| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
a Germline Development Group, Center for Animal Transgenesis and Germ Cell Research, New Bolton Center, University of Pennsylvania, Kennett Square, Pennsylvania, USA;
b Laboratory of Developmental Biology, Department of Animal Biology, University of Pavia, Pavia, Italy
Key Words. Embryo • Embryonic stem cell • Nuclear transfer • Oct4 • Reprogramming
Correspondence: Michele Boiani, Ph.D., Max Planck Institute for Molecular Biomedicine, Mendelstraße 7, D-48149 Münster, Germany. Telephone: 0049-0251-980-2864; Fax: 0049-0251-980-2992; e-mail: mboiani{at}mpi-muenster.mpg.de
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionNote Added in ProofReferences |
|---|
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionNote Added in ProofReferences |
|---|
The initial experiments on cloning in amphibia by Briggs, King, and Gurdon demonstrated that the potential of the nucleus to direct normal embryonic development decreases progressively with the differentiation stage of the donor cell [4]. This phenomenon is notably evident in mouse cloning experiments using donor nuclei from somatic cells and pluripotent embryonic stem (ES) cells [5, 6]. After birth, Sertoli cell–cloned mice tend to die prematurely of hepatic failure or tumors [7] and cumulus cell–cloned mice tend to grow obese [8], whereas ES cell–cloned mice survive in larger numbers—provided they originate from F1 nucleus donors [5]. In these cases, the epigenetic make up of the donor nucleus exerts some priming and lasting effects on the clonal phenotype. Similarly, the proportion of fibroblast-derived and cumulus cell–derived bovine clones having zygote-like epigenotypes (as defined by patterns of methylation and acetylation of histone H3 lysine 9) correlates closely with the proportion developing to the blastocyst stage [3].
Although a nucleus-driven scenario for clones is logically appealing, environmental effects should be considered as well. In isogenic Avy and Axin(Fu) mice, the coat color and tail phenotypes show extremely variable expressivity linked to the activity of an inserted retrotransposon [9, 10], and diet can alter the methylation status of the retrotransposon and cause the phenotype change [11]. Differences between embryos of mouse strains that are traditionally considered blocking and nonblocking in vitro at the two-cell stage are contingent on the culture medium osmolarity [12]. In somatic cells, cytoplasmic compartments and pools of molecules are able to reallocate in response to changes in the culture medium composition [13]. In somatic cell–derived clonal embryos, certain idiosyncrasies of myoblast and cumulus cell clones referred to as somatic cell–like features are well documented but their causes remain elusive [14]. Environmental cues seem to merely modulate the developmental competence of already established zygotes [15], but the developmental competence of somatic cell clones is not yet established at nuclear transfer. In intact mouse oocytes, nuclear remodeling activities continue to operate after activation [16–18], albeit not to the extent observed when donor nuclei are exposed to metaphase II ooplasms [19, 20]. Alteration of pronuclear remodeling by culture conditions after oocyte fertilization would lead to a suboptimal, yet still zygotic, genetic program of embryonic development. In contrast, alteration of nuclear remodeling after somatic-cell nuclear transfer may either support or prevent the establishment of a zygotic genetic program, depending on the specific cues from the environment.
Is the reprogramming of the donor nucleus also contingent on the culture environment of the clone? The focus of this study was to investigate whether developmental pluripotency after cumulus-cell nuclear transfer ensues independently from, or is conditioned by, the culture environment as tested in a variety of forms.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionNote Added in ProofReferences |
|---|
Embryo and ES Cell Culture
Handling of oocytes, nuclear transfer from ovarian cumulus cells, and activation of reconstructed oocytes were performed essentially as described [23]. Six hours after the onset of activation, the constructs were washed carefully in 
minimal essential medium (-MEM) without cytochalasin B and randomly assigned to six culture groups (CZB, Dulbecco’s modified Eagle’s medium [DMEM], G1/G2, KSOM(aa), M16, and -MEM; Table 1
[24]). Zygotes were recovered after mating and cultured in parallel as controls. All embryos were cultured in 25-µl drops overlaid with washed mineral oil (catalog no. 151694; MP Biomedicals, Irvine, CA, http://www.mpbio.com) in 35-mm Petri dishes (catalog no. 430588; Corning, Corning, NY, http://www.corning.com) under 5% CO2 in air. All culture media were supplemented with 0.4% wt/vol bovine serum albumin (Pentex, code no. 81-068, lot no. 742 [Serologicals Corporation, Protein Fractionation Manufacturing, Kankakee, IL, http://www.serologicals.com/web/homepage.asp]) and 10 µg/ml gentamicin sulphate (catalog no. 194530; MP Biomedicals). At 48 hours of culture (four-cellstage), the CZB medium was supplemented with 5.56 mM D(+)glucose, because the basal CZB is devoid of glucose [25]; G1 medium was replaced with G2 [26] for the same reason (Table 1).
|
); clonal embryos that had developed up to blastocyst stage were freed of zonae pellucidae and plated onto feeder cells to derive ES cells following a standard method [27] with certain modifications (see below).
|
cDNA synthesis was performed immediately after RNA extraction using the High Capacity cDNA Archive Kit (catalog no. 4322171; Applied Biosystems, Foster City, CA, http://www.appliedbiosystems.com) following the manufacturer’s instructions. The flow-through of a MinElute Spin Column, as described above (10–11 µl), was transferred into a 0.2-ml thin-wall polymerase chain reaction (PCR) tube and mixed with 90 µl of reverse transcription (RT) reaction mixture as follows (final concentrations): 1 x RT buffer; 1 x random hexamers; 4 mM dNTP mix; 250 IU MultiScribe RT (provided in the High Capacity cDNA Archive Kit). Binding of random hexamers was allowed for 10 minutes at 25°C, and RT was carried out for 2 hours at 37°C, per the manufacturer’s instructions, on the heating block of a GenAmp 9700 thermalcycler (Applied Biosystems). Because mRNA RT did not involve an oligo-(dT) primer, but rather random hexamers, the possibility that poly(A) of different sizes may have altered the efficiency of RT and, therefore, the detection of such transcripts can be discounted.
Total mRNA was extracted from each blastocyst separately, reverse-transcribed, and split into 12 equal volumes of cDNA aliquots (five target genes, in duplicate, and two spare volumes). All transcripts, except Oct4 mRNA, were detected by quantitative RT-PCR (Q-RT-PCR) in pooled (n = 50) cumulus cells but at less than 1% of the oocyte level (five to seven amplification cycles later; data not shown). When spare cDNA volumes from the same embryo were processed, amplifications were reproduced within ± 0.4 SDs of the crossover threshold cycle (Ct), in line with previous reports [28, 29].
Light Cycler Q-RT-PCR: Quantification of Oct4, Bex1, Gapdh, Cpt2, and Hprt Gene Transcripts in Blastocysts
Total RNA was extracted from single blastocysts in a 75-µl volume using the RNeasy Micro Kit (Qiagen Inc.) according to the manufacturer’s instructions. cDNA synthesis was performed immediately after RNA extraction using the High Capacity cDNA Archive Kit following the manufacturer’s instructions. Transcript levels were determined using the Light Cycler Real-Time PCR system (Roche, Basel, Switzerland, http://www.roche.com) with SYBR green I as the fluorochrome [30, 31].
The individual embryo cDNAs pool was split into 12 reactions (one for each of the five genes, in duplicate, and two spare reactions), and the following four negative controls were run in parallel: H2O; H2O eluted on MinElute column; H2O reverse-transcribed; and sample RNA not reverse-transcribed. Five microliters of template was added to 15 µl of the following reaction mixture in a glass capillary tube (catalog no. 1909339; Roche): 1 x LightCycler FastStart DNA Master Plus SYBR green (catalog no. 03515885; Roche), 500 nM each primer (final concentrations). Primers for Gapdh (GenBank M32599 [GenBank] ) were 5'-TGGTTCCAGTATGACTCCACTCAC-3' and 5'-GAT-GACAAGCTTCCCATTCTCG-3' (107-bp amplicon); primers for Hprt (GenBank J00423 [GenBank] ) were 5'-GAAAGACTTGCTCGA-GATGTCATG-3' and 5'-ACAGAGGGCCACAATGTGATG-3' (60-bp amplicon); primers for Cpt2 (GenBank U01163 [GenBank] ) were 5'-CGGACCCGAAGTCTGAGTATAATG-3' and 5'-GAAACCG-CACTGCAGAAACAG-3' (72-bp amplicon); primers for Oct4 (GenBank M34381 [GenBank] ) were 5'-GTTGGAGAAGGTGGAAC-CAACTC-3' and 5'-CTTCAGCAGCTTGGCAAACTG-3' (88-bp amplicon); primers for Bex1 (GenBank AF051347 [GenBank] ) were 5'-TGACCACCATGATGAGTTTTGC-3' and 5'-GGTCCCCAT-GTCATCTTCAGAG-3' (69-bp amplicon).
The LightCycler profile was as follows: first denaturation and Taq polymerase activation at 95°C for 10 minutes; first 10 cycles of amplification at 95°C for 10 seconds, 65°C to 61°C for 10 seconds (touchdown of 0.4°C per cycle), 72°C for 10 seconds; followed by 35 cycles of amplification at 95°C for 10 seconds, 60°C for 10 seconds, and 72°C for 10 seconds. SYBR green I fluorescence was measured at the end of each extension step. Melting curve analysis was undertaken to determine the specificity of the PCR products, which were further confirmed by sequencing and Basic Local Alignment Search Tool (BLAST). Comparison of all amplicons with actual targets revealed greater than 99% homology for all amplicons. Sequencing and BLASTing of a product that was occasionally amplified by Oct4 primers in the cumulus cell cDNAs pool, but not in that of zygotes nor of clones, matched an X-linked sequence (GenBank LOC385358) similar to Oct4 that may have been previously encountered [32].
Transcript amounts were expressed relative to that of Hprt using the equation 2– (Ctclonal–Ctctrl), where Ct is the number of amplification cycles at which the fluorescent signal is first detected at a level that is 10 times above that of background for zygotes (Ct ctrl) or clones (Ct clonal).
When analysis of Xist (GenBank AJ421479 [GenBank] ) was required, the RLT lysate was split into two aliquots: 37.5 µl was used for RT and PCR of Xist RNA (with DNase treatment performed on column) with primers 5'-CATAGCCCCTTCCCAATAGGTC-3' and 5'-CTACTTGGGCGTTCACTTCAGAG-3' (143-bp amplicon); 37.5 µl was used for sexing with Y chromosome (Zfy locus)-specific primers and X chromosome (DXNds3 locus)-specific primers [33].
Chromosomal Analysis of Four-Cell Embryos and Analysis of X-Chromosome Inactivation
Because the five genes analyzed are located on four different chromosomes, we postulated that altered levels of gene expression in clones might be due to chromosomal imbalance and embryo mosaicism rather than to gene regulation after reprogramming. This is based on evidence that the first two mitotic divisions are nondisjunction-prone in mouse zygotes [34, 35] and mammalian clones may be more subject to this predisposition [36]. Four cell–stage embryos were incubated for 8 hours in -MEM containing 0.5 µg/ml colcemide to arrest cells at metaphase. Embryos were incubated in hypotonic sodium (Na)-citrate solution (0.9% wt/vol in Milli-Q water) for 10 to 15 minutes. They were then transferred individually onto a glass slide in a minimal volume of Na-citrate and fixed by dropping ice-cold methanol:acetic acid (3:1 volume) on top of the hypotonic solution [37]. Chromosome spreads were air-dried, stained with Hoechst 33342 at a concentration of 1 µg/ml in Sorensen’s buffer (pH 6.8), and mounted in Vectashield H-1000 (Vector Laboratories, Burlingame, CA, http://www.vectorlabs.com). A DM-IRBE–inverted fluorescence microscope (Leica Microsystems Inc., Bannockburn, IL, http://www.leica-microsystems.com) equipped with ORCA ER camera (Hamamatsu Corp., Bridgewater, NJ, http://www.hamamatsu.com) and Openlab software (Improvision Inc., Lexington, MA, http://www.improvision.com) was used to provide images of the chromosomes under UV light. Only spreads with >40 chromosomes (referred to as hyperaneuploid) were considered indicative of aneuploidy. The ratio of hyperaneuploid spreads over the total number of embryonic mitoses displaying countable chromosomes was taken as the rate of blastomere aneuploidy.
To determine whether regulation of another epigenetic process, such as X-chromosome inactivation, was affected under conditions that altered Oct4, Bex1, and Gapdh, the expression of Xist [28, 29] was examined in clonal as well as in female and male zygotic blastocysts. Regardless of the culture medium used, Xist RNA levels in clones were in the range of those observed in female zygotes and more than 10-fold higher than those observed in male zygotes (not shown). See Light Cycler Q-RT-PCR: quantification of Oct4, Bex1, Gapdh, Cpt2, and Hprt Gene Transcripts In Blastocysts.
Total Cell Count and Differential Labeling of the Inner Cell Mass and Trophectoderm of Blastocysts
The nuclei of blastocyst inner cell mass (ICM) and trophectoderm (TE) cells were differentially labeled with polynucleotide-specific fluorochromes as described [38] after the modified immunodissection procedure of Handyside and Hunter [39]. Whereas total cell numbers were lower in clonal than in zygotic blastocysts at 96 hours of development (p < .05), total cell numbers in clonal blastocysts were similar (p > .05) across culture media (M16: 45 ± 14; CZB: 36 ± 15; KSOM(aa): 44 ± 19; -MEM: 45 ± 18), as were cell numbers in the ICM (M16: 15 ± 5; CZB: 12 ± 5; KSOM(aa): 10 ± 4; -MEM: 17 ± 7) (n > 10 blastocysts in each group).
Whole-Mount Riboprobe In Situ Hybridization of Blastocysts
Riboprobe in situ hybridization (RISH) of Oct4 mRNA was performed essentially as previously described [40].
Derivation of ES Cell Lines from Clonal Blastocysts
After digestion of the zonae pellucidae with acidic Tyrode’s solution, groups of 8 to 10 blastocysts were transferred on to STO feeder cell layers in four-well dishes and were left undisturbed for 6 to 7 days without passage. The ICMs were selectively detached (picked with a polished glass Pasteur pipette) from the trophoblastic outgrowths in the presence of 0.25% trypsin/1 mM EDTA (Invitrogen catalog no. 25200-056) and replated onto a 96-well dish covered with mitotically arrested mouse embryo fibroblast (MEF) feeders. The ES cells were grown to subconfluency and gradually expanded by serial trypsinization, dissociation, and replating (3 days) onto freshly prepared MEF feeders, scaling up to larger well sizes. The medium for blastocyst outgrowth formation and ES cell culture consisted of DMEM high glucose (Invitrogen, catalog no. 11960-044), 15% fetal calf serum (ANF190A9; HyClone, Logan, UT, http://www.hyclone.com), 0.1 mM MEM nonessential amino acids (catalog no. 11140-050; Invitrogen, Carlsbad, CA, http://www.invitrogen.com), 100 U/ml Pen-Strep, 2 mM L-glutamine, 10 µM ß-mercaptoethanol, and 104 U/ml leukemia inhibitory factor. STO or MEF feeder cells were mitotically arrested with 10 µg/ml mitomycin C (catalog No. M4287; Sigma, St. Louis, http://www.sigmaaldrich.com) in DMEM for 2 hours.
In Vivo Development of Embryos
Transfer of zygotic and clonal blastocysts to the genital tract was performed essentially as previously described [40]. When required, the cavities of zygotic blastocysts (B6C3F1 x ICR) at 96 hours of development were microinjected with 10 to 15 clonal ES cells and then transferred into the uterus of 2.5 days post coitum (dpc) pseudopregnant ICR females. Development rate and developmental contribution were assessed at 10.5 and 13.5 dpc for clones and injected blastocysts, respectively.
Data Analysis
Means, SDs, analysis of variance (ANOVA), and correlation analysis were performed using JMP 5.1 (Statistical Analysis System, SAS Institute, Cary, NC, http://www.sas.com). Differences were considered significant when p < .05.
|
RESULTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionNote Added in ProofReferences |
|---|
). Karyotype analysis of clonal four-cell stage embryos detected hyperaneuploid sets of chromosomes (>40) in 10.8% ± 3.2% of blastomeres (n = 93). This did not differ significantly (p > .05) from the rate of hyperaneuploidy observed in blastomeres of parthenogenetic (10.1% ± 3.4%; n = 79) and zygotic (8.3% ± 2.4%; n = 133) embryos cultured under the same conditions (Boiani, unpublished data). Gene/allele regulation, rather than physical absence of loci, would therefore seem to be at the root of the different patterns of gene expression expected to be found in clones. Clones and zygotes had an equivalent ooplasmic composition, B6C3, thus preventing strain-specific ooplasmic modifiers from having an unequal effect on the incoming nucleus [43].
Culture Conditions Affect the Rate of Blastocyst Formation of Genetically Identical Mouse Clones
As long as the clones are generated with the same experimental setup and only subsequently allocated to different culture media, as was the case in this study, differences among clones depend on the culture media only. We assessed the developmental program of clonal and zygotic mouse embryos by culturing them in parallel up to the blastocyst stage. To this end, strain B6C3 mouse oocytes were enucleated and transplanted with B6C3 cumulus cell nuclei to derive the clones or were fertilized in vivo with B6C3 sperm to recover one-cell zygotes from the oviducts.
At 96 hours of development in vitro (108 hours after human chorionic gonadotropin), zygotes had formed blastocysts at similarly high rates (80%–85%) in all media tested except DMEM (8%; see Fig. 2A for absolute rates). The rate of clonal blastocyst formation progressively increased (multiple t-tests, p < .05) from DMEM through G1/G2, CZB, KSOM(aa), M16 to -MEM (1% ± 0.7%, 10% ± 3.5%, 26% ± 2.3%, 32% ± 2.5%, 38% ± 2.5%, and 52% ± 2.6%, respectively; see Fig. 2B for absolute rates). Diploid parthenogenetic B6C3 embryos activated by the same stimulus as that used for clones formed blastocysts at nearly the same rate across the different media (not shown), as reported for B6D2 parthenogenetic embryos [44] and B6C3 zygotic embryos [45, 46]. These parthenogenetic and zygotic mouse embryos demonstrate an inherent ability to maintain homeostasis, which seems lacking in clonal embryos. Therefore, the transplanted somatic genome may be more susceptible to environmental cues than a parthenogenetic or zygotic genome. Because culture protocols were identical with respect to the origin of the donor nuclei, the nuclear transfer procedure, and the activation method, the observed difference in clonal development was attributed to an unequal ability of the different culture environments to support reprogramming.
|
The expression level of Oct4 and other genes was analyzed in individual blastocysts by real-time Q-RT-PCR as follows: Oct4 (chromosome [chr] 17), Bex1 a.k.a. Rex3 (chr X), Gapdh (chr 6), Cpt2 (chr 4), and Hprt (chr X). These genes were chosen for the following reasons: Bex1 is a TE-restricted gene in mouse blasto-cyst development [48] whereas Oct4 is not [47], and Gapdh and Cpt2 are metabolic genes that are relevant to the health status of the embryo in culture [49, 50]. Hprt was used for normalization purposes. In theory, differences in the gene expression levels between clonal and zygotic blastocysts might simply reflect differences in the numbers of cells in these two types of blastocysts [38] or may be due to differences in the state of activity of the genes at the time of nuclear transfer rather than to differential gene regulation due to reprogramming. Hprt was used for normalization purposes after finding it not to be altered significantly in clonal and zygotic blastocysts within the culture medium groups (see ANOVA below). Hprt expression levels are also conserved in male and female zygotic mouse blastocysts [51].
Five single clonal blastocysts and five single zygotic blastocysts grown in each medium were analyzed by Q-RT-PCR. This approach relies on the buildup of the amplification product with each PCR cycle, which increases SYBR green I light emission until it crosses a threshold; hence, the lower the template amount, the higher the number of cycles. Raw Ct and SDs of clonal and zygotic blastocysts are shown in Table 2. One-way ANOVA of Ct values for Hprt by type of embryo produces nearly identical distributions for clonal and zygotic blastocysts in concert with Hprt as a normalizer (F1,71 = 0.0274, P > F 0.8689). The raw Q-RT-PCR data indicate that Oct4 mRNA amplifies above threshold 4.11 cycles earlier than Hprt in the -MEM medium, whereas it amplifies 3.54 cycles earlier than Hprt in the CZB medium. Thus, there seems to be less Oct4 mRNA in the CZB than in the -MEM sample on an absolute basis, in agreement with the RISH data. However, Q-RT-PCR data presented in raw format are not acceptable as they may merely reflect the amount of total mRNA extracted.
|
) was conducted with the 
Ct method (ABI Prism 7.700 User Bulletin 2; PerkinElmer, Wellesley, MA, http://www.perkinelmer.com) [52]. Zygotes assumed a value of 1, and clones that exhibited a higher or lower gene transcript level than zygotes were given a value greater than or less than 1, respectively. Oct4 transcripts were detected in all embryos analyzed, and in general, based on the normalized values, Oct4 was expressed at higher levels in clones than in zygotes (1 to 5.14, +414%). There were also notable variations in Oct4 transcript levels in CZB-cultured clones (1 to 14.96, +1,396%). These results were consistent with the lower Bex1 expression levels (1 to 0.27, –73%) based on opposite regional regulation of Oct4 and Bex1 genes in the blastocyst (Oct4 expression in the ICM, Bex1 in the TE). The level of Cpt2 relative to Hprt mRNA was similar in clones and zygotes across all media (1 to 0.86, –14%), whereas Gapdh mRNA levels fluctuated. Overall, Gapdh exhibited 185% higher (1 to 2.85, +185%) expression in clones compared with zygotes, with 118% of the total increase contributed by mRNA levels in one medium—CZB. As determined by karyotype and Xist analysis, gene/allele regulation rather than physical gain or loss of chromosomes, or X-chromosome inactivation status, seemed to cause the different patterns of gene expression (see Quality Control for Cumulus Cell Nuclear Transfer and Materials and Methods).
|
) that were paired in all possible ways leaving out reciprocal duplicates (e.g., correlation between Cpt2 and Oct4 expression levels in G1/G2, CZB, KSOM(aa), M16, and -MEM), based on the formula r = covariance [series1, series2]/std dev[series1]*std dev [series2]. Correlation coefficients were as follows: Bex1-Oct4, r = 0.004; Gapdh-Bex1, r = 0.136; Gapdh-Cpt2, r = 0.248; Cpt2-Oct4, r = 0.286; Cpt2-Bex1, r = 0.492; Gapdh-Oct4, r = 0.953 (Gapdh-Gapdh, r = 1). Oct4 and Bex1 patterns (r = 0.004) were actually inversely related (Oct4 and 1/Bex1, r = 0.893). This demonstrated that genes among the set analyzed in clones respond in a variable manner to different culture conditions, particularly Oct4. This seems contradictory to the critical requirement of Oct4 for the establishment of the pluripotent founder cell population in mouse embryos [47], to the tight Oct4 regulation in mouse ES cells [53], and to the observation that Oct4 mRNA levels are rather conserved in individual ICM cells of human blastocysts [54].
Functional, But Not Spatial, Definition of the Blastocyst ICM Is Altered by Culture Conditions
Until recently, quantitative analysis of high-resolution two-dimensional protein gels [55], mRNA differential display [56], analysis of expressed sequence tags derived from libraries of various preimplantation stages [57, 58], analysis of selected genes [59], and microarray techniques [52, 60, 61] have been used in gene expression studies. Although these approaches have shed some light on the molecular basis of preimplantation development, they are based on pooled embryos and thus offer limited insight into clonal embryo development wherein embryo-embryo variability occurs, even within the same batch of clones.
Alterations in the transcript levels of genes that are locally/regionally modulated cannot be detected by any mRNA profiling method, Q-RT-PCR, or microarray technology, unless each individual cell is profiled in situ. For instance, higher Oct4 expression levels (+414% on average; Table 3) and lower Bex1 expression levels (–73% on average) in clones relative to zygotes may indicate an expansion of the ICM compartment relative to a shrunken TE. The fact that all embryos tested positive for Oct4 transcript by Q-RT-PCR where as previous studies (using riboprobes and conventional RT-PCR) found that some tested negative as well [40] reflects, in part, the different sensitivity of these assays. Furthermore, it suggests that the amount of Oct4 mRNA may be so high but in only a few cells or so minute overall as to be functionally equivalent to zero when assessed spatially in situ.
Clonal blastocysts (n = 234 in total from the various media) were analyzed individually for the presence of an ICM, as defined by localized Oct4 mRNA RISH. Differences were observed between CZB, KSOM(aa), M16, and -MEM culture media (DMEM and G1/G2 did not yield enough blastocysts to compensate for losses incurred by RISH). Hybridized blastocysts were categorized [38] into four groups: having strong Oct4 signal restricted to the ICM as opposed to weak signal; having unrestricted signal (ICM + TE); or lacking signal in both (Fig. 3A). In general, clonal blastocysts had consistently lower rates of strong, ICM-restricted distribution of Oct4 mRNA than zygotic blastocysts cultured in the same medium (p < .05, except for KSOM(aa); Fig. 3B), although intensity levels were also lower (weaker signal) for zygotes after culture in CZB (not shown). Clones cultured in CZB and KSOM(aa) media exhibited rare and limited Oct4 expression, respectively; whereas a higher proportion of clonal blastocysts exhibited a restricted pattern of Oct4 expression (Fig. 3A, type D) in M16 and -MEM media. Thus, the spatial distribution of Oct4 transcript in clonal blastocysts is neither constant nor stochastic but varies with the culture conditions used, depicting a progressive increase in the rate of Oct4 mRNA restriction to the ICM (CZB < KSOM(aa) < M16 < -MEM; -MEM 
CZB at p < .05; Fig. 3B). The differences between genetically identical clones cultured in different media are overt in Figure 3C. We would like to stress that because RISH is orders of magnitude less sensitive than Q-RT-PCR, the lack of detectable signal in situ is not equivalent to lack of mRNA. Contrast between the poor spatial localization of Oct4 mRNA (Figs. 3A, 3B) and its higher relative amount (Table 3), particularly in the CZB-cultured clones, reflects the normalization that was introduced after the Q-RT-PCR, whereby the RISH data stand as absolute values, whereas the Q-RT-PCR data are relative to Hprt mRNA. There might also be a possibility that Q-RT-PCR detects not only the parent Oct4 transcript but also transcripts originating from Oct4 pseudogene(s) [62].
|
-MEM (p > .05) (see Materials and Methods). It follows that reinstatement of pluripotency after cumulus cell nuclear transfer is a primary effect of culture conditions on reprogramming rather than a side effect of cell number allocation or faulty chromosome segregation during cleavage of clonal embryos.
Consequence of the Culture Medium–Dependent Oct4 Gene Reprogramming
Altered Oct4 levels affect the pluripotency of ES cells [53], and lack of Oct4 in the ICM negates the further development of mouse blastocysts [47]. Therefore, it seems likely that differences in Oct4 mRNA levels in clonal blastocysts, observed more frequently than in zygotic blastocysts, also have an impact on clonal development and ES cell derivation. To track this, we compared the reprogramming of genetically identical cumulus-cell nuclei in oocytes cultured in different media by measuring Oct4-GFP expression in those nuclei (Fig. 4, see legend for absolute rates). The normal timing of maternal-zygotic transition to Oct4 gene expression occurs in mice at the eight-cell stage [63]. Although the onset of Oct4-GFP expression in zygotes at the eight-cell stage was equally frequent across all media (p > .05), the proportion of eight-cell clones having Oct4-GFP expression at 55 hours was lower in CZB and G1/G2 and higher in -MEM (p < .05). This different penetrance is notable in view of the fact that CZB and -MEM clones had been activated synchronously and formed four cells at similar rates (Fig. 2). Yet, after imaging, clones formed blastocysts at variable rates, as described. ES cell rates were determined as a measure of whether formation of an ICM endowed with more or less pluripotential character (Oct4 mRNA RISH) correlated with developmental outcomes, both in vitro and in vivo (Fig. 5).
|
|
): Half of the clones were cultured in CZB and half in -MEM, in parallel, from the one-cell stage up to the blastocyst stage, followed by derivation of ES cells. These blastocysts were plated in parallel onto feeder cells (Fig. 5A), using a standard method for derivation and culture of ES cells. Cell lines were identified as ES cells by observing the expression of Oct4-GFP after six passages (Fig. 5A’) and by blastocyst injection (see below). The rates of cell line derivation were found to be 8.2% for the CZB-cultured clones (n = 5 cell lines/61 blastocysts) and 39.3% for the -MEM–cultured clones (n = 33 cell lines/84 blastocysts) (p < .05). Routine derivation from F2 (B6C3xB6C3) zygotic blastocysts in our laboratory yields 20% cell lines, regardless of the embryo culture medium used. To extend this observation, more clonal blastocysts were produced (n = 225), half in CZB and half in -MEM. These were transferred at 96 hours of development into 11 recipient females, alongside zygotic blastocysts (n = 175) transferred in 8 separate but synchronous recipients. All recipients became pregnant, and fetal development was assessed at 10.5 dpc. Fetal rates for clones did not stand up to the projections of their blastocyst rates. Clones formed in -MEM (52% blastocysts) exhibited modest implantation (18.6%) and fetal developmental (9.1%) rates, whereas clones cultured up to the blastocyst stage in CZB (26% blastocysts) exhibited higher rates of implantation and fetal development (61.4% and 25.9%, respectively). Zygotes cultured in CZB and -MEM, while not resulting in higher implantation rates (24%–47%) than the clones, resulted in significantly higher fetal rates (50%–83%).
Altogether, these data show that the patterns of Oct4 gene expression in blastocysts cannot only be assessed as done previously [40] but also arise nonrandomly according to culture medium patterns and that higher rates of blastocyst formation and ES cell potential of clones in vitro did not translate into higher rates of fetal development (Fig. 5B). To determine whether clonal ES cells would be impaired under conditions that impaired fetal development from precursor blastocysts, clonal ES cells from -MEM were injected into the cavity of blastocysts of zygotic embryos and then assessed for germline contribution at 13.5 dpc. Nine out of 13 lines tested contributed to the formation of germ cells (Figs. 5A", 5A"’). Considering ES cells as the equivalent of the ICM, these results suggest that the reprogrammability of the donor nucleus to give rise to an ICM lineage is different from that required to give rise to a TE lineage.
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionNote Added in ProofReferences |
|---|
Clones derived from the same cluster of cumulus cells, activated simultaneously in the same medium, and differing only in the postactivation medium used for culture form blastocysts and express mRNA of the stem cell marker Oct4 at different rates. Discrepancy between the success rates presented here and those previously published is due in part to the scoring criteria (rigorously blastocysts at 96 hours postactivation in this study, blastocysts up to 120 hours in Chung et al. [66], or morulae-blastocysts in Wakayama et al. [67]). Regarding the differential expression of Oct4, although the formal possibility exists that that the epigenotype of the nucleus donor cells was mixed [3] and that the proportion of each epigenotype allocated to the groups of clones did not reflect the original proportions in the donor cell population, this possibility is deemed very unlikely in view of the number of clones allocated to each group and the number of replicates performed. The clonal phenotype could not be as variable (albeit not randomly) if the cell type and the epigenotype of the nucleus donor cells were solely responsible for developmental outcome. Rather, heterogeneity would be possible if reprogramming events occurred or continued to occur subsequent to the first cell division.
Examination of Oct4-GFP fluorescence during cleavage foreshadows differences seen in clonal versus zygotic embryos after culture to the blastocyst stage in various media. Aneuploidy and cell number progression of clones resembled those typical of zygotes up to the morula stage [38] (Boiani, unpublished observations), and clonal blastocysts were composed of about the same number of cells at the blastocyst stage across culture media. Therefore, chromosome malsegregation, nonuniform retardation of cleavage, and different health status among the clones in different media are discounted as the basis for the differential Oct4 as well as other gene expression. Rather, the activation of the silent, somatic Oct4-GFP alleles unfolds differently in different media, as detected 55 hours after nuclear transfer (29%–71% penetrance), with overt differences seen in blastocysts at 96 hours. Reactivation of the silent, thymocytic Oct4-GFP took approximately 36 to 48 hours to occur, at nearly 100% penetrance, after cell hybridization with ES cells [68]. Although the use of different, not interchangeable media for oocytes and ES cells makes it hard to compare these nuclear reprogramming systems, observations suggest that the erasure of the somatic cell–specific epigenotype either follows different paths in oocytes and hybridized ES cells (possibly related to the absence/presence of a companion nucleus [69]) or is simply delayed in the former. Subcellular compartments and pools of molecules are known to reallocate within the cell in response to environmental cues [13], and interactive processes between the cytoplasm and the nucleus stretch over a longer distance in an oocyte compared with an ES cell. This may account for response delay of reprogramming in the oocyte-based system. Concerning how the nucleus-cytoplasm interaction may sense the extracellular environment, leading to different patterns of gene expression in different media, this remains to be determined. Here we speculate that the enzymatic activity of Gapdh in the cytoplasm is engaged in a metabolic feedback loop with culture medium components (e.g., glucose and pyruvate), thereby affecting ATP and ATP-dependent chromatin remodeling activities [70] that are pivotal to the reactivation of embryonic genes silent in cumulus cells, such as Oct4. Variability in Gapdh gene expression among cumulus cell mouse clones was noted earlier [49]. Unlike that of Gapdh, mRNA of other housekeeping genes was not (Hprt) or marginally (Cpt2) altered by culture conditions.
Analysis of Oct4 gene expression in clones has relevance beyond the immediate question of how the clone regulates metabolism and gene expression, as the Oct4 transcription factor is essential for the establishment of ES cells from the ICM [47] and their maintenance. An increase in the levels of Oct4 expression in mouse ES cells by approximately 50%, by experimental manipulations, leads to differentiation of these cells into endoderm and mesoderm; reduced levels result in the development to the TE, and intermediate levels preserve the pluripotent stem cell status [53]. Thus, small changes in the expression of a transcription factor could lead to widespread and dramatic changes in gene expression associated with various phenotypes. In clonal blastocysts formed in different media, the variations in Oct4 mRNA levels were not well correlated with the size of the ICM; i.e., 14.96-fold the Hprt mRNA level over the control in blastocysts with 12 ± 5 ICM cells (CZB) and 3.58-fold the Hprt mRNA level in blastocysts with 17 ± 7 ICM cells (-MEM). According to these observations, the level of Oct4 mRNA in each individual ICM cell would appear to be rather dissimilar. Although the formal possibility exists that variation is present also in the zygotes used as normalizers, this possibility is deemed unlikely in view of their degree of homogeneity in terms of developmental rates (high) and restriction of Oct4 mRNA to the ICM (frequent).
Although this assumption is now being contended [71], ES cells are often assumed to be the equivalent of ICM cells. Therefore, we attempted to derive ES cells from clonal blastocysts cultured in CZB and -MEM media and to transfer clonal blastocysts to females, anticipating a positive correlation between the rates and quality of blastocyst ICM formation and subsequent development. Clones grown in -MEM were superior to those grown in CZB with respect to ES cell derivation rates (39.3% versus 8.2%), in accord with the prevalence of ICM-restricted Oct4 mRNA distribution in the two culture media. This is notable because recent experiments using embryo carcinoma (EC) cells as donors for nuclear cloning have resulted in derivative clonal ES cells that retained the traits of the donor cells—that is, no epigenetic changes [72]. Previously, epigenetic heterogeneity of ES cell–derived mouse clones was ascribed to pre-existing heterogeneity and epigenetic instability of ES cells [73]. In contrast, we show that genetically identical cumulus cell nuclei have multiple and distinct ES cell potentials, suggestive of the implementation of different epigenetic traits depending on the extracellular environment of the clone undergoing reprogramming. The rate of ES cell derivation from the -MEM cumulus cell clones is an unprecedented high reported to date for somatic cell donors (39.3%), only lower than the rate of derivation after cloning from ES and EC cells, albeit an MEK1 inhibitor was used there to facilitate derivation [72]. Because EC and ES cells already express Oct4 and therefore derivative clones do not depend on the reinstatement of Oct4 expression, it will be of interest to test the response of these clones to different culture conditions. It will also be of interest to challenge the reprogramming-efficient -MEM environment with hard-to-clone cell types. Although it has been shown that nuclear transfer can in principle reprogram terminally differentiated B lymphocytes, this was highly inefficient [74]. In fact, the recent discovery that the olfactory receptor gene choice does not involve genetic alterations [75], unlike the case for immunoglobulin genes [74], did not rest on advances in understanding of clonal developmental biology but on the incorporation and expansion of the rare, viably reprogrammed nuclear states into ES cells.
Surprisingly, the success in ES cell formation contrasted fetal development in both media (9% and 26% for -MEM and CZB, respectively). Definitely, the efficiency of ES cell derivation does not predict the potential of clonal blastocysts to develop into fetuses. The critical requirement of Oct4 for the establishment of the pluripotent cell population precursor of germ layers in early mouse development may seem contradictory to the dual outcome observed, that is, loss of the embryo but survival of ES cells (-MEM) and survival of the embryo but loss of ES cells (CZB). Albeit in a different scenario, Oct4 has been shown to exert a dual role: cell fate determinant before embryo gastrulation and cell survival factor (for germ cells) after gastrulation [76]. The lack of a positive association between ES cell potential and fetal development of -MEM clonal blastocysts can be explained by the following possibly superimposed possibilities.
First, blastocysts may form at high rates due to efficient reprogramming up to that stage, but genes required thereafter are not reprogrammed properly. It is not clear why genes not yet expressed at the blastocyst stage would have been reprogrammed differently in the different media. However, ooplasmic epigenetic modifiers may tag genes expressed after implantation. Modifier genes may cause blastocysts with normal gene expression to express such tagged genes abnormally, thereby causing embryos to fail at midgestation [77].
Second, reprogramming errors affect all genes but may be less detrimental to the development of a preimplantation embryo than for subsequent stages due to the increasing number of functions required. This is consistent with the observation that ES cells could still be derived even if the conceptus suffered a condition that became overt at a stage following that of ES cell derivation [78, 79]; this situation is not equivalent to the ability of parthenogenetic and androgenetic mouse embryos to give rise to ES cells [80, 81]. One fundamental difference is that these uniparental embryos originate with an imprinting imbalance that per se dooms fetal development, unlike somatic cell clones that are derived from biparental nuclei.
Third, the blastocysts formed may contain a functional ICM but an impaired TE (reprogramming errors would be at genes that affect the TE but not the ICM, e.g., E-type cyclins [82] and retinoblastoma [83]), or both the ICM and the TE are impaired, but a subset of functional ICM cells is sufficient to derive ES cells whereas a subset of functional trophectodermal cells is not sufficient for successful implantation. This scenario is consistent with the lower levels of Bex1 mRNA observed in G1/G2 and -MEM clones compared with CZB and KSOM(aa) that correlate with the low rates of development up to 10.5 dpc observed in the former group. Recently published work showed that imprinted gene expression (H19 and Snrpn) was mostly preserved in the mouse embryo proper but not in the extraembryonic tissues after culture in vitro and assessment in vivo at 9.5 dpc [84]. Previous work [3] showed that bovine blastocysts produced by cloning fail to establish epigenetic asymmetry between ICM and TE for either DNA or H3-K9 histone methylation, with the TE as highly methylated as the ICM. This may underlie the substantial proportion of genes whose expression is downregulated in placentae of clonal embryos, as well as the characteristic placental abnormalities of mouse clones [85]. The trophoblast, although essential for development in utero, can hinder ICM growth during culture and make ES cell line establishment quite problematic [86].
Fourth, the lack of a positive association between ES cell potential and fetal development of -MEM clonal blastocysts is because ES cells do not correspond to ICM cells but to an early germ cell state [71]. Therefore, the ability to derive ES cells does not necessarily predict the fetal competence of the ICM.
In conclusion, although the Oct4 expression has been used by several groups as a marker of reprogramming in the strict sense that its silent state must be reversed after somatic cell nuclear transfer, this change unfolds in a variable manner depending on the embryonic environment, and development to the blastocyst stage in the presence of Oct4 is not sufficient for a broader definition of reprogramming (also Oct4–/– embryos form blastocyst-like structures [47]). Although epigenetic errors of gene expression arise very early in clonal development [49, 87], they have no inevitable ending but are tractable experimentally, as shown by the zygotic-like Oct4 gene expression obtained after clone-clone embryo aggregation [38]. Although the precise causal mechanisms of the nucleus-oocyte-medium interaction described here in have not been determined, epigenetic marks at the Oct4 locus are labile and can be experimentally manipulated in embryo-derived cell lines [88], raising the possibility that molecules active in chromatin remodeling may hijack this process if they are present in the culture medium and internalized by the embryo. Thus, new ways of reprogramming cell fate before or without transplanting nuclei into oocytes may be envisioned [89], as seen in myotubes reverting to myoblasts after exposure to myoseverin [90] or myoblasts forming osteoblasts after exposure to reversine [91]. Among the soluble, cell-permeable molecules active in chromatin remodeling, which may play a role in (clonal) embryos, the inositol precursor of nuclear inositol phosphates [92, 93] is present in the -MEM and is internalized by the mouse embryo via a sodium transporter [94]. The molecular size of inositol phosphates is compatible with a postulated role of small molecules in nuclear reprogramming after the observation that Oct4 can be activated from the silent somatic state in HeLa cells placed in connection with mouse embryo blastomeres [95].
|
NOTE ADDED IN PROOF |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionNote Added in ProofReferences |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionNote Added in ProofReferences |
|---|
|
REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionNote Added in ProofReferences |
|---|
