Metabolomics in amniotic fluid
Alexandra-Maria Michaelidou, Foteini Tsakoumaki, Maria Fotiou, Charikleia Kyrkou, and Apostolos P. Athanasiadis
Introduction
Amniotic fluid (AF) is a complex and dynamic milieu that surrounds the fetus (1,2). The biochemical composition of AF changes with the progression of pregnancy and the maturation of the fetus. In the beginning of pregnancy, AF originates from maternal plasma, which is filtered through the fetal membranes based on hydrostatic and osmotic forces. When the placental and fetal vessels start to develop, water and solutes from maternal plasma pass through the placenta to the fetal plasma and then to the AF. Hence, during the first trimester, AF composition resembles that of fetal plasma (2,3). At the fourth to fifth weeks of pregnancy, fetal kidneys start to evolve (4), and by the eighth to eleventh weeks, they begin to excrete urine into the AF (2,4,5). Furthermore, the fetus swallows AF continuously and in increasing amounts from the end of the first trimester (2,6); however, neither fetal urination nor swallowing contribute significantly to the content of AF, until the 19th–20th weeks when keratinization of fetal skin begins; at about the 25th week when the fetal skin is fully keratinized, AF is largely a product of fetal urine and lung fluid (3). It is worth mentioning that in late pregnancy, fetal swallowing prepares the gastrointestinal tract for postnatal nutrition (7), since similar nutrients and growth factors are present both in AF and human milk (7,8).
AF provides a rational compartment for studies on fetal metabolism, since its composition reflects both maternal health and fetal status. This biofluid can be examined at different time points in pregnancy to provide clinicians and pregnant women with important information for decision-making about pregnancy management (9). In fact, AF has been routinely collected through the second trimester through amniocentesis, mainly in cases of high-risk pregnancies (Annex), such as those characterized by advanced maternal age, ultrasound detection of fetal malformations (FM), or prior history of genetic abnormalities (10).
The application of metabolomics (11,12) on AF provides a snapshot of its metabolic fingerprint, and may help toward the elucidation of the mechanisms that influence intrauterine environment (13). Indeed, an embryo during its development can follow “paths” slightly different depending on the signals and messages it receives from its environment, such as nutritional stimuli (14). These changes, either heritable or not, affect how cells turn genes on or off (15). A number of recent studies suggest that changes in the epigenetic regulation of genes in embryos are central to the induction of phenotypes that persist into adulthood (14,16). Metabolic fingerprinting may serve as an excellent probe for these phenotypes, providing metabolic signatures in certain pathological and/or under physiological conditions (17).
Thus, this chapter starts by addressing the research carried out so far regarding the application of metabolomics in AF in normal pregnancy. This is followed by a summary of studies exploring the effect of maternal diet on the AF metabolome, given the general consensus that maternal dietary quality or quantity may be associated with significant shifts in the fetal environment (14). A further aim of this chapter is to present an account of the potential of metabolomics in prenatal research to implement prevention, early diagnosis, and monitoring of pregnancy complications.
Metabolomics in amniotic fluid in normal pregnancy
Amniotic fluid composition throughout pregnancy
Reliable information about gestational age (GA) is necessary for clinicians to assess fetal growth accurately and to interpret certain screening and diagnostic tests; it is also critical for the management of pre- and post-term pregnancies (18,19). Toward this direction, it was considered mandatory to track metabolic changes occurring in AF of healthy mothers over the course of a normal pregnancy (20,21).
The published research projects aiming to characterize the metabolic signature of AF in normal pregnancies are listed in Table 23.1. In an attempt to establish normative metabolite concentrations at the second and third trimesters, Cohn et al. (20) reported that betaine, succinate, creatinine, and glutamine increased significantly with advancing gestation, while numerous amino acid levels decreased; a finding that concurs with the earlier work of Bock (22), who reported a drop after the second trimester, followed by a plateau during the third trimester, for glycine, alanine, glutamate, valine, isoleucine, and the aromatic amino acids. Consistent with these findings are also the results of Pawilowicz et al. (21), who attribute the decreasing trend in the levels of amino acids to fetal growth dynamics and therefore to the increased demand for elementary building blocks.
Table 23.1 Amniotic fluid metabolic fingerprint in normal pregnancy | ||||
Reference | Population | Method | Aim of the study | Main metabolite changes |
22 | 10 samples at early second trimester—43 samples at third trimester | 1H-NMR | To compare metabolite concentrations between second and third trimester | ↑ Choline, creatinine, lactatea ↓ Succinatea, alaninea, glutamate, glycine, histidine, isoleucine, phenylalanine, tyrosine, valine |
24 | Healthy pregnancies | MS,1H-NMR, RP-LC | Metabolic profiling of AF | 60 metabolites |
20 | 23 samples at second trimester—27 samples at third trimesterb | HR-MAS | To compare metabolite concentrations between second and third trimester | ↑ Betaine, creatinine, succinate, glutamine ↓ Citrate, creatine, glucose, GPC, lactate, pyruvate, alanine, glutamate, isoleucine, leucine, lysine, valine |
18 | 24 samples at second trimester—71 samples at third trimester | HR-MAS | Development of a reliable model for the calculation of GA with the use of AF metabolite profiles | ↑ Creatinine ↓ Glucose, alanine, valine |
21 | 6 samples at second trimester—21 samples at third trimester | 1H-NMR | Track metabolic changes over the course of pregnancy | ↑ Choline, creatinine, N,N-dimethylglycine, pyruvate, succinate, urocanate, AU2 ↓ Glucose, carnitine, lactate/pyruvate, alanine, isoleucine, leucine, methionine, phenylalanine, tyrosine, valine |
Source: Ref. 20. Abbreviations: GPC, glycerophosphocholine; AU2, unassigned metabolite. a Expressed as the ratios of their peak heights to citrate peak height. b While many of the patients were noted to be diabetic, each patient was undergoing medical management for their condition and was euglycemic at the time of amniocentesis. |
The decreasing profile for AF glucose overlaps with the increasing energy demands during pregnancy progression (20,21). Furthermore, AF glucose levels, throughout gestation, reflect fetal kidney maturity, since, with the progressive maturation of the fetal kidney tubules, more of the glomerular filtrated glucose is reabsorbed and, consequently, less appears in the AF (23). Within this context, the detection and quantification of creatinine may also provide an additional biomarker for assessing fetal renal status (20).
Surprisingly, pyruvate level in AF exhibited a reverse trend against glucose. This observation in conjunction with the decrease in lactate/pyruvate ratio might, however, be attributed to the elevated oxygen consumption due to the increase in fetus aerobic metabolism (21).
As far as choline is concerned, its increased levels may be correlated with higher phospholipid demand for fetal brain development during later periods of pregnancy (21).
Taken together, all these data point to the fact that the complex metabolic information contained within AF may ensue from the by-products of fetal pulmonary, renal, and gastrointestinal development and may therefore exhibit predictability of fetal organ maturity.
As an immediate consequence of the previously discussed data, the process of mapping the compositional components of human AF onto the biological needs of the fetus is a great scientific challenge. Metabolomics may facilitate exploration of the effect of maternal diet on AF metabolome and, undoubtedly, help toward unraveling the biochemical and physiological mechanisms that could affect the trajectory of fetal growth in pregnancy complications.
Exploring effect of maternal diet on amniotic fluid metabolome
As already mentioned, there is an increasing consciousness of the pivotal role of metabolic pathways in biological processes related to health. According to the literature (25), the health of the individual and the population in general is the result of interactions between genetics and a number of environmental factors. Nutrition is confronted as an environmental factor of major importance. In this frame, maternal over- or undernutrition have been shown to significantly affect the developing fetus with important consequences later in life (13,26). However, emerging data (27,28) indicate that not only these extreme conditions but also small alterations in maternal dietary quality or quantity may be associated with significant shifts in the fetal environment.
Within this frame, conventional analytical research protocols were adopted in human (29–32) and animal studies (33–36) to explore whether maternal diet during pregnancy can alter the composition of AF (Figure 23.1). Despite the evidence that emerged from these studies, an in-depth examination of the repertoire of small molecules present in this biofluid would offer a more comprehensive description and a good perspective in the effort to explore the effect of maternal nutrition on AF composition.
Figure 23.1 Timeline of studies investigating the effect of maternal diet on the metabolic signature of amniotic fluid.
The use of “omics” techniques to investigate the effect of maternal nutrition on AF milieu is very recent (Figure 23.1). In 2008, in an effort to explore any potential association between maternal nutrient restriction during pregnancy and fetal neurodevelopment in a rat model, Shen et al. (37) employed a metabonomic (gas chromatography [GC]-mass spectrometry [MS]) and metallomic (inductively coupled plasma/MS) analytical platform. The results revealed that the AF metabolic and elemental signature was indeed influenced by maternal dietary deprivation, since variations in small-molecule metabolites as well as in trace elements related to neural induction, neuronal migration, neural tube differentiation, and synaptogenesis were recorded.
Almost 10 years later, Wan et al. (38) (Figure 23.1) adopted a metabolomic approach through 1H nuclear magnetic resonance (NMR) spectroscopy combined with biochemistry analysis to explore the effect of maternal chitosan oligosaccharide (COS) supplementation on the AF metabolic profile. Wan et al. (38) found that COS supplementation can alter AF antioxidant status by enhancing superoxide dismutase, catalase, anti-superoxide anion, anti-hydroxyl radical activities, and total antioxidant capacity, as well as AF immune status, by increasing interleukin (IL)-10, immunoglobulin G (IgG) and IgM concentrations. Apart from improving antioxidant and immune parameters, maternal COS supplementation increased AF lysine, citrate, and glucose levels, indicating that this supplementation can modulate amino acid and glucose metabolism, as well as the citric acid cycle.
In 2013, in an effort to explore the effect of maternal pre-pregnancy body mass index (BMI) on the human AF metabolic fingerprint, Athanasiadis et al. (39) employed 1H-NMR metabolomics (n = 44). The authors observed that the metabolic signature of normal pregnant women appeared to be distinct from that of overweight women, suggesting that there is a compositional trend in AF metabolic profile influenced by maternal BMI. One year later, Fotakis et al. (40) carried out a pilot study involving 27 participants in order to investigate whether maternal habitual dietary patterns influenced the composition of human AF. Interestingly, the application of 1H-NMR allowed the identification of metabolites associated with the different dietary patterns (40). These two preliminary studies, although conducted in independent samples of pregnant women, pointed out the scalability of NMR metabolomics to elucidate the role of maternal nutrition on fetal growth and development.
Capitalizing on these promising findings, in 2018, Fotiou et al. (41), using a validated food-frequency questionnaire (42) and implementing NMR-based metabolomics, explored the impact of maternal habitual diet on the human AF metabolic profile. In this study, 65 women were divided into two groups according to their dietary habits. In order to eliminate the potential overlapping with metabolic effects attributable to specific conditions or fetal/maternal disturbances and capture only the maternal diet-induced metabolic signatures in AF, the following eligibility criteria were applied by the authors: (1) singleton pregnancy, (2) absence of fetal structural malformations and/or chromosomal abnormalities, (3) delivery of an appropriate for GA infant (birth weight between the 10th and 90th percentiles), and (4) absence of obstetrical/medical complications, such as preeclampsia (PE) or gestational diabetes mellitus (GDM). In this context, Fotiou et al. (41) reported that metabolic signatures of maternal habitual dietary patterns were identified in AF, since women characterized by higher energy contributions from total and animal protein, saturated fatty acids, and higher dietary glycemic index, had higher AF glucose, alanine, tyrosine, valine, citrate, cis-acotinate, and formate levels compared to the group characterized by higher consumption of plant protein and monounsaturated and polyunsaturated fatty acids. Thus, AF metabolic modifications induced by the different maternal dietary habits were linked to amino acid and glucose metabolism and citric acid cycle (Figure 23.2).
Figure 23.2 Metabolic pathways that are possibly influenced by maternal habitual diet: (a) energy metabolism, amino acids metabolism, and urea cycle; (b) fumarate generation during purine biosynthesis; and (c) choline metabolism. (Source: Ref. 41, with permission.)
Deciphering the biological effects of such nutritional stimuli is not only challenging but also appealing, since dietary balance during pregnancy may be related to fetal adiposity and fat distribution (43).
Metabolic fingerprint of amniotic fluid in pregnancy complications
As already mentioned, a further aim of this chapter is to present an account of the potential of metabolomics in prenatal research to implement prevention, early diagnosis, and monitoring of pregnancy complications. Relevant data on FM, GDM, PE, and preterm birth (PTB) are discussed in the following sections.
Fetal malformation
The metabolomics approaches employed for the metabolic characterization of FM are presented in Table 23.2.
Table 23.2 Metabolic signature of amniotic fluid to diagnose fetal malformations | ||||
Reference | Population | Method | Main metabolite changes | |
n | Gestational age | |||
22 | 10 control samples—5 cases of spina bifida | Second trimester | 1H-NMR | Acetate, glutamate, lactatea |
46 | 17 control samples—10 cases of spina bifida | 15–39 (weeks of amenorrhea) | 1H-NMR | ↑ Succinate, glutamine ↓ Creatine, creatinine |
44 | 34 control samples—12 malformed cases | 15–24 weeks | 1H-NMR | ↑ α-fetoprotein, citrate, glycine, methionine, glycoprotein P1, succinate ↓ Ammonia, ethanol, glucose, urea, alanine, glutamine/glutamate, leucine, phenylalanine, proline, tyrosine, valine |
10 | 82 control samples—27 malformed cases | 14–25 weeks | 1H-NMR | ↑ Ascorbate, creatinine, glycoproteins, lactate, succinate, glutamine, glycine, methionine, serine, threonine ↓ α-oxoisovalerate, glucose, pyruvate, alanine, isoleucine, leucine, phenylalanine, tyrosine, valine, four unknown metabolites |
45 | 26 control samples—22 malformed cases | 15–25 weeks | UPLC-MS | ↑ Carnitine, citrate, estriol-3-sulfate-16-glucuronide, estriol-3-glucuronide/estriol-16-glucuronide, pyroglutamate ↓ glucose, polyols, glutamate/glutamine, leucine/isoleucine, lysine, tyrosine, valine |
Note: Potential changes in metabolic pathways: enhanced glycolysis (possibly under fetal hypoxia), reduced use of the respiratory chain pathway, enhanced gluconeogenesis, fetal kidney underdevelopment, disturbance in folic acid pool regulation, higher demand for protein synthesis, derangement of amino acid metabolism. a Expressed as the ratios of their peak heights to citrate peak height. |