The fusion of somatic cells with pluripotent cells results in the generation of pluripotent hybrid cells. Because the `memory' of somatic cells seems to be erased during fusion-induced reprogramming, genetic reprogramming is thought to be a largely unidirectional process. Here we show that fusion-induced reprogramming, which brings about the formation of pluripotent hybrids, does not always follow a unidirectional route. Xist is a unique gene in that it is reprogrammed to the state of somatic cells in fusion-induced pluripotent hybrids. In hybrids formed from the cell fusion of embryonal carcinoma cells (ECCs) with male neural stem cells (mNSCs), the Xist gene was found to be reprogrammed to the somatic cell state, whereas the pluripotency-related and tissue-specific marker genes were reprogrammed to the pluripotent cell state. Specifically, Xist is not expressed in hybrids, because the `memory' of the somatic cell has been retained (i.e. mNSCs do not exhibit Xist expression) and that of the pluripotent cell erased (i.e. inactivation of the partially active Xist gene of ECCs, complete methylation of the Xist region). The latter phenomenon is induced by male, but not by female, NSCs.

Introduction

It has been suggested that the fusion of embryonic stem cells (ESCs) with somatic cells is a unidirectional process giving rise to hybrid cells that retain the phenotype of ESCs but not that of somatic cells (Silva et al., 2006). Therefore, the somatic cell state seems to be a nonviable state after cell fusion. The phenotypical characteristics of ESC–somatic-cell fusion-hybrid cells include the morphological features of the pluripotent fusion partner as well as a specific epigenetic state and gene-expression profile; inactivation of tissue-specific genes expressed in the somatic cell-fusion partner; and a differentiation potential characteristic of ESCs (Do and Scholer, 2006). Investigations into the characteristics of the hybrids formed by the fusion of embryonic germ cells (EGCs) or embryonal carcinoma cells (ECCs) with somatic cells have shown that EGCs and ECCs also have the capacity to reprogram somatic cells to the pluripotent phenotype (Do et al., 2007; Tada et al., 1997). Thus, studies investigating the fusion of pluripotent cells with somatic cells have consistently demonstrated that the reprogramming process is unidirectional. That is, pluripotent cells effectively reprogram somatic cells, abolishing the gene-expression patterns of the somatic cell and resetting the gene-expression profile to that of the pluripotent state. However, the gene-expression profile of the fusion-hybrid cells is similar, but not identical, to that of the pluripotent cell-fusion partner, indicating that fusion-induced reprogramming is a unidirectional process that occurs concomitantly with the incomplete reprogramming of certain genes (Ambrosi et al., 2007). Therefore, fusion-induced reprogramming does not necessarily result in the complete genetic reprogramming of somatic cells to a phenotype identical to that of the pluripotent cell-fusion partner. Rather, the fusion-hybrid cells might still retain residual `memory' of the somatic cells, although this somatic memory might not affect the characteristics of the pluripotent cell-fusion partner of the pluripotent fusion-hybrid cells. Moreover, it is still possible that fusion-induced reprogramming is not a solely unidirectional process, because somatic cells might be capable of inducing alterations in the characteristics of their pluripotent cell-fusion partners. Here, we challenged the dogma that fusion-induced pluripotential reprogramming is a unidirectional process and we assessed the possibility that somatic cells can also induce alterations in the characteristics of their pluripotent cell-fusion partners.

Results

Fusion between ECCs and ESCs results in two different types of hybrid cells

As a first step in investigating whether fusion-induced reprogramming is solely unidirectional, we fused ESCs with ECCs – two different types of pluripotent cells with the potential to reprogram somatic cells (Do et al., 2007; Do and Scholer, 2004). Specifically, we fused OG2 ESCs (containing an Oct4-GFP transgene) with β-geoF9 ECCs (hereafter referred to as F9 ECCs; containing a neo/lacZ transgene) to obtain selectable fusion-hybrid cells. On day 2 post-fusion, GFP-positive cells were sorted by fluorescence-activated cell sorting (FACS) and cultured in ESC medium containing G418 on feeder layers. After 10 days in culture with selection medium, the majority of cells had formed ECC-like colonies (141/143; 98.6%), but a small minority had formed ESC-like colonies (2/143; 1.4%) (Fig. 1A). ECC- and ESC-like hybrid cells were cloned and cultured on gelatin-coated dishes or on feeder layers, respectively. ECC- and ESC-like hybrid cells stained positive for both GFP (derived from an OG2 ESC) and X-gal (derived from an F9 ECC). In addition, cells making up the ECC- and ESC-like colonies were tetraploid in nature (Fig. 1B), indicating that they were all ECC-ESC fusion-hybrid cells. Therefore, the vast majority of hybrids produced by the fusion of pluripotent ECCs with ESCs were reprogrammed morphologically to ECCs and only a minority to ESCs. In addition, the differentiation potential of ECC- and ESC-like hybrid cells was similar to that of F9 ECCs (rarely differentiating into cells of neuronal lineage) and ESCs, respectively (Fig. 2A,B). In the presence of retinoic acid (RA), F9 ECCs preferentially differentiate into cells of primitive or parietal endoderm but not to those of the neural lineage (Do et al., 2007). Therefore, the poor neural-differentiation potential of the ECC-like hybrid cells suggested that they might be a suitable model to investigate whether ECC-like hybrid cells exhibit the differentiation characteristics of F9 ECCs. Therefore, we assayed the ECC-like hybrid cells for the presence of an endodermal marker (Sox17) and an ectodermal marker (Tuj1), because F9 cells normally do not differentiate into cells of the neuronal lineage (Tuj1-positive cells). As expected, ECC-like hybrid cells preferentially differentiated into endodermal cells (8165±1063/10,000 FACS-sorted cells), similar to the ESC-like hybrid cells (2038±312/10,000 FACS-sorted cells). Moreover, ECC-like hybrid cells rarely differentiated into neurons (37±12/10,000 FACS-sorted cells), similar to the ESC-like hybrid cells (3257±10,000 FACS-sorted cells).

Fig. 1.

Characterization of ECC-ESC hybrid cells. (A) Hybrid cells form two different types of colonies (ECC-like and ESC-like) on feeder layers. (B) ECC-like and ESC-like hybrid cells were observed to stain positive for both GFP (OG2-ESC derived) and X-gal (β-geoF9-ECC derived), and to exhibit tetraploidy. Scale bars: 25 μm. (C) Real-time RT-PCR analysis of the hybrid cells for pluripotency-related and germ-cell-marker gene expression. y-axis values are on a logarithmic scale, and each minor gridline is one tenth the value of each major gridline.

Fig. 1.

Characterization of ECC-ESC hybrid cells. (A) Hybrid cells form two different types of colonies (ECC-like and ESC-like) on feeder layers. (B) ECC-like and ESC-like hybrid cells were observed to stain positive for both GFP (OG2-ESC derived) and X-gal (β-geoF9-ECC derived), and to exhibit tetraploidy. Scale bars: 25 μm. (C) Real-time RT-PCR analysis of the hybrid cells for pluripotency-related and germ-cell-marker gene expression. y-axis values are on a logarithmic scale, and each minor gridline is one tenth the value of each major gridline.

Fig. 2.

Immunofluorescence staining of endodermal- and neural-specific protein in differentiated ECC- and ESC-like fusion hybrids. (A) ECC-like hybrid cells preferentially differentiate into the endodermal cell lineage (Sox17-positive). (B) ECC-like hybrid cells rarely differentiate into TuJ1-positive neural cells (at a frequency of ∼1/50,000). Scale bars: 25 μm.

Fig. 2.

Immunofluorescence staining of endodermal- and neural-specific protein in differentiated ECC- and ESC-like fusion hybrids. (A) ECC-like hybrid cells preferentially differentiate into the endodermal cell lineage (Sox17-positive). (B) ECC-like hybrid cells rarely differentiate into TuJ1-positive neural cells (at a frequency of ∼1/50,000). Scale bars: 25 μm.

Reprogramming in ECC- and ESC-like hybrid cells is not a solely unidirectional process with respect to germ-cell-marker expression pattern

ESCs and ECCs share similar overall gene-expression profiles, but they differ in their absolute transcript levels. To confirm that the ECC- and ESC-like cells indeed possess the characteristics of ECCs and ESCs, respectively, we determined the transcript levels of pluripotency-related and germ-cell-marker genes; these latter genes are expressed in undifferentiated ESCs. Three hybrid clones of ECC- and ESC-like cells were used for this analysis. Although there were no striking differences in the levels of the pluripotency-related genes Oct4, Nanog, Sox2, Foxd3, Fgf4 and Klf4 (Takahashi and Yamanaka, 2006) between OG2 ESCs and F9 ECCs, there were striking differences in the expression of the germ-cell-marker genes Stella, Dazl, Stra8, Vasa and Fragilis (more than fivefold, except for Fragilis), indicating that the expression of germ-cell-marker genes is a feature that distinguishes ESCs from ECCs (Fig. 1C). The expression of Stella and Dazl in ECC-like hybrid cells is comparable to that of F9 ECCs, and the expression of Stella and Dazl in ESC-like hybrid cells is comparable to that of ESCs, suggesting that ESC- and ECC-like cells possess the characteristics of ESCs and ECCs, respectively. However, the expression of Stra8 and Vasa in ECC-like hybrid cells is comparable to that of ESCs, whereas the expression of Vasa and Fragilis in ESC-like hybrid cells is comparable to that of F9 ECCs. The genetic dominance of ECCs in ECC-ESC hybrid cells does not dictate all genes, because ECC-like hybrids cells, which were expected to display ECC gene-expression profiles, exhibited Stra8 and Vasa expression patterns that are characteristic of ESCs. In other words, the ECC-like hybrid cells are morphologically similar to F9 ECCs (reprogramming of ESCs to F9 ECCs), but the Stra8 and Vasa genes have been genetically reprogrammed to the ESC state (reprogramming of F9 ECCs to ESCs). By contrast, although ESC-like hybrids were morphologically similar to ESCs, their Vasa and Fragilis genes had been reprogrammed to the ECC state. Taken together, these data demonstrate that genetic reprogramming in ECC-ESC fusion-hybrid cells is not a solely unidirectional process, exemplified by the morphological and gene-expression profiles of both cell-fusion partners: ECCs and ESCs.

ECC–male-neural-stem-cell (mNSC) and ECC–female-NSC (fNSC) hybrid cells display a similar morphology

We next challenged the dogma that pluripotential reprogramming of somatic cells by cell fusion is unidirectional using ECC–neural stem cell (NSC) hybrid cells. Male and female neural stem cells (mNSCs and fNSCs, respectively), which were OG2+/–/ROSA26+/– double transgenic, carrying a GFP transgene under the control of the Oct4 promoter and a ubiquitously expressing neo/lacZ transgene, were fused separately with non-transgenic F9 ECCs to determine whether the sex of the somatic cell-fusion partner affects the direction of fusion-induced reprogramming. Consistent with the notion that Oct4-GFP is expressed only in pluripotent cells, NSC-ECC hybrids were found to express Oct4-GFP only after the re-establishment of pluripotency. Oct4-GFP-positive cells, which were sorted by FACS (approximately 0.1%) on day 2 post-fusion, exhibited an ECC-like morphology and stained positive for X-gal (Fig. 3A). Karyotype analysis revealed that all cells (passages 21-32) were tetraploid in nature (Fig. 3B). X-chromosome fluorescence in situ hybridization (FISH) analysis confirmed that the hybrids comprised tetraploid cells, with F9-ECC–mNSC hybrids carrying two (Fig. 3C) and F9-ECC–fNSC hybrids three X chromosomes (Fig. 3D). Both fusion-hybrid cell types displayed a similar morphology, regardless of the sex of the respective NSC fusion partner (image of F9-ECC–fNSC hybrid cell not shown).

Using purified Oct4-GFP-positive hybrid cells, we analyzed the transcript levels of Oct4, Nanog and Sox2 in F9-ECC–mNSC and F9-ECC–fNSC hybrids on days 21 and 32 post-fusion by quantitative real-time reverse-transcriptase (RT)-PCR. These three transcription factors play crucial roles in the establishment of pluripotency in mouse ESCs and ECCs (Avilion et al., 2003; Mitsui et al., 2003; Nichols et al., 1998; Niwa et al., 2000). We found that Oct4 and Nanog, which are not expressed in NSCs, were upregulated to levels characteristic of F9 ECCs (Fig. 3E). Sox2 transcript levels were similar among NSCs, ECCs and all fusion-hybrid cell samples. By contrast, expression of the neural marker genes nestin and glutamate receptor 6 (Glur6) was downregulated in hybrid cells. Collectively, these data demonstrate that pluripotency-related and tissue-specific-marker genes were unidirectionally reprogrammed to a pattern characteristic of F9 ECCs.

Xist is oppositely reprogrammed to the pluripotent state in ECC-mNSC hybrid cells

We then examined the expression of the Xist gene, a non-neural marker that is differently expressed in F9 ECCs and NSCs (Do et al., 2007). Xist RNA is transcribed only from the inactive X chromosome (Xi); it is silent on the active X chromosome (Xa) (Lee and Lu, 1999; Panning et al., 1997). Because male somatic cells have one Xa and female somatic cells have one Xa and one Xi, only female somatic cells are capable of expressing Xist. By contrast, pluripotent cells – i.e. ESCs (Lyon, 1999), ECCs (McBurney and Strutt, 1980) and EGCs (Stewart et al., 1994) – contain only Xa(s), regardless of their sex. Our quantitative real-time RT-PCR analysis showed high Xist RNA levels in fNSCs but undetectable levels in mNSCs (Fig. 4A). The Xist RNA level in F9 ECCs was 500 times lower than that in fNSCs (from Xi), suggesting that, although the X chromosome of F9 ECCs is active, it still expresses low levels of Xist RNA. To differentiate the Xa of F9 ECCs (expressing low levels of Xist) from the Xa of somatic cells, we termed the Xa of F9 ECCs as Xã. The Xist RNA level of F9-ECC–fNSC fusion-hybrid cells (day 21 post-fusion) was reduced to that of F9 ECCs. This result clearly shows that the Xi of fNSCs was reactivated not to the level of Xa but to that of Xã. Interestingly, however, Xist transcript levels were much lower in F9 ECC–mNSC hybrid cells on day 21 post-fusion compared with F9 ECCs. We therefore evaluated the Xist levels at earlier (day 9) and later (day 32) time points after fusion. We found that Xist RNA expression was gradually downregulated in F9-ECC–mNSC hybrids and finally became undetectable by day 32 post-fusion, indicating that F9-ECC–mNSC hybrids lose Xist expression, i.e. the Xã of F9 ECCs has been changed to an Xa. This finding indicates that the expression of Xist in F9 ECCs, which is expressed at low levels, is gradually downregulated and finally becomes silenced (similar to the Xa of mNSCs) after the fusion of F9 ECCs with mNSCs. The complete silencing of Xist expression occurs late after fusion, subsequent to the reprogramming of pluripotency-related genes. One possible explanation for this is that factors required for Xist reprogramming become available only after hybrid cells have achieved a stable state. Taken together, these findings indicate that the direction of Xist gene reprogramming in pluripotent hybrid cells is opposite to that of pluripotency-related and tissue-specific-marker genes. Interestingly, we found that X-linked genes are not completely reprogrammed to the mNSC state, even after Xist has been completely reprogrammed to the mNSC state (supplementary material Fig. S1). However, these data do not necessarily imply that Xist is not involved in X-chromosome inactivation, because the expression of X-linked genes might be regulated by another system, such as that involved in the maintenance of the pluripotent cell state. For example, some X-linked genes are not expressed in somatic cells, but they are expressed in pluripotent cells, even though the number of Xa is identical in both cells (male ESC and male somatic cells). Therefore, expression patterns of X-linked genes are related to the potency of cells rather than to Xist expression levels. Because ESCs have a tendency to lose one X chromosome after prolonged culture, we evaluated the X-chromosome number in both fusion-hybrid cells. More than 90% of the male-male or male-female hybrid cells (day 32) were found to have two or three X chromosomes (Fig. 3C,D), but few cells exhibited a less pinpoint signal. Therefore, X-chromosome loss on day 32 – concomitant with no Xist expression – is not expected to affect the opposite reprogramming of Xist.

Fig. 3.

Characterization of ECC-NSC hybrid cells. (A) Hybrid cells obtained by fusion of F9 ECCs with NSCs (OG2+/–/ROSA26+/–) after G418 selection. The hybrid cells stained positive for both GFP and X-gal. Scale bar: 25 μm. (B-D) The selected hybrid cells are tetraploid in nature (B); F9-mNSC hybrids contain two (C) and F9-fNSC hybrids three X chromosomes (D). (E) Real-time RT-PCR analysis of the hybrid cells shows upregulation of pluripotency marker (Oct4 and Nanog) and downregulation of tissue-specific marker (Glur6 and Nestin) gene expression, regardless of the sex of the NSCs. This is representative data of three independent experiments. The red arrow indicates that the signal was undetectable after 45 cycles, but the software considers 45 as the effective cycle threshold. y-axis values are on a logarithmic scale, and each minor gridline is one tenth the value of each major gridline. `m' refers to male and `f' to female.

Fig. 3.

Characterization of ECC-NSC hybrid cells. (A) Hybrid cells obtained by fusion of F9 ECCs with NSCs (OG2+/–/ROSA26+/–) after G418 selection. The hybrid cells stained positive for both GFP and X-gal. Scale bar: 25 μm. (B-D) The selected hybrid cells are tetraploid in nature (B); F9-mNSC hybrids contain two (C) and F9-fNSC hybrids three X chromosomes (D). (E) Real-time RT-PCR analysis of the hybrid cells shows upregulation of pluripotency marker (Oct4 and Nanog) and downregulation of tissue-specific marker (Glur6 and Nestin) gene expression, regardless of the sex of the NSCs. This is representative data of three independent experiments. The red arrow indicates that the signal was undetectable after 45 cycles, but the software considers 45 as the effective cycle threshold. y-axis values are on a logarithmic scale, and each minor gridline is one tenth the value of each major gridline. `m' refers to male and `f' to female.

Xist/Tsix RNA FISH data revealed the presence of three pinpoint signals in F9-ECC–fNSC fusion hybrids and two pinpoint signals in F9-ECC–mNSC fusion hybrids (Fig. 4B). Pinpoint signals are typically detected in pluripotent cells and might represent the expression of Tsix (Fig. 4B), because Tsix is still expressed in both hybrid cells (Fig. 4A). Xist and Tsix RNA were concomitantly detected by FISH using a double-stranded DNA probe. Because Xist RNA cannot be detected in pluripotent cells by FISH using a strand-specific RNA probe (Maherali et al., 2007), the only way to differentiate an Xa from an Xã is by quantifying the Xist RNA levels by real-time RT-PCR, DNA methylation pattern and histone modification in the Xist region. These data showed that Tsix was reprogrammed to a pluripotent cell state even after Xist had been completely silenced (similar to in somatic cells). The difference in the direction of reprogramming of Tsix and Xist is an interesting finding of the study.

We next analyzed the DNA methylation of region 1 (R1; –381 to +74) of the Xist gene (Xist R1) (McDonald et al., 1998) in the fusion-cell partner and fusion-hybrid cells (Fig. 4C). Xist R1 was found to be completely methylated in mNSCs (97.8%), but it was differentially methylated in fNSCs; i.e. one allele was completely methylated (100%), whereas the other was completely unmethylated (0%). The Xist R1 of F9 ECCs was partially methylated (71.0%), which might account for the low expression of Xist from the Xã of F9 ECCs (Fig. 4A). Xist R1 was partially methylated (∼64-74%) in F9-ECC–fNSC hybrid cells, which is comparable to that of F9 ECCs (Xã; 71.0%), confirming that the Xi of fNSCs is reactivated to an Xã following the fusion of fNSCs with F9 ECCs. In addition, it is notable that the Xa of fNSCs was also changed to an Xã, which thus resulted in XãXãXã hybrids. Most F9-ECC–fNSC hybrid cells contained partially methylated (in Xist R1) clones (corresponding to Xã). Only 3 of the 31 clones (9.7%) examined were completely methylated (corresponding to Xa) (Fig. 4C). Had the Xa not been changed to an Xã, the hybrids would have been XaXãXã, with complete methylation of 33.3% of the clones (one out of every three clones). These results indicate that, during fusion-induced reprogramming, there is not only reactivation of the Xi, but also partial inactivation of the Xist promoter on the Xa. This interpretation is also supported by the detection of a pinpoint signal of Xist/Tsix RNA expression (corresponding to an Xã) in all three X chromosomes in F9-ECC–fNSC hybrid cells (Fig. 4B). We could not detect a large Xist cloud in any of the F9-ECC–fNSC hybrid cells, which suggested that the Xã of F9 ECCs had not been reprogrammed to an Xi.

Fig. 4.

Xist reprogramming in F9-ECC–mNSC hybrid cells. (A) Real-time RT-PCR analysis shows downregulation of Xist gene expression from the Xi of fNSCs in F9-ECC–fNSC hybrids, with Xist transcript levels reaching those of F9 ECCs. Xist RNA levels were low in F9 ECCs, but they were nonexistent in F9-ECC–mNSC hybrid cells. However, Tsix expression is still observed in both types of hybrid cells. This is representative data of three independent experiments. The red arrow indicates that the signal was undetectable after 45 cycles, but the software considers 45 as the effective Ct. y-axis values are on logarithmic scale, and each minor gridline is one tenth the value of each major gridline. (B) Xist/Tsix RNA using a double-stranded DNA probe that supports the detection of low levels of Xist/Tsix RNA. Low Xist RNA levels, in pluripotent cells, cannot be detected by FISH using a strand-specific RNA probe (Wernig et al., 2007). (C) Bisulfite DNA sequencing analysis of hybrids formed by the fusion of F9 ECCs with NSCs. In F9-ECC–fNSC hybrids, partial methylation of Xist R1 is maintained, whereas, in F9-ECC–mNSC hybrids, the region is completely methylated by day 32 post-fusion. White and black circles represent unmethylated and methylated CpGs, respectively. (D) ChIP assay for acetylated histone H3 (AcH3) and acetylated histone H4 (AcH4) on the Xist region. The changes in AcH3 and AcH4 levels were measured by fold-enrichment (lower panels), which was quantified from ethidium-bromide-stained gel images (upper and middle panels). A reduction in histone H4 acetylation (AcH4) and histone H3 acetylation (AcH3) on the Xist region was detected in F9-mNSC hybrid cells. `m' refers to male and `f' to female.

Fig. 4.

Xist reprogramming in F9-ECC–mNSC hybrid cells. (A) Real-time RT-PCR analysis shows downregulation of Xist gene expression from the Xi of fNSCs in F9-ECC–fNSC hybrids, with Xist transcript levels reaching those of F9 ECCs. Xist RNA levels were low in F9 ECCs, but they were nonexistent in F9-ECC–mNSC hybrid cells. However, Tsix expression is still observed in both types of hybrid cells. This is representative data of three independent experiments. The red arrow indicates that the signal was undetectable after 45 cycles, but the software considers 45 as the effective Ct. y-axis values are on logarithmic scale, and each minor gridline is one tenth the value of each major gridline. (B) Xist/Tsix RNA using a double-stranded DNA probe that supports the detection of low levels of Xist/Tsix RNA. Low Xist RNA levels, in pluripotent cells, cannot be detected by FISH using a strand-specific RNA probe (Wernig et al., 2007). (C) Bisulfite DNA sequencing analysis of hybrids formed by the fusion of F9 ECCs with NSCs. In F9-ECC–fNSC hybrids, partial methylation of Xist R1 is maintained, whereas, in F9-ECC–mNSC hybrids, the region is completely methylated by day 32 post-fusion. White and black circles represent unmethylated and methylated CpGs, respectively. (D) ChIP assay for acetylated histone H3 (AcH3) and acetylated histone H4 (AcH4) on the Xist region. The changes in AcH3 and AcH4 levels were measured by fold-enrichment (lower panels), which was quantified from ethidium-bromide-stained gel images (upper and middle panels). A reduction in histone H4 acetylation (AcH4) and histone H3 acetylation (AcH3) on the Xist region was detected in F9-mNSC hybrid cells. `m' refers to male and `f' to female.

Fig. 5.

Xist reprogramming in ESC-mNSC hybrid cells. (A) Xist transcription was ultimately silenced in fESC-mNSC hybrids, but it was active in mESC-mNSC hybrids. However, Tsix expression is still observed in both types of hybrid cells. This is representative data of three independent experiments. The red arrow indicates that the signal was undetectable after 45 cycles, but the software considers 45 as the effective Ct. y-axis values are on logarithmic scale, and each minor gridline is one tenth the value of each major gridline. (B) Detection of Xist/Tsix RNA using a double-stranded DNA probe that can detect low levels of Xist/Tsix RNA. A strand-specific RNA probe cannot detect low Xist RNA levels in pluripotent cells by FISH. (C) Bisulfite DNA sequencing analysis of ESC and mNSC hybrid cells. Complete methylation (93.3%) of the Xist R1 in fESC-mNSC hybrids and partial methylation (75.6%) of the Xist R1 in mESC-mNSC hybrids occurred by 32 days post-fusion. White and black circles represent unmethylated and methylated CpGs, respectively. `m' refers to male and `f' to female.

Fig. 5.

Xist reprogramming in ESC-mNSC hybrid cells. (A) Xist transcription was ultimately silenced in fESC-mNSC hybrids, but it was active in mESC-mNSC hybrids. However, Tsix expression is still observed in both types of hybrid cells. This is representative data of three independent experiments. The red arrow indicates that the signal was undetectable after 45 cycles, but the software considers 45 as the effective Ct. y-axis values are on logarithmic scale, and each minor gridline is one tenth the value of each major gridline. (B) Detection of Xist/Tsix RNA using a double-stranded DNA probe that can detect low levels of Xist/Tsix RNA. A strand-specific RNA probe cannot detect low Xist RNA levels in pluripotent cells by FISH. (C) Bisulfite DNA sequencing analysis of ESC and mNSC hybrid cells. Complete methylation (93.3%) of the Xist R1 in fESC-mNSC hybrids and partial methylation (75.6%) of the Xist R1 in mESC-mNSC hybrids occurred by 32 days post-fusion. White and black circles represent unmethylated and methylated CpGs, respectively. `m' refers to male and `f' to female.

When mNSCs were fused with F9 ECCs, the Xist R1 of the resultant F9-ECC–mNSC hybrid cells showed partial methylation patterns early after fusion (79.3% and 89.2% on days 9 and 21, respectively). However, Xist RNA could no longer be detected by day 32 post-fusion (Fig. 4A) and mainly consisted of fully methylated alleles (100% and 90%); the methylation rate (∼92%) was much higher than that for Xã (∼70%) (Fig. 4C). Chromatin immunoprecipitation (ChIP) assay also confirmed a more inactive chromatin state (reduced H3 and H4 acetylation, similar to that in mNSCs) of the Xist region of F9-ECC–mNSC hybrid cells compared with that of F9-ECC–fNSC hybrid cells (Fig. 4D). Collectively, our data demonstrate that the Xã of F9 ECCs was gradually changed to an Xa after the fusion of F9 ECCs with mNSCs. These data indicate that fusion-induced reprogramming is not a unidirectional process, giving rise solely to cells of the pluripotent state. Rather, Xist is a unique gene in that it undergoes opposite reprogramming to the pluripotent cell state, which is induced by male, but not by female, NSCs.

Reprogramming of the Xã to Xa, however, was not observed in hybrid cells derived by the fusion of two different types of male pluripotent cells. Specifically, there was partial methylation of Xist R1 in F9-ECC–mESC fusion-hybrid cells (64.6% on day 32 post-fusion), similar to that of the individual cell-fusion partners, F9 ECCs and male ESCs (supplementary material Fig. S2), suggesting that the Xã state of both pluripotent cells was not affected by the fusion process. These results clearly demonstrate that only mNSCs can induce activation of the Xã of F9 ECCs to the Xa state (i.e. complete inactivation of the Xist gene of F9 ECCs) and that reprogramming of the Xã to Xa is not due to the tetraploid nature of the fusion hybrids.

Xist is oppositely reprogrammed to the pluripotent state in ESC–mNSC hybrid cells

To investigate whether opposite reprogramming of Xist could be a general phenomenon, two additional series of fusion experiments were performed; fESCs and mESCs were fused separately with mNSCs. We found that the Xist R1 of fESCs was less methylated than that of mESCs, which is consistent with a previous report of reduced DNA methylation of the Xist region in fESCs, which contain two active X chromosomes (Zvetkova et al., 2005). Downregulation of Xist expression (Fig. 5A) and complete methylation (93.3% methylation) of Xist R1 (Fig. 5C) was observed in fESC-mNSC hybrid cells. However, downregulation of Xist expression was not observed (Fig. 5A) and partial methylation of Xist R1 was still retained (Fig. 5C) in mESC-mNSC hybrids. Xist/Tsix RNA FISH data revealed the presence of three pinpoint signals in fESC-mNSC fusion hybrids and two pinpoint signals in mESC-mNSC fusion hybrids – similar to F9-fNSC and F9-mNSC fusion hybrids, respectively – which might represent Tsix expression (Fig. 5B), because Tsix is still expressed in both hybrid cells (Fig. 5A). On the basis of these results, we conclude that the fusion of somatic cells with ECCs or ESCs does not occur through unidirectional reprogramming; rather, pluripotency-related and tissue-specific-marker genes are reprogrammed to the state of pluripotent cells, whereas the Xist gene is reprogrammed to the state of mNSCs. However, this was not observed in mESC-mNSC fusion hybrids, which indicates that F9 ECCs have different properties to mESCs in respect to the ability of reprogramming the mNSC genome to pluripotency.

Discussion

The fusion of somatic cells with pluripotent cells can result in the formation of hybrid cells that retain residual memory of the somatic cell partner. When this occurs, the reprogramming process is considered to be incomplete and still follows a unidirectional route. In this study, we showed that hybrids cells inherited the `somatic memory' (Xist expression) of the male somatic cells, leading to alterations in the characteristics of their pluripotent cell-fusion partners and, thereby, maintenance of a general pluripotent-cell phenotype.

We defined this phenomenon as `pluripotential reprogramming', whereby hybrids cannot lose pluripotency after fusion. Shimazaki et al. showed that the fusion of P19 ECCs with fibroblasts resulted in the formation of hybrids that exhibited a differentiated state, wherein the Oct4 gene was silenced; i.e. the pluripotent phenotype becomes differentiated to a somatic-cell phenotype (Shimazaki et al., 1993) in a unidirectional process. Ambrosi et al. showed that the gene-expression profile of the fusion-hybrid cells is similar, but not identical, to that of the pluripotent cell partner (Ambrosi et al., 2007). Having assessed the transcriptional status of 36,601 transcripts encoded within the mouse genome, they categorized the genes in seven major clusters according to gene-expression patterns. However, these authors could not find a gene that is strongly expressed in mouse embryonic fibroblasts (MEFs) and hybrids but poorly expressed in ESCs, or a gene that is poorly expressed in MEFs and hybrids but strongly expressed in ESCs. These findings indicate that, of the 36,601 genes examined, there are no genes that were reprogrammed to the level of MEFs after cell fusion. The Xist gene, however, was not included in the investigation. Therefore, it is possible that Xist is the only gene that has been reprogrammed to the state of somatic cells after the fusion of somatic cells with pluripotent cells. Several groups have demonstrated that the Xi of somatic cells does indeed undergo reactivation after cell fusion (Do et al., 2007; Mise et al., 1996; Tada et al., 2001). These previous reports only showed that the Xist state (expression level and DNA methylation) of the Xi of somatic cells was changed to that of pluripotent cells, supporting the unidirectionality of fusion-induced reprogramming. With the present paper, we are the first to report that fusion-induced reprogramming is not a solely unidirectional process, because Xist does not behave like the other genes, which were either reprogrammed to the pluripotent cell state or remained unreprogrammed.

In a direct reprogramming system, such as that culminating in the generation of induced pluripotent stem (iPS) cells from somatic cells, any gene that remains in the somatic cell state should be referred to as having undergone `incomplete reprogramming', rather than having been reprogrammed in a single direction.

During early embryonic development, the pre-inactivated paternal X chromosome (from sperm) is maintained in an inactive state during preimplantation development and is inherited in extraembryonic tissues after implantation (Huynh and Lee, 2003). Therefore, the pre-inactivated paternal X chromosome does not assume the fate of the active maternal X chromosome. In the present study, we have shown that the pre-inactivated somatic X chromosome could change into the active X chromosome of pluripotent cells.

The opposite reprogramming of Xist expression cannot be explained by the X-chromosome counting mechanism, which reliably acts to maintain one X chromosome in an active state per two autosome sets (Lee, 2005). There is no clear explanation for the observed phenomenon, but it is possible that male somatic cells contain a factor that induces the complete inactivation of the Xist promoter. This factor might function as a `blocking factor' capable of repressing X-chromosome inactivation (Rastan and Robertson, 1985). A blocking factor has already been suggested to play a role in the counting mechanism (Rastan and Robertson, 1985). This same factor might play a totally different role in directing hybrid cells to the pluripotent cell state. Alternatively, a factor not involved in the counting mechanism might play a role in directing hybrid cells to attain the pluripotent cell state. Deciphering the most likely scenario will require further investigations into Xist gene expression and the underlying mechanism of X-chromosome inactivation, which remains largely elusive at the present time.

Materials and Methods

Mice

All mouse strains were bred and housed at the mouse facility of the Max Planck Institute (MPI) for Molecular Biomedicine or were bought from Harlan Winkelmann (Harlan GmbH). Animal handling was in accordance with the MPI animal-protection guidelines and the German animal protection laws.

Fusion-partner cell culture and fusion

ECCs (F9-ECC line) were grown on gelatin-coated (0.1% in PBS) dishes in standard ECC media, high-glucose DMEM (Gibco BRL) containing 15% fetal calf serum (FCS; Gibco BRL), 1× penicillin/streptomycin/glutamine, and 1× nonessential amino acids (Gibco BRL). Homogeneous NSCs were derived from neurospheres. Culturing of neurospheres was performed as described in detail in our previous paper (Do and Scholer, 2004). Two fESC lines were derived from blastocysts obtained by crossing OG2 with B6C3F1 mice and OG2 with ROSA26 mice. Either the E14 ESC line was used as a male fusion partner or lines derived in our laboratory from blastocysts obtained by crossing OG2 with ROSA26 mice. ESCs were cultured on mitomycin-C-treated fibroblasts (MEFs) in ESC media (ECC media containing 1000 U/ml LIF). Fusion experiments involving F9 ECCs, ESCs and NSCs were conducted with 50% polyethylene glycol (PEG1500; Roche). Cells from two fusion partners were mixed in a ratio of 1:1 and then washed in PBS. The mixture was centrifuged in conical tubes at 130 g for 5 minutes, the supernatant was removed and 1 ml of a prewarmed solution of 50% polyethylene glycol 1500 (PEG1500; Roche) was added onto the cell pellet. 20 ml of DMEM was added to the cell suspension over 5 minutes, with constant stirring. The cells were then centrifuged at 130 g for 5 minutes, washed with DMEM and cultured in ESC medium.

Karyotype analysis and X-chromosome FISH

Cells that had been treated with 3 μg/ml of Nocodazole for 4 hours were recovered by trypsinization and subjected to a hypotonic solution (0.56% KCl w/v) for 15 minutes. The cells were centrifuged at 200 g for 5 minutes, fixed by washing three times in fresh fixative (methanol:acetic acid ratio 3:1) and dropped onto clean glass slides. The slides were air-dried and stained with 3% Giemsa or Hoechst (Sigma). X-chromosome FISH was performed using STAR*FISH mouse X-chromosome-specific probe labeled with Cyanine 3 (Cambio) according to the manufacturer's protocol.

Flow cytometry

Flow cytometry was performed using the method described in our previous paper (Do et al., 2007). Briefly, the dissociated hybrid cells were washed with PBS and resuspended in ESC medium. Cell sorting was performed on a FACSAria cell sorter (Becton Dickinson) using FACSDiva software (Becton Dickinson).

X-gal staining

Culture dishes were washed with PBS, fixed with 4% paraformaldehyde at room temperature for 10 minutes, washed again with PBS, and incubated overnight at 37°C in PBS supplemented with 1 mg/ml 5-bromo-4-chloro-3-indolyl-galactosidase (X-gal; Sigma), 5 mM K2Fe(CN)6, 5 mM K4Fe(CN)6 and 1 mM MgCl2. X-gal-positive cells stained blue.

Gene-expression and DNA-methylation analysis

All data were normalized to Bact expression and calibrated on the F9 ECCs or mESCs, whose expression was considered 1 for all genes. The procedures and methods are available on request.

We are indebted to members of Dr Schöler's laboratory for fruitful discussions and valuable comments on study findings and their interpretations. We are especially grateful to Arauzo Marcos, Martin Stehling and Claudia Ortmeier for assistance with densitometry, FACS sorting and real-time RT-PCR, respectively. This work was supported by the initiative `Cell-based, regenerative medicine' (grant no. 01GN0539) of the Federal Ministry of Education and Research (BMBF).

References

Ambrosi, D. J., Tanasijevic, B., Kaur, A., Obergfell, C., O'Neill, R. J., Krueger, W. and Rasmussen, T. P. (
2007
). Genome-wide reprogramming in hybrids of somatic cells and embryonic stem cells.
Stem Cells
25
,
1104
-1113.
Avilion, A. A., Nicolis, S. K., Pevny, L. H., Perez, L., Vivian, N. and Lovell-Badge, R. (
2003
). Multipotent cell lineages in early mouse development depend on SOX2 function.
Genes Dev.
17
,
126
-140.
Do, J. T. and Scholer, H. R. (
2004
). Nuclei of embryonic stem cells reprogram somatic cells.
Stem Cells
22
,
941
-949.
Do, J. T. and Scholer, H. R. (
2006
). Cell-cell fusion as a means to establish pluripotency.
Ernst Schering Res. Found Workshop
60
,
35
-45.
Do, J. T., Han, D. W., Gentile, L., Sobek-Klocke, I., Stehling, M., Lee, H. T. and Scholer, H. R. (
2007
). Erasure of cellular memory by fusion with pluripotent cells.
Stem Cells
25
,
1013
-1020.
Huynh, K. D. and Lee, J. T. (
2003
). Inheritance of a pre-inactivated paternal X chromosome in early mouse embryos.
Nature
426
,
857
-862.
Lee, J. T. (
2005
). Regulation of X-chromosome counting by Tsix and Xite sequences.
Science
309
,
768
-771.
Lee, J. T. and Lu, N. (
1999
). Targeted mutagenesis of Tsix leads to nonrandom X inactivation.
Cell
99
,
47
-57.
Lyon, M. F. (
1999
). X-chromosome inactivation.
Curr. Biol.
9
,
R235
-R237.
Maherali, N., Sridharan, R., Xie, W., Utikal, J., Eminli, S., Arnold, K., Stadtfeld, M., Yachechko, R., Tchieu, J., Jaenisch, R. et al. (
2007
). Directly reprogrammed fibroblasts show global epigenetic remodeling and widespread tissue contribution.
Cell Stem Cell
1
,
55
-70.
McBurney, M. W. and Strutt, B. J. (
1980
). Genetic activity of X chromosomes in pluripotent female teratocarcinoma cells and their differentiated progeny.
Cell
21
,
357
-364.
McDonald, L. E., Paterson, C. A. and Kay, G. F. (
1998
). Bisulfite genomic sequencing-derived methylation profile of the xist gene throughout early mouse development.
Genomics
54
,
379
-386.
Mise, N., Sado, T., Tada, M., Takada, S. and Takagi, N. (
1996
). Activation of the inactive X chromosome induced by cell fusion between a murine EC and female somatic cell accompanies reproducible changes in the methylation pattern of the Xist gene.
Exp. Cell Res.
223
,
193
-202.
Mitsui, K., Tokuzawa, Y., Itoh, H., Segawa, K., Murakami, M., Takahashi, K., Maruyama, M., Maeda, M. and Yamanaka, S. (
2003
). The homeoprotein Nanog is required for maintenance of pluripotency in mouse epiblast and ES cells.
Cell
113
,
631
-642.
Nichols, J., Zevnik, B., Anastassiadis, K., Niwa, H., Klewe-Nebenius, D., Chambers, I., Scholer, H. and Smith, A. (
1998
). Formation of pluripotent stem cells in the mammalian embryo depends on the POU transcription factor Oct4.
Cell
95
,
379
-391.
Niwa, H., Miyazaki, J. and Smith, A. G. (
2000
). Quantitative expression of Oct-3/4 defines differentiation, dedifferentiation or self-renewal of ES cells.
Nat. Genet.
24
,
372
-376.
Panning, B., Dausman, J. and Jaenisch, R. (
1997
). X chromosome inactivation is mediated by Xist RNA stabilization.
Cell
90
,
907
-916.
Rastan, S. and Robertson, E. J. (
1985
). X-chromosome deletions in embryo-derived (EK) cell lines associated with lack of X-chromosome inactivation.
J. Embryol. Exp. Morphol.
90
,
379
-388.
Shimazaki, T., Okazawa, H., Fujii, H., Ikeda, M., Tamai, K., McKay, R. D., Muramatsu, M. and Hamada, H. (
1993
). Hybrid cell extinction and re-expression of Oct-3 function correlates with differentiation potential.
EMBO J.
12
,
4489
-4498.
Silva, J., Chambers, I., Pollard, S. and Smith, A. (
2006
). Nanog promotes transfer of pluripotency after cell fusion.
Nature
441
,
997
-1001.
Stewart, C. L., Gadi, I. and Bhatt, H. (
1994
). Stem cells from primordial germ cells can reenter the germ line.
Dev. Biol.
161
,
626
-628.
Tada, M., Tada, T., Lefebvre, L., Barton, S. C. and Surani, M. A. (
1997
). Embryonic germ cells induce epigenetic reprogramming of somatic nucleus in hybrid cells.
EMBO J.
16
,
6510
-6520.
Tada, M., Takahama, Y., Abe, K., Nakatsuji, N. and Tada, T. (
2001
). Nuclear reprogramming of somatic cells by in vitro hybridization with ES cells.
Curr. Biol.
11
,
1553
-1558.
Takahashi, K. and Yamanaka, S. (
2006
). Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors.
Cell
126
,
663
-676.
Wernig, M., Meissner, A., Foreman, R., Brambrink, T., Ku, M., Hochedlinger, K., Bernstein, B. E. and Jaenisch, R. (
2007
). In vitro reprogramming of fibroblasts into a pluripotent ES-cell-like state.
Nature
448
,
318
-324.
Zvetkova, I., Apedaile, A., Ramsahoye, B., Mermoud, J. E., Crompton, L. A., John, R., Feil, R. and Brockdorff, N. (
2005
). Global hypomethylation of the genome in XX embryonic stem cells.
Nat. Genet.
37
,
1274
-1279.