmol/L (Brouwers et al., 2015), a poor pregnancy outcome is expected, including preterm labor (Geenes and Williamson, 2009), fetal distress, fetal asphyxia, and even intra-uterine death (Fisk and Storey, 1988; Glantz et al., 2004; Williamson et al., 2004). In these cases, greater concentrations of BAs are detected in the amniotic fluid, similar to women who gave birth to infants suffering from prenatal intestinal obstruction (Délèze et al., 1977).

Mutations in some of the apical efflux transporters have been associated with pregnancy-related diseases (Jacquemin et al., 1999; Müllenbach et al., 2005; Aydın et al., 2020). As suggested by (Müllenbach et al., 2005; Aydın et al., 2020), mutations in ATP8B1/FIC1 were found to be linked to ICP in humans. This aminophospholipid translocase modulating membrane asymmetry (Tang et al., 1996) was speculated to cause or contribute to ICP, but it is still unclear how mutations in the translocase activities of ATP8B1/FIC1 (Müllenbach et al., 2005; Aydın et al., 2020), may be mechanistically involved in this disease. Clinical treatment of ICP patients using UDCA has been associated with alterations in the placental expression of some BA transporters. One study reported a significant increase in the placental expression of MRP2 (both protein and mRNA) while MRP3 protein levels were not significantly different in ICP patients treated with UCDA compared to controls (Azzaroli et al., 2007). Similarly, UDCA treatment in humans was associated with significant upregulation of placental mRNA and protein levels of ABCG2 (Azzaroli et al., 2013). On the other side, in patients with moderate ICP which were treated with UDCA, neither the mRNA expression of ABCG2 nor MRP2 was significantly different compared to controls (Ontsouka et al., 2021). Thus, the discrepancy regarding the MRP2 gene expression between the latter two studies (Azzaroli et al., 2007; Ontsouka et al., 2021) may be linked to the variable severity degree found in the ICP patients. Finally, an association between a mutation in the ABCB4 gene (located on the basal side of the syncytiotrophoblast layer) and the occurrence of ICP was also documented (Jacquemin et al., 1999), suggesting its possible mechanistic involvement in the pathogenesis of ICP. Similarly, treatment with UDCA was associated with significant downregulation of SLCO3A1 mRNA in ICP patient treated with UDCA as compared to controls (Ontsouka et al., 2021). A model summarizing the molecular mechanisms through which BAs interact with their receptors and transporters at the apical and basal sides of the syncytiotrophoblast layer is shown in Figure 2.

It is well established that the enterohepatic circuit determines serum BA levels (Hofmann, 2009). So far, the placenta has been primarily considered as a physical barrier between the maternal and fetal compartments, across which various substances including BAs and gases are exchanged (Herraez et al., 2014; Liu et al., 2021), thereby influencing the levels of these compounds in the maternal and fetal blood. Noticeably, the placenta is also a steroidogenic organ capable of the synthesis of diverse hormones such as progesterone using cholesterol as a precursor molecule (Karahoda et al., 2021). Interestingly, there is evidence that also extrahepatic tissues (e.g., the brain) are capable of synthesizing BAs and therefore contribute to the serum BA pool (Pan et al., 2017; Monteiro-Cardoso et al., 2021). In the context of pregnancy, previous studies have shown that serum BA levels are higher in pregnant compared to non-pregnant women and mice (Milona et al., 2010), regardless of the gestational age (Colombo et al., 1985; Egan et al., 2012). The concentration of BAs in healthy pregnancies is usually higher in the fetal blood than in the maternal circulation (Itoh et al., 1982; Colombo et al., 1985). In this physiological situation, the placenta serves as an exchange interface that prevents the accumulation of BAs in the fetal compartment to avoid deleterious effects on the fetus (Sewell et al., 1980; Colombo et al., 1985; St-Pierre et al., 2000). However, a study by Sasaki et al. reported that in 52.6% of the patients the BA content in the maternal serum was higher than in the umbilical cord, while in 36.8% of the cases the opposite was observed, and in 9.2% the BA content was similar (Sasaki, 1984). Overall, the vectorial transfer of the primary BAs across the placenta occurs mainly from the fetus to the mother (Figure 2), although secondary maternal BAs are transported from the mother into the fetal vessels (Colombo et al., 1987). A recent study performed in rats further suggested that also sex may play a role in BA concentrations since the total amount of placental BAs was two times higher in males than in females, but the opposite trend was observed in the fetal serum (Huang et al., 2021). Hence the factor sex adds another important variable influencing the selective allocation of BAs on both sides of the placental interface.

Given the central role of the placenta in facilitating the bi-directional transport of BAs (Figure 2), it is not surprising that placental dysfunction is a hallmark accompanying pregnancy-related metabolic diseases (e.g., preeclampsia, ICP). In ICP, particularly in severe cases (Brouwers et al., 2015), BAs are thought to be transported from the mother to the developing fetus (Meng et al., 1997; Tribe et al., 2010; Geenes et al., 2014; Liu et al., 2021; Ontsouka et al., 2021) exposing it to supraphysiological and toxic levels (Figure 2). Given the fetal over-exposure to BAs in ICP, one may question whether the placenta acts as an efficient BA protective barrier or rather as an exchange interface. It is important to note that elevated serum BAs in pregnant women do not necessarily imply ICP, since a subset of healthy pregnant women spontaneously presents levels above the mentioned cut-off value - a condition called asymptomatic hypercholanemia of pregnancy (AHP) - without suffering from ICP-related complications or other liver diseases (Pascual et al., 2002; Castaño et al., 2006). In this context, one study comparing the profiles of serum BAs between ICP and normal pregnancies including AHP suggested a shift towards a hydrophobic composition and free BAs in women with ICP (Castaño et al., 2006). This implies that the deleterious effects commonly attributed to this disease may result from the BA profile rather than the total amount of BAs. Moreover, it has been reported that some BA species —especially hydrophilic BAs such as UDCA and its taurine-conjugated derivative— exhibit beneficial health effects. Among others, they promote cytoprotective effects by inhibiting cellular oxidative stress and apoptosis (Perez and Britz, 2009). Serum BA profiles on both sides of the placenta are markedly different in healthy pregnancies (Sasaki, 1984; Geenes et al., 2014), highlighting the potential involvement of the placenta in the selective distribution of BA species between the maternal and fetal blood. Sasaki et al. (Sasaki, 1984) compared the BA distribution in maternal and umbilical cord sera in healthy pregnancies and showed that GCDCA was more prominent in the mother, while TCDCA was predominant in the umbilical cord (Sasaki, 1984). In contrast, the most abundant species in the fetal circulation in ICP is CA, which accounts for approximately 70%–80% of all circulating BAs (Tribe et al., 2010; Geenes et al., 2014). The observed differences imply an important role of the placenta in the transport, selection, and metabolism of the BAs arriving from both the mother and the fetus. Moreover, the placenta potentially complements the arriving BA pool with own locally synthesized BAs given its intrinsic placental steroidogenic machinery (see chapter 2.4.2.).

In a recent study, we reported the presence of mRNA transcripts of most of the BA synthesizing enzymes in the placentas of healthy mice and humans and proposed this organ as a potential new source of BA (Ontsouka et al., 2023). In this study, a species-specific expression profile of BA synthesizing enzymes was observed in placental tissues. Among others, CYP7A1, CYP46A1, and BAAT were undetected in human placental tissue, but their homologs were detected in the mouse placenta. Conversely, mRNA of Cyp8b1 and Hsd17b1 were missing in the mouse placenta, while their homologs were expressed in the human placenta. CYP39A1/Cyp39a1 and cholesterol 25-hydroxylase (CH25H/Ch25h) mRNA were detected in both species. It should be also noted that, even though mRNA transcripts of CYP7A1 were not consistently detected in the human placenta, they were found in isolated primary cytotrophoblast and syncytiotrophoblast cells. The apparent discrepancy between placental tissues and cell-based data could have resulted from a dilution effect occurring in placental tissue, which contains a myriad of cell types. The findings reported by (Ontsouka et al., 2023) are in line with studies where another extrahepatic organ —the brain —was found to exhibit BA biogenetic capabilities (Pan et al., 2017; Monteiro-Cardoso et al., 2021). To date, the majority of studies focused on investigating the role of the placenta exclusively as an interface across which BAs are transported between the mother to the fetus (Sewell et al., 1980; Itoh et al., 1982; Colombo et al., 1985; Meng et al., 1997; St-Pierre et al., 2000; Tribe et al., 2010; Geenes et al., 2014; Liu et al., 2021; Ontsouka et al., 2021). Available data at https://www.proteinatlas.org, supplemented with mRNA data from (Ontsouka et al., 2023), are compiled in Table 1 and provide indications regarding the potential capacity of the human placenta to synthesize BAs. For most of the enzymes involved in BA biosynthesis, an overlap exists between their expression profiles in the human liver, brain, and placenta (Table 1). However, it should be considered that the BA enzymatic machinery has been described so far only at the mRNA level (Ontsouka et al., 2023) and confirmation at the protein level is still lacking. At this stage, it is important to decipher whether primary and secondary BAs that cross the placenta are only of hepatic/intestinal origin or originate in situ in the placenta because BAs locally synthesized in the placenta may directly affect placental performance and fetal development (Sepúlveda et al., 1991; Egan et al., 2012; De Aguiar Vallim et al., 2013; Lofthouse et al., 2019; Li et al., 2020) (Figure 3). Considering both the recent molecular evidence implying potential placental BA biogenesis (Ontsouka et al., 2023), and the fact that serum BA levels are higher in pregnant compared to non-pregnant women and mice (Milona et al., 2010), the previous belief that solely the enterohepatic axis determines the serum concentration of BAs (Hofmann, 2009) should be re-evaluated. We propose that serum concentrations of BAs during healthy pregnancies (including AHP) are probably maintained thanks to the contribution of all three compartments (mother-placenta-fetus) of the entero-hepato-placental axis (Figure 3). Although the fetal liver and the placenta may contribute to the circulating BA levels within the entero-hepato-placental axis in addition to the maternal organs (Figure 3), their respective relative contributions must still be elucidated.