Antonio Ragusa, Alessandro Svelato, and Sara D’Avino
Metabolomics in normal pregnancies
The definition of metabolomics is still in flux, and hence, different researchers provide slightly different definitions. Metabolomics can generally be defined as the study of global metabolite profiles in a system (cell, tissue, or organism) under a given set of conditions (1).
Metabolomics studies have been increasing the knowledge and understanding of metabolic adaptaions of the human organism to pregnancy and diseases.
In a normal pregnancy, metabolomics help unveil metabolic details and define normal metabolic trajectories of pregnancy. Pinto et al. have been able to highlight specific lipoprotein/protein metabolic aspects of pregnancy with impact on the excreted metabolome and, therefore, provide an interesting direction for further understanding pregnancy metabolism (2). They collected blood plasma samples for healthy nonpregnant and pregnant women in their first, second, and third trimesters (T) of pregnancy. In this way they were able to demonstrate that some plasma macromolecules (high-density lipoprotein [HDL] and low-density lipoprotein [LDL] + very-low-density lipoprotein [VLDL] levels) increased throughout pregnancy at different rates. The study suggests an inverse relationship between lipoproteins (HDL and LDL + VLDL) and albumin, and correlation of plasma albumin with particular circulating and excreted metabolites. Metabolomics profiling can also be used to better understand the physiological changes of the complex interdependencies of the mother, the placenta, and the fetus during pregnancy, Orczyk-Pawilowicz et al. were able to describe metabolic changes occurring in amniotic fluid and plasma of healthy mothers over the course of pregnancy; they demonstrated that in amniotic fluid, the transition from the second to the third trimester was associated with decreasing levels of glucose, carnitine, and amino acids, and increasing levels of creatinine, succinate, pyruvate, and choline (3). In dairy cows, the combination of metabolomics and new network methods proved to be able to generate rich biochemical insight into possible biological modules related to early pregnancy (4). Thanks to the use and integration of these new techniques, in the next few years, metabolomics will also enable us to better clarify the normal metabolic changes that occur during human gestation and consequently to improve our capacity to predict, diagnose, and monitor diseases and illnesses.
Metabolomics and preterm birth
Preterm delivery (PP) affects about 9% of pregnancies every year. Etiology is multifactorial and involves maternal genomics, fetal genomics, and environmental factors (5). The pathophysiology of PP recognizes four mutually nonexclusive paths (6,7): infection and inflammation, decidual hemorrhage, uterine overdistension, and activation of the hypothalamus/maternal or fetal hypophysis axis (stress) (Figure 22.1). Upstream of these biological pathways, multiple exogenous and endogenous factors can trigger a multitude of biomolecular pathways that remain undefined and that can ultimately result in a preterm birth (8).
Figure 22.1 Pathogenic pathway leading to preterm birth. The activation of one or more pathways leads to the activation of the chorion-decidual interface, which, in turn, can lead to cervical effacement, uterine contractions, and preterm premature rupture of membranes. The kind of damage also depends on the spatial location of the infectious/inflammatory insult.
Although preterm birth is a multifactorial process, metabolomics studies seek to identify biomarkers that will allow the development of new therapies to prevent, recognize, and treat it. In a study on samples of amniotic fluid taken at 22 and 35 weeks of gestation, Romero et al. showed that carbohydrates, amino acids, and xenobiotic compounds are excellent predictors and allow identification of patients at risk (9). Other studies that analyzed the amniotic fluid and urine of women who had preterm births reported different levels of amino acids compared to levels seen during controls (10). Studies have shown about 116 biomarkers in preterm birth; however, to date, no single biomarker has been identified that can reliably predict which woman’s pregnancy will result in a preterm birth (11). This is because there are numerous factors that limit their use as predictive biomarkers of preterm birth, the most important factor being the high intraindividual variability. In fact, the pathophysiological response of the host to a given etiological factor (intrauterine infection, decidual hemorrhage, fetal stress, etc.) can vary according to genotype, epigenetic mechanisms, and exogenous environmental exposure. Recently, a great deal of research has been carried out on the intestinal microbiome. It is known that during pregnancy, for every woman, exposure to chronic or acute stress factors influences the risk of preterm birth (12) depending not only on the woman’s genotype, but most likely, also on the composition of her microbiome, which plays an important role in the activation of the maternal and fetal immune system, influencing the stress response (13). Studies have identified metabolites and mediators in the immune response (14,15), and through metabolomics analysis, new insights have emerged regarding the inflammatory and immune responses to infection, which offers an opportunity to study intrauterine infection, inflammation, and relationships between the degradation of the extracellular matrix, the metabolism of the estrogen, stress, and fetal abnormalities, all factors that contribute to preterm births (16). Until now, research has shown numerous alterations in maternal biofluids in women whose pregnancies conclude with preterm birth; for this reason, metabolomics can be considered a strategy for research to identify how disparate stimuli are made operative by means of metabolites and metabolic pathways associated with the triggering of the parturition in some women, but not in others. As with other pathophysiological aspects, one of the main problems of metabolomics studies also for the study of preterm birth is the use of small and few samples. The use of larger samples will allow us to have a better understanding of the metabolic mechanisms involved in the origin of preterm birth (8).
Metabolomics and gestational diabetes
Pregnancy is characterized by a reduction in insulin sensitivity. Pregnancy causes insulin resistance progressively during the second and third trimesters, and this is considered secondary to the increase in maternal adipose tissue and hormonal variations typical of pregnancy. Pregnant women have a higher insulin secretion, compared to that of nonpregnant women (17). Gestational diabetes (GDM) is defined as any degree of glucose intolerance that begins or is seen for the first time during pregnancy (18). The prevalence of GDM varies globally, ranging from 1% to 14% of all pregnancies, in relation to the diagnostic criteria used and the studied population (17). The increase in the prevalence of GDM is probably related to the increase in obesity rates in women of reproductive age.
The diagnosis of GDM is characterized by the plasma response of high glucose levels, often at the end of the second quarter; however, there is currently no unanimous agreement on diagnostic criteria (19).
Women with GDM have an increased risk of developing type 2 diabetes in the years following pregnancy. The cumulative incidence of diabetes after a diagnosis of GDM is 2.6% after 2 years, 8.1% after 5 years, 17.3% after 10 years, and 25.8% after 15 years (20).
In addition, women with GDM may have an increased risk of developing high blood pressure, metabolic disorders, and cardiovascular diseases (17).
GDM also has a profound impact on the short- and long-term health of the unborn child. Maternal glucose, unlike insulin, can cross the placenta. This causes an increase in the glucose load for the fetus, with a consequent increase in the production of insulin by the fetal pancreas, which in turn promotes the growth and accumulation of adiposity in the fetus (17). There is a considerable amount of data supporting the association between maternal hyperglycemia and risks to future fetal health, such as obesity, insulin resistance, reduced glucose tolerance, type 2 diabetes, risk factors for cardiovascular disease, and autism (17).
Metabolomics has been used to analyze the metabolic profile and identification of new biomarkers associated with insulin resistance. The use of these new technologies will help to understand the etiology and pathogenesis of GDM. The metabolic profile of women with GDM compared to that of healthy controls is characterized by the presence of very different metabolites, both qualitatively and quantitatively.
Some studies have found high levels of fatty acids (17,21–23) and low levels of glycerophospholipids (21–25) in subjects with GDM compared with healthy controls. The mechanism that leads to this could be explained by the increase in insulin resistance and the functioning deficit of the β cells of the pancreas, following the intracellular accumulation of toxic lipid derivatives, oxidative stress, inflammation, and mitochondrial dysfunction (17,26). The finding of reduced circulating levels of glycerophospholipids during pregnancy (17,25,27) and postpartum (17,23,25) are in agreement with the data obtained from studies conducted on type 2 diabetes (28). This concordance suggests that the pathogenesis of GDM could be related to maternal pancreatic β-cell dysfunction (21), to alteration in enzymatic activity (e.g., lower levels of cytosolic calcium-dependent phospholipase-A2 isoform) (17), and an unbalanced pro-inflammatory versus anti-inflammatory ratio (21).
A study by Powe et al. highlights the heterogeneity in the pathogenesis of GDM (29). In fact, there are four subgroups: GDM with an insulin secretion defect, GDM with an insulin sensitivity defect, GDM with both defects, and normal glucose tolerance based on results from a fasting 75 g OGTT (oral glucose tolerance test) administered at 24–30 gestational weeks. The GDM with an insulin sensitivity defect has an increased risk of cesarean sections and greater neonatal weight, even after correcting the data for maternal body mass index (BMI). Furthermore, this group has higher levels of leptin and lower levels of adiponectin. These findings bring to light the physiological heterogeneity within GDM subtypes, a concept that is not addressed by current methods of diagnoses (30). In addition, metabolomics may increase our capacity for early identification of GDM or classify the risk of subsequent cardiovascular disease among women and their newborns.
Early identification of women at risk of developing GDM would enable early treatment, with the ability to reduce the negative impact that hyperglycemia has on the mother and unborn child. Unfortunately, studies on the metabolic profile of women in early stages of pregnancy in order to predict the risk of developing GDM have, so far, given inconsistent results. Some studies report a reduction in blood creatinine, trimethylamine-N-oxide, and betaine (17) and an increase of urinary N-methyl nicotinamide and choline urine; other studies have found an increase in glucose and a reduction of glutamate in the amniotic fluid (17). The identification of metabolic differences between women with GDM and healthy women during the third trimester of pregnancy may lead to a better understanding of the associated complications and potential influences on fetal metabolic development.
Two studies considered the metabolism of babies born to mothers with GDM in order to evaluate the effect of maternal pathology on the development of the newborn. Logan et al. (31) hypothesized that the metabolic profile of infants born to mothers with GDM was different from that of infants born to healthy mothers. The statistical analysis of the data allowed discrimination between the two groups of newborns, and the most important differences with regard to the urinary metabolites were those detected for glucose, formate, fumarate, citrate, and succinate—all metabolites involved in the tricarboxylic acid (TCA) cycle. These preliminary data indicate differences in the metabolism of infants born to diabetic mothers and control infants in the newborn period.
In the study conducted by Dani et al. (32), who studied the metabolome of infants born to mothers with GDM, the authors found a lower level of glucose and a higher level of pyruvate, histidine, alanine, valine, methionine, arginine, lysine, hypoxanthine, lipoproteins, and lipids than controls, but they did not find any clinical differences. The authors concluded that prolonged fetal exposure to hyperglycemia during pregnancy can change the neonatal metabolomic profile at birth without affecting the clinical course.
Also in the case of the GDM, as for premature birth, there are several possible explanations of the inconsistency of some results obtained so far. First, the studies in this field use different biological samples. It is not often known whether the subject is fasting at the time the sample is collected. The equipment and technologies used to analyze collected samples are often different among the various studies. The lack of unanimity on which diagnostic criteria to use for the GDM, which are consequently different among the various studies, makes everything even more complex. The age and ethnicity of the study populations are often not considered. Most studies used small sample sizes. Moreover, few studies included information on dietary intake. Statistical methods used to compare metabolic profiles and identify significant differences between groups were often highly inconsistent.
Metabolomics and fetal growth restriction
Intrauterine growth retardation (IUGR) is characterized by a fetus that does not reach its growth potential, with birth weight and BMI below normal for gestational age (33). The incidence of IUGR ranges between 4% and 8% of newborns in industrialized countries and 6% and 30% in developing countries (34). Etiology may depend on genetic abnormalities, congenital infections, and maternal and placental pathologies. The intrauterine environment is the main factor that influences fetal growth and development. Metabolomics analyses looked for biomarkers in maternal plasma, amniotic fluid, urine, cervicovaginal secretions, placenta, and umbilical cord blood, and provided information on the transfer of placental nutrients (35). Some studies evaluated an increase in different metabolites (sphingolipids, phospholipids, carnitines, and fatty acids) in maternal plasma and a decrease in umbilical cord blood in underdeveloped fetuses, demonstrating a failure of placental transfer to the fetus of these nutrients (36).
A study allowed the identification of the molecules responsible for the differences between the various metabolic profiles, including myoinositol, where urine content was higher in IUGR fetuses compared to controls (37). The increase in myoinositol concentration in plasma and urine, which has often been associated with glucose intolerance and insulin resistance in adults, could also be considered a valid marker of impaired glucose metabolism during fetal development in IUGR fetuses (38), but further studies are needed to draw definitive conclusions.
Various studies in literature show the correlation between the different metabolites, all of which are involved in the same metabolic pathway, the tricarboxylic acid (TCA) cycle. This could be explained by the fact that insulin plays an important role in the production of energy through the oxidation of acetyl-CoA and promotes the conversion of glucose into pyruvate in the TCA cycle. In IUGR there is a reduced sensitivity to insulin; therefore, insulin resistance or glucose intolerance can lead to altered intermediate metabolites of the TCA cycle (glutamine, alanine, leucine, and aspartate) (39). A good example of the close relationship between environment, genotype, and phenotype are the small for gestational age (SGA) newborns or IUGR (in this text, for reasons of space, we do not examine the difference between the two conditions, which, in fact, exists and is very important). These infants have a low birth weight because the uterine environment was very poor and did not provide much energy for different reasons. Only fetuses who could adjust to this hostile environment could survive. After birth, if these newborns grow up in an environment rich in energy, they develop cardiovascular disease in a higher percentage than newborns with normal birth weight (40).
In recent years, several biomarkers have been proposed for IUGR prediction, but none has been demonstrated as sufficiently precise to be recommended as a predictor of IUGR in clinical practice (41).
Metabolomics can help identify the modifiable risk factors that contribute to changes in the concentrations of disease-related metabolites. In particular, during prenatal life, the organism is highly sensitive to environmental stimuli, such as nutrients, chemicals, drugs, infections, and other stress factors. The in utero exposure to toxic substances can, in addition to clinical manifestations such as death, birth defects, and low birth weight, also cause functional changes in gene expression, leading to an increased risk of disease in adulthood (cardiovascular diseases, obesity, type 2 diabetes, dysfunction of the reproductive system, brain, and immune system) (35). Therefore, the metabolomic approach can have an interesting impact on prenatal care, in an attempt to prevent perinatal outcomes, but also in the understanding and management of adult chronic diseases.
Metabolomics and metabolic syndrome
Metabolic syndrome (MetS) is defined as the association of obesity, insulin resistance (IR), hypertension, and dyslipidemia (42). MetS predisposes the individual to severe metabolic diseases such as type 2 diabetes and cardiovascular disease in adulthood (43). Obesity is considered the main cause of the increase in prevalence of MetS, in association with higher plasma triglyceride (TG) levels, lower high-density lipoprotein-cholesterol (HDL-C) levels, hyperglycemia, and increased cardiovascular risk (42,43). The prevalence of obesity has been increasing dramatically worldwide over the past three decades (44). The increased prevalence of obesity is one of the main factors responsible for the increase in obstetric complications, such as preeclampsia (PE), GDM, large and small gestational age fetuses, cesarean section, and iatrogenic preterm birth (45). Obesity is the result of a complex interaction between predisposing genetic factors and changes in environmental factors such as diet and pollutants.
In order to investigate the environmental factors related to obesity, Pietilainen et al. conducted a study on monozygotic twins discordant for obesity. A change in the overall serum composition of the metabolites between obese and nonobese twins was observed (46).
It is recommended that obesity be the primary target of intervention for MetS.
Insulin resistance is currently thought to be the central part of the pathogenesis of MetS (Figure 22.2). Abdominal obesity is considered the initial factor of insulin resistance. Obesity, insulin resistance, and MetS are inseparable (47). We must, however, bear in mind that not all obese individuals have MetS (44). It is important to point out that 10%–30% of obese individuals are insulin sensitive and have normal plasma lipid profile and blood pressure, thus being considered as obese but metabolically healthy (44). In order to better understand this fact, there is a growing interest in the discovery of biomarkers that enable differentiation of individuals who are obese and metabolically healthy from those who are obese and metabolically ill (48). The increase in the branched chain amino acids (BCAAs; leucine, isoleucine, valine) and aromatic amino acids (AAAs; phenylalanine, tyrosine) are often reported in these populations (49). Studies of differences in acylcarnitine levels generally report increased levels of the BCAA-associated species (i.e., propionylcarnitine, butyrylcarnitine, isovalerylcarnitine) (49). Less consistent results were obtained for other amino acids. The increase in plasma BCAAs in metabolically unwell and diabetic individuals has both research and clinical implications.
Figure 22.2 Factors in the pathogenesis of metabolic syndrome.