Umbilical Cord miRNAs in Small-for-Gestational-Age Children and Association With Catch-Up Growth: A Pilot Study

Abstract

Context

Catch-up growth in infants who are small for gestational age (SGA) is a risk factor for the development of cardiometabolic diseases in adulthood. The basis and mechanisms underpinning catch-up growth in newborns who are SGA are unknown.

Objective

To identify umbilical cord miRNAs associated with catch-up growth in infants who are SGA and study their relationship with offspring’s cardiometabolic parameters.

Design

miRNA PCR panels were used to study the miRNA profile in umbilical cord tissue of five infants who were SGA with catch-up (SGA-CU), five without catch-up (SGA-nonCU), and five control infants [appropriate for gestational age (AGA)]. The miRNAs with the smallest nominal P values were validated in 64 infants (22 AGA, 18 SGA-nonCU, and 24 SGA-CU) and correlated with anthropometric parameters at 1 (n = 64) and 6 years of age (n = 30).

Results

miR-501-3p, miR-576-5p, miR-770-5p, and miR-876-3p had nominally significant associations with increased weight, height, weight catch-up, and height catch-up at 1 year, and miR-374b-3p, miR-548c-5p, and miR-576-5p had nominally significant associations with increased weight, height, waist, hip, and renal fat at 6 years. Multivariate analysis suggested miR-576-5p as a predictor of weight catch-up and height catch-up at 1 year, as well as weight, waist, and renal fat at 6 years. In silico studies suggested that miR-576-5p participates in the regulation of inflammatory, growth, and proliferation signaling pathways.

Conclusions

Umbilical cord miRNAs could be novel biomarkers for the early identification of catch-up growth in infants who are SGA. miR-576-5p may contribute to the regulation of postnatal growth and influence the risk for cardiometabolic diseases associated with a mismatch between prenatal and postnatal weight gain.

Small for gestational age (SGA) is a frequent medical problem the incidence of which varies from 5% to 25% among infants born worldwide (1). Most infants who are SGA (up to 85%) experience a spontaneous recovery of growth (catch-up). This catch-up produces a remodeling of body composition and adipose tissue development, regardless of the presence of obesity (2). Both adipose tissue mass and its distribution are widely related to the development of metabolic disorders and cardiovascular diseases. Among the abdominal fat depots, perirenal fat was shown to be best associated with carotid intima–media thickness (cIMT) in children (3). A rapid catch-up in the first years of life may thus be responsible for the subsequent metabolic alterations observed in this population, including metabolic syndrome, type 2 diabetes, and cardiovascular diseases (4–7).

Recent studies have demonstrated a potential role of miRNAs in the pathogenesis of several pregnancy-related diseases, including maternal obesity (8, 9), gestational diabetes mellitus (10, 11), preeclampsia (12, 13), intrauterine growth restriction (14, 15), and SGA (16) in peripheral blood and placenta. Other pregnancy-related tissues such as umbilical cord have been less extensively studied despite their importance in maternal–fetal communication (17, 18).

The umbilical cord is physiologically and genetically part of the fetus and is the link between the developing fetus and the placenta, allowing the transfer of oxygen and materials to and from the maternal blood (19). It is composed of Wharton’s jelly, which is made up of mucopolysaccharides and contains some fibroblasts, macrophages, and mesenchymal stem cells (20). It has been traditionally studied as a reflection of the fetal state, as functional and morphological alterations have been described in the umbilical cord of pregnancies with several outcomes, including fetal growth restriction (19, 21).

Catch-up growth in infants who are SGA has been shown to be important in the programming of later metabolic disease risk; however, the basis and mechanisms underpinning this fetal programming are unknown. We aimed to identify umbilical cord miRNAs associated with catch-up growth in infants who are SGA and to study their relationship with offspring’s cardiometabolic parameters.

Materials and Methods

Study population and auxology

The study was carried out in a cohort of mother–newborn pairs, recruited in the Hospital Dr. Josep Trueta of Girona, Spain. The inclusion criteria were: (i) infants born at term (37 to 42 weeks) from singleton pregnancies, (ii) umbilical cord collected at delivery, and (iii) informed written consent obtained. The exclusion criteria were: (i) maternal disease (hypertension, preeclampsia, gestational diabetes, or preexisting type 1 and type 2 diabetes mellitus), alcohol abuse, or drug addiction, and (ii) fetal malformations or complications at birth.

Fifteen mother–newborn pairs [five appropriate for gestational age (AGA), five infants who were SGA without catch-up (SGA-nonCU), and five infants who were SGA with catch-up (SGA-CU)] were used for identification of the miRNA profile (discovery cohort), and 64 additional mother–newborn pairs were used for the validation study (validation cohort) (Fig. 1). The 15 infants in the discovery set were randomly selected among the studied subjects who fulfilled the following criteria: birth weight range between −0.5 and +0.5 SD (AGA) and below −2.0 SD (SGA), and the presence of catch-up growth at 12 months into SGA-CU, SGA-nonCU, and control infants (AGA). The sampling method was stratified to ensure that the sample would be representative of the validation cohort subjects with regard to gestational age, maternal body mass index (BMI), and newborn’s sex. The 64 infants in the validation cohort were selected among those studied subjects who fulfilled the following criteria: birth weight range between −0.5 and +0.5 SD (AGA) and between −3.0 and −1.0 SD (SGA), and the presence [SGA-CU (n = 24)] or absence [SGA-nonCU (n = 18) and AGA (n = 22)] of catch-up growth at 12 months. We selected all SGA-CU and SGA-nonCU from the study cohort, and the sampling method to select AGA infants was stratified to ensure that this group would be similar to the others with regard to maternal (age, prepregnancy BMI, and gestational weight gain) and delivery characteristics (gestational age and newborn’s sex).

Catch-up growth is defined as an increase >0.67 SD in weight and/or height during the first 2 years of life (22). However, in our study, this criterion was used to define catch-up growth during the first year of life, as most infants who are SGA and who are born at term experience catch-up growth during this time period. Weight catch-up was calculated as change in weight 12-month SD score (SDS) = weight 12-month SDS − birth weight SDS, and height catch-up was calculated as change in height 12-month SDS = height 12-month SDS − birth length SDS.

A follow-up was done in the offspring when they were in their first grade of school. A total of 30 mothers agreed to participate, so a subpopulation of 30 infants (10 AGA, 10 SGA-nonCU, and 10 SGA-CU) was studied. The study thus included 5- and 6-year-old children whose mothers had been previously enrolled in the studied prenatal cohort. The exclusion criteria were: evidence for congenital anomalies; abnormal kidney, thyroid, or liver function; and presence of chronic illness or receiving any chronic medication. In case of acute minor illness, enrolment was delayed until the child had recovered from that minor illness. Parents signed a written informed consent for the inclusion of their child in the study.

Clinical assessments

Women were followed up throughout pregnancy (one visit per trimester), and prenatal screening tests, including clinical examinations, fetal ultrasounds, and blood analysis, were performed. Medical and reproductive features were retrieved from the mothers’ clinical records along with labor and delivery information. The last menstrual period was used to calculate the gestational age and was corroborated by ultrasound when needed. Maternal venous blood sampling was obtained under fasting, at the end of the second trimester (between 24 and 28 weeks of gestation). Maternal prepregnancy weight, height, and gestational weight gain were obtained from medical records. BMI was calculated weight/height squared (kg/m²). Parity, smoking, and paternal data were obtained by questionnaire.

Weight and length of the newborn were measured after delivery and shown as SDS adjusting for gestational age and sex according to regional normative data. The ponderal index was calculated as follows: [(weight/length³) × 100].

Umbilical cord was collected immediately after childbirth. Upon delivery, umbilical cord was clamped and cut, and a section of the cord (a piece measuring 1 to 4 cm in length) was preserved in RNAlater and immediately stored at −80°C. For miRNA extraction, a section of the umbilical cord free of blood vessels and containing only Wharton’s jelly was used.

At age 6 years, children were examined in the morning, under fasting, and a venous blood sample was obtained. A calibrated scale and Harpenden stadiometer were used to determine weight and height, respectively. Weight SDS, height SDS, and BMI SDS were adjusted for age and sex regional normative data. Waist and hip circumferences were determined in the supine position at the level of umbilicus and of the buttocks, respectively. An electronic oscillometer (Dinamap^® Pro 100; GE Medical Systems, Chalfont St. Giles, UK) with appropriate cuff size was used to measure blood pressure. It was measured on the right arm, after a 10-minute rest, and the patient was lying horizontally in the supine position. Data are presented as the average of two measurements. High-resolution ultrasonography (MyLab™ 25; Esaote, Florence, Italy) was used to measure cIMT and subcutaneous, preperitoneal, intra-abdominal, and perirenal fat according to our previous studies (3). cIMT was assessed 1 cm away from the bifurcation of the right distal common carotid artery. Subcutaneous and preperitoneal fat were assessed in longitudinal abdominal scans at the subxiphoid level. Intra-abdominal fat was assessed in a transverse abdominal scan just above the umbilical level as the width of the layer from the internal face of the abdominal muscle to the posterior wall of the aorta. Perirenal fat was assessed laterally, with the surface of the kidney almost parallel to the skin, as the thickness of fat from the inner side of the abdominal musculature to the surface of the kidney. Data are presented as the average of three measurements. All measurements were performed by the same observer who was unaware of the clinical and laboratory characteristics of the subjects. The method showed a high repeatability with within-operator variability <6%.

Analytical methods

Fasting serum insulin, glucose, and lipids were assayed as previously described (8). Serum glucose was analyzed through the hexokinase method (Cobas C; Roche Diagnostics, Indianapolis, IN). The detection limit was 2.0 mg/dL and coefficients of variation (CVs) were <3%. Immunochemiluminescence (Immulite 2000; Diagnostic Products, Los Angeles, CA) was used to determine immunoreactive insulin in serum. The detection limit was 0.4 mIU/L and the CVs were <10%. Fasting insulin sensitivity was calculated using homeostatic model assessment of insulin resistance = [(fasting insulin in mU/liter) × (fasting glucose in mM)]/22.5. Total serum triacylglycerol was measured by monitoring the reaction of glycerol phosphate oxidase and peroxidase. The detection limit was 5.0 mg/dL, and CVs were <5%. High-density lipoprotein cholesterol was quantified following the homogeneous method of selective detergent with accelerator. The detection limit was 2.5 mg/dL, and CVs were <4%. High-sensitivity C-reactive protein was assessed with the CRP Vario immunoassay (Sentinel Diagnostics, Abbott, Milan, Italy). The detection limit was 0.2 mg/L, and CVs were <3%.

miRNA extraction and RT

Total RNA was extracted from 50 to 100 mg of umbilical cord tissue free of blood vessels, using a Qiagen miRNeasy mini kit (Qiagen, Madrid, Spain). The quantity of isolated RNA was determined with a Nanodrop ND-1000 spectrophotometer (Nanodrop Technologies, Wilmington, DE). Fifty nanograms of RNA was reversed transcribed in 50-μL reactions using a miRCURY LNA™ universal RT miRNA PCR cDNA synthesis kit (Exiqon, Vedbaek, Denmark).

Umbilical cord miRNA profile

The miRNA profile in umbilical cord was studied in 15 samples (5 SGA-CU, 5 SGA-nonCU, and 5 AGA) with the miRCURY LNA™ universal RT miRNA PCR human panels I and II (Exiqon). The amplification was performed in a LightCycler^® 480 real-time PCR system (Roche Diagnostics, Pleasanton, CA), and data were analyzed with Roche LC software.

Data analyses of the miRNA profile were performed by the biostatistics department of Exiqon (Bionova Científica, Madrid, Spain). The amplification efficacy was calculated using algorithms (LinReg software). All assays were inspected for distinct melting curves, and the melting temperature was checked to be within known specifications for the assay. Furthermore, assays must be detected with five quantification cycles (Cq) fewer than the negative control, and with Cq <37 to be included in the data analysis. Data that did not pass these criteria were omitted from any further analysis. Cq was calculated as the second derivative. All data were normalized to the average of assays detected in all samples.

Umbilical cord miRNA validation

Differentially expressed miRNAs were validated in 64 infants who were not part of the discovery cohort (24 SGA-CU, 18 SGA-nonCU, and 22 AGA). The expression of miRNAs was assessed by individual real-time PCR using specific primers sets (miRCURY LNA™ universal RT miRNA PCR LNA^TM PCR primers set; Exiqon). Reactions were run on a LightCycler 480 real-time PCR System (Roche Diagnostics). The cycling conditions comprised 10-minute polymerase activation at 95°C, 45 cycles of 10 seconds at 95°C for denaturation, and 10 minutes at 60°C for annealing and extension. According to NormFinder software, miR-421 was the most stable miRNA in all samples and was used as the endogenous control. Relative miRNA levels were calculated according to the 2-ACT method.

miRNA target prediction

The miRSystem database, a Web-based tool that combines many databases (KEGG, BioParta, Pathway Interaction Database, Reactome, and GO molecular function), was used to predict the genes and pathways targeted by the studied miRNAs.

Statistical analysis and ethics

Statistical analyses were carried out with SPSS 22.0 (SPSS, Chicago, IL). Results are shown as mean ± SEM. Nonparametric data were logarithmically transformed before analyses.

One-way ANOVA and Bonferroni post hoc tests were used to compare differences in continuous variables between groups. Differences with a P value <0.05 were considered statistically significant.

In the discovery analysis, given that few differential miRNAs were obtained (umbilical cord miRNA profile), no correction for multiple comparisons was made, making these analyses all exploratory, and all miRNA with a P value <0.05 were evaluated in the validation cohort.

In the validation analysis, bivariate correlations were done to study the relationship of umbilical cord miRNA with growth parameters at 12 months and 6 years of life. In these analyses, correction for multiple comparisons was made by considering a P value ≤0.004 [0.05 divided by 12 (number of studied miRNAs)]. Multiple linear regression analyses with all of the miRNA in the same model and adjusting by confounding variables were done to determine which miRNA was most associated with the dependent variables. Differences with a P value <0.05 were considered statistically significant.

Receiver operating characteristic (ROC) curve analysis was done by a nonparametric method using SPSS 22.0 (23) to evaluate the usefulness of miR-576-5p to predict catch-up growth. Using the area under the ROC curve (AUC) we measured how well miR-576-5p could distinguish between the two diagnostic groups (SGA-CU and SGA-nonCU).

The study has an 80% power to detect a nominally significant Pearson correlation coefficient of at least 0.15 between miRNA expression and catch-up growth parameters, accepting an α risk of 0.05 in a bilateral contrast (GRANMO version 7.12, Hospital del Mar Medical Research Institute).

The study was approved by the Institutional Review Board of Dr. Josep Trueta Hospital, and informed written consent was obtained from all parents.

The data discussed in this publication have been deposited in an online repository (24).

Results

Maternal and offspring characteristics

Clinical characteristics of pregnant women enrolled in the study and their offspring at birth and at the 12-month follow-up are summarized in Table 1 (discovery cohort) and Table 2 (validation cohort).

The validation cohort consisted of healthy pregnant women with normal weight and adequate gestational weight gain. They were mainly primiparous and with a rather high rate of gestational smoking in women delivering SGA vs AGA children. No differences in clinical characteristics were observed between the study groups, except for maternal glucose that was lower in women delivering SGA children compared with AGA children. All newborns were born at term, and SGA children were symmetrical (ponderal index >10th percentile). As expected, significant differences were noticed both at birth and at 12 months of life due to inclusion criteria. At birth, SGA children were smaller than AGA children (lower birth weight and height) and had smaller placentas (lower placental weight). At 12 months, SGA-nonCU children were smaller than both AGA and SGA-CU children [lower weight, height, weight catch-up (change in weight 12-month SDS), and height catch-up (change in height 12-month SDS)] whereas SGA-CU children were bigger than AGA children (higher weight, height, weight catch-up, and height catch-up) (Table 2).

The discovery cohort consisted of healthy pregnant women with similar characteristics to the validation cohort. Women had normal weights as well as adequate gestational weight gain, they were mainly primiparous, and there was a rather high rate of smoking during gestation in women delivering SGA-CU vs AGA children. All newborns were born at term, and SGA children were symmetrical. As expected, SGA children had lower birth weight and height compared with AGA children. At 12 months, SGA-nonCU children had lower weight, height, weight catch-up, and height catch-up compared with AGA and SGA-CU children (Table 1).

Clinical characteristics of mother–newborn pairs with (n = 30) and without (n = 34) follow-up at 5 to 6 years of age are shown in an online repository (25). The characteristics of the group with follow-up did not differ from the group without follow-up.

miRNA profile

Of the 752 human miRNAs present on the miRCURY LNA™ universal RT miRNA PCR human panels I and II (Exiqon), 658 miRNAs were detected in AGA children, 593 miRNAs in SGA-nonCU children, and 601 miRNAs in SGA-CU children. Among all of the detected miRNAs, some were unique for each group: 27 miRNAs in AGA children, 10 miRNAs in SGA-nonCU children, and 24 miRNA in SGA-CU children, whereas 551 miRNAs were common among all groups (Fig. 2).

Of the 551 common miRNAs, 12 miRNAs were differentially expressed between SGA-CU and SGA-nonCU children (miR-128-3p, miR-222-5p, miR-300, miR-374b-3p, miR-501-3p, miR-548c-5p, miR-576-5p, miR-628-5p, miR-770-5p, miR-873-5p, miR-876-3p, and miR-940) (all nominal P < 0.05, Table 1). Even though the rate of differentially expressed genes did not exceed the expected false discovery rate, these miRNAs were chosen as candidates for miRNAs regulating the catch-up growth in SGA children and were validated by individual RT-PCR in the study subjects.

Validation of differential miRNA

The validation of the 12 differential miRNAs by RT-PCR demonstrated that miR-128-3p, miR-501-3p, miR-576-5p, miR-628-5p, miR-770-5p, and miR-876-3p were upregulated in SGA-CU vs SGA-nonCU children (all nominal P ≤ 0.05). Moreover, miR-501-3p and miR-576-5p were upregulated in SGA-CU vs AGA children, and miR-628-5p was downregulated in SGA-nonCU vs AGA children (all nominal P ≤ 0.05; Table 2; Fig. 3A–3G). miR-576-5p showed significant differences after correcting for multiple comparisons. Individual RT-PCR failed to demonstrate differences in miR-222-5p, miR-300, miR-374b-3p, miR-548c-5p, miR-873-5p, and miR-940.

Association with growth parameters at 12 months

Most of the studied miRNAs showed associations with offspring growth parameters at 12 months of life (Table 3). miR-501-3p and miR-576-5p correlated positively with weight SDS (nominal P = 0.03 and P = 0.01), height SDS (nominal P = 0.002 and P = 0.004), weight catch-up (nominal P = 0.03 and P = 0.004), and height catch-up (nominal P = 0.02 and P < 0.0001); miR-770-5p correlated positively with weight SDS (nominal P = 0.03) and height SDS (nominal P = 0.02); and miR-876-3p correlated positively with weight SDS (nominal P = 0.004) and weight catch-up (nominal P = 0.04). The strongest associations were observed for miR-576-5p, which was the only miRNA with significant associations with these variables after correcting for multiple comparisons.

In a multiple linear regression analysis model with all miRNAs and adjusting for newborn’s sex, gestational age, birth weight and height, gestational smoking, parity, maternal BMI, and maternal and paternal height, only miR-576-5p showed significant associations with weight, height, catch-up weight, and catch-up height at 12 months and explained between 7% and 29% of their variances (Table 4). Moreover, in ROC curve analyses, miR-576-5p was found to have a low-to-moderate discriminatory accuracy for catch-up growth prediction (area under the ROC curve of 0.75) (Fig. 3H).

When we analyzed the associations within each study group, the results showed that only in SGA-CU children, miR-576-5p was associated with weight, height, weight catch-up, and height catch-up at 12 months (all nominal P < 0.05). The association between miR-576-5p and height catch-up remained significant after correction for multiple comparisons. In multiple linear regression analysis adjusting for confounding variables (newborn’s sex, gestational age, birth weight and height, gestational smoking, parity, maternal BMI, and maternal and paternal height), miR-576-5p was an independent predictor of height and catch-up height and explained ∼20% of their variances (data not shown). No significant associations were found for the other studied miRNAs in AGA or SGA-nonCU groups.

Association with growth parameters at 6 years

Clinical and laboratory assessments of the offspring follow-up at 6 years are shown in Table 2.

As expected, SGA-CU children showed similar anthropometric measures and were similar to AGA children but showed significant differences when compared with SGA-nonCU children, including more weight, height, BMI, waist, hip, and perirenal fat (all P < 0.05).

Table 3 shows the associations of differential miRNA and growth parameters at 6 years of life. miR-374b-3p and miR-548c-5p correlated positively with waist SDS (nominal P = 0.02 and P = 0.03) and hip (nominal P = 0.004 and P = 0.02), and miR-576-5p correlated positively with weight SDS (nominal P = 0.001), height SDS (nominal P = 0.003), waist SDS (nominal P = 0.003), hip (nominal P = 0.02), and renal fat (nominal P = 0.03). The association of miR-576-5p with weight SDS, height SDS, and waist SDS remained significant after correction for multiple comparisons.

In a multiple linear regression analysis model with all miRNAs and adjusting for children’s age and sex, maternal BMI, and maternal and paternal height, miR-576-5p was an independent predictor of weight, waist, and renal fat at 6 years of age and explained 8%, 9%, and 20% of their variance, respectively (Table 4).

Other ultrasound and blood measurements did not have significant associations with the miRNA expression patterns.

Predicted miRNA pathways and targets

In silico analysis of miR-576-5p using the miRSystem database showed that the four main signaling pathways associated with this miRNA were insulin, IGF-1, PDGFR-B, and mTOR signaling. miR-576-5p targets six genes in insulin signaling (EXOC5, GRB10, NCK1, PIK3CA, PTPN1, RPS6KB19), five genes in IGF-1 signaling (GRB10, IGF1, PIK3CA, PTPN1, RPS6KB1), eight genes in PDGFR-B signaling (FOS, GRB10, ITGAV, MAP2K4, NCK1, NCKAP1, PIK3CA, PTPN1), and five genes in mTOR signaling (IGF1, PIK3CA, RICTOR, RPS6KB1, VEGFA) (Table 5).

Similar in silico analysis with the 24 miRNAs uniquely expressed in SGA-CU children showed that a group of nine miRNAs (miR-302e, miR-34c-3p, miR-490-5p, miR-508-5p, miR-561-3p, miR-650, miR-767-5p, miR-875-5p, miR-885-3p) was related with MAPK, PDGFR-B, IGF-1, and insulin signaling pathways (25).

Discussion

We have studied, to our knowledge for the first time, the miRNA profile in umbilical cord tissue of SGA and AGA children and provide evidence that several umbilical cord miRNAs may be associated with catch-up growth in infants who are SGA. Moreover, miR-576-5p was associated with offspring anthropometric parameters and was a predictor of increased weight, waist circumference, and renal fat at 6 years of age.

The causes of SGA are multifactorial and include maternal lifestyle and obstetric factors, placental dysfunction, and fetal (epi)genetic abnormalities. However, little is known about the mechanisms underpinning catch-up growth in SGA children.

Recent data suggest that it could be related to fetal programming (26). We have studied epigenetic differences (miRNA expression) between SGA-CU and SGA-nonCU children. Our data showed 12 miRNAs dysregulated in SGA-CU vs SGA-nonCU children, of which 6 were preliminarily validated in all study subjects (miR-128-3p, miR-501-3p, miR-576-5p, miR-628-5p, miR-770-5p, and miR-876-3p). Among them, miR-501-3p, miR-576-5p, miR-770-5p, and miR-876-3p were associated with increased weight, height, weight catch-up, and height catch-up at 1 year of age in the whole group. Moreover, significant associations between miR-576-5p and increased height and height catch-up at 1 year of age were observed in the SGA-CU subgroup.

As far as we know, these miRNAs have not been previously related with pregnancy disorders; however, they have been related with the promotion of cell proliferation, migration, and invasion in several types of cancer (27–31) and with the regulation of glucose and lipid metabolism. Specifically, miR-128-3p regulates adipogenesis and lipolysis in 3T3-L1 cells (32), miR-501-3p was described to stimulate insulin secretion in the pancreas from diabetic rats (33), miR-628-5p was dysregulated in platelets from patients with type 2 diabetes (34), miR-770-5p was associated with hyperglycemia in patients with type 1 diabetes (35), and miR-876-3p was shown to regulate glucose homeostasis and insulin sensitivity by targeting adiponectin in adipocytes differentiated from human mesenchymal stem cells (36). It is known that early catch-up growth may confer an increased risk of later metabolic disturbances—particularly of glucose and lipid metabolism—in individuals born SGA (37). We proposed that epigenetic mechanisms, including the miRNAs described in this work, may be involved in this regulation. In fact, other epigenetic mechanisms, such as DNA methylation of key enzymes influencing insulin action, have been described to regulate glucose metabolism in SGA children compared with AGA children (38, 39). However, further studies in vitro or with animals are required to validate the findings of this study.

We have also studied the association of these miRNAs with offspring anthropometric parameters. Our results showed that miR-576-5p was an independent predictor of weight, waist circumference, and renal fat at 6 years of life. Waist circumference is a well-established measure of central adiposity in children and a clinically relevant marker of cardiometabolic risk factors (40). Renal fat is considered to be an important abdominal fat depot and has been recently associated with cardiovascular risk in children (3). These data suggest that miR-576-5p may influence the risk for cardiometabolic diseases associated with postnatal growth, but in this study, it was not associated with any clinically relevant abnormalities in laboratory measures.

Little is known about the function of miR-576-5p. It has been related to the regulation of cell proliferation and migration in several types of cancer (41, 42) as well as the proinflammatory immune response in chronic kidney disease and pertussis disease (43, 44), and a recent manuscript showed that it was upregulated in human nonalcoholic fatty liver disease (45), a health problem closely associated with metabolic and cardiovascular diseases (46). Moreover, in vitro studies with hepatocytes transfected with miR-576-5p showed that this miRNA participates in modulation of the mTOR signaling pathway (45). These results are in agreement with our in silico analysis showing that miR-576-5p targets genes of the insulin, IGF-1, PDGFR-B, and mTOR signaling pathways. The insulin signaling pathway is critical in regulating metabolism, whereas the IGF-1 pathway plays a role in regulating growth. Both pathways are strong activators of the phosphatidylinositol 3-kinase gene that recruits AKT to the plasma membrane where is phosphorylated. mTOR, a serine/threonine kinase, can also phosphorylate AKT and increase its activity (47). These signaling pathways are dysregulated in many human diseases, including growth disorders, diabetes, and obesity (48). In silico studies with the miRNAs uniquely expressed in SGA-CU children pointed to the same signaling pathways as the most representative among these miRNAs; however, the validation of their expression in a larger group of SGA-CU children is needed to confirm these results.

Novel data suggest that miRNAs are secreted in the circulation and act as signaling molecules mediating intercellular communication among many distance tissues (49). Umbilical cord–secreted miRNAs have been implicated in the regulation of several processes, including myofibroblast formation in the skin (50), bone regeneration (51), and endothelial cell proliferation and migration (52). We suggest that our newly described umbilical cord miRNAs would be secreted by umbilical cord cells and influence the phenotype of other fetal cells (e.g., adipose tissue and liver) involved in the regulation of catch-up growth. In vitro studies using these cells would help to validate our results and elucidate the physiological effects of these miRNAs.

We acknowledge several limitations of the present study. The main limitation of our study was the small sample size, which may have precluded finding statistically significant results after correction for multiple comparisons. Another limitation is that, even though the miRNA profile was performed in SGA children with birth weight below −2.0 SD, the validation in a larger cohort was performed in children with birth weight below −2.0 SD and children with birth weight between −1.0 SD and −2.0 SD. This cutoff allowed us to get three homogeneous groups. This could be the reason why only 6 of the 12 dysregulated miRNAs showed significant differences in the validation study, and the false discovery rate was not exceeded for the number of miRNAs found to be altered. Thus, larger studies in a cohort of SGA children with birth weight below −2.0 SD are required to validate these findings. In our study, we identified SGA children with catch-up growth during the first year of life. Even though catch-up growth is predominantly defined as an increase >0.67 SD in weight and/or height during the first 2 years of life, it mainly occurs during the first 6 postnatal months (53, 54); therefore, we do not think this classification might have biased our results. To identify potential biomarkers that can be easily assessed for the early identification of catch-up growth, we decided to study the miRNA profile in Wharton’s jelly because it is the most available part of the umbilical cord and no extra processing steps are required. However, given the wide variety of cell types present in this tissue, further studies are needed to know exactly the origin of these miRNAs. Similar studies in additional tissues are needed to perform confirmatory analysis of the pathways and to elucidate the physiological effect of these miRNAs.

In conclusion, our data suggest that umbilical cord miRNAs could be novel biomarkers for the early identification of catch-up growth in infants who are SGA. miR-576-5p may contribute to the regulation of postnatal growth and potentially influence the risk for cardiometabolic diseases associated with postnatal growth.

Acknowledgments

The authors are grateful to all the pregnant women and their offspring who took part in the study.

Financial Support: This study was supported by a grant from the Ministerio de Ciencia e Innovación, Instituto de Salud Carlos III, Madrid, Spain (PI13/01257 to A.L.-B. and PI14/01625 and PI17/00557 to J.B.), projects co-funded by Fondo Europeo de Desarrollo Regional (FEDER). B.M.-P. is an investigator of the PERIS (SLT002/16/00065, Generalitat de Catalunya). S.X.-T. is an investigator of the Sara Borrell Fund (CD15/00162, National Institute of Health Carlos III, Spain). F.d.Z. is a clinical investigator of the Clinical Research Council of the University Hospital Leuven (Leuven, Belgium). L.I. is a clinical investigator of Centro de Investigación Biomédica en Red de Diabetes y Enfermedades Metabólicas Asociadas. A.L.-B. is an investigator of the I3 Fund for Scientific Research (Ministry of Science and Innovation, Spain). J.B. is an investigator of the Miguel Servet Fund (CPII17/00013, National Institute of Health Carlos III, Spain).

Author Contributions: B.M.-P. and S.X.-T. contributed to data acquisition and drafted the manuscript; A.B., G.C.-B., and A.P.-P. contributed to the design and acquisition of data; E.L.-M. contributed to the interpretation of data and reviewed the manuscript; J.M.M.-C. contributed to data acquisition and to discussion of the paper; F.d.Z. and L.I. contributed to the interpretation of data and reviewed the manuscript; and A.L.-B. and J.B. contributed to conception and interpretation of data and reviewed the manuscript.

Disclosure Summary: The authors have nothing to disclose.

Data Availability: All data generated or analyzed during this study are included in this published article or in the data repositories listed in References.

Abbreviations:

Abbreviations:
 
• AGA

  appropriate for gestational age

 
• BMI

  body mass index

 
• cIMT

  carotid intima–media thickness

 
• Cq

  quantification cycle

 
• CV

  coefficient of variation

 
• ROC

  receiver operating characteristic

 
• SDS

  SD score

 
• SGA

  small for gestational age

 
• SGA-CU

  infants who are SGA with catch-up

 
• SGA-nonCU

  infants who are SGA without catch-up

References and Notes

Author notes