Metabolomics and perinatal complications

Flaminia Bardanzellu, Moshe Hod, and Vassilios Fanos

Introduction

Metabolomics is a highly promising tool in the early diagnosis of several fetal and perinatal conditions, through the detection of specific markers. Studies in this field really enriched our knowledge during the last years and, up to now, many studies and preliminary results are available.

Metabolomic models also provide promising results in disease progression monitoring; moreover, they could help to ensure the best therapeutic approach, allowing a “tailored” management of the patient. Finally, drug-related toxicity and the response to the performed approach could be, in the future, evaluated through metabolomics.

This chapter summarizes the knowledge regarding metabolomics studies in perinatal acute and chronic inflammation and infections, such as chorioamnionitis, sepsis, congenital cytomegalovirus (CMV) infection, and bronchodysplasia (BPD). We systematically analyzed available literature found on MEDLINE and reported the findings, grouped according to the analyzed samples (amniotic fluid [AF] and neonatal biofluids). However, before novel diagnostic and predictive biomarkers are introduced, it will be necessary to confirm and validate them on larger studies and population samples.

Chorioamnionitis

Chorioamnionitis or intra-amniotic inflammation (IAI) can be caused by several aerobic or anaerobic pathogens, and its incidence can reach the value of 50% among premature births (1,2).

It represents an inflammatory status occurring between the maternal tissues and the fetal membranes (choriodecidual space) or in fetal annexes (chorioamniotic membranes, AF, umbilical cord) (3). A persistent, uncontrolled inflammatory status may affect fetal development, especially brain growth.

IAI, determining an increase in oxidant status and potentiating inflammation, exposes the fetus to negative factors, resulting in immediate (funisitis, chorionic vasculitis, vertical sepsis, preterm delivery, necrotizing enterocolitis, death) and tardive consequences (encephalopathy, and/or cerebral palsy, late-onset sepsis) (3,4).

IAI diagnosis is often delayed and based on clinical manifestations; in fact microbial identification in AF, when possible, can be obtained in almost 7 days (1).

Clinical chorioamnionitis occurring at term (TCC) determines higher morbidity and mortality in mother and fetus (4).

Mothers affected by chorioamnionitis could benefit from a tailored treatment, and neonates, if correctly diagnosed, can also receive a personalized therapy, also avoiding antibiotic prophylaxis when unnecessary (3).

Amniotic fluid

In a rat model, a metabolomics evaluation was performed to detect potential correlations between IAI and neonatal brain maturation (5).

In AF metabolome and evaluating fetal and neonatal rat brain, differences related to fetal sex metabolites were detected, potentially representing a pathogenetic prenatal explanation for different sexual incidence of neurobehavioral disorders.

Laboratory-induced inflammation determined acute metabolic AF changes after 6 hours, since amino acids and purine significantly increased, returning to the steady state after 48 hours (5).

Precocious IAI-related biomarkers of inflammation have been measured, through liquid chromatography-mass spectroscopy/mass spectroscopy (LC-MS/MS), by Cháfer-Pericás and coworkers (1,2).

AF samples from n = 23 women (gestational age [GA] 19+6−35+5 weeks), divided into n = 10 IAI and n = 13 controls were collected. As a result, glutathione sulfonamide (GSA) and 3-chloro-tyrosine (3Cl-Tyr) (markers of inflammation) and 8- hydroxy-2′-deoxyguanosine (8OHdG) (oxidative stress related) were increased in IAI; such pregnancies also showed lower GA, birth weight (BW), and AF glucose content.

The increase in GSA and 3Cl-Tyr seems also to predict IAI severity. These potential biomarkers of IAI should be thoroughly validated (1,2).

Two relevant studies have been performed by Maddipati and coworkers (4,6), that evaluated AF lipidome in TCC using LC/MS, taking into account the great role played by lipids in inflammatory pathways.

In the first study, they compared TCC samples characterized by microbial invasion of the amniotic cavity (TCC-MIAC) and absent microbial invasion of the amniotic cavity (TCC-noMIAC) (4).

In all TCC samples, a proinflammatory status occurred; anti-inflammatory/proresolution lipids were significantly decreased compared to term labor birth, suggesting new targets for innovative therapies.

Anti-inflammatory bioactive lipids’ reduction, represented by modifications in acyl-lipidomic pathways, occurred in both TCC types but mostly characterized samples without microbial invasion.

They retrospectively evaluated n = 35 samples from healthy full-term births, n = 24 TCC (divided into n = 12 TCCMIAC and n = 12 TCC-noMIAC). Linoleic, arachidonic, eicosapentaenoic, and docosahexaenoic acid produced by cyclooxygenase (COX), lipoxygenase, and the epoxygenase pathways of polyunsaturated fatty acids (PUFAs) metabolism were evaluated.

Globally, among 144 PUFAs, 51 were present in more than 50% of the samples in any group. As a result, 13,14-dihydro-15-keto PGF2α increased in TCC-MIAC compared to term birth. The ω-3 and ω-6 PUFAs 12-HETE (HETE, hydroxyl eicosatetraenoic acid), 15-HETE, and 11-HEPE (HEPE, hydroxy eicosapentaenoic acid) resulted in higher TCC-MIAC than in term birth, while all the hydroxyl fatty acids derived from ω-3 PUFAs were significantly lower in TCC-noMIAC than in term birth.

The ω-3 PUFA-derived hydroxy fatty acids (HOTrEs, HEPEs, and HDoHEs) and ω-6 PUFA-derived hydroxy fatty acids (HOTrE-γ, HEDE, and HETEs) were lower in TCC-noMIAC than in term birth.

Epoxy fatty acids derived from PUFAs in the epoxygenase pathway, from ω-3 PUFAs (EpETEs and EpDPEs) were significantly reduced in TCC than term birth, while those derived from ω-6 PUFAs (Ep- OMEs and EpETrEs) were lower in TCC-noMIAC than in term birth.

The physiologic inflammatory response undergoing spontaneous resolution, which can be observed in spontaneous term labor, results impaired in TCC.

As a result, anti-inflammatory epoxygenase and lipoxygenase metabolites of ω-3 PUFAs (linolenic, eicosapentaenoic, and docosahexaenoic acids) and proresolution mediators hydroxy fatty acids of ω-3 PUFAs (i.e., HEPEs and HDoHEs) are reduced in TCC, especially without microbial invasion.

Discriminating infective and sterile IAI is very helpful in correct management, and metabolomics seems a promising tool (4).

In the second study of Maddipati and coworkers (6), LC-MS was used to evaluate n = 35 full-term healthy controls, n = 25 showing IAI and preterm labor (PTL) divided into two subgroups (n = 15 PTL-IAI-noMIAC, n = 10 PTL-IAI-MIAC), n = 24 TCC (divided into n = 12 TCC-IAI-MIAC, n = 8 TCC-SI, n = 4 TCC-NI, NI = no intra-amniotic inflammation), and n = 28 sonographic short cervix (SCX).

Pro-inflammatory lipid mediators of the 5-lipoxygenase pathway, LTB4 and 5-hydroxyeicosatetraenoic acid (5-HETE) significantly increased in case of microbial involvement, thus LTB4 can represent a useful marker of microbial presence, identifying patients potentially requiring antibiotics.

Inflammatory lipids derived from arachidonic acid by 5-lipoxygenase, including 5-HETE, 5-oxoeicosatetraenoic acid (5-oxoETE), LTB4, 5(S),12(S) dihydroxyeicosatetraenoic acid (5[S],12[S]-diHETE), and lipoxin B4 (LXB4) were higher in case of TCC with microbial presence.

In premature deliveries, the results were similar, since 5-HETE and LTB4 were significantly higher in case of microbial presence. LC-MS AF analysis could represent a diagnostic instrument that detects the IAI pathway rapidly (less than 1 hour) and early (6).

In the study of Revello et al. (7), several interleukins modified their levels in chorioamnionitis; in particular, interleukin (IL)-4, -10, -12, and -8 allowed the identification of funisitis in women predisposed to chorioamnionitis (7).

AF metabolic patterns in chorioamnionitis and the possible specific variations occurring in case of perinatal and subsequent neurological impairment have been studied by Dudzik and collagues (8).

N = 28 AF samples were collected. Among them, n = 13 were healthy controls, and n = 15 represented pathological cases, divided into two subgroups according to the presence of neurological damage (intraventricular hemorrhage or periventricular leukomalacia). All of the samples were evaluated through liquid chromatography (LC/MS).

Statistically significant differences emerged comparing samples from mothers affected and not affected by chorioamnionitis; in particular, the concentration of glycerophospholipids choline metabolites (lysophosphatidylcholines 16:0, 18:0, 18:1, 18:2, and 20:4, lysophosphatidylethanolamines 16:0, 18:0, 18:1, 18:2, 20:4, lysophosphatidylserine 16:0 and 18:1, and phosphatidylcholine 16:0 and 18:1), sphingolipids (sphingomyelins 16:0 and 18:1, lactosylceramides 16:0 and 18:1), and mediators derived from bile acids metabolism increased; values of vitamin D3 derivatives, creatine, glucuronide conjugates, and metabolites involved in pyruvate metabolisms (butynedioic acid) were notably altered and glucose levels decreased. Variations mostly regarded metabolites taking part in signaling, proinflammatory patterns, neuroinflammatory disease, and apoptosis in amniotic cells.

It was speculated that bile acids compounds increased since inflammation impaired their excretion, while glucose seems to decrease due to its utilization by bacteria.

In conclusion, metabolites such as lactosylceramides 16:0 and lysophosphatidylethanolamines 16:0 were highly predictive for chorioamnionitis, while lysophosphosphatidylcholine 18:0, sulfocholic acid, and especially trioxocholenoic acids seem highly associated with neurological complications (8).

Metabolomic AF analysis performed by Romero et al. (9) from mothers undergoing preterm labor and IAI, showed a decrease in hexoses (mannose and galactose), potentially deriving from bacterial request of carbohydrates (9).

Moreover, Prince and coworkers (10) measured metabolic pathways in placental membranes of mothers delivering preterm and affected by chorioamnionitis. The result was an increase in glycerophospholipid and production of arachidonic acid, promoting inflammation (10).

Neonatal biofluids

A study performed in our group is the first reported study that evaluated metabolomics pathways in early collected urine samples (<24 hours of life) of preterm neonates (<35 weeks of GA and BW ≤ 1500 grams) whose mother was affected by chorioamnionitis.

Gas chromatography-mass spectrometry (GC-MS) was used to evaluate samples from n = 12 cases and n = 24 healthy controls, finding significant differences involving 29 metabolites, and therefore pointing out metabolic modifications occurring after chorioamnionitis. Detected metabolites, responsible for sample separation, reduced their concentrations in cases instead of controls. Interestingly, only gluconic acid (an oxidation product of glucose) showed an inverse trend.

In the urine of cases, glutamate metabolism, mitochondrial electron transport chain, citric acid cycle (i.e., tricarboxylic acid cycle [TCA]), galactose metabolism, and fructose and mannose degradation metabolism underwent the greatest variations (3).

Moreover, lower values of succinic, citric, and malic acid in cases could point out a neonatal mitochondrial dysfunction, potentially related to systemic inflammatory cascade consequent to microbial infection. Moreover, bacterial metabolism could play a relevant role and should be further investigated (3).

An animal model performed by Dedja et al. (11), allowed the researchers to deduce that L-citrulline (L-Cit) can improve alveolar and vascular growth, offering a better pulmonary outcome in rats showing chorioamnionitis during fetal life, via reducing the lung inflammatory response measured through hematic mediators (11).

Sepsis

Sepsis, potentially caused by microorganisms such as viruses, fungi, or gram-negative or gram-positive bacteria, is a major cause of neonatal morbidity and mortality and shows an incidence of 0.76–0.77 of every 1,000 live births in the United States, with a fatality rate as high as 24.4%. It represents a high risk in neonates due to their immature immune system, especially in the vulnerable category of premature births. “Early onset” sepsis (EOS) occurs within the first 72 hours of life, while “late-onset” sepsis (LOS) is characterized by an onset between 72 hours and 6 days of life (12).

Although early diagnosis helps in the prompt and correct management, improving survival and outcome, well-defined biomarkers are still lacking. Actual available markers could be quite inaccurate. LOS is related to high mortality, with rates of 20%–40% in neonates with less than 32 weeks of GA, especially when caused by gram-negative pathogens, which can be strongly reduced through antibiotic treatment. Surviving neonates undergo a higher risk of neurological and other morbidities, i.e., BPD (13,14).

Blood microbiological culture is the gold standard in sepsis diagnosis, but it requires a time-window of (24–72 hours); moreover, some cases of negative blood culture sepsis can appear (15).

Some classic biomarkers are currently used, showing different levels of sensibility, such as C-reactive protein (CRP), IL-6, IL-8, procalcitonin, urokinase-type plasminogen activator receptor (SuPAR) (12), white blood cell count (WBC) and differential count, micro-erythrocyte sedimentation rate (ESR) (16).

Moreover, according to recent data, the most interesting biomarkers for managing neonatal sepsis seem to be represented by soluble CD14 subtype presepsin (sCD14-ST), lipopolysaccharide binding protein (LBP), angiopoietins (Ang)-1 and -2, triggering receptor expressed on myeloid cells (TREM1), soluble urokinase-type plasminogen activator receptor (suPAR), platelet-activating factor (PAF), and calprotectin (17).

High values of sTREM-1 in neonatal EOS have also been associated with higher mortality (18).

An early diagnosis could help in the management of such a condition, and metabolomics are promising. Metabolomic analysis helps to monitor modifications induced by sepsis, such as hypoxia, oxidative stress, and increased energy requests (modulating glucose and oxidative metabolism of fatty acids) (12,15).

In the literature, there are few available data regarding metabolomics studies on neonates to precociously detect the occurrence of sepsis. Among these studies, possible limitations are small samples of septic newborns, differences in sampling collection, data acquisition, time of sampling, and analysis parameters.

However, clinical relevance of the detected biomarkers still requires additional evaluation and confirmation.

Metabolomics results are highly promising in the perspective of an early diagnosis without any false-positive or false-negative results (17), tailored management, and even for monitoring disease progression and antibiotic toxicity (16,19,20).

Neonatal biofluids

Urinary metabolomics between septic neonates and healthy controls has been for the first time compared by our research group, demonstrating (through 1H-NMR [nuclear magnetic resonance] and GC-MS) statistically significant differences. In particular, urinary metabolomic from n = 9 infected neonates with EOS and LOS and n = 16 healthy controls, collected at a single time point, showed increased levels of acetone ketone bodies, including hydroxybutyrate and acetoacetate, glucose, maltose, lactate, and acetate, and lower levels of the TCA cycle, such as citrate, ribitol, ribonic acid, pseudouridine and 2-ketogluconic acid, 3,4-dihydroxybutanoic acid, and 3,4,5-trihydroxypentanoic acid in sepsis. Therefore, it seems that sepsis modifies the metabolism of glucose and acetone ketone bodies, in addition to modifying oxidative stress pathways (21).

Serafidis et al. (13) evaluated through 1H-NMR and LC-MS/MS urine samples from neonates with LOS to describe their metabolic profile and detect potential biomarkers. Urine samples were collected at diagnosis and after 3 and 10 days from n = 9 LOS neonates and n = 7 neonates with suspected LOS. A well-defined separation occurred between healthy controls and septic neonates samples (including both confirmed or suspected LOS), which disappeared at the end of the symptoms, important results for both diagnosis and therapy response.

Stay updated, free articles. Join our Telegram channel

Join