HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Supplemental Data All Versions of this Article: 2006-0532v1 25/5/1104    most recent Alert me when this article is cited Alert me if a correction is posted Citation Map Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints/Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Ambrosi, D. J. Articles by Rasmussen, T. P. Search for Related Content PubMed PubMed Citation Articles by Ambrosi, D. J. Articles by Rasmussen, T. P.

EMBRYONIC STEM CELLS: CHARACTERIZATION SERIES

Genome-Wide Reprogramming in Hybrids of Somatic Cells and Embryonic Stem Cells

^aCenter for Regenerative Biology,
^bDepartment of Molecular and Cell Biology,
^dDepartment of Animal Science, University of Connecticut, Storrs, Connecticut, USA;
^cDepartment of Genetics and Developmental Biology, University of Connecticut Medical School, Farmington, Connecticut, USA

Key Words. Fusion • Reprogramming • Stem cells • Genomics

Correspondence: Theodore P. Rasmussen, Ph.D., Center for Regenerative Biology, 1392 Storrs Road, U-4243 University of Connecticut, Storrs, Connecticut 06269-4243, USA. Telephone: (860) 486-8339; Fax: (860) 486-8809; e-mail: theodore.rasmussen{at}uconn.edu

Received August 25, 2006; accepted for publication January 23, 2007.
First published online in STEM CELLS EXPRESS   February 1, 2007.

	ABSTRACT

Top Abstract Introduction Materials and Methods Results Discussion Disclosure of Potential... Acknowledgments References

 
Recent experiments demonstrate that somatic nuclei can be reprogrammed to a pluripotent state when fused to ESCs. The resulting hybrids are pluripotent as judged by developmental assays, but detailed analyses of the underlying molecular-genetic control of reprogrammed transcription in such hybrids are required to better understand fusion-mediated reprogramming. We produced hybrids of mouse ESCs and fibroblasts that, although nearly tetraploid, exhibit characteristics of normal ESCs, including apparent immortality in culture, ESC-like colony morphology, and pluripotency. Comprehensive analysis of the mouse embryonic fibroblast/ESC hybrid transcriptome revealed global patterns of gene expression reminiscent of ESCs. However, combined analysis of variance and hierarchical clustering analyses revealed at least seven distinct classes of differentially regulated genes in comparisons of hybrids, ESCs, and somatic cells. The largest class includes somatic genes that are silenced in hybrids and ESCs, but a smaller class includes genes that are expressed at nearly equivalent levels in hybrids and ESCs that contain many genes implicated in pluripotency and chromatin function. Reprogrammed genes are distributed throughout the genome. Reprogramming events include both transcriptional silencing and activation of genes residing on chromosomes of somatic origin. Somatic/ESC hybrid cell lines resemble their pre-fusion ESC partners in terms of behavior in culture and pluripotency. However, they contain unique expression profiles that are similar but not identical to normal ESCs. ESC fusion-mediated reprogramming provides a tractable system for the investigation of mechanisms of reprogramming.

Disclosure of potential conflicts of interest is found at the end of this article.

	INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion Disclosure of Potential... Acknowledgments References

 
Each somatic cell contains a complete complement of genetic information, but only a characteristic subset of genes is expressed within any given cell type. Nuclear reprogramming can reset the transcriptional state of the genome to a configuration that is compatible with developmental pluripotency. Reprogramming to a state of totipotency was first demonstrated by nuclear transplantation experiments that resulted in the production of clonally derived frogs [1]. Only much later was it shown that mammals, too, can be generated through the process of somatic cell nuclear transfer (SCNT) [2, 3]. However, mammalian cloning is an inefficient process as judged by the production of live-birth clones, and it has been suggested that these difficulties are a result of incomplete reprogramming of somatic chromatin [4]. In mice, SCNT-derived blastocysts from immune-compromised Rag2-null mice yielded ESC lines whose Rag2^–/– defect was then corrected by targeted gene replacement, prior to their differentiation in vitro, to produce hematopoietic cells that were immunologically matched to the donor [5]. Although possible in theory, the production of patient-matched human ESCs by SCNT has yet to be achieved. Furthermore, the relative unavailability of large numbers of donated human oocytes and ethical concerns may render this approach impractical in a future clinical setting. For this reason, much effort has been focused on attempts to reprogram somatic nuclei through methods that do not rely upon SCNT.

One such approach involves the reprogramming of the somatic genome through the fusion of somatic cells with ESCs, which, like oocytes, also contain reprogramming activities [6]. A number of studies have shown that cells of embryonic origin contain activities that can reprogram somatic nuclei. Developmental pluripotency has been demonstrated in hybrids made between murine ESCs and somatic cells [7, 8], as well as thymocytes [9]. In addition, embryonal carcinoma and even embryonic germ cells have been shown to contain reprogramming activities as demonstrated by fusion experiments. Hybrids contain reconfigured chromatin [10], are pluripotent [11], and express stem cell markers such as Oct4 and Nanog [12]. The reprogramming activity of ESCs has been shown to reside within the nuclear compartment because karyoplasts, but not cytoplasts, derived from ESCs can bring about reprogramming of somatic cells upon fusion [13]. Recently, human fibroblasts have been successfully reprogrammed when they were fused with human ESCs [14]. In addition, human myeloid precursor cells have been reprogrammed to a pluripotent state by fusion with human ESCs [15].

Hybrids between somatic cells and ESCs are entirely new cell types. It is reasonable to suppose that cell type identity is accurately measured through analysis of the transcriptome. Because the transcriptomes of differentiated somatic cells and ESCs are widely divergent, we investigated the somatic/ESC transcriptome in detail. We show here that the mouse embryonic fibroblast (MEF)/ESC hybrid transcriptome resembles that of the ESC transcriptome, but with interesting distinctions. Data are presented to show that differences exist from gene to gene in their susceptibilities to the reprogramming process. The results indicate that MEF/ESC hybrids constitute a tractable system for the investigation of reprogramming mechanisms and the identification of reprogramming factors. Furthermore, because this approach avoids the use of oocytes, it may form the basis of future efforts to produce pluripotent cells for human cell-based therapies in a way that is both expedient and free from controversy.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion Disclosure of Potential... Acknowledgments References

 
General Cell Culture Methods
Mouse ESCs were grown in standard embryonic stem cell media consisting of Dulbecco's modified Eagle's medium (Cellgro; Mediatech Inc., Herndon, VA, http://www.cellgro.com/), nonessential amino acids (Cellgro), penicillin/streptomycin (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com), 0.0004% β-mercaptoethanol (Sigma-Aldrich), 15% fetal bovine serum (HyClone Laboratories, Logan, UT, http://www.hyclone.com), and 1,000 units of leukemia inhibitory factor (LIF)/ml on 60-cm² plastic tissue culture vessels coated with 0.2% gelatin. ESCs and resulting hybrids were grown on drug-resistant 4 (DR4) MEFs (Jackson Immunoresearch Laboratories, West Grove, PA, http://www.jacksonimmuno.com) that were mitotically inactivated with mitomycin C (Roche Diagnostics, Basel, Switzerland, http://www.roche-applied-science.com) in HERAcell incubators (Thermo Scientific; Kendro, Waltham, MA, http://www.thermo.com) (37°C, 5% CO₂). Somatic parental fusion cells consisted of MEFs derived from E13.5 embryos that were heterozygous for the ROSA26-β-geo transgene. Transgenic MEF lines were identified by polymerase chain reaction (PCR) with primers F150 5'-GGCTTAAAGGCTAACCTGATG-3' and 315 5'-GCGAAGAGTTTGTCCTCAACC-3' designed by Jackson Immunoresearch Laboratories, confirmed to be resistant to the drug G-418, and blue staining in the presence of 5-bromo-4-chloro-3-indolyl-β-D-galactoside (X-gal). Mouse wild-type J1 ESCs (129/SvJae) were made resistant to puromycin by the introduction of the plasmid pPGKpuro (Peter Laird, University of Southern California, Los Angeles, CA) by electroporation using a Bio-Rad GenePulser Xcell (Bio-Rad, Hercules, CA, http://www.bio-rad.com) set to 400 V and 25-µF capacitance in a 0.4-cm gap cuvette. Transformed cells were plated in standard ESC medium on DR4 MEFs, and puro^R lines were established by growth in 2 µg/ml puromycin (Sigma-Aldrich).

Cell Fusion and Hybrid Selection
J1-puro^R ESCs (5 x 10⁵) were mixed with 5 x 10⁵ ROSA26-β-geo MEFs in standard ESC medium lacking serum at 37°C. The mixture was gently pelleted and resuspended in 1 ml of 50% polyethylene glycol 1500 (PEG 1500) in HEPES (Roche) for 3 minutes at 37°C. Then, 1 ml of prewarmed serum-free ESC medium was added slowly, and the suspension was incubated for 1 minute. Then, 3 ml of prewarmed serum-free ESC medium was gently added, followed by incubation for 3 minutes at 37°C. Finally, 10 ml of serum-free ESC medium was added, and the suspension was incubated for 5 minutes at 37°C. The cells were then pelleted, resuspended in ESC medium, and plated on mitotically inactivated DR4 feeder cells. After approximately 24 hours, selection with ESC medium supplemented with 2 µg/ml puromycin and 50 µg/ml G-418 was initiated. After 12 days, colonies were picked, briefly trypsinized, and seeded into gelatinized 2-cm² wells on DR4 feeder layers. The resulting hybrid cell lines were then passaged, expanded, and frozen using standard ESC culture methods.

Karyotype Analysis
Cells were arrested in metaphase by adding colcemid (0.02 µg/ml) (Gibco, Grand Island, NY, http://www.invitrogen.com) in ESC medium for 3 hours at 37°C, 5% CO₂, trypsinized to single cells, washed in phosphate-buffered saline (PBS), resuspended for 6 minutes at room temperature in 5 ml of 0.56% KCl added dropwise, and then fixed in 3:1 methanol/acetic acid and dropped on slides to create chromosome spreads. Slides were stained in a 1:20 dilution of Giemsa (Sigma-Aldrich) stain for 60 minutes, rinsed twice with distilled water, coverslipped with Cytoseal 60 (Richard-Allan Scientific, Kalamazoo, MI, http://www.rallansci.com/), and assessed with light microscopy.

X-Gal Staining for β-Galactosidase Expression
Cells were grown in 3.8-cm² tissue culture plates, then washed in PBS, and fixed in 2% paraformaldehyde/0.25% glutaraldehyde in PBS for 30 minutes at 4°C. Wells were then washed five times with PBS and stained in 2 mM MgCl₂, 5 mM potassium ferricyanide, 0.1% sodium deoxycholate, 0.01% Nonidet P40, and 1x PBS supplemented with 4% X-gal in solution in dimethylformamide until color developed. Stained cells were fixed in 2% paraformaldehyde/0.25% glutaraldehyde.

Microarray Analysis
For each sample, 5 x 10⁶ cells were harvested from feeder-free cultures and lysed in 1 ml of TRIzol reagent, and RNA was prepared according to the manufacturer's instructions. Five µg of total RNA was amplified using a T7 RNA polymerase-based protocol [16] and biotinylated using a One Cycle target labeling kit (Affymetrix, Santa Clara, CA, http://www.affymetrix.com). In brief, 5 µg of RNA was used for first- and second-strand cDNA synthesis, and the double-stranded cDNA was amplified in the presence of biotinylated ribonucleotides to generate biotinylated cRNA, which was then purified by ion exchange column chromatography, fragmented, and hybridized to mouse MG 430 2.0 GeneChips (Affymetrix). After 16 hours of hybridization at 45°C, the arrays were stained in an Affymetrix Fluidics Station using a two-stage signal amplification protocol based on detection of the biotinylated targets by streptavidin-phycoerythrin, according to Affymetrix instructions. Signal was quantified by detection of bound phycoerytherin using an Affymetrix GeneChip Scanner 3000. Array hybridization signal levels were normalized using the invariant set normalization method, which is a part of the dChip software [17]. Statistical analysis was performed on quantile-normalized data using the perfect match model and analysis of variance (ANOVA) function in the dChip program [18] and on robust multiarray analysis-normalized data using R/MAANOVA [19] with an ANOVA model consisting of the factors sample and cell type (hybrid lines A3, A4, A7, MEFs, and ESCs). Differentially regulated genes were organized into clusters with the hierarchical clustering module of dChip, with the clustering parameter set to (a) Euclidean distance and (b) clustering by both genes and samples. For genomic arrangement of genes found to be differentially expressed, the gene list and the annotation files containing the chromosome information for each gene in the analysis were uploaded into the dChip software. The plot was created using the chromosome option of the software, which graphically displays the association of the chromosome information for each of the differentially regulated genes.

Expressed Single Nucleotide Polymorphism Analysis
Parental fusion partner cell lines differed in genetic background (J1-puro are 129Sv/Jae, and MEFs are BALB/c). We used the Mouse SNP Query component of the Mouse Genome Database [20] to identify candidate-expressed single nucleotide polymorphisms (SNPs) that could distinguish transcription from chromosomes of both parental origins for genes in clusters 1 to 3. We verified by DNA sequencing that chromodomain helicase DNA binding protein 1-like (Chd1l) (5'-CCATGAGAGGCCTGAGAGAC-3' and 5'-TTCGGCTTGGATAAACTGCT-3') and piwi-like homolog 2 (Piwil2) contained SNPs that differed between MEFs and ESCs. Samples of RNA from the same cells used in microarray analysis (ESCs, MEFs, hybrids A3, A4, and A7) were used to produce cDNA using a Superscript III reverse transcription-polymerase chain reaction (RT-PCR) kit with random hexamers (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) following the manufacturer's protocol. Intron-spanning primers Chd1l, 5'-CTGAAGTCCATCCTGGGAGA-3' and 5'-TCACCAGCTGCTCAAATGAC-3', and Piwil2, 5'-AATGTCCGAACCATTGGTCA-3' and 5'-GATTTATGCTGGCCACGAAG-3', were used to amplify cDNA products, and these were assessed for SNP content by direct sequencing using primers Chd1lS (5'-CTGAAGTCCATCCTGGGAGA-3') and Piwil2S (5'-AAACTGGGTGGTGAGCTCTG-3').

RNA Fluorescence in Situ Hybridization
An RNA fluorescence in situ hybridization (FISH) probe for nascent Oct4 transcripts was derived from 3 kilobases (kb) of an intronic sequence spanned by primers 5'-CACGAGTGGAAAGCA-ACTCA-3' and 5'-CTGCAGAATTGACCAGACGA-3', designed from the combined use of Primer3 [21] and Repeat Masker [22]. PCR was performed on 100 ng of genomic DNA at 63°C in the presence of HiFi Taq Polymerase (Roche) according to the manufacturer's modified protocol for 3-kb templates. PCR product was purified in a 100-column Microcon (Millipore, Billerica, MA, http://www.millipore.com), and PCR product (1 µg) was labeled with dig-11-dUTP using the Dig High Prime Kit (Roche). Five hundred nanograms of probe was precipitated with RNA FISH cocktail containing 20 µg of mouse Cot-1 DNA for 30 minutes at –80°C. Following centrifugation for 20 minutes, hybridization cocktail DNA was resuspended in 17 µl of 100% deionized formamide and 17 µl of 2x hybridization mix [23]. Hybridization cocktail DNA was denatured at 70°C for 10 minutes and incubated at 37°C for 30 minutes to allow blocking of repeat sequences in the probe. Feeder-free cells were plated on eight-well chamber slides for 18 hours in ESC, fusion, or MEF media, respectively. Slides were processed in cytoskeletal buffer for 2.5 minutes, followed by formaldehyde fixation [24].

Following overnight hybridization at 37°C, slides were washed three times in 50% formamide/2x standard saline citrate (SSC) at 45°C for 5 minutes each, followed by 1x SSC (prewarmed to 60°C) at 45°C for three washes of 5 minutes each. Detection using anti-dig-rhodamine was then performed [25]. Cells were counterstained with 4,6-diamidino-2-phenylindole, mounted in antifade and viewed on a Leica DM6000B equipped with a DFC350FX CCD camera (Leica, Heerbrugg, Switzerland, http://www.leica.com), and analyzed on a CW4000 Cytogenetics Image Analysis Workstation (Leica).

Hybridization signals were scored for a minimum of 50 cells per well [26], with the exception of MEF cells, which contained no positive signal and for which 25 cells were scored. For statistical purposes, ESC and hybrid cells with no hybridization signals were excluded from subsequent analyses, since it could not be determined whether hybridization was absent because of the condition of the cell, its timing in the cell cycle, or to bona fide gene expression variation [27]. Chi-square statistical tests were performed to determine whether the observed peak signal distribution of four hybridization signals within each hybrid cell line was significantly different than the peak signal distribution of ESCs (2) or MEFs (0).

EB Assays
EBs were generated from J1 puro^R ESCs and MEF/ESC hybrid cell lines as follows. Cells were expanded to near-confluence, and EB cultures were initiated by trypsinizing to yield cell aggregates. As soon as cell clusters began to detach, trypsinization was stopped by the addition of 5 ml of ESC medium without LIF, and cells were transferred onto bacterial cell culture dishes in the appropriate media (G-418, puromycin for hybrids, and puromycin for ESCs). After 5 days, 50 to 100 nonadherent EBs were plated onto 10-cm tissue culture dishes coated with 0.2% gelatin. RNA samples were collected immediately after trypsinization (d0), and 3, 6, 9, and 34 days after initiation of EB differentiation. Random hexamer-primed first-strand cDNA was prepared with a SuperScript III reverse transcriptase kit (cat no. 18080-051; Invitrogen) according to manufacturer's instructions. RT-PCR was performed with the following primer pairs: Fgf5: 5'-GCTGTGTCTCAGGGGATTGT-3' and 5'-CACTCTCGGCCTGTCTTTTC-3'; Hbb-bh1: 5'-TGGACAACCTCAAGGAGACC-3' and 5'-TGCCAGTGTACTGGAATGGA-3'; Myh6: 5'-CAGAGGAGAAGGCTGGTGTC-3' and 5'-CTGCCCCTTGGTGACATACT-3', Sox17: 5'-GGAGGGTCACCACTGCTTTA-3' and 5'-AGATGTCTGGAGGTGCTGCT-3', Oct4, 5'-ATGGCATACTGTGGACCTCA-3' and 5'-CCTGGGAAAGGTGTCCTGTA-3', and GAPDH: 5'-AACTTTGGCATTGTGGAAGG-3' and 5'-ACACATTGGGGGTAGGAACA-3'.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion Disclosure of Potential... Acknowledgments References

 
Somatic/ESC Hybrids
Hybrid cells were created by fusing 5 x 10⁵ puromycin-resistant ESCs with 5 x 10⁵ G-418-resistant ROSA26-β-geo MEFs in the presence of PEG 1500 (Roche) (Fig. 1A). After 12 days of puromycin and G-418 selection, 14 colonies were identified and picked for further culture. From these, we established four stable cell lines, which could be repeatedly passaged and frozen. The colonies were named A3, A4, A7, and B9. We have subjected these cell lines to over 24 passages and have yet to detect senescence or decreased growth rates. No G-418^R,puro^R colonies were observed in control fusions consisting of only MEFs or ESCs fused to themselves.

View larger version (73K): [in this window] [in a new window]   Figure 1. MEF/ESC hybrid cell lines. (A): Hybrid production strategy: puromycin-resistant ESCs and MEFs harboring the ROSA26-β-geo transgene were fused in the presence of polyethylene glycol, and hybrids were selected in ESC medium supplemented with puromycin and G-418. (B): Genotype assays for ROSA26-β-geo transgene. Lane 1: PuroR ESC, lane 2: ROSA26-β-geo MEF, lane 3: ROSA26-β-geo mouse, lane 4: MEF/ESC line A3, lane 5: MEF/ESC line A4, and lane 6: MEF/ESC line A7. (C): Colony morphologies of ESCs and MEF parental fusion partner cell lines (top panels). MEF/ESC hybrid lines form colonies of ESC-like morphology on feeder layers but stain blue in the presence of 5-bromo-4-chloro-3-indolyl-β-D-galactoside because they express the ROSA26-β-geo transgene. Abbreviation: MEF, mouse embryonic fibroblast.

Karyotypic Stability of MEF/ESC Hybrids
We analyzed the chromosomal content of the MEF/ESC hybrid lines, and normal ESCs as a diploid control (Fig. 2). The MEF/ESC hybrid lines contained near-tetraploid chromosome complements, with a higher variability in the number of chromosomes present within spreads than the diploid control (Fig. 2A). Hybrid line A3 had a modal number of 78 chromosomes, whereas in A4 the mode was 80, in A7 the mode was 79, and in B9 the mode was 79 (Fig. 2). The J1-puro ESC line contained chromosomes with a modal number of 40, as did MEFs (data not shown). We observed four out of 50 spreads from line A3 with approximately 40 chromosomes, which likely originated from contamination of mitotically inactivated feeder cells.

View larger version (53K): [in this window] [in a new window]   Figure 2. Karyotype analysis of MEF/ESC hybrid lines. (A): Histograms of the number of chromosomes in individually scored chromosome spreads from normal ESCs and MEF/ESC lines A3, A4, and A7 at passage 13. (B): Representative karyotypes from each cell line.

We operationally defined reprogramming simply as the alteration of steady-state levels of transcripts in comparisons of one or more cell lines. Therefore, we subjected the array data to statistical analyses to identify genes that were differentially expressed between one or more of 15 samples (representing five cell lines, each in triplicate). We surmised that reprogramming events would likely include cases in which a gene is actively transcribed in one or more cultures, but silent in others. For this reason, we chose not to use "present or absent" calls, but rather to subject the entire expression database to ANOVA-based approaches that would identify genes with statistically different levels of expression between one or more of the five cell lines. This approach allowed us to identify genes that are (a) expressed in some cell lines but silenced in others or (b) expressed at statistically different levels in comparisons of two or more cell lines.

We subjected the complete dataset to two separate ANOVA software programs to find reprogrammed genes in MEF/ESC hybrids as compared to their parental fusion partners. We first subjected the array data to the dChip software program, which was developed specifically for the analysis of Affymetrix arrays [17, 29]. The dChip ANOVA module assesses array data calculating statistical F-tests that determine the likelihood of differential expression values across an exhaustive matrix of pairwise comparisons. dChip ANOVA analysis identified 4,258 differentially expressed genes at a significance threshold of p = .05. The second program, R/MAANOVA, is an open-source software program written in R-language [19] that uses a battery of four independent F-tests for analysis of microarray data. Using the R/MAANOVA software and robust multiarray analysis normalization, we identified 2,309 genes with a p value less than .015 that are expressed differentially in the hybrid cells compared with one or both parental cell lines.

The R/MAANOVA and dChip algorithms use different normalization, background correction, and F-test procedures that lead to different false-positive and false-negative error rates in the identification of significant differences. For this reason, we compiled a list of genes that were detected as differentially expressed by both methods, which contained 1,694 differentially expressed target transcripts, each associated with a unique probe on an MG 430 GeneChip. The 1,694 targets correspond to 1,597 GenBank accession numbers and 1,041 unique Entrez ID numbers. The 1,694 differentially expressed target transcripts do not correspond one-to-one with individual genes because many genes have multiple transcripts (with corresponding Entrez IDs that result from alternative splicing and alternative promoter usage). The 1,694 target transcripts therefore represent the subset of at least 1,041 differentially regulated genes for which there is the highest confidence of differential expression (as identified by both ANOVA applications). We observed that RNA from MEF/ESC hybrid lines was 60-fold down to 170-fold upregulated with respect to RNA from MEFs. In addition, steady-state expression levels spanned a range from 10-fold down to fivefold upregulated when we compared RNA from MEF/ESC hybrid lines with RNA of ESC origin.

Hierarchical Clusters of Similarly Reprogrammed Genes
The ANOVA analysis resulted in a dataset of 1,694 targets that were differentially expressed in at least one of the 15 RNA samples representing ESCs, MEF/ESC hybrids, and MEFs. We subjected this dataset to further analysis by applying it to the hierarchical clustering algorithm of the dChip software application. This approach organized the transcripts into clusters with similar patterns of expression (Fig. 3). Overall, the results indicated that the differentially expressed genes in MEF/ESC hybrids are nonstochastic, with MEF/ESC hybrid gene expression being much more similar to the ESC fusion partner as opposed to the MEF fusion partner. The dChip module also generated a dendrogram that parsed the 15 individual RNA samples (each from independent cultures) into clades based on the degrees of similarity of expression of all 1,694 target transcripts. This expression-based cladistic analysis indicated that the MEF transcriptome was quite divergent from either ESCs or MEF/ESC hybrids. Inspection of a second dendrogram revealed that two major classes of reprogrammed expression exist: (a) transcripts that are expressed at lower levels in MEFs compared with either ESCs or MEF/ESC hybrids and (b) transcripts that are expressed at higher levels in MEFs compared with either ESCs or MEF/ESC hybrids.

View larger version (29K): [in this window] [in a new window]   Figure 3. Hierarchical clusters of genes differentially expressed in ESCs, MEF/ESC hybrids, and MEFs. RNA was gathered from three independent cultures of ESCs (J1.1, J1.2, and J1.3), each MEF/ESC hybrid line (A3.1, A3.2, A3.3; A4.1, A4.2, A4.3; A7.1, A7.2, and A7.3), and MEFs (M1.1, M1.2, and M1.3). Independent transcriptional profiles were collected from each RNA sample. Genes (1,694) were found to be differentially expressed (bright green represents low or silenced expression, whereas bright red represents high levels of transcription). Differentially expressed gene data were subjected to a dChip hierarchical clustering algorithm (shown here), which computationally identified ESCs, hybrids, and MEFs as individual clades. Abbreviations: Hi, high; Lo, low; MEF, mouse embryonic fibroblast.

View larger version (23K): [in this window] [in a new window]   Figure 4. Clusters of reprogrammed and differentially expressed genes. (A): Transcriptional array data from all MEF, HYB, and ESC cultures were averaged and then subjected to a dChip hierarchical clustering algorithm, resulting in a composite clustergram. The data indicate that the hybrid transcriptome is more similar to the ESC transcriptome than the MEF transcriptome. (B): Seven major clusters of reprogrammed or differentially expressed genes were identified. Diagrams depict the relative patterns of expression of genes within each of the three cell types within a given cluster. Abbreviations: Hi, high; HYB, MEF/ESC hybrid line(s); Lo, low; MEF, mouse embryonic fibroblast.

We were interested in the distribution of genomic loci of the 1,694 genes upon the mouse genome. We used chromosome location abilities of the dChip software to determine the relative chromosome locations of each of the genes as well as the directions of transcription. Although this program does not allow us to place the genes on scaled chromosome ideograms for each mouse chromosome, the results indicate that the general distribution of all differentially expressed genes is random, and there is no apparent correlation between the direction of transcription (relative to centromeres) of reprogrammed genes.

Allele-Specific Assays of Reprogrammed Transcription in MEF/ESC Hybrids
The genes in hierarchical clusters 1, 2, and 3 are expressed at levels in hybrids that are similar or greater than those found in normal ESCs. To determine the chromosomal origins of transcription for genes in these clusters, we made use of known differences in the genetic backgrounds of the parental fusion partner cell lines (129Sv/Jae for ESCs and BALB/c for MEFs). We searched mouse genome informatics databases [20] to identify expressed SNPs in transcribed regions of genes from clusters 1, 2, and 3. This search yielded several SNP-tagged genes. We assessed allele-specific transcription for the class 3 genes Piwil2 (piwi-like homolog 2) and Chd1l (chromodomain helicase DNA binding protein 1-like) by direct sequencing of intron-spanning reverse transcriptase PCR products (Fig. 5A). All three hybrid lines expressed Piwil2 and Chd1l from chromosomes of both ESC and MEF origin.

View larger version (29K): [in this window] [in a new window]   Figure 5. Gene expression from chromosomes of MEF origin in MEF/ESC hybrids lines. (A): Expressed single nucleotide polymorphism (SNP) analyses of Piwil2 and Chd1l within hybrid line A3 and their ESC and MEF parental fusion partners. A diagram for each gene shows SNPs that distinguish transcription from ESC and MEF chromosomes, as well as the location of intron-spanning reverse transcription-polymerase chain reaction primers and sequencing primers to detect SNPs in cDNA. Sequence traces show that Piwil2 and Chd1l are expressed from chromosomes of both MEF and ESC origin in hybrids. (Similar expression results were obtained from lines A4 and A7, not shown.) (B): Intranuclear sites of Oct4 transcription as detected by intron-FISH for nascent Oct4 transcripts in MEFs, ESCs, and hybrid lines A3, A4, and A7. The histogram shows the percentage of cells with 0, 1, 2, 3, 4, or more sites of transcription in each cell type. (Asterisks denote cell types with four signals at a significance threshold of p < .001.) Representative microscopy images show sites of Oct4 transcription (red) merged with 4,6-diamidino-2-phenylindole in MEFs, ESCs, and hybrid line A3. Abbreviations: FISH, fluorescence in situ hybridization; MEF, mouse embryonic fibroblast.

Pluripotency of MEF/ESC Hybrids
We found that MEF/ESC hybrid lines, like normal ESCs, readily formed EBs (Fig. 6A). We determined that reprogrammed MEF/ESC hybrids could differentiate into the three primordial germ layers, even though they were tetraploid (Fig. 6B). To do this, we prepared RNA from pooled EBs collected at varying time points. We investigated transcription of the mesoderm markers Hbb-bh1, which encodes a β-like chain of fetal hemoglobin Z, and Myh6, which encodes myosin heavy chain 6. In wild-type ESCs, Hbb-bh1 message was first detected in day 9 EBs but was silenced in late day 34 EBs. Hbb-bh1 mRNA was first induced at day 6 in MEF/ESC lines A3 and A4 but was not induced until day 9 in line A7. Hbb-bh1 mRNA was greatly reduced if not absent from day 34 EBs from all three hybrid cell lines. This pattern of induction followed by silencing is reminiscent of the normal expression pattern of fetal hemoglobin genes in development in vivo, which is characterized by switching from fetal to adult globin genes. We found that expression of Myh6 was also induced during EB differentiation, although the kinetics of induction were slightly precocious in EBs prepared from MEF/ESC hybrids compared with J1 EBs. Myh6 expression was upregulated at day 9 in J1 EBs and day 6 for all hybrid cell-derived EBs, with persistent expression at day 34 in all samples. The endoderm specific marker SRY-box containing gene 17 (Sox17) was also upregulated upon differentiation, with similar expression patterns in both embryonic stem and hybrid cell-derived EBs. Up regulation was observed at day 6 for J1, A4, and A7 EBs, and at day 3 for A3 EBs. By day 34, Sox17 expression was reduced. Finally, we investigated an ectoderm-specific marker, Fgf5 (fibroblast growth factor 5). We were surprised to find Fgf5 expression throughout the course of EB formation for all cell lines. This may be a result of some ectodermal differentiation at the periphery of early-stage EBs. In addition, we found that Oct4 expression persisted throughout the course of the EB experiment, possibly because late EBs may harbor undifferentiated ESCs in their interiors. Nonetheless, the results demonstrate that MEF/ESC hybrid lines exhibit a remarkable degree of developmental potential to form all three germ layers.

View larger version (34K): [in this window] [in a new window]   Figure 6. Developmentally regulated gene expression in EBs from MEF/ESC hybrid lines. Embryoid bodies were formed from wild-type diploid embryonic stem cells (J1) and tetraploid MEF/ESC lines A3, A4, and A7. (A): Morphology of representative day 7 EBs derived from wild-type ESCs (J1) and MEF/ESC hybrid lines (A3, A4, and A7). (B): RNA was extracted from EBs at 0, 3, 6, 9, and 34 days after induction of differentiation. RNA was assessed for content of specific markers of germ layer induction, whereas GAPDH was used as a universal expression control. Abbreviations: d, day; GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion Disclosure of Potential... Acknowledgments References

The observed patterns of gene expression found within each cluster make it clear that the MEF/ESC hybrid lines contain a transcriptome that is not simply the additive sum of the expression patterns of the parental fusion partners. Cluster 6 is by far the largest and contains genes that are highly expressed in MEFs but not in ESCs or MEF/ESC hybrid lines. Thus, cluster 6 probably arises because of the silencing of genes on chromosomes of MEF origin, which is by far the most frequent reprogramming event observed, and is consistent with observations in hybrids between human fibroblasts and ESCs [14]. Inspection of gene annotation for cluster 6 reveals a broad array of functions for these genes, many of which are characteristic of differentiated cell types. In contrast, cluster 3 contains genes that are expressed at low levels in MEFs but at relatively high levels in both ESCs and hybrids. Inspection of cluster 3 revealed a very high content of genes with roles in the maintenance of pluripotency (including Oct4, Sox2, Rex1, and many other genes with annotations suggestive of stem cell function). In addition, cluster 3 contains a surprisingly high content of genes implicated in chromatin management and epigenetic function, such as Jumonji members and Polycomb group proteins, and other genes involved in the post-translational modification of histones, and Dnmt3l. Cluster 1 contains genes that are poorly (if at all) expressed in MEFs, moderately expressed in ESCs, but expressed at modestly higher levels in MEF/ESC hybrids. This class (of only 47 genes) includes the mouse Nanog gene and may represent genes that are transcriptionally activated upon chromosomes of MEF origin as a consequence of fusion with ESCs. Clusters 2 and 4 may contain genes that are recalcitrant to the reprogramming process since the levels of expression in the hybrids are intermediate between those observed in either of the parental cell types. Clusters 5 and 7 contain a surprising pattern of expression in which expression of the genes is lower in MEF/ESC hybrids compared with either parental fusion partner. This phenomenon could merely represent stochastic variation in the data, or possibly, that there may be mechanisms that cause gene silencing to occur upon chromosomes of both ESC and MEF origin in the hybrids.

The production of reprogrammed MEF/ESC hybrids is a low-frequency event. We obtained only 14 colonies that were resistant to both of the appropriate selective agents from 1 million parental cells. Of the 14 colonies, only four ESC-like lines were established. Although the efficiency of this process may be improved in the future, it is currently far less efficient than somatic cell nuclear transfer. This relatively low efficiency may be a consequence of comparatively low levels of reprogramming activity in a single ESC compared with an oocyte.

The results show that ESCs (like oocytes) contain endogenous biochemical activities that can reprogram the transcriptional status of somatic nuclei to a pluripotent state. Recent evidence suggests that the bulk of such reprogramming activity in ESCs is nuclear, because ESC-derived karyoplasts, but not cytoplasts, can activate an Oct4 promoter-driven green fluorescent protein reporter [13]. Animal oocytes contain reprogramming factors that are not associated with chromatin or the meiotic spindle; however, reprogramming activities within preimplantation embryos at the pronuclear stage may be sequestered within nuclei since the transfer of differentiated nuclei into enucleated zygotes results in poor preimplantation development to the blastocyst stage in vitro [31]. It may be that the subcellular location of reprogramming factors remains nuclear until the formation of the inner cell mass, from which ESCs are derived. Clearly, it is important to determine the identity of reprogramming factors and understand the mechanisms by which they function. However, these goals have been difficult to achieve in analyses of mammalian oocytes due to the exceedingly small amount of material available. Analysis of reprogramming factors in ESCs may be more fruitful. In a recent study, it was shown that overexpression of Nanog can increase the recovery of reprogrammed ESC/neurosphere hybrid cell lines [30]. Fusion-mediated reprogramming with ESCs may provide a much more tractable system to identify reprogramming factors, and at a minimum, this system should shed considerable light on the mechanisms involved in the reprogramming process.

Nuclear transfer can, in theory, produce pluripotent ESCs that are immunologically matched to a prospective patient that supplies somatic cells for purposes of reprogramming. In this process, nuclei from easily obtained somatic cells (such as skin fibroblasts) could be transplanted into oocytes that are devoid of the maternal nuclear genome, and the resulting embryos could be cultured in vitro for purposes of human ESC derivation. This process has already been shown to be feasible in an animal model [5]. Although this approach may conceptually be adapted to a human system, it would be difficult to implement on even a modest scale since every patient-matched ESC line derived from SCNT would likely require dozens of human oocytes.

Consequently, much interest has been placed on the identification of alternative methods to generate pluripotent cell types that are immunologically similar, if not indistinguishable, from somatic cells supplied by prospective patients. Recent advances demonstrate that ESCs contain potent activities that can reprogram the somatic nucleus to a pluripotent state. Cellular fusion is one method that can bring the reprogramming activities resident within ESCs into contact with somatic cells. This approach produces cells that, although pluripotent to a possibly useful degree, would likely not be a suitable source of engraftable cells, because they are nearly tetraploid and contain somewhat unstable chromosome contents. In addition, a problematic feature of somatic/ESC hybrid lines is that the configuration of transplantation antigens in such cells are at least in part determined by the ESC genome. Because the ESC genome would not be matched to a patient, cells from such hybrids would likely be rejected. Therefore, a reasonable next step would be to eliminate the ESC genome from somatic/ESC hybrid lines. Possible approaches include the removal of the ESC genome prior to fusion or the introduction of ESC-based extracts into somatic cells. The ideal patient-matched reprogrammed cells would be immortal, pluripotent, immunologically matched, and diploid so that they could be subjected to directed differentiation in vitro prior to engraftment.

Recently, it has been shown that a pluripotent state can be induced when a group of four transgenes are introduced into fibroblasts by retroviral transfection [32]. This select group of transgenes consists of Oct4, Sox2, c-Myc, and Klf4. Inspection of the list of differentially expressed genes presented in this study revealed that Oct4 and Sox2 were identified by the fusion approach, whereas c-Myc and Klf4 are absent from our lists of differentially expressed genes. Overall, the results suggest that there may be many combinations of gene expression that can lead to the establishment of pluripotent states and that levels of expression as well as gene identity may constitute important considerations.

	DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST

Top Abstract Introduction Materials and Methods Results Discussion Disclosure of Potential... Acknowledgments References

 
The authors indicate no potential conflicts of interest.

	ACKNOWLEDGMENTS

Top Abstract Introduction Materials and Methods Results Discussion Disclosure of Potential... Acknowledgments References

 
We thank Evan Barry and Therese Doherty for critical reading of this manuscript. This research was supported in part by NIH Grant RO1AG23687 and the Robert Leet and Clara Guthrie Patterson Trust for Biomedical Research. R.J.O. and C.O. were supported by the National Science Foundation (0093250) and the University of Connecticut Research Advisory Council.

	REFERENCES

Top Abstract Introduction Materials and Methods Results Discussion Disclosure of Potential... Acknowledgments References

 

1. Gurdon JB, Elsdale TR, Fischberg M. Sexually mature individuals of Xenopus laevis from the transplantation of single somatic nuclei. Nature 1958;182:64–65.[CrossRef][Medline]

2. Campbell KH, McWhir J, Ritchie WA et al. Sheep cloned by nuclear transfer from a cultured cell line. Nature 1996;380:64–66.[CrossRef][Medline]

3. Wilmut I, Schnieke AE, McWhir J et al. Viable offspring derived from fetal and adult mammalian cells. Nature 1997;385:810–813.[CrossRef][Medline]

4. Jaenisch R, Eggan K, Humpherys D et al. Nuclear cloning, stem cells, and genomic reprogramming. Cloning Stem Cells 2002;4:389–396.[CrossRef][Medline]

5. Rideout WM 3rd, Hochedlinger K, Kyba M et al. Correction of a genetic defect by nuclear transplantation and combined cell and gene therapy. Cell 2002;109:17–27.[CrossRef][Medline]

6. Ambrosi DJ, Rasmussen TP. Reprogramming mediated by stem cell fusion. J Cell Mol Med 2005;9:320–330.[Medline]

7. Matveeva NM, Shilov AG, Baiborodin SI et al. [Hybrids between murine embryonal stem and somatic cells maintain their pluripotency]. Russian. Dokl Akad Nauk 1996;349:129–132.

8. Matveeva NM, Shilov AG, Kaftanovskaya EM et al. In vitro and in vivo study of pluripotency in intraspecific hybrid cells obtained by fusion of murine embryonic stem cells with splenocytes. Mol Reprod Dev 1998;50:128–138.[CrossRef][Medline]

9. Tada M, Takahama Y, Abe K et al. Nuclear reprogramming of somatic cells by in vitro hybridization with ES cells. Curr Biol 2001;11:1553–1558.[CrossRef][Medline]

10. Kimura H, Tada M, Nakatsuji N et al. Histone code modifications on pluripotential nuclei of reprogrammed somatic cells. Mol Cell Biol 2004;24:5710–5720.[Abstract/Free Full Text]

11. Tada M, Morizane A, Kimura H et al. Pluripotency of reprogrammed somatic genomes in embryonic stem hybrid cells. Dev Dyn 2003;227:504–510.[CrossRef][Medline]

12. Hatano SY, Tada M, Kimura H et al. Pluripotential competence of cells associated with Nanog activity. Mech Dev 2005;122:67–79.[CrossRef][Medline]

13. Do JT, Scholer HR. Nuclei of embryonic stem cells reprogram somatic cells. STEM CELLS 2004;22:941–949.[Abstract/Free Full Text]

14. Cowan CA, Atienza J, Melton DA et al. Nuclear reprogramming of somatic cells after fusion with human embryonic stem cells. Science 2005;309:1369–1373.[Abstract/Free Full Text]

15. Yu J, Vodyanik MA, He P et al. Human embryonic stem cells reprogram myeloid precursors following cell-cell fusion. STEM CELLS 2006;24:168–176.[Abstract/Free Full Text]

16. Van Gelder RN, Von Zastrow ME, Yool A et al. Amplified RNA synthesized from limited quantities of heterogeneous cDNA. Proc Natl Acad Sci U S A 1990;87:1663–1667.[Abstract/Free Full Text]

17. Li C, Wong WH. Model-based analysis of oligonucleotide arrays: Expression index computation and outlier detection. Proc Natl Acad Sci U S A 2001;98:31–36.[Abstract/Free Full Text]

18. Barberi T, Klivenyi P, Calingasan NY et al. Neural subtype specification of fertilization and nuclear transfer embryonic stem cells and application in parkinsonian mice. Nat Biotechnol 2003;21:1200–1207.[CrossRef][Medline]

19. Kerr MK, Martin M, Churchill GA. Analysis of variance for gene expression microarray data. J Comput Biol 2000;7:819–837.[CrossRef][Medline]

20. Eppig JT, Bult CJ, Kadin JA et al. The Mouse Genome Database (MGD): From genes to mice—a community resource for mouse biology. Nucleic Acids Res 2005;33:D471–D475 (Database issue).[Abstract/Free Full Text]

21. Rozen S, Skaletsky H. Primer3 on the WWW for general users and for biologist programmers. Methods Mol Biol 2000;132:365–386.[Medline]

22. Smit AFA, Hubley R, Green P. RepeatMasker 1996–2005 Available at http://repeatmasker.org.

23. Ferreri GC, Marzelli M, Rens W et al. A centromere-specific retroviral element associated with breaks of synteny in macropodine marsupials. Cytogenet Genome Res 2004;107:115–118.[CrossRef][Medline]

24. Nutt SL, Vambrie S, Steinlein P et al. Independent regulation of the two Pax5 alleles during B-cell development. Nat Genet 1999;21:390–395.[CrossRef][Medline]

25. Brown JD, Strbuncelj M, Giardina C et al. Interspecific hybridization induced amplification of Mdm2 on double minutes in a Mus hybrid. Cytogenet Genome Res 2002;98:184–188.[CrossRef][Medline]

26. van Raamsdonk CD, Tilghman SM. Optimizing the detection of nascent transcripts by RNA fluorescence in situ hybridization. Nucleic Acids Res 2001;29:E42.[CrossRef][Medline]

27. Ohno M, Aoki N, Sasaki H. Allele-specific detection of nascent transcripts by fluorescence in situ hybridization reveals temporal and culture-induced changes in Igf2 imprinting during pre-implantation mouse development. Genes Cells 2001;6:249–259.[Abstract]

28. Zambrowicz BP, Imamoto A, Fiering S et al. Disruption of overlapping transcripts in the ROSA beta geo 26 gene trap strain leads to widespread expression of beta-galactosidase in mouse embryos and hematopoietic cells. Proc Natl Acad Sci U S A 1997;94:3789–3794.[Abstract/Free Full Text]

29. Li C, Wong HW. DNA-Chip analyzer (dChip). In: Parmigiani G, Garrett ES, Irizarry R, Zeger SL, eds. The Analyses of Gene Expression Data: Methods and Software.New York: Springer Verlag,2003;.

30. Silva J, Chambers I, Pollard S et al. Nanog promotes transfer of pluripotency after cell fusion. Nature 2006;441:997–1001.[CrossRef][Medline]

31. Wakayama T, Tateno H, Mombaerts P et al. Nuclear transfer into mouse zygotes. Nat Genet 2000;24:108–109.[CrossRef][Medline]

32. Takahashi K, Yamanaka S. Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors. Cell 2006;126:663–676.[CrossRef][Medline]

This Article Abstract Full Text (PDF) Supplemental Data All Versions of this Article: 2006-0532v1 25/5/1104    most recent Alert me when this article is cited Alert me if a correction is posted Citation Map Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints/Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Ambrosi, D. J. Articles by Rasmussen, T. P. Search for Related Content PubMed PubMed Citation Articles by Ambrosi, D. J. Articles by Rasmussen, T. P.