Background

Embryonic stem cell (ESC) research has made remarkable progress since the establishment of the first human embryonic stem cell line in 1998 [1]. The pluripotent nature of ESCs makes them valuable as a tool to model embryonic development and for regenerative medicine in vitro. They are also valuable as a cell resource for transplantation. However, the ethical issues surrounding the derivation of ESCs from embryos hinders the clinical application of ESCs and many countries limit or ban their use [2].

In 2006, Yamanaka brought pluripotent cell research into a new era by showing that over-expression of four key transcription factors, Oct4, Sox2, Klf4 and c-Myc, could reprogram mouse somatic cells into ESC-like cells that showed similar morphology and pluripotent nature to that of ESCs [3]. They named these ESCs-like cells “induced pluripotent stem cells” (iPSCs). Research into iPSCs has since proceeded at an astonishing pace and has included the establishment of human iPSCs and high efficiency induction of iPSCs with fewer transcription factors in combination with microRNAs (miRNAs) or small compounds [4–9]. With ongoing advances in miRNA biology, these findings may lead to a nonviral, nontranscription-factor mediated procedure for generating iPSCs for use not only in basic stem cell biology studies, but also in high throughput generation of human iPSC clones from large patient populations.

Despite the robustness of iPSCs technology, human somatic cell nuclear transfer (SCNT) research remains an important approach for regenerative medicine [10, 11]. The recent establishment of human pluripotent ESCs by SCNT has been long-anticipated as an approach for generating patient-matched nuclear transfer (NT)-ESCs for studies of disease mechanisms and for develo** specific therapies [12]. Since the initial discovery in amphibians in 1962, SCNT success in a range of different mammalian species has demonstrated that such reprogramming activity is universal [12–14]. Direct comparisons between iPSCs and NT-ESCs in the mouse indicated that SCNT-based reprogramming is more efficient in resetting the epigenetic identity of parental somatic cells [15, 27–30]. Interestingly, the miR-302 family was represented in group 4 and was highly expressed in iPSCs, indicating that iPSCs might be more highly dependent on miR-302 expression than pluripotent cells produced by the SCNT method.

It has been clearly shown that various types of cell vary not only in the expression of their coding genes, but also in the expression of their noncoding genes. In the present study, we compared differences in miRNAs expression between MEFs and ESCs and among MEF-derived iPSCs, NT-ESCs and NT-iPSCs to identify pluripotent specific miRNAs. The 50 top differentially expressed miRNAs were assigned to four clusters which were almost all highly expressed in pluripotent cells. Among them, miR-290 and miR-302 clusters were identified in previous studies to play key roles in pluripotency maintenance. The consistency of our results with those in the literature demonstrate the reliability of our sequence data. Furthermore, the target gene analysis showed that four miRNAs clusters mainly targeted genes involved in cancers and signal transduction pathways. The common characteristics of cell proliferation and immortalization are shared between cancer cells and pluripotent cells, as are activated signal transduction pathways.

Conclusions

In conclusion, we first report differentially expressed miRNAs among ESCs, MEFs and three kinds of reprogrammed cells. The unique expression of miRNAs in pluripotent cells mainly represents acquired expression of miRNAs, while the higher and lower expression levels of miRNAs in ESCs compared with reprogrammed cells may reflect the difference between naturally pluripotent cells and reprogrammed cells. Finally, the variation in miRNA expression among reprogrammed cells derived using different reprogramming strategies suggests different characteristics induced by nuclear transfer and iPSC generation, as well as different developmental potential among NT-ESCs, iPSCs and NT-iPSCs.

Methods

Cell culture

Mouse ES cell line (E14) was maintained in our lab. iPSCs were obtained by infecting MEFs (C57B6/129SvJae F1) with a dox-inducible lentivirus carrying the four reprogramming factors (Oct4, Sox2, Klf4 and c-Myc). NT-ESCs were established by reprogramming MEFs into ESCs using nuclear transfer. To establish NT-iPSCs, the nucleus of an iPSC was transferred into an enucleated oocyte. NT-iPSCs were established to reflect the combination of nuclear transfer and iPS technologies. iPSCs, NT-ESCs, and NT-iPSCs were derived from the same MEF cells. All the cells were cultured and maintained as described previously [16]. All experiments were approved by the Ethics Committee of Shanghai Institute of Biochemistry and Cell Biology.

RNA preparation and sequencing

Total RNA was isolated using TRIzol reagent. RNA quality was assessed with an Agilent 2100 bioanalyzer. Total RNAs from MEFs, ESCs and the three reprogrammed cell types were subjected to Solexa sequencing, performed by BGI-Shenzhen, Shenzhen, China. The sequencing data have been deposited with the Gene Expression Omnibus repository (http://www.ncbi.nlm.nih.gov/geo) under accession number GSE52950.

Data analysis and statistics

After removal of adaptors, low quality tags and contaminants from the sequenced tags, clean reads were annotated. MiRNA reads were analyzed using the DESeq package [31] in R language [32]. Normalization and variance stabilizing transformations (VST) were performed before further analysis. Differently expressed miRNAs and sample distances between any two kinds of samples were calculated by DESeq. MiRNAs with variance stabilizing transformed values of more than 10 were clustered using the gplots package [33].

We present a hypothesis that miRNAs contributing to pluripotency should meet at least two criteria. First, these miRNAs should be highly expressed in ESCs and expressed at lower levels in MEF cells. Second, these miRNAs should also show relatively high levels of expression in iPSCs. Based on these criteria, the expression profiles of the top 50 miRNAs that were more highly expressed in ESCs than in MEF cells were grouped by k-means clustering using the Vegan package [34]. The genome context of each miRNA was extracted from miRbase [35–38] and presented using the Ensembl Genome Browser. Target genes of miRNAs were predicted and enriched in KEGG pathways using mirPath [39]. The enrichment results are presented in a “bubble plot” using the ggplot2 package [40].

To identify iPSC-specific miRNAs, the top 50 miRNAs differently expressed between ESCs and each reprogrammed cell type were screened and 34 miRNAs that were commonly differentially expressed between ESCs and all reprogrammed cell types were identified as iPSC-specific miRNAs. These miRNAs were grouped by k-means clustering. The target genes of these miRNAs were also mapped to KEGG pathways.

To identify differentially expressed miRNAs in the three reprogrammed cell types, miRNAs with a VST value more than 10 in at least one reprogrammed cell type were analyzed. MiRNAs with an adjusted p value less than 0.05 (ANOVA) were identified as differentially expressed miRNAs in these different reprogrammed cells, and were grouped by k-means clustering using the Vegan package.