Sex-Comparative Analysis of the miRNome of Human Amniotic Mesenchymal Stem Cells During Obesity

Experimental evidence indicates differences between women and men in medical areas such as drug response, susceptibility to breast cancer, and autoimmune, neurodegenerative, and metabolic diseases. Two X chromosomes and hormonal and environmental factors all contribute to sexual dimorphism by influencing gene expression, also through microRNA (miRNA)-regulated epigenetic mechanisms [1]. miRNAs are small noncoding RNA molecules that regulate gene expression by inhibiting translation or degradation of their mRNA target. miRNA expression and functions are tissue specific and these molecules have been implicated in physiological and pathological cellular processes such as the fetal programming of obesity, mostly investigated in animal models [2]. We previously found significant alterations of the miRNome and proteome in at-term placentas and in human amniotic mesenchymal stem cells (hA-MSCs) from obese (Ob-) versus normal weight (Co-) women [3,4]. The aim of this study was to look for gender-related differences in the miRNA regulation of gene expression in Ob-hA-MSCs that might be involved in metabolic changes. To this aim, we compared the miRNomes of 7 Co-hA-MSCs and 13 Ob-hA-MSCs from women who gave birth to boys (3 M-Co, 6 M-Ob-) or girls (4 F-Co-, 7 F-Ob-).

The hA-MSCs-miRNome was analyzed by small RNA deep sequencing (Small RNA-Seq) (Supplementary Data; Supplementary Data are available online at www.liebertpub.com/scd). Most miRNAs were similarly expressed in the Co-hA-MSCs regardless of gender. One exception was miR-125b-5p, which was significantly up-expressed (P < 0.04) in F-Co-hA-MSCs versus M-Co-hA-MSCs. Small RNA-Seq comparisons, after multiple test correction, showed 15 (6 up- and 9 down-expressed) and 9 (4 up- and 5 down-expressed) differently expressed miRNAs in F-Ob-hA-MSCs versus F-Co-hA-MSCs and in M-Ob-hA-MSCs versus M-Co-hA-MSCs, respectively (P < 0.01) (Fig. 1A, B). In terms of gender differences in the Ob-hA-MSC miRNomes, 13 miRNAs were down-expressed and miR-146a was up-expressed in F-Ob-hA-MSCs versus M-Ob-hA-MSCs (P < 0.01) (Fig. 1C).

OPEN IN VIEWER

FIG. 1.

Gender-related comparison of the miRNomes of hA-MSCs isolated from normal weight and obese women who gave birth to males or females. miRNA expression was obtained by small RNA deep sequencing (Supplementary Data). Differentially expressed (DE) miRNAs both from edgeR and Fisher's test were selected. The distribution of DE miRNA levels in all the samples is indicated as log10 of normalized read counts. The boxes represent the median (black middle line) limited by the 25th (Q1) and 75th (Q3) percentiles. The whiskers are the upper and lower adjacent values, which are the most extreme values within Q3 + 1.5 (Q3−Q1) and Q1-1.5 (Q3−Q1), respectively. The black dots represent outliers. (A) Read counts distribution of DE miRNAs in hA-MSCs from Ob- versus Co-women, indicated by gray and white boxes, respectively, who gave birth to females (F-hA-MSCs). Six miRNAs were up-expressed and nine down-expressed. (B) Read counts distribution of DE miRNAs in hA-MSCs from Ob- versus Co-women, indicated by gray and white boxes, respectively, who gave birth to males (M-hA-MSCs). Four miRNAs were up-expressed and five down-expressed. (C) Read counts distribution of DE miRNAs in hA-MSCs from obese women (Ob-hA-MSCs) who gave birth to females (F) and males (M), indicated by white and gray boxes respectively. Only miR-146a-5p was up-expressed and 13 miRNAs were down-expressed between F-Ob- versus M-Ob-hA-MSCs. Interestingly, 2 of 13 down-expressed miRNAs in F-Ob-hA-MSCs versus M-Ob-hA-MSCs (miR-3182 and miR-7641) were up-expressed in M-Ob-hA-MSCs versus M-Co-hA-MSCs, which supports the gender-related miRNA expression association with obesity. (D) Semantic similarity-based graph (http://revigo.irb.hr) reporting the top scoring enriched gene ontology (GO) biological processes of the down- and up-expressed miRNA targets of F-Ob-hA-MSCs versus M-Ob-hA-MSCs. Positioning in the semantic space indicates semantic and, hence, functional similarity of GO terms. Colors of the nodes are scaled in accordance with the P value of the enrichment (darker the color, lower is the P value). (D-1) Down-expressed miRNAs are predicted to derange three annotation clusters: (i) the transcriptional and post-transcriptional RNA machinery, together with chromatin remodeling and protein metabolisms, (ii) the mitotic phase of the cell cycle, and (iii) the response to DNA damage. (D-2) Over-represented GO terms of the targets of the only one up-expressed miRNA, namely miR-146a, in F-Ob- versus M-Ob-hA-MSCs involve immune and inflammatory response. GO terms of the targets of miRNAs depicted in panels A and B (comparison: F-Ob-hA-MSCs vs. F-Co-hA-MSCs and M-Ob-hA-MSCs vs. M-Co-hA-MSCs, respectively) are reported in Supplementary Table S1. hA-MSCs, human amniotic stem cells; miRNA, microRNA.

Analysis of the gene targets of the gender-deregulated miRNAs predicted miR-146-mediated alterations in inflammatory and immune responses. Down-expressed miRNAs may impair three major metabolic clusters: (i) transcriptional and post-transcriptional RNA machinery, together with chromatin remodeling and protein metabolisms, (ii) mitotic phase of cell cycle, and (iii) response to DNA damage (Fig. 1D and Supplementary Table S1).

Thus, in healthy pregnancy, the miRNome of hA-MSCs is not affected by the gender of the offspring, except for the androgen-regulated miR-125b-5p that is up-expressed in F-Co- versus M-Co-hA-MSCs. In contrast, we observed that in hA-MSCs from obese women, altered miRNA expression was more frequent when the offspring was female as opposed to male (15 vs. 9 deregulated miRNAs, respectively). Notably, some deregulated miRNAs are under direct estrogen control (miR-21-5p/miR-92b-3p/miR-31-5p) or located on chromosome X (miR-221-3p/miR-222-3p). In a previous study, we found that several proteins from predicted impaired pathways were targeted by gender-specific miRNAs and were deregulated in the proteome of Ob- versus Co-hA-MSCs [4] (Supplementary Table S2).

In conclusion, although the postpartum implications of the gender-related miRNA differences identified in Ob-hA-MSCs are far from being completely elucidated, we provide the first evidence that the obesogenic environment can affect fetal hA-MSC epigenetics in a gender-specific manner, which also implies that caution be exerted in interpreting miRNA expression data when gender is not reported.