Placental serotonin signaling, pregnancy outcomes, and regulation of fetal brain development^†

Abstract

The placenta is a transient organ but essential for the survival of all mammalian species by allowing for the exchanges of gasses, nutrients, and waste between maternal and fetal placenta. In rodents and humans with a hemochorial placenta, fetal placenta cells are susceptible to pharmaceutical agents and other compounds, as they are bathed directly in maternal blood. The placenta of mice and humans produce high concentrations of serotonin (5-HT) that can induce autocrine and paracrine effects within this organ. Placental 5-HT is the primary source of this neurotransmitter for fetal brain development. Increasing number of pregnant women at risk of depression are being treated with selective serotonin-reuptake inhibitors (SSRIs) that bind to serotonin transporters (SERT), which prevents 5-HT binding and cellular internalization, allowing for accumulation of extracellular 5-HT available to bind to 5-HT(2A) receptor (5-HT(2A)R). In vitro and in vivo findings with SSRI or pharmacological blockage of the 5-HT(2A)R reveal disruptions of 5-HT signaling within the placenta can affect cell proliferation, division, and invasion. In SERT knockout mice, numerous apoptotic trophoblast cells are observed, as well as extensive pathological changes within the junctional zone. Collective data suggest a fine equilibrium in 5-HT signaling is essential for maintaining normal placental structure and function. Deficiencies in placental 5-HT may also result in neurobehavioral abnormalities. Evidence supporting 5-HT production and signaling within the placenta will be reviewed. We will consider whether placental hyposerotonemia or hyperserotonemia results in similar pathophysiological changes in the placenta and other organs. Lastly, open ended questions and future directions will be explored.

Introduction

The conceptus is most vulnerable to developmental perturbation during its early stages when the placenta is becoming established and the orchestrated series of events that create functional fetal organ systems are being launched [1]. Besides serving as a conduit for solutes, the placenta possesses a tightly regulated endocrine system that is, of itself, vulnerable to chemicals that interfere with the production of or signaling of such hormones or other factors. The hemochorial placenta constantly encounters potentially toxic and teratogenic agents [2–5] and is thought to play a crucial role in buffering the fetus from the effects of these compounds. As such, it possesses enzymes, binding proteins, and transporters that are capable of confronting such threats but can also be targets in of themselves.

In all mammals, the initial hallmark event of placental development is segregation of trophectoderm from future epiblast and hypoblast cells at blastocyst formation [6, 7]. Thereafter, the processes somewhat diverge. Both the human and rodent placentae [7, 8] are invasive, but in the latter, maternal blood vessels penetrate the fetal labyrinth region around 9.5 days post-coitus (dpc), and the labyrinth layer only becomes fully operational by 12.5 dpc of a ~ 20 day pregnancy, whereupon the fetal trophoblasts (TB) are in direct contact with maternal blood [9]. In the human, the villous placenta begins to form proportionately much earlier [7]. Additionally, while both human and mouse placentae are hemochorial, in mice there are three TB cell layers surrounding the maternal blood, uninucleated cytoTBs (CTB), synctioTB (STB) layer I (STB1), and STB layer II (STBII); whereas, in human, there is only a single STB layer. In the human, TB projections, called villi, covered by a single STB layer are directly surrounded by maternal blood. Villous CTB continue to divide and provide a progenitor population for the overlying STB, [10, 11]. Another population of cells, known as extravillous TB (EVTB), which are positive for human leukocyte antigen G (HLA-G) and originate at the tips of anchoring villi, invade deeply into the uterine wall, thereby enlarging the feto-placental bed. A sub-population of EVTB occupies maternal blood vessels (endovascular invasion) and, modifies the maternal spiral arteries (vascular remodeling). Improper perfusion of the placenta by maternal blood or the presence of a small placenta may compromise the health of the fetus [12], and the mother as well. Hence, any compound that alters the developmental trajectory of the placenta is likely to be a threat to maternal and fetal health.

An integral role for placental serotonin (5-HT) in maintaining normal pregnancies but also potential involvement in gestational diseases has been postulated since the 1960s [13–17]. Several studies support the notion that the fetal placenta in mice and humans produces 5-HT [18–22]. Within the fetal placenta, 5-HT can induce autocrine and paracrine effects through binding and activation of membrane-bound 5-HT(2A) receptor (5-HT(2A)R) [23]. Deficiencies in placental 5-HT have been linked with autism spectrum disorders (ASD) [24, 25], fetal growth restriction (FGR) [26, 27], and anxiogenic behaviors [28, 29]. Selective serotonin-reuptake inhibitors (SSRIs), which bind to serotonin transporters (SERT) and thereby block cellular reabsorption of 5-HT are commonly prescribed antidepressants for pregnant women. By binding to trophoblastic SERT, such pharmaceutical agents can disrupt normal placental function, as shown with in vitro placental cell line studies [30–32]. Animal model studies also support the notion that a fine balance of 5-HT signaling within the placenta is necessary to preserve the normal architecture of this organ and associated functions [33, 34]. In this review, we will examine the synthesis and metabolism of 5-HT, evidence to date that this neurotransmitter is produced by the mouse and human placenta, effects of pharmacological inhibition or genetic suppression of SERT or disruptions in other components of 5-HT signaling. Limitations of the current studies and future directions will also be considered.

Synthesis and metabolism of serotonin and placental 5-HT production and signaling pathway

As shown in Figure 1, L-tryptophan is needed for 5-HT synthesis. Tryptophan hydroxylase (TPH1/2) hydrolyzes this compound to 5-hydroxytryptophan (5-HTP). In turn, 5-HTP is converted by amino acid decarboxylase (AAAD) to 5-HT. Monoamine oxidase A (MAO-A) then catabolizes and presumably inactivates 5-HT to 5-hydroxyindoleacetic acid (5-HIAA).

The majority of studies reveal that the fetal placenta directly synthesizes 5-HT [18–22]. Comparison of 7.5–9.5 dpc mouse fetal placenta shows that the genetic machinery to produce 5-HT is upregulated during this window of gestation [18]. To induce autocrine and paracrine effects within TB cells, 5-HT binds to 5-HT(2A)R within the cellular membrane (Figure 2A). TB cells express both 5HT(2A)R and 5-HT(1A)R [22, 35]. Binding of 5-HT to SERT (SLC6A) terminates its actions. Villous CTB and STB in normal human placenta express both mRNA and protein for SLC6A (SERT) and HTR2A (5-HT(2A)R) [23]. On the plasma membrane of TB, SERT regulates extracellular 5-HT concentrations and may thus prevent 5-HT inducing vasoconstriction of placental blood vessels [34]. Synthesis of placental 5-HT and expression of SERT/5HT(2A)R can be suppressed by maternal gestational diabetes mellitus and obesity [23, 36]. MAO-A is expressed and active within the fetal placenta and can catabolize 5-HT [37–39].

Effects of SSRI and ketanserin, serotonin receptor antagonist, on placental cells and whole placenta

Approximately, 8–10% of pregnant women are currently prescribed SSRI to combat depression [40, 41]. Examples of such drugs include fluoxetine, norfluoxetine, citalopram, sertraline, and venlafaxine. Such drugs act by binding to SERT within the plasma membrane, which in the central nervous system prevents the presynaptic cells from taking up this neurotransmitter. Blockage of this SERT in this manner results in increased concentrations of 5-HT remaining in the synaptic space and available to stimulate cognate receptors on the post-synaptic cells.

However, such drugs may also bind SERT within placental TB and thereby result in an increase in extracellular 5-HT concentrations (hyperserotonemia) and resultant signaling in the placenta, as shown in Figure 2B. In vitro studies have thus begun to examine the effects of SSRI on various placental cell lines [30–32]. JEG-3 and HIPEC TB cells lines treated with fluoxetine or sertaline (10 μM) demonstrate reduced cell proliferation, and treatment of these cells with these two and other SSRI altered gene expression patterns for MMP9, TIMP1, and ADAM10 with the directionality depending upon the specific SSRI and dose [32]. Sertraline and venlafaxine increased fusion of human primary TB cells [30]. In BeWo CTB cells, norfluoxetine (up to 10 μm tested) increased chorionic gonadotropin beta and gap junction protein alpha expression, both of these proteins are considered biomarkers of syncytialization [30]. Norfluoxetine treatment (3 μm) of individual H295R cells (fetal adrenocortical cells) caused an increase in CYP19 (aromatase) activity, but this same treatment reduced CYP19 activity in BeWo cells [31]. This SSRI (at 1 and 3 μm) reduced E2 in co-cultures of these above cells. In contrast, fluoxetine (1 μm) induced CYP19 activity in both H295R and BeWo cells but did not alter E2 levels in monocultures of these cells. Co-treatment of BeWo cells with fluoxetine (3 μm) and 5-HT(2A)R antagonists, ketanserin or ritanserin, resulted in equivalent CYP19 activity as control cell lines. Both fluoxetine and norfluoxetine (1 μm) predictably suppressed SERT activity in BeWo cells [31].

Treatment of human placental vascular smooth muscle with 5-HT induced vasoconstriction, but co-treatment with ketanserin abolished this effect [42]. 5-HT vasoconstriction with non-competitive inhibition of the contractions by ketanserin has also been detected in capacitance blood vessels in the human placenta [43]. The 5-HT(2A)R agonist, (+/−)-2,5-dimethoxy-4-iodoamphetamine hydrochloride (DOI) increased placental cell viability for BeWo and JEG-3 cells, but this effect was reversed by ketanserin. DOI stimulation of 5HT(2A)R enhanced cellular progression to G2/M and S phase of DNA synthesis for BeWo and JEG-3 cells, respectively [44]. In both cell lines treated with DOI, MEK-ERK1/2 and JAK2-STAT3 signaling pathways were activated. Taken together, findings from this study suggest that stimulation of 5HT(2A)R in placental cells maintains cell survival, division, differentiation, and stimulates key signaling pathways [44].

To test the in vivo effects of an SSRI, timed mated pregnant Sprague Dawley rats were orally treated daily with venlafaxine hydrochloride (0, 3, 10, 30, or 100 mg/kg/day) spanning from gestation day 8 to 20. On gestation day 21, dams and conceptuses were euthanized [33]. While these treatments did not alter the number of live pups, fetal weight, cause outward fetal disfigurement, or induce morbidity in the dams, endophentoypes were evident in the heart and placenta. In rats, in utero exposure to venlafaxine decreased fetal placental weight and increased the incidence of cardiac congenital abnormalities [33]. Slc6a/SERT mRNA and protein expression, respectively, were reduced in the placenta of venlafaxine treated dams; whereas, they were increased in the heart of female fetuses but not males. Fetal hearts from treated dams exhibited increased expression of 5-HT(2B)R (Htr2b) an fibroblast growth factor 8 (Fgf8). This pharmaceutical agent could induce direct effects in the placenta and heart through binding of SERT, but some of the cardiac pathologies may also be secondary to placental disturbances.

To test the placental effects induced by stimulation of 5-HT signaling, pregnant spontaneously hypertensive rats (SHR) and normotensive Wistar Kyoto rats (WKY) were treated with ketanserin (12, 24, 48 mg/kg), which reduced systolic blood pressure and placental blood flow [45]. Fetal weight showed a dose dependent decrease in WKY but not SHR individuals. This treatment reduced placenta weight in both, which is likely a direct effect of the decreased blood flow to this organ as 5-HT presumably induces vasoconstriction to increase overall blood pumped and remaining within the placenta.

The above studies provide strong support that disruptions in placental 5-HT signaling can lead to deleterious effects on TB cells and the placenta as a whole. Being vital for conceptus survival, pathologies within the placenta may lead to downstream affects in other organs, including the heart and brain. Transgenic mouse models lacking one of the components needed for 5-HT signaling have also been employed to elucidate the role of this neurotransmitter in the placenta.

Placental abnormalities in mice with disrupted 5-HT signaling

To understand how 5-HT signaling deficits affect the placenta, another group examined placental morphology in mice either unable to synthesize 5-HT due to deficiency in TPH1 (TPH1 knockout-KO) mice or those lacking SERT needed to bind and terminate the actions of 5-HT, i.e., SERT knockout (SERT-KO) mice [34]. Placenta from SERT-KO and TPH1-KO mice had similar reductions in placental weight relative to wild-type (WT) counterparts. Apoptosis, as measured by TUNEL staining, was increased in both groups with 49- and 8-fold greater number of TUNEL positive cells in SERT-KO and TPH1-KO mice, respectively. The junctional zone (spongiotrophoblast region) in SERT-KO had considerable areas of necrosis, hemorrhage, and fibrosis (Figure 3). Results thus suggest a fine equilibrium of 5-HT signaling is needed to maintain this fetal placental area. Proliferation index, as determined by MKI67 staining, was not different in any of the KO groups relative to WT placenta [34]. The collective findings from these two transgenic mouse models suggest that hyposerotonemia, due to deficiencies in TPH1, or hyperserotonemia, subsequent to SERT ablation, may disrupt normal placental architecture with the latter condition leading to the most pronounced changes.

Effects of placental 5-HT disruption on fetal brain development and other disorders

Deficiencies in placental 5-HT have been linked with ASD [24, 25], FGR [26, 27], and anxiogenic behaviors [28, 29]. These ensuing neural disorders may be due to the fact that the placenta is the sole source of 5-HT during early brain development [19, 20]. 5-HT is required during fetal brain development to stimulate cell division, neuronal migration, cell differentiation, and synaptogenesis [25]. Hyperserotonemia during fetal development may trigger a negative feedback loop that ultimately suppresses the 5-HT terminal. Increased levels of 5-HT may also reduce oxytocin concentrations in the paraventricular nucleus of the hypothalamus, but elevate calcitonin gene related peptide in the central nucleus of the amygdala. Both of these factors regulate social behaviors and consequently, are implicated in ASD. Hyposerotonemia may impact sensory, motor, and cognitive abilities, and thus, reductions in placenta-derived 5-HT may also contribute to ASD etiology [25].

Preeclampsia (PE) is typified by gestational hypertension and proteinuria and generally diagnosed after 20 weeks of gestation in women who were previously symptom-free [46]. The genesis of this disease remains uncertain as it is like influenced by both genetic and extrinsic factors [47, 48]. Once the placenta is delivered or removed, full reversal of the disease symptoms results, suggesting the fetal placenta is the initiating organ. Early onset PE has been attributed to improper remodeling of the uterine spiral arteries by the invasive EVTB (detailed in the Introduction) [49]. Given the potential importance of 5-HT in the placenta, there has been interest in determining whether this placental signaling pathway is distorted and may contribute to PE development. One study tested the contractile responses induced by 5-HT in branches of chorionic artery and veins in normotensive and PE placenta [50]. The contraction in PE blood vessels was significantly reduced. To verify these responses involved 5-HT binding to 5HT(2A)R, ketanserin was concurrently tested. This antagonist suppressed blood vessel contraction in normal blood vessels and to a lesser extent in those from PE placenta. HTR2A and 5HT(2A)R expression was similar in both placental groups. Thus, in PE placenta, the vascular responses to 5-HT are blunted in the absence of a change in its cognate receptor. An earlier study found that norepinephrine transporter and monoamine transporter (MET) were downregulated in PE placenta but SLC6A4 levels were not altered [51]. These other METs may confer protective mechanisms against vasoconstriction of placental blood vessels, which facilitates a stable blood flow to the fetus. Previous work though suggests that placenta and umbilical cord blood vessels from PE pregnancies shows reduced vasoconstrictive properties in response to 5-HT [52].

Placenta from normotensive pregnancies exhibits greater metabolism of 5-HT than PE placental homogenates, suggestive of disruptions in MAO-A [53]. Consistent with this finding, 5-HT levels are substantially elevated in PE placenta, which also show reduced immunohistochemical staining for MAO-A within TB cells [38]. According to one report, MAOA, mRNA levels and activity were reduced in PE placenta [39], but another study suggested that only the activity was suppressed [54]. SLC6A4 levels were also not changed in the PE placenta tested in the former study [39].

Conclusions

The current studies provide conclusive evidence that the fetal placenta in the mouse and human produces 5-HT [18–22]. To induce autocrine and paracrine effects within the placenta, 5-HT binds to 5-HT(2A)R and possibly 5-HT(1A)R within the plasma membrane of TB cells [22, 35]. However, the effects of 5-HT cease upon binding to SERT and subsequent intracellular transport [23]. On the plasma membrane of TB, SERT regulates extracellular 5-HT and may thus prevent 5-HT inducing vasoconstriction of placental blood vessels [34]. MAO-A is present and active in high amounts to catabolize 5-HT to its inactive form, 5-HIAA [37–39]. In placenta from PE pregnancies, the enzyme may show both reduced expression and activity [38, 39, 54] with hyperserotonemia resulting in the placenta and possibly extending to the developing fetus. Both hypo- and hyperserotonemia have been proposed as potential risk factors of ASD [24, 25]. In several cohort studies, PE has also been linked to the development of ASD [55–60]. Whether, this linkage is due to disruptions in placental 5-HT or other placental/maternal factors remains uncertain.

While SSRI have been helpful in treating depression, such drugs may also tamper with 5-HT signaling within the fetal placenta. The collective findings testing various placental lines and one published study with pregnant rats treated with an SSRI suggest that this class of drugs may alter various placental genes, including those essential for synctialization, CYP19 activity in the case of fluoxetine, suppress TB cell proliferation, and in the whole animal model, cause reductions in placental weight [30–33]. In contrast, activation of 5-HT(2A)R agonist by DOI increases placental cell viability, proliferation, differentiation, and stimulates other cellular signaling pathways [44]. SHR and normotensive WKY treated with the 5-HT(2A)R antagonist show reduced placental weights. Thus, overstimulation of 5-HT(2A)R that may result from SSRI binding to SERT and accumulation of 5-HT in the extracellular space and antagonistic inhibition of this receptor both result in reductions in placental weight. It is not clear if such effects are due to reduced 5-HT vasoconstriction and blood flow into fetal placental vasculature and/or suppression of TB proliferation.

Apoptosis is increased in placenta of TPH1-KO and SERT-KO mice, although effects are more striking in the latter mouse model [34]. The middle fetal placental layer, junctional zone (or spongiotrophoblast layer) exhibits extensive necrosis, fibrosis, and hemorrhage in SERT-KO mice. The findings from these two transgenic mice further implicates both hyposerotonemia and hyperserotonemia in destroying the normal placental morphology, although elevations in 5-HT may lead to more pronounced pathologies.

To understand further the role of 5-HT signaling in the placenta and abnormalities that result from hyposerotonemia and hyperserotonemia, the pathological changes throughout gestation should be tracked in rodents where 5-HT signaling is pharmacologically or genetically disrupted, whether this is through reduced synthesis, SERT binding/deficiency or agonism/antagonism of 5-HT(2A)R. While past in vitro studies have examined candidate genes in TB cells treated with SSRI, global transcriptome profiles should be analyzed in placenta from these above conditions. It would be of interest to determine whether similar signature patterns of placental genes result from under- or over-stimulation of this pathway. Single cell RNA-seq approach will permit elucidation of how individual TB cell responds to distortions in 5-HT signaling. However, the soon to be released Visium Spatial Transcriptomics (ST) technology [61, 62] that allows for mRNA quantification in the spatial context of intact tissue may be preferred over single cell RNA-seq. The ST data will also assumingly permit inferences to be drawn as to how gene expression changes in different TB cell lineages relate to each other.

In sum, the fetal placenta is a major source for 5-HT where it can induce local effects. Placental derived 5-HT may also be the sole source of this neurotransmitter for the developing brain. The placenta and fetal brain are thus vulnerable to agents that suppress or overstimulate 5-HT production and signaling. Too much or too little of placental derived 5-HT may yield similar placental and neurobehavioral disorders, such as ASD. Recent developments in high throughput RNA-seq methods, in particular ST analysis, will likely extend our understanding of the role of placental 5-HT in health and disease.

Conflict of interest

The author does not have personal, professional, or financial conflicts.

References

Author notes