Both Copies of the Nanog Gene Are Expressed in Embryonic Stem Cells


Commonly held ideas are sometimes held because there is a great of evidence to substantiate them. However, other times, an idea is commonly held because simply because it has been repeated over and over and over even though the evidence for it is poor. Thus, when new evidence come to light showing the commonly held believe to be untrue, it becomes incumbent on us to readjust what we think.

When it comes to embryonic stem cells and the genes that keep them pluripotent, the transcription Nanog plays a very critical role in the self-renewal of embryonic stem cells and there is a great deal of evidence for this assertion. However, the expression of the gene that encodes Nanog was thought to follow the same mode of expression as some of the other pluripotency promoting genes. Namely, that only one of the copies of the Nanog gene were thought to be expressed in embryonic stem cells. This turns out to be probably false.

First a little background. In 2007, Ian Chambers and others published a paper in the journal Nature that examined the expression and function of Nanog in embryonic stem cells. Chambers and others found that Nanog expression levels in individual embryonic stem cells from a culture derived from a single cell varied wildly.  The figure from the Chambers et al paper is shown below.

Immunofluorescence of TNG cells for Oct4 and Nanog. Individual signals from 4,6-diamidino-2-phenylindole (DAPI), GFP, anti-Oct4 and anti-Nanog are shown on the left alongside a combined view of GFP with the stainings from anti-Oct4 and anti-Nanog.
Immunofluorescence of TNG cells for Oct4
and Nanog. Individual signals from 4,6-diamidino-2-phenylindole (DAPI),
GFP, anti-Oct4 and anti-Nanog are shown on the left alongside a combined
view of GFP with the stainings from anti-Oct4 and anti-Nanog.

The reason for this fluctuation in Nanog levels was uncertain, but Chambers and others showed that Nanog could be deleted from mouse embryonic stem cells without affecting their ability to contribute to various sundry embryonic tissues during mouse development, even though they do not make functional gametes (eggs and sperm).  In fact, mouse embryonic stem cells can self-renew under particular conditions without a functional copy of the Nanog gene even though they are prone to differentiation.  From this, Chambers and others concluded that Nanog stabilized rather than promoted pluripotency of embryonic stem cells by “resisting or reversing alternative gene expression states.”

Fast forward to 2012 and another Nature paper by Yusuke Miyanari and Maria-Elena Torres-Padilla from the IGBMC in Strasbourg, France, which showed that before mouse embryos implanted into the uterus, only one copy of the Nanog gene was expressed, but after implantation, both copies of the Nanog gene was expressed.  Miyanari and Torres-Padilla also made mouse embryonic stem cells that had copies of the Nanog gene labeled with different glowing proteins.  This ingenious experiment showed confirmed that Nanog levels were variable, but also showed that only one copy of the Nanog gene was expressed in growing embryonic stem cells in culture.

a, Schematic of the Nanog knock-in reporter NGR. A PEST motif in the carboxy terminus of the fluorescent proteins allows monitoring of dynamic Nanog expression. iHyg, internal ribosome entry site (IRES) hygromycin; iNeo, IRES neomycin; mChe, mCherry; NLS, nuclear localization signal; tGFP, TurboGFP. b, Representative image of NGR ES cells cultured with LIF or 2i/LIF. Scale bar, 10 µm. c, The incidence of allelic switching of Nanog expression in ES cells. Cells were classified into four groups: monoallelic (TurboGFP-positive, green), monoallelic (mCherry-positive, red), biallelic (TurboGFP- and mCherry-positive, yellow) and no expression (black). The proportion of cells undergoing a transition between these four groups during a single cell cycle is indicated. Overall, 47% of cells showed a colour change in this period. n, number of cells analysed. d, The asymmetric replication of Nanog in ES cells cultured with LIF changes to symmetric replication upon treatment with 2i. The cell nuclei were classified as single/double (SD), single/single (SS) and double/double (DD) according to DNA-FISH signals5. n, number of nuclei analysed. *, P < 4 × 10−7; **, P < 1.4 × 10−3 (Fisher’s exact test). e, Representative image of DNA-FISH for Nanog (arrowheads) and Oct4 in ES cells cultured with LIF or 2i/LIF. Scale bar, 2 µm. f, Nanog allelic expression is unaffected in the absence of DNA methyltransferase activity. Quantification of RNA-FISH for Nanog in wild-type (WT) ES cells and ES cells lacking all three DNA methyltransferases (TKO) cultured with LIF or 2i/LIF. g, ChIP for H3K4me3, MED12 or NIPBL along the Nanog locus (black line, top) in ES cells cultured with LIF or 2i/LIF. The position of the ChIP amplicons is depicted by the thick boxes below the line, the TSS by an arrow, the first exon by the black box on the line, and the distal enhancer by the blue box on the line. The Oct4 promoter region (Oct4 Pro) and distal enhancer (Oct4 DE) were positive controls (right)20. The mean ± s.d. of three independent biological replicates is shown.
a, Schematic of the Nanog knock-in reporter NGR. A PEST motif in the carboxy terminus of the fluorescent proteins allows monitoring of dynamic Nanog expression. iHyg, internal ribosome entry site (IRES) hygromycin; iNeo, IRES neomycin; mChe, mCherry; NLS, nuclear localization signal; tGFP, TurboGFP. b, Representative image of NGR ES cells cultured with LIF or 2i/LIF. Scale bar, 10 µm. c, The incidence of allelic switching of Nanog expression in ES cells. Cells were classified into four groups: monoallelic (TurboGFP-positive, green), monoallelic (mCherry-positive, red), biallelic (TurboGFP- and mCherry-positive, yellow) and no expression (black). The proportion of cells undergoing a transition between these four groups during a single cell cycle is indicated. Overall, 47% of cells showed a colour change in this period. n, number of cells analysed. d, The asymmetric replication of Nanog in ES cells cultured with LIF changes to symmetric replication upon treatment with 2i. The cell nuclei were classified as single/double (SD), single/single (SS) and double/double (DD) according to DNA-FISH signals5. n, number of nuclei analysed. *, P < 4 × 10−7; **, P < 1.4 × 10−3 (Fisher’s exact test). e, Representative image of DNA-FISH for Nanog (arrowheads) and Oct4 in ES cells cultured with LIF or 2i/LIF. Scale bar, 2 µm. f, Nanog allelic expression is unaffected in the absence of DNA methyltransferase activity. Quantification of RNA-FISH for Nanog in wild-type (WT) ES cells and ES cells lacking all three DNA methyltransferases (TKO) cultured with LIF or 2i/LIF. g, ChIP for H3K4me3, MED12 or NIPBL along the Nanog locus (black line, top) in ES cells cultured with LIF or 2i/LIF. The position of the ChIP amplicons is depicted by the thick boxes below the line, the TSS by an arrow, the first exon by the black box on the line, and the distal enhancer by the blue box on the line. The Oct4 promoter region (Oct4 Pro) and distal enhancer (Oct4 DE) were positive controls (right)20. The mean ± s.d. of three independent biological replicates is shown.

Now if we fast forward one more year and a paper from the journal Cell Stem Cell and a letter to the same edition of this journal, we have an article by Dina Faddah and others from the laboratory of Rudolf Jaenisch at the Whitehead Institute (MIT, Cambridge, MA), and a supporting letter from Adam Filipczyk and others from Germany and Switzerland.  In this article, the authors also double-labeled mouse embryonic stem cells and examined multiple cells and showed that BOTH copies of Nanog were expressed, and that the range of variability of Nanog expression was approximately the same as other pluripotency genes.

Filipczyk and others used a similar approach to examine the expression of Nanog in mouse embryonic stem cells and they came to the same conclusions as those of Faddah and others.

What is the reason for the differences in findings?  Faddah and others did an important experiment to answer this question.  Some of the cells that with labeled copies of the Nanog gene disrupted the production of a functional Nanog protein.  The constructs used in the papers by Faddah and others and by Filipczyk and others did not disrupt Nanog protein production.  When Faddah and others tested these other constructs that disrupted Nanog protein production to determine is the amount of glowing protein tracked with the amount of Nanog protein produced, it was clear that the amount of Nanog protein made by the cells did not reflect the amount of glowing protein produced.  According to Dina Faddah, “The way the reported was inserted into the DNA seems to disrupt the regulation of the alleles so that when the reported said Nanog isn’t being expressed, it actually is.”

Jaenisch sees this as an instructional tale for all stem cell scientists.  He noted: “Clearly, the conclusions for this particular gene need to be reconsidered.  And it raises the question for other genes.  For some genes, there might be similar issues.  For other genes, they might be more resistant to this type of disturbances caused by a reporter.”

Bottom line – read the materials and methods part of the paper carefully because the way these experiments are done can determine if the results are trustworthy.

One-Step iPS cells


Three years ago, Kyoto University scientist Shinya Yamanaka made induced pluripotent stem cells (iPSCs) by inserting four genes into skin cells from laboratory mice. To get these genes into cells he used genetically engineered retroviruses, which insert copies of genes directly into the chromosomes of the cells they infect.  In order to convert the skin cells into iPSCs, he had to infect the cells with four different recombinant viruses, and subject the cells to four different insertion events.  Because retroviruses randomly insert into the gene, there is the possibility that they could insert their DNA right in the middle of an active gene.  Thus Yamanaka’s procedure left people worried about making mutations in tumor suppressor genes or other genes necessary for life.  If iPSCs were going to be used for clinical purposed, then how could such a procedure be used in the clinic if it was so potentially dangerous?

Now two labs have shown that you can do the same experiment by inserting one piece of DNA to do the same job. Rudolf Jaenisch’s lab at MIT and Konrad Hochedlinger at Harvard University have combined the four mouse reprogramming genes onto a single piece of DNA, known as a cassette, and inserted it at a single locus in the mouse genome. The mice with the insert were bred, and their somatic cells were transformed into iPS cells following the addition of the antibiotic doxycycline, which triggers the cassette to express the four reprogramming genes. This innovation saves time and money and brings iPSCs one step closer to clinical trials.

One group of experiments that scientists hope to do with these newly made mouse strains is test the ability of iPSCs to form particular cell types versus traditionally made embryonic stem cells (ESCs).  In a study published earlier this year in Cell Stem Cell, hundreds of genes are differentially expressed between iPSCs and ESCs (Chin, M. H. et al. (2009) Cell Stem Cell 5, 111-123).  Another study in Nature showed that iPS cells are not as efficient as embryonic stem cells at differentiating into all cell types (Zhao, X.-Y. et al. (2009) Nature 461, 86-90).