miR-138/miR-222 Overexpression Characterizes the miRNome of Amniotic Mesenchymal Stem Cells in Obesity

Abstract

Clinical findings and data obtained in animal models indicate that nutrient uptake and exposure to environmental agents during pregnancy may affect fetal/newborn gestational programming, thereby resulting in obesity and/or obesity-related disorders in offspring. Human amniotic mesenchymal stem cells (hA-MSCs) differentiate into adipocytes and are thus a suitable model to investigate adipocyte functions in obesity. The aim of this study was to elucidate the miRNome of hA-MSCs and its contribution to obesity in pregnancy. To this aim we used the following: (i) high-resolution small RNA sequencing to characterize the microRNA (miRNA) profiles of hA-MSCs of 13 obese (Ob-) and 7 control (Co-) pregnant women at delivery; (ii) multiple-method integrated bioinformatics to predict the metabolic pathways potentially miRNA deregulated in Ob-hA-MSCs; and (iii) microarray mRNA expression profiling to verify obese-associated mRNA alterations. In summary, 12 miRNAs were differentially expressed between Ob-hA-MSCs and Co-hA-MSCs, with a multiple-methods bioinformatic consensus on miR-138-5p and miR-222-3p, which were overexpressed in Ob-hA-MSCs versus Co-hA-MSCs. The top 20 significant pathways predicted to be deregulated through miR-138-5p and/or miR-222-3p/target interaction included fat cell differentiation and deposits, lipid/carbohydrate homeostasis, response to stress, metabolic syndrome, heart disease, and ischemia. In conclusion, our finding of miR-138-5p/miR-222-3p overexpression in Ob-hA-MSCs, together with the transcriptomic data, suggests that these miRNAs in obese pregnancy could derange metabolic pathways previously found impaired in tissues from obese adults or in obesity-associated disorders and concur to modify gestational programming as has been demonstrated in animal models. This raises the possibility of using diet-based strategies to normalize the perinatal miRNome in obesity.

Introduction

In particular, in pregnant animal models, high calorie diet induced miRNA-mediated variations in gene expression, and so affected the development of offspring [1,2]. In line with this finding, we previously reported a different miRNA profile in the amnion of obese women versus that of control women and predicted the deregulation of several pathways concurring to placental growth and function that could enhance the newborn's risk of obese-associated metabolic diseases [7]. Deregulated metabolic pathways have also been described in subcutaneous and visceral adipose tissues from obese adults in association with altered miRNA profiles [8,9].

The amnion is also a source of human amniotic mesenchymal stem cells (hA-MSCs). These cells have two advantages for research: (i) they have a close ontogenic relationship with embryonic stem cells but, unlike the latter, are accessible without ethical problems, as placenta is usually discarded at birth [10]; (ii) they are multipotent stem cells that have the potential to differentiate into a wide range of cells, including adipocytes. Consequently, hA-MSCs are a suitable cellular model with which to investigate the metabolic functions of adipocytes during obesity [10].

The expression of miRNAs has been investigated in MSCs isolated from such sources as adipose tissue, bone marrow, and amniotic fluid [11], mostly using a microarray or candidate gene approach [12]. To our knowledge, hA-MSCs isolated from obese women have not yet been studied using small RNA sequencing (small RNA-Seq). This technique is a powerful and a widely used tool with which to identify known and unknown miRNAs and to quantify miRNA levels at a higher resolution than that possible with the array technique [13]. Furthermore, thanks to powerful software, we can avoid false-positive results in the small RNA-Seq of miRNA profiles and predict, with high statistical significance, the putative effect of miRNAs in deregulating cell functions.

The aim of this study was to elucidate the miRNome of hA-MSCs from obese and control women and look for abnormalities associated with obesity in pregnancy.

Materials and Methods

Patients and controls

Thirteen morbidly obese (Ob) and seven lean control (Co) pregnant women aged [mean ± standard error mean (SEM)] 33.7 ± 1.0 and 30.7 ± 1.5 years, respectively, with a prepregnancy body mass index (kg/m²) >30 and <25 kg/m², respectively, were recruited at the Dipartimento di Neuroscienze e Scienze Riproduttive ed Odontostomatologiche, University of Naples Federico II. The clinical and anamnestic data of the enrolled women were recorded upon hospitalization during a medical interview conducted by an expert. Mean (SEM) gestational age was 38.8 (0.3) and 38.7 (0.2) in obese and control women, respectively. Five of the thirteen (46%) obese women and 1/7 (14%) control women were at their first pregnancy. The exclusion criteria were neoplasia, viral infection, diabetes, and metabolic syndrome.

All patients and controls gave their informed consent to the study that was performed according to the Helsinki II Declaration and approved by the Ethics Committee of the School of Medicine, University of Naples Federico II (authorization n. 248/08, 23/02/2009; amendment n. 248/08/ES1, 1/10/2014). A second cohort of 14 placental tissue sections (amnion) from nine obese and five control pregnant women was used for validating purposes [7].

Sample collection

Two maternal blood samples were collected in the morning after a 12-h fast from all enrolled women immediately before delivery. One sample was used for DNA extraction, whereas the other was centrifuged and the serum was stored at −80°C until required for the measurement of the main biochemical parameters and leptin/adiponectin concentrations by routine methods or immunoassay (Bio-Rad, Hemel Hempstead, Herts, United Kingdom), respectively. Term placentas were collected at delivery by cesarean section and immediately processed to isolate the hA-MSCs, according to Parolini et al. [14].

Briefly, we removed the maternal decidua and separated the chorion from amnion. Then, we washed the amnion five times in 40 mL phosphate-buffered saline (PBS) containing 100 U/mL penicillin, 100 μg/mL streptomycin, and 250 μg/mL amphotericin B (all from Sigma-Aldrich, St. Louis, MO) and mechanically minced it into small pieces. The latter was digested overnight at 4°C in ACCUMAX^® reagent (Innovative Cell Technology, San Diego, CA), a combination of DNase, protease, and collagenolytic enzymes [15] containing 100 U/mL penicillin, 100 μg/mL streptomycin, and 250 μg/mL amphotericin B. Digestion enzymes were then inactivated with complete culture medium constituted by low-glucose Dulbecco's modified Eagle's medium (DMEM) (Sigma-Aldrich) supplemented with 10% of heat-inactivated bovine serum [fetal bovine serum (FBS)], 1% of nonessential amino acids, and 2% of UltraGlutamine (all from Lonza, Basel, Switzerland).

After centrifugation at 300 g for 10 min, cell pellets and digested tissue fragments were seeded in a cell culture dish (BD Falcon, New York, NY) in complete culture medium and incubated at 37°C in 5% CO₂ for 7 days. We removed digested tissue fragments and isolated cells that formed distinct fibroblast colony-forming units. Colonies reaching 70% confluence were washed with PBS, detached with trypsin/ethylenediaminetetraaceticacid (EDTA; Sigma-Aldrich), and counted and reseeded in complete medium for expansion at a concentration of about 5,000/cm² [15]. hA-MSCs were further expanded, and when 70%–80% confluent cultures reached about four population doublings [15], they were detached with trypsin/EDTA and resuspended in PBS with 10% FBS.

The fetal origin of hA-MSCs was verified by typing DNA obtained from hA-MSCs, amnion, and mother's peripheral blood leucocytes with the AmpFlSTR^® Identifiler™ PCR Amplification Kit (Applied Biosystems, Foster City, CA). DNA was extracted using the Nucleon BACC2 extraction kit (Illustra DNA Extraction Kit BACC2; GE Healthcare, Calfont St. Giles, Bucks, United Kingdom) and its concentration evaluated with a NanoDrop^® ND-1000 UV-Vis spectrophotometer (NanoDrop Technologies, Wilmington, DE). Polymerase chain reaction (PCR) was performed with a GeneAmp PCR System 9700 (Applied Biosystems) instrument, and PCR primers and conditions were according to the manufacturer's instructions. Analysis of PCR products by capillary electrophoresis performed with the ABI Prism 3130 Genetic Analyzer (Applied Biosystems) confirmed the fetal origin of all the isolated hA-MSCs [15].

To immunophenotype hA-MSCs, we analyzed 38 hematopoietic, mesenchymal, endothelial, epithelial, and no-lineage membrane antigens (CD14, CD15, CD16, CD19, CD28, CD33, CD34, CD45, and CD117; CD9, CD10, CD13, CD26, CD29, CD44, CD47, CD49d, CD54, CD56, CD58, CD71, CD81, CD90, CD99, CD105, CD151, CD166, CD200, HLA-ABC, PECAM-1/CD31; thrombospondin receptor/CD36, Bp50/CD40, Prominin-1/CD133, MDR-1/CD243, NGFR/CD271, ABCG-2/CD338, and HLA-DR), using an unmodified Becton Dickinson FACSCanto II flow cytometer and FACS-DiVa software (Becton Dickinson, San Jose, CA) as previously reported [15]. Monoclonal antibodies were from Becton Dickinson, except anti-CD338-APC (R&D, Minneapolis, MN), anti-CD-133-PE, and anti-CD271-APC (Miltenyi Biotec, Bergisch Gladbach, Germany). The immunophenotype characterization confirmed the mesenchymal origin of the isolated hA-MSCs (Supplementary Table S1; Supplementary Data are available online at www.liebertpub.com/scd).

Small RNA deep sequencing and analysis

Total RNA was purified from hA-MSCs using the mirVana^™ miRNA isolation kit (Ambion, Austin, TX) and quantified with the Nanodrop 2000c spectrophotometer (Thermo Fisher Scientific, Waltham, MA). RNA quality was evaluated with the Agilent 2100 BioAnalyzer (Agilent, Santa Clara, CA). Small RNA library preparation and sequencing were carried out according to the manufacturer's instructions (Illumina, San Diego, CA). Briefly, 1 μg of total RNA was ligated to specific adapters, required for library amplification and sequencing, then retrotranscribed and amplified to obtain double-stranded cDNAs. These cDNAs were loaded on a 6% polyacrylamide gel electrophoresis for size purification of small RNAs. The gel band corresponding to fragments ranging from 140 to 160 bp was cut, ethanol-precipitated, and eluted in ultrapure water.

The libraries were then evaluated for quality on the Agilent 2100 BioAnalyzer (Agilent) and sequenced using the MiSeq instrument (Illumina).

Bioinformatics

The quality of raw sequencing data was evaluated with FastQC (http://bioinformatics.babraham.ac.uk/projects/fastqc). Poor-quality ends of reads were trimmed using a toolset integrated into the Galaxy online data analysis platform [16 –18]. FastQ sequences were mapped and annotated using two application tools for miRNA identification: iMir [19] and miRDeep* [20], which are modular pipelines for the comprehensive analysis of small RNA-Seq data. iMir integrates in its pipeline the miRAnalyzer stand-alone tool [21] that detects known miRNAs. miRAnalyzer has two alignment options based on the presence of adapter sequences and it uses a machine-learning approach based on random forest algorithm implemented in WEKA [22] to predict known/unknown miRNAs. miRDeep uses the MegaBLAST/BLAST algorithm to map the reads, and RNAfold and Bayes’ theorem to evaluate the secondary structure of candidate RNAs. In both cases, low-quality reads were filtered out and known adapter sequences were trimmed from raw data.

Unique sequence reads of 15 nucleotides or more in length, with a minimum read count of three, were aligned, without any mismatch, with precursor human miRNA sequences from miRBase version 20 [23]. Differentially expressed miRNAs were identified using three statistical methods: Fisher's exact test, DESeq [24], and edgeR [25]. Read counts were normalized to minimize systematic technical bias. Quantile normalization and Fisher's exact test were performed with the iMir pipeline. The two Bioconductor packages (DESeq and edgeR) were chosen as a negative binomial distribution model with which to test the differential expression between Ob- and Co-hA-MSCs. In all analyses, a P-value <0.05 was considered significant. Hierarchical biclustering of samples, based on annotated miRNAs and their normalized read counts, was performed using MATLAB's Clustergram algorithm [26,27]. Venn diagrams were obtained using the Jvenn viewer [28].

Target prediction and validation

We used the miRanda [29,30] and TargetScan [31] tools to predict putative targets of the miRNAs that differed significantly between Ob- and Co-hA-MSCs. TargetScan computationally predicts miRNA gene targets by searching for 8-mer and 7-mer sites matching the seed region of each miRNA, whereas miRanda target prediction incorporates current knowledge on target rules and uses a compendium of mammalian miRNAs. The targets predicted by miRanda were filtered according to two main criteria: the score was set to ≥130 and the free energy ≤−16 kCal/mol. The targets predicted by TargetScan were filtered according to a total context score ≤−0.30.

We used MirTarBase [32] to search for experimentally validated interactions. Functional annotation of targets was performed using the Database for Annotation, Visualization and Integrated Discovery (DAVID) [33]. This Functional Classification Tool is a rapid means with which to organize large lists of genes into functionally related groups and thus to help unravel the biological content captured by high-throughput technologies. The following annotation categories were selected: Gene Ontology, Genetic Association DB Disease, OMIM disease, BIOCARTA, KEGG pathway, and PANTHER pathway. Statistics related to over-representation of functional categories was based upon a False Discovery Rate statistic methodology. A P-value <0.05 was considered significant.

Validation of miRNA expression by TaqMan miRNA assay

Real-time (RT) PCR was performed on the 7900 HT Real-Time PCR System (Applied Biosystems) by using the TaqMan miRNA Assay Protocol to assess the relative expression of miR-138-5p and miR-222-3p. For RT reactions, 10 ng of total RNA was used in each reaction (15 μL) and mixed with the corresponding TaqMan miRNA assays RT primer (3 μL). The RT reaction conditions were as follows: 16°C for 30 min; 42°C for 30 min; 85°C for 30 min; and then hold at 4°C. After the RT reaction, 1.33 μL of cDNA was used for the PCR together with TaqMan primer (2 μL). The PCR was carried out at 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and 60°C for 60 s. The miRNA expression values were normalized to RNU48, and then the relative expression values were obtained versus Co-hA-MSCs using the ΔΔCT method (Relative Quantification, RQ = 2^−ΔΔCT) with Sequence Detection System v2.3 and RQ Manager 1.2 software (Applied Biosystems).

The investigated parameters were expressed as mean (SEM). Student's t test was used to compare group means, and a P-level <0.05 was considered statistically significant. Statistical analyses were carried out with the PASW package for Windows (Ver.18; SPSS, Inc., Headquarters, Chicago, IL).

Microarray analysis

Total RNA from Co- and Ob-hA-MSCs was hybridized on a Human HT-12V4.0 Gene Expression Bead Chip (Illumina) containing 47,732 probes, and procedures were according to the manufacturer. The binary intensity data files were read with Illumina's GenomeStudio software and converted into intensity matrices. Probes without a significant detection P-value <0.05 in at least one sample were removed. Differential expression analysis (obese subjects vs. controls) was performed using the limma package [34]. Genes were considered to be differentially expressed at a significant level in case of Benjamini–Hochberg multiple testing correction adjusted P-value <0.05 and with a log₂FC >0.5 or <−0.5. Gene lists were analyzed using the DAVID [33].

Results

The clinical and biochemical characteristics of obese and control pregnant women are reported in Table 1. The prepregnancy body mass index was higher (mean [SEM] 41.5 [2.2] vs. 22.4 [1.0] kg/m²), P < 0.0001, and the weight gain in pregnancy lower (mean [SEM] 10.0 [1.1] vs. 14.3 [1.8] kg) in obese than in control pregnant women, as recommended by guidelines [35]. Clinical and biochemical characteristics were similar in the two groups of pregnant women, whereas serum leptin (P < 0.0001) and the leptin to adiponectin ratio were significantly higher (P < 0.05), and serum adiponectin lower, although not significantly, in obese women than in control women.

Table 1.

Clinical and Biochemical Characteristics of Obese and Control Pregnant Women at Delivery

Parameters	Ob-pregnant women (n = 13)	Co-pregnant women (n = 7)
Age (years)	33.7 (1.0)	30.7 (1.5)
Weight (kg)^a	109.9 (5.6)	65.2 (3.6)
Height (m)	1.62 (0.0)	1.69 (1.7)
Prepregnancy BMI (kg/m²)^a	41.5 (2.2)	22.4 (1.0)
Weight gain in pregnancy	10.0 (1.1)	14.3 (1.8)
Gestational age	38.8 (0.3)	38.7 (0.2)
Number of women at:
First pregnancy	5	1
Second pregnancy	5	3
Third pregnancy	2	2
Fourth pregnancy	1	1
Systolic blood pressure (mmHg)	126.7 (3.2)	117.1 (5.1)
Diastolic blood pressure (mmHg)	84.5 (2.5)	74.2 (2.0)
Frequency cardiac	81.0 (1.6)	79.0 (3.7)
Glucose (mmol/L)	4.2 (0.1)	4.0 (0.3)
Total cholesterol (mmol/L)	7.0 (0.5)	7.3 (0.1)
Triglycerides (mmol/L)	2.7 (0.3)	2.3 (0.3)
AST (U/L)	23.7 (5.2)	14.8 (0.7)
ALT (U/L)	22.0 (5.2)	12.1 (1.1)
ALP (U/L)	128.0 (12.2)	115.0 (12.6)
GGT (U/L)	10.8 (1.6)	8.8 (1.5)
Leptin (L) (ng/mL)^a	38.7 (2.6)	13.6 (3.4)
Adiponectin (A) (μg/mL)	6.6 (0.9)	7.9 (1.6)
L/A^b	7.5 (0.7)	2.4 (0.6)

Data are expressed as mean (SEM). Biochemical parameters were evaluated in serum.

Statistically significant difference at Student t test: ^a P < 0.0001, ^b P < 0.05.

AST, aspartate aminotransferase; ALT, alanine aminotransferase; ALP, alkaline phosphatase; GGT, γ-glutamyl transferase; Ob, obese; Co, control; BMI, body mass index; SEM, standard error of mean.

miRNA sequences from small RNA libraries

The miRAnalyzer detected from 147 to 411 known miRNAs, whereas miRDeep* identified from 136 to 320 miRNAs, the range of the reads number being 78,093–3,588,828 (Table 2). In all samples, the miRAnalyzer detected more miRNAs than miRDeep* did, therefore leading to an increased risk of false-positive results, which we subsequently reduced considering only the miRNAs identified as being differently expressed with both tools. At Spearman correlation analysis, there were no significant differences between Co- and Ob-hA-MSC miRNAs (Supplementary Fig. S1). We carried out a hierarchical clustering analysis using the miRNA expression data of each individual sample, excluding those miRNAs whose read count was equal to 0 in more than 15 samples (Fig. 1A, B, miRNAs identified by MiRAnalyzer and by miRDeep, respectively).

FIG. 1.

Hierarchical clustering of Ob- and Co-hA-MSCs based on miRNA expression data of each individual sample, identified using miRAnalyzer (A) or miRDeep* (B). Note the high variability within Ob- and Co-hA-MSCs and the lack of clusters in the samples. miRNAs whose read count was equal to 0 in more than 15 samples were excluded. Ob, obese; Co, control; hA-MSCs, human amniotic mesenchymal stem cells; miRNA, microRNA.

Table 2.

microRNAs Detected by miRAnalyzer and miRDeep ^* in Control and Obese Human Amniotic Mesenchymal Stem Cell Samples

Sample	Reads number	Number of miRNAs detected by miRAnalyzer	Number of miRNAs detected by miRDeep ^*
Co-01	1,108,278	337	297
Co-02	477,191	271	237
Co-03	645,715	305	276
Co-04	478,945	288	171
Co-05	218,489	242	172
Co-06	156,310	205	182
Co-07	78,093	147	136
Ob-01	496,107	261	210
Ob-02	695,807	299	237
Ob-03	475,608	292	228
Ob-04	737,742	306	232
Ob-05	878,053	334	252
Ob-06	954,778	322	241
Ob-07	1,717,220	411	300
Ob-08	2,126,002	410	320
Ob-09	359,398	267	213
Ob-10	3,588,828	410	299
Ob-11	712,424	285	232
Ob-12	160,191	190	166
Ob-13	401,036	254	208

Ob, obese; Co, control; miRNA, microRNA.

There was a high variability among Ob- and Co-hA-MSC miRNAs, and no miRNA cluster was detected in any sample. We grouped the identified miRNAs in families, according to the MiRBase database, and calculated the read frequencies by dividing specific miRNA read counts by the sum of all miRNA read counts for each sample. We conducted hierarchical clustering for all the miRNA families that represented at least 5% of the total miRNA reads in at least one sample. miRNA families did not differ significantly between Ob- and Co-hA-MSCs.

miRNA expression profiling analysis

The number of differentially expressed (P < 0.05) miRNAs was calculated by three methods: Fisher's exact test, DESeq, and edgeR. miRAnalyzer identified 224, 13, and 41 differentially expressed (P < 0.05) miRNAs, using one-by-one Fisher's exact test, and the DESeq and edgeR methods (Fig. 2A). Only six miRNAs were identified by all the three methods. Thirty-four miRNAs were identified by Fisher's exact test and the edgeR method and four miRNAs by Fisher's exact test and the DESeq method (Fig. 2A, B). The miRdeep* tool identified 210, 13, and 21 differentially expressed (P < 0.05) miRNAs using Fisher's exact test, the DESeq method, and the edgeR method, respectively. Eight miRNAs were identified by all the three procedures, two miRNAs were identified by Fisher's exact test and the DESeq method, and 12 miRNAs were identified by Fisher's exact test and the edgeR procedure (Fig. 2C, D).

FIG. 2.

Venn diagrams of miRNAs identified by miRAnalyzer (A, B) and by miRDeep* (C, D) found to be differentially expressed by three statistical methods. The miRAnalyzer tool identified 224, 13, and 41 differentially expressed (P < 0.05) miRNAs, using one by one Fisher's exact test, DESeq, and edgeR. Thirty-four miRNAs were identified by the Fisher's and edgeR tests and four miRNAs by the DESeq and Fisher's tests (A). Only six miRNAs were identified by all three methods (A, B). miRDeep* identified 210, 13, and 21 differentially expressed (P < 0.05) miRNAs, using Fisher's exact test, DESeq, or edgeR, respectively. Two miRNAs by Fisher's test and DESeq, and 12 miRNAs by edgeR and Fisher's test (C). Eight miRNAs were identified by all three analyses (C, D). NA, not available.

Intersection of the miRNAs identified with both miRAnalyzer and miRDeep* and, differently expressed (P < 0.05) by all the three statistical methods, resulted in consensus for two miRNAs: miR-138-5p and miR-222-3p, both overexpressed in Ob- versus Co-hA-MSCs. It should be noted that the mature fragment miR-138-5p, as reported in miRBase, could arise from two different stem loop precursors, miR-138-1 and miR-138-2. While miRAnalyzer correctly identified miR-138-5p (Fig. 2B), miRDeep* assigned the reads to the precursors (Fig. 2D).

We extracted the fasta sequences of the reads matching miR-138-1 and those matching miR-138-2, based on the miRDeep* results. We found that all the reads aligned to miR-138-2 by BLAST local alignment mapped to the mature miRNA miR-138-5p, and the reads aligned to miR-138-1 mapped mostly to miR-138-5p, and to a lesser extent to miR-138-1-3p. Thus, we confirmed that the mature miR-138-5p was identified by both miRAnalyzer and miRDeep. Regarding miR-222-3p, miRDeep* also detected the stem loop precursor miR-222, and we checked that most of the reads, according to BLAST, mapped to the mature fragment miR-222-3p, in accordance with the data obtained by miRAnalyzer.

Validation of miR-138-5p and miR-222-3p expression by quantitative RT-PCR

We measured the expression levels of miR-138-5p and miR-222-3p also using TaqMan miRNA assay and found that the mean RQ was higher in Ob-hA-MSCs than in Co-hA-MSCs [RQ = 2^−ΔΔCT (SEM): 5.77 (1.08) and 4724.93 (1252.22), for miR-138-5p and for miR-222-3p, respectively]. This finding confirms that both miRNAs are overexpressed in Ob-hA-MSCs in agreement with the data obtained using the small RNA-Seq approach (Supplementary Fig. S2).

We also evaluated the expression of miR-138-5p and miR-222-3p in a second cohort of 14 placental tissue sections (amnion) from nine obese and five control pregnant women using TaqMan miRNA assay. We obtained in Ob- respect to Co-amnion the following mean RQ (SEM) values: miR-138-5p: 5.3 (2.0) and miR-222-3p: 1.1 (0.1). Consequently, the expression of miR-138-5p confirmed the trend observed in Ob-hA-MSCs, whereas the expression of miR-222-3p was similar in Ob- and Co-samples.

Identification of miR-138-5p and miR-222-3p validated gene targets

Validated gene targets of miR-138-5p and miR-222-3p were identified using the mirTarBase database. We investigated in greater depth the meaning of the top 20 significant biological processes most pertinent to hA-MSC epigenetic regulation (enriched terms) (Tables 3 and 4, respectively). The enrichment analysis of the validated targets of miR-138-5p revealed terms related to fat cell differentiation, fat deposits, lipid and carbohydrate homeostasis, metabolic syndrome, heart disease, ischemia, as well as terms involved in response to stress, cytoskeleton, and the p53 signaling pathway. The enrichment terms of miR-222-3p referred to several cell functions: cell differentiation, transport, DNA replication, DNA damage response, signal transduction by p53, and stress response.

Table 3.

Top Twenty Significant Biological Processes (Enriched Terms) Involving the miR-138-5p Validated Gene Targets as Reported by miRTarBase Tool

Enriched terms for miR-138-5p	P value	miR-138-5p validated gene targets	Fold enrichment
h_vobesityPathway:Visceral Fat Deposits and the Metabolic Syndrome	0.0065	LPL, PPARG, NR3C1	22.45
GO:0050873∼brown fat cell differentiation	0.0083	CEBPA, PPARG, FABP4	21.41
GO:0045444∼fat cell differentiation	0.0000	CEBPA, CCND1, PPARG, FABP4, CBY1, SIRT1	19.39
GO:0008013∼beta-catenin binding	0.0116	CDH1, CBY1, CDH2	18.03
hsa05216:Thyroid cancer	0.0140	CCND1, PPARG, CDH1	15.94
GO:0030111∼regulation of Wnt receptor signaling pathway	0.0023	CCND1, SOX4, ZEB2, CBY1	14.89
GO:0033500∼carbohydrate homeostasis	0.0028	HIF1A, PPARG, SERPINE1, NEUROD1	13.98
h_prc2Pathway:The PRC2 Complex Sets Long-term Gene Silencing Through Modification of Histone Tails	0.0172	SUZ12, EZH2, EED	13.82
heart disease, ischemic	0.0004	LPL, HIF1A, PPARG, SERPINE1, MMP3	13.54
GO:0042393∼histone binding	0.0241	DEK, H2AFX, SIRT1	12.29
kidney cancer	0.0232	CCND1, HIF1A, MMP3	12.05
skin cancer, nonmelanoma	0.0247	CCND1, PPARG, NR3C1	11.65
REACT_7970:Telomere Maintenance	0.0040	HIST1H2BK, HIST1H2BJ, H2AFX, TERT	11.56
hsa04115:p53 signaling pathway	0.0008	PPM1D, CCND1, CASP3, CCND3, SERPINE1	11.33
GO:0003707∼steroid hormone receptor activity	0.0294	PPARG, RARA, NR3C1	11.04
lymphoma, non-Hodgkin	0.0296	CCND1, CASP3, H2AFX	10.59
GO:0055088∼lipid homeostasis	0.0349	LPL, PPARG, FABP4	10.07
GO:0033293∼monocarboxylic acid binding	0.0351	PPARG, FABP4, RARA	10.02
GO:0080135∼regulation of cellular response to stress	0.0003	TMEM189-UBE2V1, SEMA4C, UBE2 V1, H2AFX, ZEB2, TMEM189, SIRT1, MAP3K11	9.98
GO:0000079∼regulation of cyclin-dependent protein kinase activity	0.0387	CCND1, CASP3, CCND3	9.51

Table 4.

Top Twenty Significant Biological Processes (Enriched Terms) Involving the miR-222-3p Validated Gene Targets as Reported by miRTarBase Tool

Enriched terms for miR-222-3p	P value	miR-222-3p validated gene targets	Fold enrichment
GO:0006607∼NLS-bearing substrate import into nucleus	0.0001	TRPS1, IPO5, KPNA6, KPNA2, KPNB1	17.40
GO:0048070∼regulation of pigmentation during development	0.0233	KITLG, KIT, BCL2L11	12.34
h_EfpPathway:Estrogen-responsive protein Efp controls cell cycle and breast tumor growth	0.0231	ESR1, TP53, CDK6	11.84
GO:0045646∼regulation of erythrocyte differentiation	0.0068	ETS1, STAT5A, CDK6, FOXO3	10.05
GO:0008088∼axon cargo transport	0.0068	APP, KIF4A, KIF3B, MACF1	10.05
GO:0006268∼DNA unwinding during replication	0.0370	MCM7, TOP3A, HMGA1	9.70
GO:0042771∼DNA damage response, signal transduction by p53 class mediator resulting in induction of apoptosis	0.0420	BBC3, HIPK2, TP53	9.05
GO:0001836∼release of cytochrome c from mitochondria	0.0105	BBC3, TP53, MYC, SOD2	8.62
GO:0000184∼nuclear-transcribed mRNA catabolic process, nonsense-mediated decay	0.0040	PAN2, EXOSC10, PNRC2, RNPS1, CASC3	7.54
GO:0034623∼cellular macromolecular complex disassembly	0.0189	C12ORF65, MAZ, STMN1, HMGA1	6.96
endometrial cancer	0.0018	MGMT, ESR1, TP53, MDM2, PTEN, MMP1	6.51
GO:0000956∼nuclear-transcribed mRNA catabolic process	0.0071	PAN2, EXOSC10, PNRC2, RNPS1, CASC3	6.46
GO:0010970∼microtubule-based transport	0.0253	APP, KIF4A, KIF3B, MACF1	6.24
hsa03030:DNA replication	0.0086	MCM7, RFC1, POLD2, POLE, MCM3	6.04
GO:0008637∼apoptotic mitochondrial changes	0.0302	BBC3, TP53, MYC, SOD2	5.84
h_ctcfPathway:CTCF: First Multivalent Nuclear Factor	0.0314	CDKN1B, TP53, MDM2, PTEN	5.53
GO:0045639∼positive regulation of myeloid cell differentiation	0.0355	ETS1, STAT5A, KITLG, FOXO3	5.48
GO:0031047∼gene silencing by RNA	0.0355	SRRT, DICER1, PIWIL1, TNRC6B	5.48
brain cancer	0.0120	MGMT, TP53, MDM2, PTEN, MMP1	5.43
GO:0034976∼response to endoplasmic reticulum stress	0.0383	DERL2, VCP, TP53, CCDC47	5.32

Gene expression profiling analysis

We used microarray-based expression profiling to identify the genes differentially expressed between Ob- and Co-hA-MSCs. We found that 241 genes were upregulated and 33 downregulated probes (Benjamini–Hochberg adjusted P-value <0.05, log₂FC >0.5 or <−0.5) in obese versus controls, some of which were targeted by miR-138-5p and/or miR-222-3p as predicted by miRanda, TargetScan, and miRTarbase (Supplementary Table S2). The differently expressed genes were functionally annotated with DAVID (Tables 5 and 6), and only those with a significant P-value (<0.05) were considered. The 20 significant biological processes involving upregulated genes included cytoskeleton organization, blood vessel development, and immune response, while those involving downregulated genes included steroid and lipid metabolism (Tables 5 and 6).

Table 5.

Significant Biological Processes (Enriched Terms), Including Differently Expressed Genes (Upexpressed) in Obese Versus Controls Human Amniotic Mesenchymal Stem Cells as Evaluated by Microarray

Term	P value	Genes	Fold enrichment
Enriched terms of upexpressed genes
GO:0007595∼lactation	0.0184	CAV2, CAV1, CREB1	14.21
GO:0010212∼response to ionizing radiation	0.0001	DCLRE1C, CAV1, XRCC2, FAM175A, XRRA1, C9ORF80	13.26
hsa05340:Primary immunodeficiency	0.0250	DCLRE1C, AIRE, RFXAP	11.78
GO:0070507∼regulation of microtubule cytoskeleton organization	0.0392	CAV2, CAV1, ROCK2	9.47
GO:0051495∼positive regulation of cytoskeleton organization	0.0444	CAV2, CAV1, TPM1	8.84
GO:0010638∼positive regulation of organelle organization	0.0241	CAV2, CAV1, ROCK2, TPM1	6.39
GO:0014706∼striated muscle tissue development	0.0121	CAV2, CAV1, TGFBR3, TPM1, GJC1	5.57
GO:0060537∼muscle tissue development	0.0143	CAV2, CAV1, TGFBR3, TPM1, GJC1	5.31
GO:0005925∼focal adhesion	0.0424	CAV2, CAV1, LIMS1, ARHGAP24	5.11
GO:0005924∼cell-substrate adherens junction	0.0467	CAV2, CAV1, LIMS1, ARHGAP24	4.92
GO:0009314∼response to radiation	0.0038	DCLRE1C, CAV1, XRCC2, FAM175A, MBD4, XRRA1, C9ORF80	4.64
GO:0001525∼angiogenesis	0.0248	IL18, CCBE1, ANGPT1, ARHGAP24, POFUT1	4.48
GO:0048514∼blood vessel morphogenesis	0.0050	CAV1, IL18, CCBE1, ANGPT1, ARHGAP24, POFUT1, GJC1	4.40
GO:0001568∼blood vessel development	0.0024	CAV1, IL18, CCBE1, TGFBR3, ANGPT1, ARHGAP24, POFUT1, GJC1	4.33
GO:0001944∼vasculature development	0.0027	CAV1, IL18, CCBE1, TGFBR3, ANGPT1, ARHGAP24, POFUT1, GJC1	4.23
GO:0051130∼positive regulation of cellular component organization	0.0466	CAV2, CAV1, ROCK2, TPM1, AHR	3.66
GO:0009628∼response to abiotic stimulus	0.0202	DCLRE1C, CAV1, XRCC2, FAM175A, IL18, MBD4, XRRA1, C9ORF80	2.88
GO:0019898∼extrinsic to membrane	0.0126	CAV2, GOLGA8B, ICA1, DAPP1, ROCK2, AAK1, TAOK3, YRDC, PLA2G4B, NKTR	2.64
GO:0009986∼cell surface	0.0489	ALCAM, CAV2, CAV1, ULBP2, TGFBR3, GREM1, ALPP	2.62
GO:0006955∼immune response	0.0138	LILRB1, DCLRE1C, CCR6, IL18, LILRB3, CD46, ULBP2, AIRE, LIME1, TGFBR3, TNFSF14, IL10	2.31
GO:0003677∼DNA binding	0.0316	ZNF430, ZNF486, XRCC2, C8ORF45, ZNF674, RFXAP, ZNF652, BLZF1, MCM8, ZNF738, ZNF223, PRDM10, ZNF394, LRRFIP1, ZNF577, ZNF69, SNAPC1, CREB1, HNRNPA2B1, MBD4, ZNF669, ZNF549, AHR, ZNF786, AIRE, DMC1, ZNF483	1.47

Table 6.

Significant Biological Processes (Enriched Terms), Including Differently Expressed Genes (Downexpressed) in Obese Versus Controls Human Amniotic Mesenchymal Stem Cells as Evaluated by Microarray

Term	P value	Genes	Fold enrichment
Enriched terms of downexpressed genes
hsa00100:Steroid biosynthesis	0.0007	DHCR7, LSS, DHCR24	69.03
GO:0016126∼sterol biosynthetic process	0.0011	DHCR7, LSS, DHCR24	57.98
GO:0006694∼steroid biosynthetic process	0.0062	DHCR7, LSS, DHCR24	23.87
GO:0008203∼cholesterol metabolic process	0.0073	DHCR7, LSS, DHCR24	22.06
GO:0016125∼sterol metabolic process	0.0087	DHCR7, LSS, DHCR24	20.09
GO:0016628∼oxidoreductase activity, acting on the CH-CH group of donors, NAD or NADP as acceptor	0.0303	DHCR7, DHCR24	62.12
GO:0008202∼steroid metabolic process	0.0321	DHCR7, LSS, DHCR24	10.05
REACT_602:Metabolism of lipids and lipoproteins	0.0455	DHCR7, LSS, DHCR24	7.55

Discussion

Experimental evidence obtained in animal models suggested that the placental epigenome is involved in gestational programming and that gestational programming of obesity results in offspring obesity and/or in obesity-related disorders [2,3]. Given the ban on experimentation with human embryo tissue, studies to evaluate whether intrauterine maternal obesity in humans is related to obesity in childhood or in adult life would entail long-term follow-up monitoring. Thus, given their capability to differentiate in adipocytes, hA-MSCs could represent a surrogate cellular model with which to investigate the impact of maternal obesity on the programming of adipose tissue dysfunctions. In the attempt to shed light on the role of miRNAs in human pregnancy during obesity, we used high-resolution small RNA-Seq to unravel the entire miRNA expression profiles of Ob- and Co-hA-MSCs.

Integrated bioinformatics based on a consensus method previously showed that the expression of miR-138-5p and miR-222-3p was significantly higher in Ob-hA-MSCs than in Co-hA-MSCs [20 –22]. Using quantitative RT-PCR, we confirmed the higher expression of these two miRNAs in Ob-hA-MSCs and that of miR-138-5p also in amnion from obese pregnant women, thus enlarging the panel of miRNAs previously found deregulated in this multicellular placental tissue [7]. Interestingly, the most significant pathways predicted to be deregulated by miR-138-5p and/or miR-222-3p included such obesity-associated functions and disorders as fat cell differentiation, fat deposits, lipid and carbohydrate homeostasis, stress response, metabolic syndrome, heart disease, and ischemia.

Transcriptome analysis of hA-MSCs highlighted a panel of differently expressed mRNAs in Ob- versus Co-hA-MSCs and some of them were targeted by miR-138-5p and/or miR-222-3p. The most significant mRNAs were engaged in cytoskeleton organization, vessel development, immune response, and steroid and lipid metabolism, all of which are involved in adipose tissue remodeling in obesity [36]. Thus, our hA-MSC miRNome data indicate that miRNAs could play a role in the adverse metabolic consequences associated with obesity.

Support for this hypothesis comes from our previous observation of miR-138-5p overexpression in subcutaneous adipose tissue and of miR-222-3p overexpression in both subcutaneous and visceral adipose tissues from obese adults [8,9]. Interestingly, both miR-138-5p and miR-222-3p are involved in the regulation of adipogenesis, obesity, and inflammation [37,38]. In particular, miR-222-3p was reported to be overexpressed in inflammation-induced adipocytes and macrophages, and in subcutaneous adipose tissue from obese adults [38]. Notably, activation of the inflammatory and immune systems, together with oxidative stress, is associated with abdominal obesity and has been implicated in the pathogenesis of obesity-related metabolic disorders [39,40]. miR-222-3p was also associated with adipose tissue oxidative stress, apoptosis, and inflammation, and endothelial dysfunction and adhesion, with infiltration of inflammatory cells into the endothelial space [38,41].

In addition, overexpression of miR-138 during obesity, induced by secreted proinflammatory cytokines, causes endothelial cell dysfunction, which is the most significant predisposing risk factor for cardiovascular disease [42]. In this context, it is noteworthy that our enrichment analysis of miR-138-5p targets resulted in the terms “stress response,” “heart disease,” and “ischemia.”

The prediction effect of miR-222-3p and miR-138-5p on deregulating gene expression in hA-MSCs during obesity is in agreement with the altered protein profile that we recently reported in the same cohort of hA-MSCs [43], which suggests that the response to stress and to inflammation is impaired in obesity. In particular, we found several heat shock proteins downexpressed, including HPSA8, which is a target of miR-222-3p [43].

Accordingly, in animal models, intrauterine maternal obesity has been shown to represent a stress state for embryo that leads to fetal programming of obesity and/or of obese-associated diseases [44]. In detail, maternal obesity was associated with increased oxidative stress and reduced antioxidants in placental and fetal tissues, thereby impairing mitochondrial biogenesis and glucose tolerance [44,45]. Furthermore, obesity-related alterations of oxidative stress markers have been demonstrated, both prenatally and postnatally, in offspring muscle, liver, pancreas, and adipose tissues [46].

In conclusion, our finding of miR-138-5p/miR-222-3p overexpression in Ob-hA-MSCs, together with the transcriptomic data, suggests that these miRNAs in obese pregnancy could derange metabolic pathways found to be impaired in tissues from obese adults and in obesity-associated disorders [37,38,41,42] and concur to modify gestational programming as has been demonstrated in animal models. Taken together, these findings raise the possibility of using diet-based strategies to normalize the perinatal miRNome in obesity.

Footnotes

Acknowledgments

The authors thank Jean Ann Gilder (Scientific Communication srl, Naples) for editing the text and Vittorio Lucignano, CEINGE–Biotecnologie Avanzate for technical assistance related to graphics. This work was supported by grant POR CAMPANIA FSER 2007–2013 Project DIAINTECH and grant CAMPUS-Bioframe, from Regione Campania, Italy; grant PON02_00677 (BIOGENE) Pot. lab.8 A/B-2012 from the Italian Ministry of University and Research. Centro di Ricerca per lo studio di malattie genetiche dell'uomo e loro modelli cellulari e animali–Progetto: Terapia genica delle malattie metaboliche.

Author Disclosure Statement

No competing financial interests exist.

References

Desai

, Jellyman

and Ross

. (2015). Epigenomics, gestational programming and risk of metabolic syndrome. Int J Obes (Lond), 39:633–641.

Alfaradhi

, Fernandez-Twinn

, Martin-Gronert

, Musial

, Fowden

and Ozanne

. (2014). Oxidative stress and altered lipid homeostasis in the programming of offspring fatty liver by maternal obesity. Am J Physiol Regul Integr Comp Physiol, 307:R26–R34.

Januar

, Desoye

, Novakovic

, Cvitic

and Saffery

. (2015). Epigenetic regulation of human placental function and pregnancy outcome: considerations for causal inference. Am J Obstet Gynecol, 213:S182–S196.

Arner

and Kulyté

. (2015). MicroRNA regulatory networks in human adipose tissue and obesity. Nat Rev Endocrinol, 11:276–288.

, Zhang

, Deng

, Miao

, Guo

, Gao

and Cui

. (2008). An analysis of human microRNA and disease associations. PLoS One, 3:e3420.

Bandiera

, Hatem

, Lyonnet

and Henrion-Caude

. (2010). microRNAs in diseases: from candidate to modifier genes. Clin Genet, 77:306–313.

Nardelli

, Iaffaldano

, Ferrigno

, Labruna

, Maruotti

, Quaglia

, Capobianco

, Di Noto

, Del Vecchio

, et al. (2014). Characterization and predicted role of the microRNA expression profile in amnion from obese pregnant women. Int J Obes (Lond), 38:466–469.

Martinelli

, Nardelli

, Pilone

, Buonomo

, Liguori

, Castanò I

, Buono

, Masone

, Persico

, et al. (2010). miR-519d overexpression is associated with human obesity. Obesity (Silver Spring), 18:2170–2176.

Capobianco

, Nardelli

, Ferrigno

, Iaffaldano

, Pilone

, Forestieri

, Zambrano

and Sacchetti

. (2012). miRNA and protein expression profile of visceral adipose tissue reveal miR-141/YWHAG and miR-520e/RAB11A as two potential miRNA/protein target pairs associated with severe obesity. J Proteome Res, 11:3358–3369.

10.

Pappa

and Anagnou

. (2009). Novel sources of fetal stem cells: where do they fit on the developmental continuum?. Regen Med, 4:423–433.

11.

Mattar

and Bieback

. (2015). Comparing the immunomodulatory properties of bone marrow, adipose tissue, and birth-associated tissue mesenchymal stromal cells. Front Immunol, 6:560.

12.

Huat

, Khan

, Abdullah

, Idris

and Jaafar

. (2015). MicroRNA expression profile of bone marrow mesenchymal stem cell-derived neural progenitor by microarray under the influence of EGF, bFGF and IGF-1. Genom Data, 5:201–205.

13.

Wang

, Gerstein

and Snyder

. (2009). RNA-Seq: a revolutionary tool for transcriptomics. Nat Rev Genet, 10:57–63.

14.

Parolini

, Alviano

, Bagnara

, Bilic

, Bühring

, Evangelista

, Hennerbichler

, Liu

, Magatti

, et al. (2008). Concise review: isolation and characterization of cells from human term placenta: outcome of the first International Workshop on Placenta Derived Stem Cells. Stem Cells, 26:300–311.

15.

Iaffaldano

, Nardelli

, Raia

, Mariotti

, Ferrigno

, Quaglia

, Labruna

, Capobianco

, Capone

, et al. (2013). High aminopeptidase N/CD13 levels characterize human amniotic mesenchymal stem cells and drive their increased adipogenic potential in obese women. Stem Cells Dev, 22:2287–2297.

16.

Goecks

, Nekrutenko

, Taylor

and Galaxy Team. (2010). Galaxy: a comprehensive approach for supporting accessible, reproducible, and transparent computational research in the life sciences. Genome Biol, 11:R86.

17.

Blankenberg

and Hillman-Jackson

. (2014). Analysis of next-generation sequencing data using Galaxy. Methods Mol Biol, 1150:21–43.

18.

Giardine

, Riemer

, Hardison

, Burhans

, Elnitski

, Shah

, Zhang

, Blankenberg

, Albert

, et al. (2005). Galaxy: a platform for interactive large-scale genome analysis. Genome Res, 15:1451–1455.

19.

Giurato

, De Filippo

, Rinaldi

, Hashim

, Nassa

, Ravo

, Rizzo

, Tarallo

and Weisz

. (2013). iMir: an integrated pipeline for high-throughput analysis of small non-coding RNA data obtained by smallRNA-Seq. BMC Bioinformatics, 14:362.

20.

, Lai

, Lehman

and Nelson

. (2013). miRDeep*: an integrated application tool for miRNA identification from RNA sequencing data. Nucleic Acids Res, 41:727–737.

21.

Hackenberg

, Sturm

, Langenberger

, Falcón-Pérez

and Aransay

. (2009). miRanalyzer: a microRNA detection and analysis tool for next-generation sequencing experiments. Nucleic Acids Res, 37,(Web Server issue)::W68–W76.

22.

Hall

, Frank

, Holmes

, Pfahringer

, Reutemann

and Witten

. (2009). The WEKA data mining software: an update. SIGKDD Explor, 11:10–18.

23.

Griffiths-Jones

, Saini

, van Dongen

and Enright

. (2008). miRBase: tools for microRNA genomics. Nucleic Acids Res, 36,(Database issue)::D154–D158.

24.

Anders

and Huber

. (2010). Differential expression analysis for sequence count data. Genome Biol, 11:R106.

25.

Robinson

, McCarthy

and Smyth

. (2010). edgeR: a bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics, 26:139–140.

26.

Bar-Joseph

, Gifford

and Jaakkola

. (2001). Fast optimal leaf ordering for hierarchical clustering. Bioinformatics, 17,(Suppl. 1)::S22–S29.

27.

Eisen

, Spellman

, Brown

and Botstein

. (1998). Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci U S A, 95:14863–14868.

28.

Bardou

, Mariette

, Escudié

, Djemiel

and Klopp

. (2014). jvenn: an interactive venn diagram viewer. BMC Bioinformatics, 15:293.

29.

Betel

, Wilson

, Gabow

, Marks

and Sander

. (2008). The microRNA.org resource: targets and expression. Nucleic Acids Res, 36(Database issue):D149–D153.

30.

Betel

, Koppal

, Agius

, Sander

and Leslie

. (2010). Comprehensive modeling of microRNA targets predicts functional non-conserved and non-canonical sites. Genome Biol, 11:R90.

31.

Lewis

, Burge

and Bartel

. (2005). Conserved seed pairing, often flanked by adenosines, indicates that thousands of human genes are microRNA targets. Cell, 120:15–20.

32.

Chou

, Chang

, Shrestha

, Hsu

, Lin

, Lee

, Yang

, Hong

, Wei

, et al. (2016). miRTarBase 2016: updates to the experimentally validated miRNA-target interactions database. Nucleic Acids Res, 44,(D1)::D239–D247.

33.

Huang da

, Sherman

and Lempicki

. (2009). Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc, 4:44–57.

34.

Ritchie

, Phipson

, Wu

, Hu

, Law

, Shi

and Smyth

. (2015). Limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 43:e47.

35.

Rasmussen

, Catalano

and Yaktine

. (2009). New guidelines for weight gain during pregnancy: what obstetrician/gynecologists should know. Curr Opin Obstet Gynecol, 21:521–526.

36.

Choe

, Huh

, Hwang

, Kim

and Kim

. (2016). Adipose tissue remodeling: its role in energy metabolism and metabolic disorders. Front Endocrinol (Lausanne), 7:30.

37.

Yang

, Bian

, Zhou

, Huang

, Wang

, Liao

and Zhao

. (2011). MicroRNA hsa-miR-138 inhibits adipogenic differentiation of human adipose tissue-derived mesenchymal stem cells through adenovirus EID-1. Stem Cells Dev, 20:259–267.

38.

Ortega

, Moreno

, Mercader

, Moreno-Navarrete

, Fuentes-Batllevell

, Sabater

, Ricart

and Fernández-Real

. (2015). Inflammation triggers specific microRNA profiles in human adipocytes and macrophages and in their supernatants. Clin Epigenet, 7:49.

39.

Chawla

, Nguyen

and Goh

. (2011). Macrophage-mediated inflammation in metabolic disease. Nat Rev Immunol, 11:738–749.

40.

Ouchi

, Parker

, Lugus

and Walsh

. (2011). Adipokines in inflammation and metabolic disease. Nat Rev Immunol, 11:85–97.

41.

Hulsmans

, De Keyzer

and Holvoet

. (2011). MicroRNAs regulating oxidative stress and inflammation in relation to obesity and atherosclerosis. FASEB J, 25:2515–2527.

42.

Sen

, Most

and Peppel

. (2014). Induction of microRNA-138 by pro-inflammatory cytokines causes endothelial cell dysfunction. FEBS Lett, 588:906–914.

43.

Capobianco

, Caterino

, Iaffaldano

, Nardelli

, Sirico

, Del Vecchio

, Martinelli

, Pastore

, Pucci

and Sacchetti

. (2016). Proteome analysis of human amniotic mesenchymal stem cells (hA-MSCs) reveals impaired antioxidant ability, cytoskeleton and metabolic functionality in maternal obesity. Sci Rep, 6:25270.

44.

Malti

, Merzouk

, Loukidi

, Karaouzene

, Malti

and Narce

. (2014). Oxidative stress and maternal obesity: feto-placental unit interaction. Placenta, 35:411–416.

45.

Kamel

, Helmy

, Hanafi

, Mahmoud

, Elfetooh

and Badr

. (2014). Maternal obesity and malnutrition in rats differentially affect glucose sensing in the muscles and adipose tissues in the offspring. Int J Biochem Res Rev, 4:440–469.

46.

Saad

, Abdelkhalek

, Haiba

, Saleh

, Hanafi

, Tawfik

and Kamel

. (2016). Maternal obesity and malnourishment exacerbate perinatal oxidative stress resulting in diabetogenic programming in F1 offspring. J Endocrinol Invest, 39:643–655.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.15 MB

0.03 MB

0.02 MB

0.03 MB