Abstract
The canonical and noncanonical NFκB signaling pathways regulate a variety of cellular activities; however, their functions in human embryonic stem cells (hESCs) have not been fully investigated. Expression studies during hESC differentiation indicated a significant increase in the expression of two key components of the canonical NFκB pathway (p50 and Ser529 phosphorylated form of p65) as well as a significant reduction in expression of key components of the noncanonical NFκB pathway [v-rel reticuloendotheliosis viral oncogene homolog B (RELB), p52, NIK]. Inhibition of canonical NFκB resulted in hESC apoptosis, changes in cell cycle distribution, and reduced hESC proliferation. In addition, inhibition of canonical NFκB was associated with significant changes in NANOG and OCT4 expression, suppression of differentiation toward all primitive extraembryonic and embryonic lineages with the exception of primitive ectoderm and ectodermal lineages. Inhibition of noncanonical NFκB via small interfering RNA-mediated downregulation of RELB resulted in reduced hESC proliferation and opposite changes to expression of key differentiation lineage markers genes when compared with downregulation of canonical NF-κB. Chromatin immunoprecipitation assays indicated binding of p65 and RELB to regulatory regions of key differentiation marker genes suggesting a direct transcriptional role for both branches of this pathway in hESC. These findings coupled with opposing trends in expression of key components during hESC differentiation, suggests a fine and opposing balance between the two branches of NFκB signaling pathways and their involvement in two distinct processes: the canonical pathway regulating hESC differentiation and the noncanonical pathway maintaining hESC pluripotency.
Introduction
NFκB was identified 24 years ago as a nuclear factor that binds the κ light-chain enhancer in B cells is a cardinal regulator of inflammatory and innate immune responses [1]. A large body of research suggests that NFκB activation is implicated in various aspects of cell proliferation, migration, cell cycle progression, and apoptosis [2, 3]. Activation of NFκB and its ability to confer cell survival is linked to several types of cancer including hematological and epithelial malignancies [4]. On the other hand, inhibition of NFκB might increase the incidence of cancers of the liver and skin suggesting cell-specific effects [5–7].
The mammalian NFκB family is composed of five subunits, namely p65 (RelA), v-rel reticuloendotheliosis viral oncogene homolog B (RelB), c-Rel, p52, and p50, [8] which combine to generate a variety of heterodimeric and homodimeric transcription factors. The five subunits share a Rel homology domain, which is responsible for DNA binding, dimerization, and nuclear translocation. The p65, RelB, c-Rel proteins are synthesized as mature forms and contain a transactivation domain at their C-terminus enabling them to directly stimulate transcription. The p50 and p52 subunits are generated from precursors, p105 and p100, and lack a transactivation domain and cannot stimulate transcription unless partnered with p65, RelB, or c-Rel. Interestingly, the C-terminal region of p105 and p100 contains a series of ankyrin repeats that are characteristic of inhibitors of NFκB (IκB) [9, 10].
In resting cells, the majority of NFκB subunits are associated with a family of proteins called IκB, which includes IκBα, IκBβ, IκBε, IκBγ, IκBζ, BCL3, p100, and p105 [11]. These proteins act as chaperones of the NFκB subunits and prevent their binding to DNA. On receiving relevant cell stimulation, the IκBs are degraded and NFκB is released to translocate to the nucleus where it binds to κB DNA binding motifs in the regulatory region of genes controlling inflammation and cell proliferation, differentiation, and fate.
There are currently three known mechanisms of NFκB activation [12]. The canonical pathway culminates in the degradation of IκBα and is achieved via activation of the IkappaB kinase (IKK) complex. The core IKK complex contains two kinases, IKK1/α and IKK2/β along with a noncatalytic subunit called IKKγ, more commonly known as NEMO. Activation of the IKK complex leads to IKKβ-mediated serine phosphorylation of IκBα triggering polyubiquitination and proteasome-mediated degradation. In addition to IKK-mediated IκB degradation many other events are essential for canonical activation, the best characterized so far being phosphorylation of p65. This can be achieved by many kinases such as PI3K/v-akt murine thymoma viral oncogene homolog (AKT), NFκB inducing kinase (NIK), p38, protein kinase C (PKC), protein kinase A (PKA), etc [13]. The phosphorylation of p65 is thought to enhance nuclear transport, increase DNA binding of NFκB and stabilization of p65 by inhibiting its binding to IκBα.
The second mode of NFκB activation, the so-called noncanonical pathway is independent of IKKβ and IKKγ and is crucial to lymphoid organogenesis [14]. A small number of stimuli such as lymphotoxin-β and B-cell activating factor (BAFF) are known to activate this pathway via an upstream kinase called NIK. A different IKK complex composed of IKKα dimers phosphorylates p100, (a sequestrator of RelB), which is partially processed to generate the p52:RelB dimer. The third pathway of NFκB signaling is activated in response to DNA damage caused by agents such as UV or doxorubicin that results in IκB degradation in IKK-independent fashion [13]. This results in dimerization of free NFκB subunits that are then mobilized in a similar way to canonical NFκB signaling.
A number of recent publications suggest that NFκB signaling plays a role in stem cell proliferation and survival [14]. A very relevant example is provided by the activation of the canonical NFκB pathway, which is shown to stimulate proliferation of neural stem cells via upregulation of the bona fide target gene cyclin D1 [15]. Genetic loss of both p65 and p50 NFκB subunits also results in reduced numbers of neural progenitors and an increased proportions of neurons suggesting that the effects of NFκB signaling are likely to affect both the proliferation and differentiation of neural stem cells [16]. An overexpression of the NFκB components and targets are also found to be a major characteristic of hematopoietic progenitor cells found in umbilical cord blood [17], although the functional significance of this pathway remains to be evaluated in this cell type. NFκB pathway also plays an important role as a survival factor for lymphoid progenitor cells and recent findings have revealed that inhibition of this pathway impairs the generation of lymphoid cells from adult bone marrow and fetal liver hematopoietic stem cells [18] suggesting that the activation of this pathway is important for the differentiation of adult hematopoietic stem cells.
Virtually all members of the NFκB pathway are expressed in embryonic, trophoblast, and uterine cells in a developmental stage and cell-specific manner [19]. It has been suggested that the antiapoptotic effect of NFκB in these embryonic stages is indispensable for proper development during organogenesis and is important for resistance toward stress-induced processes. More recently, it has been shown that murine and human embryonic stem cells (hESCs) possess a low level of NFκB activity that increases significantly during the differentiation process [20, 21]. The low NFκB activity in hESC does not, however, exclude a role for this pathway. Indeed, the work done in our group suggests that the NFκB pathway plays a crucial role in the maintenance of viability in hESC [22]. In accordance with this, it has been shown that karyotypically abnormal hESC and embryonic carcinoma cells express CD30, a member of tumor necrosis receptor subfamily that is responsible for the activation of the canonical NFκB pathway [23]. In addition, BAFF, which is known to induce the noncanonical NFκB signaling pathway is used in defined medium together with other growth factors for pluripotent culture of hESC [24].
In this article, we have investigated the expression and role of the canonical and noncanonical NFκB pathways in proliferation, viability, and differentiation capability of hESC. Our results suggest that canonical NFκB is involved in hESC differentiation and hESC viability, while noncanonical NFκB is mainly implicated in maintenance of hESC pluripotency. Binding of both p65 and RELB to promoters of several lineage-specific genes indicates that some of these effects are likely to be mediated at the level of gene transcription.
Materials and Methods
Culture and Differentiation of hESC
The hESC lines H9 and H1 (WiCell, Madison, Wisconsin, USA) were routinely passaged and maintained in hESC media on mitotically inactivated mouse embryonic fibroblast (MEF) feeder layers and differentiation was achieved by forming embryoid bodies (EBs) as described by Stojkovic et al. [25]. In 1–2 passages prior to experiments, hESCs were transferred to Matrigel-coated plates with feeder-conditioned media as previously described [25]. In all figures (unless indicated in figure legends), average data from both cell lines is presented.
Reverse Transcription Polymerase Chain Reaction
This method is described in Supporting Information Annex A.
Cell Cycle Analysis
This method is described in Supporting Information Annex A.
Apoptosis Assay
This method is described in Supporting Information Annex A.
Western Blotting
Lysates were subjected to electrophoresis on a 10% SDS-polyacrylamide (PAGE) gel and transferred to a polyvinylidene difluoride membrane (Hybond-P, Amersham, Buckinghamshire, UK). Membranes were blocked in Tris-buffered saline with 5% milk and 0.1% Tween. Blots were probed with primary antibodies overnight and revealed with horseradish peroxidase-conjugated secondary antibodies. Primary antibodies used include NFκB p105/p50 antibody (Abcam, Cambridge, UK; ab32360), NFκB p100/52 (Santa Cruz sc-7386), NFκB p65 (Santa Cruz sc-372), RelB Antibody (Cell Signaling #4954), c-Rel (Santa Cruz sc-70), NIK Antibody (Cell Signaling #4994), IKKα Antibody (Cell Signaling #2682), NFκB p65 (phospho S529) antibody (Abcam ab10684), Phospho-NFκB p65 (Ser536) Antibody (Cell Signaling #3031), OCT4 [POU5F1] antibody, clone 7F9.2 from (Millipore, Watford, UK; # p20263), NANOG antibody (Aviva Systems Biology, San Diego, California, USA; # P100591_P050), PAX6 antibody (Millipore # P26367), BETA ACTIN antibody (b-Actin [C4] from Santa Cruz Biotech. (Heidelberg, Germany) # sc-47778), and GAPDH antibody (# Abcam ab9485). Antibody-antigen complexes were detected using ECL (Amersham Biosciences) and images were acquired using a luminescent image analyser FUJIFILM and LAS-3000 software (FUJI).
Small Interfering RNAs and Transfection
hESCs were cultured under feeder-free conditions with feeder conditioned media free of antibiotics for at least 4 days prior to transfections. Small interfering RNAs (siRNAs) for RelB and p65 were obtained from Invitrogen (Paisley, UK). The siRNA sequences are shown in Supporting Information Table 2. Transfection with scrambled control siRNAs with similar GC content to gene-specific siRNA sequences provided by the same company were used as a negative control. Transfection of siRNA into hESC was carried out using the high efficiency nucleofection kit L from Amaxa (Cologne, Germany) and 80 pmol siRNA (in 2 ml medium) as outlined in manufacturer's instructions (program A-023).
Luciferase Reporter Assay
Plasmid DNA was prepared using Maxiprep kit (Qiagen, Sussex, UK). The IκBα-Luc promoter reporter gene construct has been described elsewhere [26]. hESC differentiation was induced with differentiation medium (80% knockout Dulbecco's modified Eagle's medium, 20% foetal calf serum (FCS), 1X nonessential amino acids, 1X L-glutamine/penicillin-streptomycin) and removing ESC from feeder layers. At various differentiation time points, differentiated hESC and their undifferentiated counterparts were dissociated to single cells. Equal cell numbers from all samples were transfected with 1 μg of reporter plasmid DNA and 10 ng of control Renilla plasmid pRL-TK (Promega, Southampton, UK) using the Amaxa Cell Line Nucleofector Kit L according to their instructions (program A-023). After 24 hours, cells were lysed using the lysis buffer provided in the dual luciferase detection kit (Promega) and following the manufacturer's instructions. The firefly and Renilla luciferase activities were measured in turn using the LARII and Stop Glow solution and the ratio between the two was calculated.
Alkaline Phosphatase Staining
This method is described in Supporting Information Annex A.
Flow Cytometry Analysis of hESC
This method is described in Supporting Information Annex A.
Measurement of Cell Proliferation Using 5-Ethynyl-2′-Deoxyuridine (EdU) Incorporation Method
This method is described in Supporting Information Annex A.
Inhibitor Studies
IKK2 inhibitor VI (Calbiochem, Nottingham, UK; 401483) at concentration of 20 μM, domain (NEMO-binding domain [NBD]) binding peptide wild-type (Calbiochem 480025, concentration of 10 μM) or mutated negative control (Calbiochem 480030, concentration of 10 μM) were used throughout this study. The hESCs were plated in Matrigel in the presence of MEF-conditioned medium as described in [25]. Inhibitors were added the following day and medium was changed every day for the following 6 days of culture.
Chromatin Immunoprecipitation Assays
First bioinformatics analysis using Genomatix software was performed on each gene to identify potential NFkB family binding sites. DNA oligonucleotides were designed with the aim of amplifying genomic areas with potential binding sites (PBS) and as negative control areas without binding sites (NPBS) were used (Supporting Information Table 3). In brief, cells were harvested at 70%–80% confluence and chromatin immunoprecipitation (ChIP) was carried out essentially as in Dahl and Collas [27]. Sonication was optimized to give chromatin fragments of 100–500 bp in length and DNA from each immunoprecipitation was purified using the Qiaquick DNA Purification kit (Qiagen, West Sussex, U.K.) prior to quantitative polymerase chain reaction (qPCR) analysis. Also included in the experiment was an IgG antibody control immunoprecipitate to detect any background which, if present, was subtracted from each immunoprecipitate within that experiment.
Statistical Analysis
Two tailed pair wise Student's t test was used to analyze results obtained from two samples with one time point. The results were considered significant if p < .05.
Results
Canonical NFκB Pathway Is Stimulated During hESC Differentiation
To monitor canonical NFκB activity in hESC and during the differentiation process, we used an IκBα-Luc promoter reporter, which is induced by the p65 transactivating subunit of NF-kappa B [28]. IκBα-Luc [29] was cotransfected with Renilla luciferase into undifferentiated and differentiated hESC at day 10, 20, and 30 of EB-induced differentiation. Compared with undifferentiated hESCs, differentiated cells generated elevated IκBα-Luc activity (Supporting Information Fig. 1), indicating that canonical NFκB activity is stimulated during hESC differentiation. This result is in accordance with a recent report with another hESC line, SNUhES3 [21].
Differential Expression of NFκB Pathway Components During hESC Differentiation
To investigate the expression of components of the NFκB system in hESC differentiation, we used the EB-differentiation method and collected cell samples at 10-day intervals till day 30. Our previous work has shown that during this differentiation time period, hESC lose the expression of key pluripotency markers (such as OCT4, NANOG, SOX2, TERT) and acquire characteristics of differentiated cells derived from the three germ layers marked by the increased expression of AFP, PAX6, and KDR [22, 30]. Quantitative reverse transcription polymerase chain reaction (qRT-PCR) analysis indicated a significant and continuous increase in p50/p105 and p65 expression during hESC differentiation, although transcripts encoding these proteins were present in hESC (Fig. 1). Increased expression of p50/p105 and p65 was further confirmed by Western blotting (Fig. 2). No changes in phosphorylation of p65 (Ser536) were observed, although a slight decrease of this phosphorylated form was noticed by flow cytometry during differentiation of hESC [22], which may reflect a difference in the sensitivity of these two techniques. Of interest is a sudden increase in phosphorylation of p65 (Ser 529) at day 20 of differentiation followed by a decrease at day 30 (Fig. 2). The observed upregulation of p65 and p50/p105 expression is consistent with the enhancement of canonical NFκB activity during hESC differentiation observed with the IκBα-Luc promoter reporter assay reported in Supporting Information Figure 1.
Expression analysis of various components of NFκB pathway during human ESC (hESC) differentiation by quantitative real time reverse transcription polymerase chain reaction. The value for the hESC was set to 1 and all other values were calculated with respect to that. Data are presented as mean ± SEM (n = 4). Statistical analysis was performed using Student's t test where each differentiation time point was compared with hESC. Abbreviations: c-REL, v-rel reticuloendotheliosis viral oncogene homolog (avian); GAPDH, glyceraldehyde-3-phosphate dehydrogenase; IKK, IkappaB kinase; RELB, v-rel reticuloendotheliosis viral oncogene homolog B; NAK, NF-kappa-B-activating kinase; NIK, NFκB inducing kinase.
Expression analysis of various components of NFκB pathway during human ESC differentiation by Western blotting. The images are representative of at least three independent experiments. Abbreviations: c-REL, v-rel reticuloendotheliosis viral oncogene homolog (avian); GAPDH, glyceraldehyde-3-phosphate dehydrogenase; IKK, IkappaB kinase; RELB, v-rel reticuloendotheliosis viral oncogene homolog B; NIK, NFκB inducing kinase.
qRT-PCR analysis also indicated a significant decrease in expression of inhibitors IκBβ and IκBε as well as a time course specific increase in expression of activators, IKK2 (day 30), NEMO (day 10), and BCL3 (day 20 and 30), which may contribute to NFκB activity enhancement (Fig. 1). In addition, IκBζ expression was upregulated. IκBζ preferentially associates with the NFκB subunit p50 and suppress NFκB activity [31], so its upregulation may serve as negative feedback of NFκB signaling. The expression of IκBα did not change significantly during hESC differentiation and a similar result was obtained for IKK1 (Fig. 1).
In contrast to the increased expression of canonical pathway subunits, the expression of RELB, NIK, and p52 decreased significantly during hESC differentiation at both the mRNA and protein level (Figs. 1, 2). No significant changes were observed in expression of IKK1 transcript (Fig. 1); however, a decrease in IKK1 protein expression was suggested by Western blot (Fig. 2). These findings suggest that noncanonical NFκB is suppressed during hESC differentiation. There are potentially two mechanisms underlying this process and these include direct downregulation of expression of p100/p52, NIK, IKK1, and RELB and inhibition of p100 proteasomal processing through downregulation of NIK and subsequent decrease of IKK1 activity [32].
Inhibition of Canonical NFκB Signaling Results in Loss of hESC Viability and Modulation of Differentiation Potential
We next investigated the function of canonical NFκB signaling in hESC initially employing a selective IKK inhibitor (Calbiochem IKK inhibitor VI). Inhibition of IKK led to loss of tight and compact hESC colony morphology compared with hESC cultures treated with vehicle control, dimethyl sulfoxide (DMSO) (Fig. 3A). The IKK inhibitor-treated hESC cultures developed gaps indicative of cell death. Flow cytometry analysis confirmed this idea, with increases in the percentage of early (Annexin V+/ 7-AAD−) and late (Annexin V+/7-AAD+) apoptotic cells compared with the control group (Fig. 3B). In addition, suppression of IKK caused a decrease in hESC proliferation indicated by reduced numbers of cells in S-phase (Fig. 3C). This was further confirmed by propidium iodide flow cytometry, which showed a higher percentage of cells in G2/M phase of the cell cycle (data not shown) compared with DMSO control group. Increasing evidence suggests that NFκB plays a pivotal role in regulating apoptotic responses by transactivating the expression of antiapoptotic genes in a variety of cell types including primary B cells [33–38], hepatocytes [39, 40], and primary rat cortical neurons [41, 42]. Our results in the hESC system corroborate these findings and suggest that inhibition of canonical NFκB reduces hESC viability.
Impacts of inhibition of canonical NFκB signaling on human ESC (hESC) viability, proliferation, and gene expression. (A): Microphotograph of hESC treated for 48 hours with vehicle alone (DMSO, left-hand side panel) and IKK2 inhibitor VI (right-hand side panel). (B): Schematic graph showing an increase in the number of early apoptotic cells (Annexin V+/ 7 AAD−) and apoptotic cells (Annexin V+/7-AAD+) as result of application of IKK inhibitor. (C): Schematic graph showing a decrease in number of proliferating hESCs as result of application of IKK inhibitor. (D): Changes in expression of pluripotent markers estimated by quantitative real time reverse transcription polymerase chain reaction (RT-PCR) 48 hours after treatment with IKK inhibitor. The value for the control cells (treated with DMSO only) was set to 1 and all other values were calculated with respect to that. (E): Changes in expression of lineage markers estimated by quantitative real time RT-PCR 48 hours after treatment with IKK inhibitor. The value for the control cells (treated with DMSO only) was set to 1 and all other values were calculated with respect to that. (F): Changes in expression of lineage markers in embryoid bodies (EBs; day 7) made from hESC treated with IKK inhibitor or DMSO prior to differentiation process. Note that neither DMSO nor IKK inhibitor are present in the culture medium during the whole 7 days of differentiation. The value for the control cells (treated with DMSO only) was set to 1 and all other values were calculated with respect to that. (G): Changes in the expression of lineage markers in EBs (day 7) made from untreated hESC. Note that DMSO and IKK inhibitor are added for the whole duration of differentiation to the culture medium. The value for the control cells (treated with DMSO only) was set to 1 and all other values were calculated with respect to that. (B–G): Data is presented as mean ± SEM (n = 3). Statistical analysis was performed using Student's t test. Abbreviations: 7ADD, 7 amino-actinomycin D; DMSO, dimethyl sulfoxide; EdU, 5-ethynyl-2′-deoxyuridine; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; IKK, IkappaB kinase.
qRT-PCR analysis indicated higher NANOG and OCT4 expression in hESC treated with IKK inhibitor (Fig. 3D), this is consistent with a recent report in murine ESC, where inhibition of NFκB lead to a detectable increase in Nanog and Oct4 expression [20]. Alkaline phosphatase staining assays showed no effect of IKK inhibition on numbers of positively stained colonies and this was corroborated by the flow cytometry analysis showing no changes in the percentage of cells staining with cell surface marker SSEA4 (data not shown). Notwithstanding this, the mean intensity of staining for SSEA4 was significantly higher in cells treated with the IKK inhibitor (p = 5.49 E −05; n = 3; IKK inhibitor VI mean staining 59,831 ± 1,698; DMSO-treated group: 41,562 ± 417). qRT-PCR revealed downregulation of expression of trophoectodermal marker (CDX2), primitive endodermal marker (GATA6), mesodermal marker (BRACHYURY), and endodermal marker (IHH) in hESC treated with IKK inhibitor (Fig. 3E). We also observed a marked increase in the expression of FGF5 and PAX6 on inhibition of IKK (Fig. 3E); this may be suggestive of acquisition of primitive ectoderm or ectodermal features. We used Western blotting technique to confirm some of these findings at the protein level. Although the increase in PAX6 expression was confirmed, a significant decrease in protein level for both OCT4 and NANOG was observed (Supporting Information Fig. 2) suggesting a more complex regulation at post-transcriptional level for OCT4 and NANOG on inhibition of canonical NFκB signaling, which merits further investigation. A decrease in these markers without noticeable changes in expression of SSEA4 and alkaline phosphatase staining suggest the emergence of different cell subpopulation with propensity to differentiate into ectodermal lineages. Similar changes in marker expression to the ones we have identified have been reported during the 72 hours of neuralization of hESC [43] and or their commitment to an epidermal pathway [44] and suggest that inhibition of canonical NFκB signaling could be clinically relevant for directing differentiation of hESC to ectodermal derivative lineages.
To confirm our results obtained with the IKK inhibitor and build a stronger case for function of canonical NFκB as a regulator of hESC differentiation, we treated hESC with the canonical NBD inhibitor (Calbiochem 480025), which blocks IKKβ activation of NF-κB. We again observed loss of compact hESC colony formation and development of gaps between cells (Supporting Information Fig. 3A) consistent with increased apoptosis, which was confirmed by flow cytometry analysis (Supporting Information Fig. 3B). These effects were not observed with a control mutant NBD peptide. RT-PCR analysis indicated an increase in NANOG, OCT4, FGF5, and PAX6 expression in NBD-treated hESC, again corroborating data obtained with the Calbiochem IKK inhibitor (Supporting Information Fig. 3C). To further verify these data, we specifically perturbed canonical NFκB by siRNA-mediated knockdown of p65 in hESC (Supporting Information Fig. 3D). Depletion of p65 was associated with increased cell death (25% of dead cells in p65 siRNA group compared with 12.5% in control siRNA group, n = 3) and with a significant increase in FGF5 and PAX6 expression (Supporting Information Fig. 3D).
To observe whether the effects of suppression of canonical NFκB signaling are reversible, we treated hESC with IKK inhibitor for 48 hours and on its removal we induced differentiation using the EB method for 7 days. RT-PCR analysis suggested a suppression of differentiation toward primitive trophoectoderm, primitive endoderm, endoderm, and mesoderm as shown by diminished induction of CDX2, GATA6, IHH, and BRACHYURY, respectively (Fig. 3F). In contrast, FGF5 and PAX6 were upregulated in these EBs compared with control group, suggesting a preferential differentiation toward ectodermal lineages on inhibition of canonical NFκB signaling for a limited time at hESC stage (Fig. 3F).
To investigate the role of canonical NFκB signaling during the early stages of hESC differentiation, we made EBs and subjected them for 7 days to IKK inhibitor or DMSO. Although the two groups of EBs appeared morphologically similar (data not shown) qRT-PCR analysis suggested a lesser differentiation toward primitive trophoectoderm, primitive endoderm, endoderm, and mesoderm (Fig. 3G). Again a preferential differentiation to primitive ectoderm and neuroectoderm was suggested by a higher upregulation of FGF5 and PAX6 markers, respectively (Fig. 3G).
In summary, canonical NFκB signaling appears to be required in the maintenance of viability of hESC and may play an important role in hESC differentiation. These results could be explained by a direct effect of p65/p50 binding on the promoters of the affected genes or by a more general effect of canonical NFκB signaling on hESC viability, which results in lower cell numbers and loss of hESC niche characteristics. To investigate the first hypothesis, we performed ChIP assays using a p65 ChIP grade antibody. Several positive controls such as c-MYC, IKKα, and IKKβ, shown to be bona fide transcriptional targets of p65 were included in addition to a negative control gene, TIMP1 (Fig. 4A). Statistical analysis of the results using Student's t test comparison methods between each gene tested and the negative control, TIMP1 gene indicated significant binding to the promoters of three positive control genes as expected. Significant binding was also noticed at the p52/p100 locus (Fig. 4A).
ChIP assays using p65 antibody. (A): Bar chart showing enrichment of IKKα, IKK, p52/p100, and c-MYC promoter fragments after chromatin immunoprecipitation with p65 antibody in human ESC (hESC). The data represent the mean ± SEM from four experiments carried out in H9 cell line. Student t test was carried out to assess significant binding above the level observed for the negative control TIMP1 gene. (B): Bar chart showing enrichment of CYCLIN D1, FGF5, GATA6, CDX2, and PAX6 promoter fragments after chromatin immunoprecipitation with p65 antibody in hESC. The data represent the mean ± SEM from four experiments carried out in H9 cell line. Student t test was carried out to assess significant binding above the level observed for the region predicted not to contain any p65 binding sites. Abbreviation: ChIP, chromatin immunoprecipitation.
To analyze binding of p65 to key pluripotency and lineage markers, we first performed bioinformatics analysis to identify genomic areas with predicted p65 binding sites, which we named PBS and also control genomic areas within the same gene but without predictive binding sites, which we named NPBS. This analysis revealed potential p65 binding sites within NANOG, SOX2, GATA6, BRACHYURY, CDX2, PAX6, and CYCLIN D1 (CCND1); however, we were unable to find potential binding sites within the 5′regulatory region of OCT4. ChIP was performed in hESC and Student's t test comparison method was performed between areas defined as PBS and NPBS (Fig. 4B). Significant binding of p65 was noticed in the CYCLIN D1 promoter, which may be the causative factor behind reduced cell proliferation on inhibition of canonical NFκB signaling. In addition, significant binding was noticed in the regulatory regions of FGF5, GATA6, CDX2, and PAX6, which fits well with data presented in Figure 3 and Supporting Information Figure 3. Together, these finding would suggest a role for p65 in maintenance of pluripotency and lineage differentiation by direct suppression at transcriptional level for FGF5 and PAX6, direct activation at transcriptional level for CDX2 and GATA6 as well as a role in cell proliferation by direct binding to CYCLIN D1 promoter.
Downregulation of RELB Impacts Proliferation of hESC and Gene Expression of Lineage Markers
During hESC differentiation, expression of RELB and p52 was significantly and gradually suppressed (Fig. 1), suggesting that the noncanonical NFκB signaling pathway may play a critical role in hESC renewal and pluripotency. To investigate this, we used RNA interference, which resulted in fivefold reduction in expression of RELB resulting in remaining of 20% of expression compared with control siRNA sample (Fig. 5A, 5B). This was also confirmed by Western blotting (data not shown). qRT-PCR analysis indicated that RELB knockdown did not result in changes in expression of two pluripotentency markers (NANOG and OCT4, Fig. 5B) but caused a significant downregulation of SOX2. Notwithstanding, downregulation of RELB caused an increase in expression of mesodermal marker (BRACHYURY), trophoectodermal marker (CDX2), primitive endodermal marker (GATA6), and a reduction in expression of primitive ectoderm marker (FGF5) and ectodermal marker (PAX6). Upregulation of GATA6 was also confirmed by immunocytochemistry (Supporting Information Fig. 4). Compared with the control group, there were no significant changes in percentages of apoptotic cells (Fig. 5C), percentage of cells staining with the cell surface marker SSEA4, or the percentage of pluripotent colonies detected with alkaline phosphatase staining on knockdown of RELB (data not shown). A small but significant reduction in numbers of proliferating hESC was observed on knockdown of RELB (Fig. 5D) using EdU incorporation combined with flow cytometry analysis. These results were reproduced by propidium iodide staining and flow cytometry analysis (data not shown). Together these data suggest that RELB is likely to have an impact on hESC proliferation and activation and repression of key lineage markers.
Impacts of RELB inhibition on human ESC (hESC) viability, proliferation, and gene expression. (A): Microphotograph of hESC treated for control siRNA (left-hand side panel) and RELB siRNA (right-hand side panel) at 48 hours post-transfection. (B): Changes in expression of lineage markers estimated by quantitative real time reverse transcription polymerase chain reaction 48 hours post-transfection of RELB siRNA in hESC. The value for the control siRNA-treated cells was set to 1 and all other values were calculated with respect to that. (C): Schematic graph showing no change in the number of apoptotic cells (Annexin V+/7-AAD+) as result of RELB knockdown by RNA interference at 48 hours post-transfection. (D): Schematic graph showing a decrease in the number of proliferating hESCs as result of RELB knockdown by RNA interference at 48 hours post-transfection. (B-D): Data is presented as mean ± SEM (n = 3). Statistical analysis was performed using Student's t test. (E): ChIP assays using RELB antibody. Bar chart showing enrichment of IKKα, IKKβ, and p52/p100 promoter fragments after ChIP with RELB antibody in hESC. The data represent the mean ± SEM from four experiments carried out in H9 cell line. Student t test was carried out to assess significant binding above the level observed for the negative control TIMP1 gene. (F): ChIP assays using RELB antibody. Bar chart showing enrichment of SOX2, CDX2, BRACHYURY, and PAX6 promoter fragments after ChIP with RELB antibody in hESC. The data represent the mean ± SEM from four experiments carried out in H9 cell line. Student t test was carried out to assess significant binding above the level observed for the region predicted not to contain any RELB binding sites. Abbreviations: 7ADD, 7 amino-actinomycin D; ChIP, chromatin immunoprecipitation; EdU, 5-ethynyl-2′-deoxyuridine; NPBS, negative control areas without binding sites; PBS, potential binding sites; RELB, v-rel reticuloendotheliosis viral oncogene homolog B.
To investigate whether these effects are mediated by direct binding of RELB to pluripotency or differentiation markers, we performed ChIP assays. Several positive controls including IKKα, IKKβ, and p52/p100 shown to be bona fide transcriptional targets were included in addition to negative control gene, TIMP1 (Fig. 5E). Statistical analysis of the results using the Student's t test comparison methods between each gene tested and the negative control, indicated significant binding to the promoters of the three positive control genes as expected.
Significant binding of RELB was noticed in the regulatory regions of SOX2, BRACHYURY, PAX6, and CDX2 that fits well with data presented in Figure 5B. Together, these finding would suggest a role for RELB in maintenance of pluripotency and lineage differentiation by direct suppression at transcriptional level for BRACHYURY and CDX2 and direct activation at transcriptional level for SOX2 and PAX6.
Discussion
Because of their feature of unlimited self-renewal and capacity to differentiate into cells of all the three germ layers, hESC have been proposed for regenerative medicine and tissue replacement after injury or disease. In view of this, the identification and study of signaling pathways required for hESC proliferation and differentiation is important for understanding early human embryonic developmental biology and for clinical cell therapy. The NFκB signaling pathway plays an important role in inflammatory and immune responses, apoptosis, transformation, cell adhesion, oxidative stress responses, embryo development, hematopoiesis, as well as neuronal functions via the induction of certain growth and transcription factors [45–47]. But the impacts of NFκB signaling pathway on hESC have not been fully examined and have formed the topic of this article.
In this study, H1 and H9 hESCs were used as a model system to study the activity of canonical and noncanonical NFκB signaling. We found that canonical NFκB activity was stimulated and expression of p65 and p50 were enhanced during hESC differentiation. Our studies showed that p65 was able to bind at the regulatory regions of several genes that mark the development of primitive ectoderm (FGF5), ectoderm (PAX6), trophoectoderm, (CDX2) and primitive endoderm (GATA6). It is of interest to note that binding of p50 and p65 to the CDX2 promoter has also been observed in adenocarcinomas and esophageal cancer cells [48] and FGF5 has also shown to be induced in human fibroblasts as result of activation of canonical NFκB signaling [49]. Although no published data exist for direct interaction between canonical NFκB signaling and GATA6 or PAX6, our results showing transcriptional changes for all four genes in response to inhibition of this signaling branch coupled with ChIP assays argue for a direct binding of p65/p50 complex at the regulatory regions of these genes. This would, however, need to be further investigated by ChIP assays with coactivators and corepressors of the pathway and is currently ongoing in our group. Inhibition of the canonical NFκB signaling resulted in a significant increase in transcription of NANOG and OCT4 expression as well as decreases at the protein level, although our ChIP assays did not show direct binding of p65 to the regulatory regions of these genes. We are more inclined to believe that changes in the expression of these two pluripotency genes reflect suppression of cell differentiation to several primitive and definitive lineages (such as differentiation to primitive trophoectoderm and endodermal lineages) and appearance of a more homogenous and ectodermally poised cell population in culture. Previously published data in the murine ESC system has shown that Nanog binds to NFκB proteins, and this binding inhibits NFκB transcriptional activity, resulting in maintenance of murine ESC pluripotency [20]. It seems that in hESC the mechanisms by which NFκB interacts with pluripotency markers such as OCT4 and NANOG are more complex and need to be explored further.
Despite the molecular mechanisms that may be involved in regulation of gene transcription by canonical NFκB signaling, our findings have direct implications for directing differentiation to particular lineages. Of outermost relevance is enhancement of ectodermal differentiation on inhibition of canonical NFκB signaling that can be used in defining conditions for neuronal, retinal, as well as epidermal differentiation studies from pluripotent stem cells. This also has implications for genetic diseases and cancer-related studies that have revealed an association between these diseases and NFκB signaling but have failed to identify NFκB-regulated genes that may underline the disease phenotype. For example, mutations at the EDARADD locus completely abrogate NFκB activation and result in ectodermal dysplasias [49, 50]. Equally, constitutive activation of NFκB signaling is often found in lung, liver, pancreatic, and esophageal cancers [4], and given the new interactions, we have revealed here between NFKB and GATA6 trancription factors often known to be involved in gut and cardiovascular development, new insights can be gathered toward new cancer treatments. For example, some Gata factors such as GATA4 and GATA5 [51] have shown to be epigenetically silent in gastric cancers, thus changing the balance between cell differentiation and proliferation. In view of this, it would be interesting to examine the expression of GATA6 in the stem cell compartment of endodermal type cancers that show activation of canonical NFκB signaling.
Another aspect of our studies is the increased expression of p105/p50 together with dramatic upregulation of Bcl3 that is known to act both as coactivator or corepressor through association with p50 homodimers. Furthermore, Bcl3 expression can be induced by NFκB and this which forms a part of the autoregulatory loop that controls the nuclear residence of p50 NFκB. Future studies should therefore be directed at determining the role of Bcl3 as a regulator of NFκB activities in differentiating hESC.
Our observation for a dramatic decrease in expression of the noncanonical factors RELB, p52, and NIK during hESC differentiation raises the possibility that this pathway way may be a key determinant for the switch between hESC proliferation and differentiation. It is indeed noteworthy that BAFF, a known activator of noncanonical NFκB is included in the media used to maintain pluripotent hESC cultures. We were able to demonstrate binding of RelB to lineage-specific marker genes and knockdown of RelB was associated with increased expression of these differentiation markers supporting our proposal that noncanonical NFκB may function as a suppressor of hESC differentiation. Of note, the p52 component of noncanonical NFκB has previously been shown to either stimulate or repress cell proliferation depending on the cell type studied [7], it will therefore be instructive to determine those cell cycle genes that are targets for p52 in pluripotent hESC. An interesting observation concerning the diminished expression of RelB and p52 with hESC differentiation was the more dramatic loss at protein versus transcript level. This may be indicative of post-transcriptional control mechanisms either at the level of increased rates of protein degradation or possibly translational control via a micro RNA-related mechanism. Given our limited understanding of the control of NFκB subunit expression, the hESC differentiation process provides an interesting model system to study mechanisms of regulation of subunit expression.
Conclusion
In summary, we conclude that canonical and noncanonical NFκB signaling is multifunctional in the control of hESC fate, and moreover, that hESC differentiation is accompanied by a switch from noncanonical to canonical NFκB signaling that may be critical for survival and differentiation in the postpluripotent state. Further investigations into the specific functions of the NFκB subunits and their coregulators should improve our understanding of the transcriptional regulation of hESC pluripotency, differentiation, and survival.
Acknowledgements
We are grateful to Ian Dimmick and Dr. Rebecca Stewart for help with the flow cytometry analysis, Dennis Kirk for technical assistance and Mann's lab for providing most of the reagents needed for this work. This study was supported by One North East Regional Developmental Agency, and funds for research in the field of Regenerative Medicine through the collaboration agreement from the Conselleria de Sanidad (Generalitat Valenciana) and the Instituto de Salud Carlos III (Ministry of Science and Innovation).
Disclosure of Potential Conflicts of Interest
The authors indicate no potential conflicts of interest.
References
Author notes
Author contributions: C.Y.: collection and assembly of data, data analysis and interpretation, manuscript writing and final approval of manuscript; S.P.A.: collection and assembly of data, data analysis and interpretation, manuscript writing and final approval of manuscript; F.V.: collection and assembly of data and final approval of manuscript; M. Lloret: collection and assembly of data and final approval of manuscript; L.A.: conception and design, financial support, data analysis and interpretation, manuscript writing and final approval of manuscript; D.A.M.: conception and design, financial support, data analysis and interpretation, manuscript writing and final approval of manuscript; M. Lako: conception and design, collection and assembly of data, data analysis and interpretation, manuscript writing and final approval of manuscript. C.Y. and S.P.A. contributed equally this article.
First published online in STEM CELLS EXPRESS September 29, 2010.
Disclosure of potential conflicts of interest is found at the end of this article.
Telephone: 44-191-241-8688; Fax: 44-191-241-8666
