Induction of microRNA resistance and secretion in differentiating human endometrial stromal cells

Dear Editor, A key event in the preparation for pregnancy is the differentiation of human endometrial stromal cells (hESCs) into epithelioid decidual cells. This process, termed decidualization, is initiated in the stromal cells of the superficial endometrial layer 9 days after ovulation. Upon decidualization, hESCs acquire the ability to resist oxidative stress, to modulate immune responses to fetal alloantigens, and to control trophoblast invasion. Differentiating primary hESCs recapitulate many changes in gene expression observed upon decidualization in vivo with progesterone and cAMP signaling pathways cooperating in hESCs at several levels [see review (Gellersen and Brosens, 2003)]. MicroRNAs (miRNAs) are a diverse class of non-coding small RNAs, which posttranscriptionally regulate gene expression by interacting with sites in the 3′UTRs of mRNAs. miRNA synthesis begins with the transcription of miRNA genes to produce primary miRNAs (pri-miRNAs). In the nucleus, the pri-miRNAs are cleaved by the Drosha/DGCR8 complex to produce 70 nt pre-miRNAs. Exportin-5 transfers pre-miRNAs to the cytoplasm where Dicer removes the loop to give duplex mature miRNA of 22 nt. Argonaute proteins facilitate the miRNA–mRNA interactions and mediate silencing. In addition to their diverse roles in the parental cell, miRNAs can act as paracrine or endocrine signals and are present in the plasma as well as in the conditioned medium of cultured cells (Kosaka and Ochiya, 2011). Endometrial miRNA expression is cycle dependent and regulated by ovarian hormones (Pan et al., 2007). With miRNA microarray analysis we identified 16 significantly (P , 0.05) differentially expressed miRNAs in primary hESCs that had been vehicle treated or decidualized with 8-BrcAMP and medroxyprogesterone acetate (C + M) for 8 days (Supplementary Table S1). Real-time qPCR was used to monitor the expression of four miRNAs (miR-29b, miR-29c, miR-100, and miR-143) in independent primary hESC cultures treated either with vehicle or with C + M for 2, 4, or 8 days. These miRNAs were selected for study as they showed the most significant changes in expression and were likely to target genes important for decidual function including DNA methyltransferase 3b (DNMT3B), Tribbles 2 (TRIB2) and prokineticin 1 (PROK1). Figure 1A shows that the expression of miR-29b, miR-29c, and miR-100 increases as expected by 8 days of C + M treatment. In contrast, the levels of mature miR-143 gradually declined, which is also in keeping with the array data. The changes in the pri-forms of the four miRNAs during decidualization largely mirrored the changes observed in mature miRNA levels (Supplementary Figure S1), indicating that transcriptional regulation of miRNA genes occurs during decidualization. Next, we examined whether key components of the miRNA synthesis pathway are altered upon decidualization of hESCs. Parallel primary cultures treated with C + M for various time points were harvested for total protein and RNA. The levels of Drosha, Dicer, Ago1 and Ago2 were assessed at the protein and transcript levels (Figure 1B and Supplementary Figure S2). Drosha levels modestly decline in response to C + M treatment. Conversely, Dicer is up-regulated and again this response seemed more pronounced at the protein level. Strikingly, C + M treatment profoundly inhibited the abundance of Ago1 and Ago2, especially at the later time points. Densitometric analysis comparing three independent samples from day 0 and day 8 confirmed the changes observed (Figure 1B). siRNA-mediated knockdown of Dicer did not block the induction of the classical decidual marker genes prolactin (PRL) and insulin-like growth factor binding protein 1 (IGFBP1) during decidualization (Supplementary Figure S3). The robust induction of these genes indicated that Dicer-depleted hESCs are still able to differentiate upon treatment with C + M. This result is in keeping with the data from the conditional Dicer KO mouse where female mice lacking Dicer in the reproductive tract were infertile due to defective oviducts but exhibited histologically normal decidual responses (Nagaraja et al., 2008). Although Dicer knockdown did not block hESC differentiation, we speculated that regulated miRNAs target and modulate the expression of specific decidual genes. An experimentally validated target of the miR-29 family is DNMT3B (Fabbri et al., 2007). To test if DNMT3B expression is inhibited upon decidualization as a consequence of miR-29b/c induction, primary cultures were first transfected with anti-miR-29b/c and then treated with vehicle or C + M. As shown in Figure 1C, DNMT3B expression is indeed down-regulated upon decidualization at the protein level. Furthermore, the transfection of anti-miR-29b/c enhanced DNMT3B protein levels in undifferentiated cells, although this response was entirely lost upon treatment with C + M. In decidualizing hESCs, the effect of anti-miR-29b/c transfection on DNMT3B expression was indistinguishable from transfection negative control anti-miR or non-targeting anti-miR100. Similar findings were also observed with respect to DNMT3B mRNA levels, for TRIB2, a predicted target of both miR-29b/c and miR-100, and for PROK1, a predicted target of miR-100 (Supplementary Figure S4). We employed a reporter assay to monitor the generic miRNA activity in both doi:10.1093/jmcb/mjs058 Journal of Molecular Cell Biology (2013), 5, 67–70 | 67 Published online October 25, 2012