Identification of factors involved in the maintenance of embryonic stem cell self renewal and pluripotency 2

Chapter2 Results and Discussion 2.1 ES cell secretion of LEFTY2 maintains pluripotency 2.1.1 Introduction The molecular basis of self-renewal and the maintenance of pluripotency in ES cells have yet to be fully elucidated Here, I identify additional factors and delineate additional pathways important for “stemness” using a candidate gene approach was adopted ES cells are derived from the ICM of blastocysts and are similar (although not completely identical) to the pluripotent cells of the ICM Hence, to shortlist candidates in an informed and intelligent manner, extensive perusal of literature, particularly pertaining to genomic studies on blastocysts, preimplantation embryos, the inner cell mass and ES cells were carried out Some published literature that were exploited to shortlist candidates include MPSS data by Wei et al (2005), ChIP-PET data by Loh et al (2006) performed on mouse ES cells, ChIP-chip data by Boyer et al (2005) performed on human ES cells and other genomic data such as gene expression profile changes during mouse preimplantation development (Hamatani et al., 2004) Several criteria were used in the selection of genes The first criterion was the expression level of the candidate genes in pluripotent cell types of both mouse and human, such as in ES cells in the in-vitro scenario, primitive ectoderm, epiblast and the inner cell mass etc when we talk about relevance to in vivo physiology Genes that are highly expressed in a particular cell type may play a critical role in the regulation of cellular 41 processes specific to that cell The in depth transcriptome profiling by Wei et al (2005) using MPSS also highlighted a list of genes with differential expression patterns between ES cells and differentiated cells and these may be prime candidates to examine further for a role in ES cell self renewal Of primary interest are genes whose expressions are restricted to the embryonic stages, as pluripotency is characteristic of embryo derived stem cells The similarity in co-expression pattern between the candidate genes and Oct4, Nanog or Sox2 was also used as a selection criterion, as the co-expression suggests that the genes can potentially regulate or are regulated by Oct4, Sox2 and Nanog Little is currently known about what is upstream of Oct4, Sox2 and Nanog The studies by Loh et al and Boyer et al have however revealed the putative downstream targets of OCT4, SOX2 and NANOG These targets were of particular interest as it is hypothesized that these genes form key nodes in the OCT4/SOX2/NANOG network and plays major functional roles in ES cells, possibly as gatekeepers against differentiation by repressing differentiation inducing genes such as Gata6 and Lhx1 Genes linked to the WNT pathway such as Zfx were also included This is because WNT is a highly critical conserved pathway that may be critical in maintaining ES cell self renewal The use of a synthetic pharmacological inhibitor of GSK3 in ES cells has been reported by Sato et al to sustain the short term pluripotency and self-renewal via activation of the Wnt pathway (Sato et al., 2004) 42 Using these criteria, 70 candidate genes were shortlisted An RNAi screen in mouse ES cells (E14) using OCT4 and Nanog promoter activity as readouts was then performed in joint effort with my colleagues to further narrow down the choice of candidate genes This thesis documents only the candidate genes for which I was solely responsible for the cloning of their shRNAs and measurements of OCT4 / Nanog promoter activity in their knocked down cells (See Appendix) The rationale behind the use of the luciferase assay is as such: In wild type ES cells, OCT4 and Nanog promoters are active because the factors that regulate it are present However, when the ES cells differentiate, there is a consequential shutdown of the OCT4 and Nanog promoters This would lead to a reduction in measured luciferase activity If any gene is important for the maintenance of the undifferentiated state of the ES cells or regulation of Oct4, Sox2 or Nanog, ablating it by RNAi would reduce the luciferase activity Co-transfection of the OCT4 or Nanog promoter luciferase construct with short-hairpin RNA (shRNA) overexpression constructs can hence enable us to rapidly and effectively screen for candidate genes that have a biological effect in maintaining the undifferentiated cell state (Figure 2) One interesting factor identified from the screen is Lefty2, a secreted protein well known for its role in regulating Nodal signaling during embryo development Data that are a validation to the functional importance of Lefty2 in ES cell biology are presented in this chapter The RNAi screen also identified Lefty1, the homolog of Lefty2 to be a potential candidate, since its depletion by RNAi also perturbed Oct4 and Nanog promoter activities Here, data that show a role for Lefty1 in ES cell self renewal is presented The 43 mechanism via which Lefty and operates to maintain ES cell self renewal is also addressed 2a pOct4-Luciferase DE PP PE PE CR2 CR4 CR3 TGA TSS Human OCT4 promoter CR1 Firefly Luciferase PolyA ATG Oct4/Sox2 Figure 2a OCT4-promoter luciferase reporter comprising kb of the Human OCT4 promoter upstream of the transcription start site (TSS) for firefly luciferase The OCT4SOX2 binding site is indicated (DE- distal enhancer; PE- proximal enhancer; PPproximal promoter; CR- human-mouse conserved region, ATG- translation start codon; TGA- translation stop codon; PolyA- polyadenylation signal) 2b pNanog-Luciferase Murine Nanog promoter TGA TSS Firefly Luciferase Oct4/Sox2 PolyA ATG Figure 2b Nanog-promoter luciferase reporter comprising mouse Nanog sequences (from positions -289 to +117 relative to the transcription start site) upstream of firefly luciferase reporter Position of OCT4-SOX2 binding site within Nanog promoter is indicated Nanog transcription start site is defined by the position of the furthest 5' EST in public database (NCBI) (ATG- translation start codon; TGA- translation stop codon; PolyApolyadenylation signal) 44 Figure 2c A schematic representation of the pSuper.puro vector which was used for the overexpression of shRNA The vector consists of a H1 RNA polymerase III promoter that drives the endogenous production of short hairpin RNA (shRNA) Oligo duplexes were inserted between the BglII and HindIII restriction sites The vector consists of a PGK-pur cassette that enables for positive selection with puromycin (Pur- puromycin) 45 2.1.2 Results 2.1.2.1 Identification of Lefty2 as a candidate for self renewal and pluripotency from a high throughput RNAi screen To ensure reliability and specificity of the silencing effect, 2-4 shRNA constructs were designed for each gene to be knocked down The RNAi experiment was performed on feeder-free E14 mouse ES cell cultured in serum based ES cell growth medium supplemented with LIF Puromycin selection was used as a means to select for cells transfected with the shRNAs (Figure 3a) The endogenous levels of the targeted genes could be reduced to less than 50% in most cases by third day of puromycin selection It was found that at least 50% of the candidate genes selected for knockdown were able to reduce the Oct4 luciferase activity, by 25% or more Most of the genes which perturbed Oct4 promoter activity such as Rex2, Lefty1 and Lefty2 were also able to the same for Nanog promoter activity (Figure 3b and 3c), with the exception of Zfx Of these genes, some were transcription factors such as Zfx and Rex2 and some were signaling molecules like the Lefty proteins which are involved in the Nodal pathway This chapter first describes the functional dissection of Lefty2, followed by that of its homolog, Lefty1 46 Figure 3a The application of shRNA coupled with puromycin drug selection in E14 mouse ES cells was able to reduce the endogenous levels of targeted gene candidates to less than 50% in most cases, as evidenced by quantitative PCR analysis of mRNA level of candidates Mean values ± standard error of the mean (SEM) are plotted as percentages relative to scrambled shRNA control (100%) for qPCR measurement of gene candidates Transfection experiments were performed in technical triplicates for each biological sample and in biological duplicates (n=2) 47 250 200 150 100 50 Vector Non-silencing Rnai Lef Rnai Lef Rnai Lef Rnai Dot 1l Rnai Dot 1l Rnai Catnb Rnai Catnb Rnai Catnb Rnai Catnb Rnai Catna1 Rnai Catna1 Rnai Catna1 Rnai Zfx Rnai Zfx Rnai Zfx Rnai Zfx Rnai lefty2 rnai lefty2 rnai lefty2 rnai Lefty1 RNAi Lefty1 RNAi Lefty1 RNAi Lefty1 RNAi Gata3 RNAi Gata3 RNAi Gata3 RNAi Gata3 RNAi Rex Rnai Rex Rnai Rex Rnai Rex Rnai Oct4 Rnai Luciferase activity normalized to wrt non sil ctrl (%) Oct4 luciferase activity non-silencing control (%) 300 Figure 3b Knockdown of factors such as Lefty2, Lefty1, Zfx and Rex2 perturbed Oct4 promoter activity The strategy of RNAi knockdown of ES cell-specific genes coupled with an Oct4 promoter luciferase reporter was employed to test candidate genes for their ability to perturb OCT4 promoter activity when depleted At least two shRNA constructs were made for each targeted gene and the cell line used is E14 Oct4 RNAi was used as a positive control Luciferase values were normalized to a non-targeting control shRNA construct Transfection experiments were performed in biological triplicates (n=3) 48 120 100 80 60 40 20 Vector Non-silencing RNAi Catb RNAi4 Catb RNAi1 Catna RNAi4 Zfx RNAi4 Zfx RNAi2 Lefty2 RNAi3 Lefty2 RNAi1 Lefty1 RNAi4 Lefty1 RNAi1 Gata3 RNAi4 Rex2 RNAi3 Rex2 RNAi2 Nanog RNAi Oct4 RNAi Nanog luciferase activity normalized to non silencing control (%) 140 Figure 3c Genes such as Lefty2, Lefty1 and Rex2 that perturbed OCT4 promoter activities also affected Nanog promoter activity The strategy of RNAi knockdown of ES cell-specific genes coupled with a Nanog promoter luciferase reporter was employed to test candidate genes which can perturb Nanog promoter activity when depleted Oct4 and Nanog RNAi served as positive controls Luciferase values were normalized to a non-targeting control shRNA construct Transfection experiments were performed in biological triplicates (n=3) and the cell line used is E14 49 Activin A DMSO Control SB43154 pCNA P-Smad2 Figure 41a P-SMAD2 level is upregulated by Activin A treatment and is decreased in the presence of SB431542, as shown by western blot analysis Nodal/Activin signaling gradient was generated using Activin A treatment and 10µM SB431542, with the DMSO control as basal level PCNA was used as loading control for this western blot analysis Protein was harvested for analysis after18 hours of Activin A treatment 2.2.2.9 Nodal target genes respond to different signaling thresholds through differential occupancy of regulatory elements by SMAD2 It is clear now that ES cell fate decisions are mediated by the level of Nodal signaling How different level of signaling actually translates to different cell fates remains elusive One possible explanation is that differing level of pathway activities are interpreted by the cells to give distinct transcriptional responses It is hypothesized that Nodal signaling can achieve its concentration dependent effects on gene expressions in part through the differential binding of its intracellular effectors, the receptor regulated SMADs (SMAD2 and/or SMAD3) to regulatory elements of target genes at the various thresholds To confirm this hypothesis, the R-SMADs binding occupancy of regulatory elements of known Nodal responsive genes has to be correlated to the magnitude of 142 Nodal signaling To quantify the interaction between SMAD2 and Nodal target gene regulatory elements, a p-SMAD2 chromatin immunoprecipitation (chIP) assay was used The decision to perform a pSMAD2 chIP was based on the knowledge that SMAD2 appears to be the master transcription factor that relays Nodal signals during differentiation events in ES cells and early developmental processes of embryos Knockout of the Smad2 gene in mice results in early embryonic lethality at embryonic day 7.5-12.5 This lethality is in part attributed to the failure of Smad2-/- embryos to gastrulate, form primitive streaks or head folds or develop mesodermal cells (Brown et al., 2007) There is also evidence that ES cell commitment into mesodermal cell type in vitro is also SMAD2 dependent For example, Zeineddine et al has recently demonstrated that upregulation of Oct3/4 expression is specifically required for mesodermal commitment and cardiac commitment of ES cells and this Oct3/4 upregulation is specifically SMAD2 dependent (Kitamura et al., 2007; Zeineddine et al., 2006) Graded Nodal signals were generated using the Tet-on/ALK4* on TAG1 ES cells and SB431542 To control for the effect of DMSO- the solubilizing agent for SB431542, a third sample treated with DMSO was generated as a basal signaling level control(Figure 41b) Briefly, the samples for P-SMAD2 ChIP were generated by subjecting TAG1 ES to induction with 1ug/ml doxycyclin, inhibition with 10uM SB or treatment with DMSO for 15hours in chemically defined, LIF free media was then carried out The sample size was 50 million cells each The doxycyclin treated, DMSO treated and SB431542 treated samples represent the three signaling levels of the gradient - elevated above basal, basal 143 and repressed state of Nodal signaling respectively The samples were crosslinked with formaldehyde, and the fragmented chromatin lysates were then subjected to chIP using a phospho-SMAD2 (Ser465/467) antibody from Abcam A mock chromatin immunoprecipitation with rabbit IgG (rIgG) antibody was used to control for non specific binding events To detect SMAD2 DNA occupancy, primers flanking the asymmetric enhancer (ASE) regions of Lefty2 and Pitx2 were designed and the ASE regions were scanned using qPCR The rationale behind the choice of these genomic locations for pull down with SMAD2 antibody is explained as follow ALK4* On DMSO Control SB431542 Actin P-Smad2 High Low Basal Nodal signaling gradient Figure 41b P-SMAD2 level is upregulated during ALK4* induction in TAG1 ES cells and is decreased in the presence of SB431542, as shown by western blot analysis Nodal gradient was generated using ALK4* induction and 10µM SB431542 with the DMSO control as basal level Actin was used as loading control for this western blot analysis Protein was harvested for analysis at 15th hour of ALK4* induction 144 Several lines of evidence point to the fact that the ligand responsible for activation of Lefty2 and Pitx2 during developmental processes such as left right determination is Nodal In vivo, Nodal signals are mediated in part through FAST2 It is known that FAST2 has to form a complex with SMAD2/4 to elicit Nodal induced response at the transcriptional level FAST2 binding sites have since been located in the asymmetric enhancer (ASE) region of Lefty2 (Saijoh et al., 2000) The Lefty2 ASE contains two FAST binding sites (AATCCACAT) essential for its enhancer activity Using a lacZ reporter gene assay, FAST2 has been confirmed by Saijoh et al to be able to transactivate Lefty2 asymmetric enhancer activity in vivo in response to Nodal signals in mouse embryos Mutagenesis studies have also verified that these FAST2 binding sites are essential to fully activate Lefty2 ASE In the ASE of Lefty2, two potential SMAD binding elements (SBE) with the consensus sequence GTCT have been located 10bp and 50 bp downstream of the 3’ FAST binding site The proximity between the SBE and FAST2 binding sites suggests that these two factors may interact and cooperate with each other in the regulation of Lefty2 and Pitx2 ASE activity Indeed, Saijoh et al have demonstrated that Lefty2 ASE is activated by a complex that includes FAST2, SMAD2 and SMAD4, although definitive binding by SMAD2 itself to the ASE region has not yet been shown Similarly for Pitx2, the ASE contained within its fifth intron contains three FAST binding sites-AATACACA, TGTGGATT, and AATAGACA (Shiratori et al., 2001) Shiratori et al has previously demonstrated that the Nodal initiation of Pitx2 ASE activity was through FAST binding sites Similar to Lefty2, FAST binding sites were found to be sufficient for the induction of Pitx2 ASE activity Also, the cooperative action of SMAD2 and FAST is essential for initiation of Pitx2 transcription For these reasons, we 145 postulated that the Pitx2 and Lefty2 ASE are potential genomic regions that may be bing SMAD2 in response to Nodal signals For analysis of DNA from chIP by qPCR, a cut-off of 4-fold enrichment of genomic sequences was used to define whether the Lefty2 and Pitx2 ASE were bound by P-SMAD2 Fold enrichment here represents the abundance of enriched DNA fragments over a control region that does not bind P-SMAD2 The shoulder regions of the two ASEs were used as the control region A mock chIP using rabbit IgG showed no significant enrichment over the entire surveyed region for both Lefty2 and Pitx2 ASEs For the doxycyclin and DMSO treated samples, the binding profiles showed a sharp peak at the Lefty2 and Pitx2 ASE regions, suggesting P-SMAD2 binding For doxycyclin treated samples, the fold enrichment was 17.9 and 59.4 above background at the Lefty2 ASE and Pitx2 ASE respectively As for DMSO treated samples, the fold enrichment was 6.1 and 16.3 times above background at the Lefty2 ASE and Pitx2 ASE respectively The fold enrichment for both Lefty2 and Pitx2 ASEs were higher for the doxycyclin treated samples as compared to the DMSO samples This suggests greater recruitment of SMAD2 to the Lefty2 and Pitx2 ASE as the level of Nodal signaling was increased above its basal threshold consistent with the fact that Lefty2 and Pitx2 are regulated positively by Nodal The binding profiles of the SB431542 treated sample showed specific abolishment of SMAD2 binding when Nodal signaling was reduced The fold of enrichment was 2.6 and 3.6 above background at the Lefty2 ASE and Pitx2 ASE respectively This suggests that there was no significant P-SMAD2 binding over both the Lefty2 and Pitx2 ASE regions interrogated for the samples treated with 10µM SB431542 146 This observation agrees well with the fact that 10µM SB431542 was sufficient to inhibit Nodal signaling, and hence the abolishment of expression of Nodal target genes Also, this observation confirms the specificity of the SMAD2 binding events observed in the doxycyclin and DMSO treated samples (Figure 42a and 42b) Here, the actual binding localities of SMAD2 on the Pitx2 and Lefty2 ASEs were identified for the first time This is also the first study that demonstrates the differential occupancy of Nodal responsive regulatory elements by P-SMAD2 in response to graded Nodal signaling levels 147 1k Figure 42a Schematic illustrating the Lefty2 genomic locus in mouse The exons are shown as brown boxes and promoters as white circles The arrow indicates the transcriptional direction ASE refers to asymmetric enhancer ChIP analysis showed differential levels of SMAD2 binding on the Lefty2 asymmetric enhancer located upstream of mouse Lefty2 in response to graded Nodal signals Control IP was performed with IgG antibody, and did not show any enrichment Thirteen primer pairs were designed to scan the region of analysis (Dox, DMSO and SB represent the three different signaling levels, that is, the elevated above basal, basal and repressed state of Nodal signals respectively.) 148 c ASE 1k Figure 42b Schematic illustrating the Pitx2 genomic locus in mouse The exons are shown as brown boxes and promoters as white circles The arrow shows the transcriptional direction ASE refers to asymmetric enhancer ChIP analysis showed differential SMAD2 binding on the Pitx2 asymmetric enhancer located in Intron5 in response to graded Nodal signals Control IPs was performed with IgG antibody, and did not show any enrichment Fourteen primer pairs were designed to scan the region of analysis (Dox, DMSO and SB represent the elevated above basal, basal and repressed state of Nodal signaling respectively.) 149 2.2.3 Discussion Cell fate decisions made during early embryo development are controlled by signaling pathways that acts to activate or repress specific genetic programs The general dogma is that specific factor or pathway specifies particular cell fates At the level of microenvironmental sensing and signal processing, this dogma however appears fallible It cannot account for how these factors and pathways can be multifunctional, playing roles in cell fate decisions in developmental processes and diseases Signal transduction pathways such as Nodal and WNT also appear to have promiscuous and seemingly irreconcilable roles in both the self renewal and lineage commitment of ES cells This baffles the interpretation of the roles of the various pathways and pathway components in ES cells, and hinders their applications In vivo hypomorphic alleles studies of Nodal and pathway components like Smad2 and Smad3 have taught us that Nodal, as a potent morphogen has concentration dependent effects and elicits cellular response in a dose dependent manner physiologically I therefore questioned if this principle holds true for ES cell fate determination in the in-vitro scenario as well Hence, a series of experiments were carried out to determine if ES cells respond to Nodal signals in a dose dependent manner and what the responses at the various thresholds are Interestingly, graded Nodal signals were found to govern distinct fates of mouse ES cells An increase of Nodal signals above basal level causes differentiation into predominantly the mesendodermal lineage while a repression of Nodal signals induced loss of pluripotency and subsequent fate 150 respecification into trophectoderm A moderate level is required to maintain ES cell self renewal, and a negative feedback mechanism that operates in part via LEFTY2 is active in ES cells to maintain this self renewing threshold of signals It seems that mouse ES cells can generate at least three cell fate decisions according to the level of Nodal signals (Figure 43) Whether graded Nodal signals generate similar responses in human ES cells awaits confirmation Nodal is well known for its role in mesoderm and endoderm formation as evidenced by studies in zebrafish and mouse models (Dunn et al., 2004; Schier and Shen, 2000) The Nodal pathway also takes part in the generation of an intermediate mesodermrelated cell population from undifferentiated human ES cells (Dvash et al., 2007) It has also been shown that regulated Nodal signaling promotes differentiation of the definitive endoderm and mesoderm from ES cells (Takenaga et al., 2007) Such evidence collected by other groups to date supports our observation that upregulated ALK4/5/7 signaling as achieved with Activin A treatment, ALK4* induction and Nodal transgene overexpression promotes differentiation of mouse ES cells into mesendodermal cell types My finding that depletion of Nodal signaling drives trophectodermal differentiation concurs with Wieduwilt’s findings in 2003 As revealed by his microarray analysis, genes representing greater than 10 fold change in teratomas generated using Smad2-/- ; Smad3 / double mutant ES cells over wildtype teratomas were almost exclusively of placental/trophoblast origin (Smad2-/- ; Smad3 / embryos used to generate teratomas were null for Smad2 and homozygous conditional for Smad3 Smad2/3 151 compound mutants are embryonic lethal and it is inherently difficult in generating ES cells from these embryos.) Examples of upregulated marker reported were Proliferin and prolactin like peptides that are well established marker for trophoblast giant cells Wieduwilt also frequently observed giant cells that stained positive for prolactin-like proteins A and B in the double mutant teratomas and rarely noticed a similar phenomenon in control teratomas -feto protein, generally taken as a visceral endoderm marker was also highly represented in the double null teratomas generated by Wieduwilt Taken together, Wieduwilt’s data indicate that Smad2/3 double mutant ES cells were by default directed to extraembryonic fates If we were to realize that the use of SB431542 can in fact be likened to a chemical means of generating a “Smad2/3 null” state, it is not difficult to see why data presented in this thesis is onsistent with that of Wieduwilt’s Together, the data conclusively demonstrated that blockage of Nodal/Activin pathway is sufficient to direct spontaneous differentiation of pluripotent ES cells toward a trophoblast fate Whether intermediate signaling thresholds generates other cell fates, or that it just attenuates the differentiation process is currently still unclear and awaits further investigation This can be achieved by studying the consequences of intermediates levels of Nodal and ALK4* activation, or by supplementing ES cells with intermediate levels of Activin A and SB431542 inhibitor It would also be very interesting to carry out a quantitative analysis on other pathway with effects on ES cells such as the BMP signaling mediated by SMAD1/5/8 using methodologies similar to those described in this thesis 152 Our study illustrates the sophistication of Nodal pathway regulation of ES cell pluripotency and lineage commitment The consequent importance of quantitative analysis of pathway effect on ES cell fates is also highlighted through our study The fact that determination of ES cell fate at the signal transduction level is not a binary on-off event is highlighted by our work This thesis agrees well with the implications of Niwa et al’s work on how the quantitative expression of the transcription regulator, Oct3/4 defines differentiation or self renewal of ES cells It is reported that the precise level of Oct3/4 governs three distinct fates of ES cells (Niwa et al., 2000) A less than twofold increase in Oct3/4 expression causes differentiation into primitive endoderm and mesoderm In contrast, repression of Oct3/4 induced trophectodermal differentiation The implications that can be made from Niwa et al’s study is that Oct3/4 must be kept at its basal level and deviation from this critical threshold induces differentiation Recently, Rodriguez et al (2007) also observed a similar dosage effect of OCT4 in human ES cells Zhang et al (2006) has also very recently reported that Sall4, a transcriptional regulator of Oct3/4 gave rise to trophectodermal and primitive endodermal differentiation of ES cells respectively when manipulated above and below its critical threshold This further supports Niwa’s observation and adds strength to the thesis that ES cell fate decisions are not on-off, but rather dependent on the dose of the regulator Nodal signaling can be mediated by SMAD2 or SMAD or both Here, an attempt to dissect the individual contributions of SMAD2 and SMAD3 to the various cell fates was made Inhibition of SMAD3 activity using the pharmalogical inhibitor, SIS3 blocks Nodal/Activin signaling induced mesendodermal specification, suggesting that 153 mesendodermal specification of ES cells is a SMAD3 dependent process It is currently unclear if SMAD3 work in concert with SMAD2 during mesendodermal differentiation as selective inhibition of SMAD2 activity has yet to be performed Nonetheless, Activin treatment on cells transfected with Smad2 RNAi cells or Smad2 knockout ES cells can be performed to evaluate the role of SMAD2 in mesendodermal specification On the other hand, inhibition of SMAD3 activity with SIS3 does not lead to ES cell differentiation towards trophectoderm Since inhibition of both SMAD2 and SMAD3 with SB431542 induces trophectodermal differentiation, while selective inhibition of SMAD3 does not yield the same phenotype, it appears that it is SMAD2, and not SMAD3 that is responsible for suppressing the exit of ES cells from the self renewing state towards trophectoderm To confirm this point, Smad2 RNAi can be carried out Otherwise, the same experiment could be repeated with a Smad2-/- ES cell line and observed for trophectodermal differentiation As with work carried out with pharmacological inhibitors, there is always the argument as to how specific the SIS3 inhibitor is To further validate the data obtained is not due to non-specific effects of the SIS3 inhibitor, Smad3 RNAi can be performed instead of using SIS3 Alternatively, the same experiment can be performed with Smad3 knockout cells instead to see if it yields similar experimental outcome We are also beginning to dissect the mechanism by which ES cells interprets Nodal gradient Here, it is shown that graded activation of Nodal signal transduction pathways is translated into quantitative differences in the level of SMAD2 phosphorylation It may be that graded Nodal signaling results in distinct cell fates 154 observed at the various thresholds as a result of differing P-SMAD2 dependent transcription Indeed, we have already shown differential recruitment of P-SMAD2 to binding sequences in a signaling dependent manner Mapping the target genes set of SMAD2 at the various thresholds, using chIP-chip or chIP-Seq technologies will further clarify how quantitative Nodal signals are translated qualitatively into distinct cell fates at the molecular level In conclusion, this work has highlighted the fact that graded Nodal signaling is critical for ES cell fate decisions Such an understanding will enable us to have a more robust control over ES cell fate and may eventually prove to be useful in advancing the application of ES cells to cellular therapies 155 Figure 43 Nodal/Activin signaling elicits a dose dependent effect on ES cell fates Nodal signals have to be tightly regulated at a critical threshold to maintain ES cells in the undifferentiated state Increasing the magnitude of signaling above basal level induces mesendodermal specification whereas reducing the magnitude below basal level leads to trophoblast specification 156 .. .2. 1 ES cell secretion of LEFTY2 maintains pluripotency 2. 1.1 Introduction The molecular basis of self- renewal and the maintenance of pluripotency in ES cells have yet to be... secreted into and deposited in the ES cell microenvironment 79 2. 1 .2. 8 LEFTY2 maintains ES cell self renewal by inhibiting Nodal signaling Next, the mechanism of LEFTY2 action was dissected The fact... analyses against the V5 tag detected the presence of both the precursor (42kDa) and processed forms of LEFTY2 (34kDa and 62 28kDa) in the medium conditioned by the transfected E14 cells (Figure 12b)