Functional Analysis of Tcl1 Using Tcl1-Deficient Mouse Embryonic Stem Cells

Tcl1 is highly expressed in embryonic stem (ES) cells, but its expression rapidly decreases following differentiation. To assess Tcl1’s roles in ES cells, we generated Tcl1-deficient and -overexpressing mouse ES cell lines. We found that Tcl1 was neither essential nor sufficient for maintaining the undifferentiated state. Tcl1 is reported to activate Akt and to enhance cell proliferation. We found that Tcl1 expression levels correlated positively with the proliferation rate and negatively with the apoptosis of ES cells, but did not affect Akt phosphorylation. On the other hand, the phosphorylation level of β-catenin decreased in response to Tcl1 overexpression. We measured the β-catenin activity using the TOPflash reporter assay, and found that wild-type ES cells had low activity, which Tcl1 overexpression enhanced 1.8-fold. When the canonical Wnt signaling is activated by β-catenin stabilization, it reportedly helps maintain ES cells in the undifferentiated state. We then performed DNA microarray analyses between the Tcl1-deficient and -expressing ES cells. The results revealed that Tcl1 expression downregulated a distinct group of genes, including Ndp52, whose expression is very high in blastocysts but reduced in the primitive ectoderm. Based on these results, we discuss the possible roles of Tcl1 in ES cells.


Introduction
To elucidate the key molecules involved in the pluripotency of mouse embryonic stem (ES) cells, we compared expressed sequence tag (EST) counts between embryonic stem (ES) cells and somatic tissues using digital differential display (http://www. ncbi.nlm.nih.gov/UniGene/info_ddd.html) [1]. The T-cell lymphoma breakpoint 1 gene, Tcl1, was one of the genes we identified using this method. This gene is expressed at high levels in ES cells. The normal expression of Tcl1 in mice is restricted to early embryogenesis [2], fetal tissues (liver, thymus, bone marrow, and yolk sac) [3], developing lymphocytes [4], and adult testis [5], suggesting that it functions in stem cells and progenitor cells. The human ortholog, TCL1A, is responsible for T-cell tumors caused by chromosomal rearrangements involving 14q32 [6]. Thus, Tcl1 may have a positive role in cell proliferation and/or survival, an idea that is supported by the occurrence of T-cell leukemia in mice carrying a TCL1 transgene under control of the lck promoter [7]. On the other hand, an analysis of Tcl1-null mutant mice indicated that Tcl1 is important for the development of preimplantation embryos; a lack of maternally derived Tcl1 impairs the embryo's ability to undergo normal cleavage and develop to the morula stage, especially in vitro [2].
Glover et al. [8] identified genes whose expression changes when ES cells are induced to differentiate. Tcl1 is one of seven genes that showed a rapid decrease in expression concurrent with a decrease in the frequency of undifferentiated cells. Genetic manipulations that affect the undifferentiated state of ES cells are often reported to downregulate Tcl1 together with other pluripotency-related genes, such as Dppa3, Klf2, and Zfp42 [9,10]. Matoba et al. [11] identified Tcl1 as a downstream target of Oct3/4 using the ZHBTc4 ES cell line, in which the expression of Oct3/4 (encoded by Pou5f1) can be downregulated by tetracycline [12]. They showed that Oct3/4 binds to the promoter region of the Tcl1 gene to activate its transcription, and, using ES cells in which Tcl1 was knocked down by shRNA, they showed that Tcl1 is involved in regulating proliferation, but not differentiation. However, the effect of complete loss of the Tcl1 gene on the state of ES cells has not been reported. In the present study, we generated Tcl1deficient and -overexpressing ES cell lines and compared the undifferentiated phenotypes and gene expression patterns between them.

Generation of -deficient ES Cells
We first examined the Tcl1 expression during ES cell differentiation into trophectoderm using the ZHBTc4 ES cell line, in which the expression of Oct3/4 can be downregulated by tetracycline [10]. As shown in Figure S1, Tcl1 expression decreased with similar kinetics as Fgf4, a well-known target of Oct3/4 [13], consistent with the report by Matoba et al. [11] that Tcl1 is a downstream target of Oct3/4.
We also examined the effects of Tcl1 deficiency on apoptosis by immunostaining cleaved caspase 3 ( Figure 3B). The Tcl1 2/2 ES clones (#4 and #5) showed 1.8-and 1.5-fold increase in percent cleaved caspase 3-positive cells, respectively, compared with wildtype ES cells. Thus, Tcl1 overexpression in these Tcl1 2/2 ES cells clearly reduced the frequency of apoptotic cells, which was even lower than that in wild-type ES cells.
We next tested the in vivo growth capacity of these cells by performing teratoma formation assays ( Figure 3C). The Tcl1 2/2 ES cells formed smaller tumors than did wild-type ES cells, consistent with their slower growth rate. To our surprise, however, the Tcl1 2/2 (CAG-Tcl1) ES cells produced only barely recognizable tumors. Considering that the Tcl1 2/2 (CAG-Tcl1) ES cells showed a similar or even better proliferation rate in vitro than did wild-type ES cells, this result might have been owing to an effect of Tcl1 overexpression on the differentiation capacity or status of the ES cells. Thus, we examined the teratomas derived from the Tcl1 2/2 and Tcl1 2/2 (CAG-Tcl1) ES cells histologically. Within individual tumors, neural rosettes (ectoderm), epidermis (ectoderm), cartilage (mesoderm), adipose tissue (mesoderm), and gutlike epithelium (endoderm) were found, indicative of the differentiation into cells fated for each of the three germ layers ( Figure 3D).

Effect of Tcl1 on Wnt-b-catenin Signaling in ES Cells
To gain further insight into Tcl1's function, we looked for signaling pathways that might be affected by it. Akt is known to phosphorylate GSK3b in insulin signaling [16], and GSK3b also serves as a component of the canonical Wnt pathway involving bcatenin. In addition, other reports have indicated a functional link between Akt and b-catenin [18,19]. Since the Wnt pathway has been implicated in maintaining the undifferentiated state of ES cells, we explored whether Tcl1 acts as a bridge between the Akt signaling and Wnt pathways. In canonical Wnt signaling, briefly, when extracellular Wnt is absent, b-catenin is phosphorylated by casein kinase I and GSK3b. Phosphorylated b-catenin is recognized by E3 ligase and targeted for degradation. Upon Wnt-signaling activation, the phosphorylation of b-catenin is inhibited, and the accumulated b-catenin translocates to the nucleus, where it drives the expression of target genes through an association with Tcf/Lef [20].
To examine the effect of Tcl1 on Wnt-b-catenin signaling, we performed western blot analyses to determine the phosphorylation levels of Akt, GSK3, and b-catenin ( Figure 4A). Interestingly, the phosphorylation of b-catenin was dramatically reduced by Tcl1 overexpression, but the total amount of b-catenin appeared unchanged, indicating there was a reverse correlation between the Tcl1 expression and b-catenin phosphorylation level. At the same time, the Akt and GSK3 phosphorylation levels did not correlate with the b-catenin phosphorylation level, implying that Tcl1 was closely, perhaps directly, involved in regulating b-catenin. We next examined whether Tcl1 overexpression led to an increase in nonphosphorylated b-catenin in the nuclear fraction, and found that the nonphosphorylated active b-catenin levels were enhanced in the nuclei ( Figure 4B).
To measure the canonical Wnt signaling directly, we used a Tcf/b-catenin reporter system, the TOPflash assay [21]. The reporter activity was not affected by the Tcl1 knockout, but Tcl1 overexpression enhanced it approximately 1.8-fold compared with wild-type ES cells ( Figure 4C). These results suggested that bcatenin signaling is normally repressed in ES cells and is enhanced by Tcl1 overexpression. We further examined the expression levels of well-known canonical Wnt target genes, such as c-myc, Axin2, Lef1, and Dll1 (http://www.stanford.edu/rnusse/pathways/ targets.html). Since Tcl1 positively regulates Wnt signaling, the expression levels of these genes might be lower in Tcl1 2/2 cells and higher in Tcl1 2/2 (CAG-Tcl1) cells than in wild-type ES cells. However, our quantitative RT-PCR results showed that the Tcl1 expression level did not significantly affect these Wnt target genes (data not shown). The only exception was Gbx2, a recently identified Wnt-b-catenin signaling target [22], which showed elevated expression in the Tcl1-overexpressing ES cells (see below).

Effects of Tcl1 Deficiency on the gene Expression Pattern of ES Cells
To our knowledge, the Wnt targets in ES cells have not been systematically investigated, and there are reports of Tcf/Lefindependent targets in some other cell types. To find genes that respond to Tcl1 overexpression in ES cells possibly through Wnt/ b-catenin signaling, we compared the gene expression profiles by dual-channel DNA microarray analysis between Tcl1 2/2 #4 and Tcl1 2/2 (CAG-Tcl1) #1, between Tcl1 2/2 (CAG-EGFP) #6 and Tcl1 2/2 (CAG-Tcl1) #4, between Tcl1 2/2 #4 and wild-type ES cells, and between Tcl1 2/2 #5 and wild-type ES cells. The genes whose expression levels were consistently affected by Tcl1 expression by more than 1.7 fold were listed in Table 1. We found 16 genes (Pem, Ndp52, Tmem64, Dppa3, Tcstv1, Fbxo15, Ephx2, Mlana, Zfp42, Jam2, Morc1, Tcfcp2l1, Psx1, Psx2, Myl7, and Plac8) that were consistently downregulated by Tcl1 expression. On the other hand, only two genes, Gbx2 and Fndc4, showed elevated expression in Tcl1-expressing cells compared with Tcl1-deficient cells and the elevation of their expression by Tcl1 was 2.7 and 2.2 fold, respectively. Quantitative RT-PCR analysis was performed for these affected genes, using Tcl1 2/2 #5, Tcl1 2/2 (CAG-Tcl1) #11, Tcl1 2/2 (CAG-Tcl1) #14, and wild-type ES cells. As shown in Figure 5A, the expression of Gbx2 and Fndc5 was significantly upregulated by Tcl1 overexpression, although the  expression levels of the Gbx2 gene were not considerably different between Tcl1 2/2 #5 and wild-type ES cells. All of the 16 genes were shown to be downregulated by Tcl1 expression in agreement with the DNA microarray data ( Figure 5B; see Figure 2 for Zfp42).

Discussion
In the current study, we sought to elucidate the molecular pathways in which Tcl1 is involved, and the physiological role of Tcl1 in ES cell management. One well-documented function of Tcl1 is to bind Akt and increase its kinase activity [23]. Important roles of Akt and/or its upstream signal molecule, PI3K, in the self-renewal of ES cells have been reported [24][25][26][27]. Matoba et al. [11] showed that Tcl1downregulation leads to a reduction in Akt phosphorylation in ES cells. Ema et al. [28] showed that Krüppel-like factor 5 (Klf5) is essential for the normal self-renewal of mouse ES cells using Klf5knockout ES cells, and that Tcl1 is downregulated in the Klf5knockout ES cells. They also showed that the Akt phosphorylation is reduced in these ES cells. These reports support the idea that Tcl1 regulates ES-cell proliferation via Akt phosphorylation. In fact, we observed that Tcl1 overexpression in ES cells clearly incresed the cell proliferation and reduced the frequency of apoptotic cells as shown in Figure 3. However, these effects of Tcl1 could not be accounted for by an increase of Akt phosphorylation. Our data showed that Tcl1 expression was not correlated with the global phosphorylation levels of Akt and GSK3b ( Figure 4A). Although we do not know the reason for this discrepancy, it should be noted that high-level Tcl1 expression does not necessarily lead to Akt phosphorylation, because Akt phosphorylation is undetectable in seminomas and CD4 + CD56 + blastic tumors of dendritic cell origin, in which Tcl1 is highly expressed [2,29]. In any case, we believe that our Tcl1-deficient andoverexpressing ES cells are ideal tools for identifying Tcl1 targets in ES cells.
Interestingly, our data showed that Tcl1 expression was correlated with the phosphorylation level of b-catenin ( Figure 4A) in ES cells. Tcl1 may regulate b-catenin through a hitherto unknown pathway related or unrelated to Akt or GSK3b. bcatenin plays a major role in the canonical Wnt pathway. It was reported that Wnt-pathway activation by 6-bromoindirubin-39oxime (BIO), a specific pharmacological inhibitor of GSK-3, helps maintain the undifferentiated phenotype in human and mouse ES cells [30]. However, Wnt's precise role in mouse ES cells has been debated [31][32][33].
Stimulation of the canonical Wnt pathway through the binding of Wnt ligand to its receptor causes a repression of GSK3 activity, which further inhibits b-catenin degradation, resulting in bcatenin's recruitment to the cell membrane to associate with Ecadherin or in its localization to the nucleus. Nuclear b-catenin associates with Tcf1 and activates its transcriptional activator function. It was reported that c-myc, a downstream b-catenin target, is involved in maintaining the undifferentiated state [34]. However, it is not known to what degree such transcriptional activation contributes to the maintenance of pluripotency in ES cells. In addition, b-catenin binds to Tcf3 and inhibits its repressor function. Because Tcf3 interacts with core pluripotency-associated transcription factors, such as Oct3/4, to repress their transcriptional activity, its binding to b-catenin may stabilize pluripotency [31]. It is possible that different threshold levels of these two action modes operate in ES cells, in which a low level of Wnt signaling is sufficient to direct a derepression of gene expression through the inhibition of Tcf3 repressor function, whereas high levels are required for the activation through Tcf1. The overexpression of Tcl1 led to a nearly complete loss of b-catenin phosphorylation, but not to a parallel enhancement of Tcf/Lef reporter activity and activation of target genes. The canonical Wnt activity might be actively repressed or maintained in a narrow range in ES cells. In addition, the nuclear partners of b-catenin in stem cells might be different from those in well-studied somatic cells.
As a negative modulator of b-catenin phosphorylation, Tcl1 may well be involved in anterior-posterior determination and gastrulation development, especially considering that many reports point to the importance of b-catenin at this stage [35][36][37]. The poor teratoma formation elicited by the Tcl1 2/2 (CAG-Tcl1) ES cells may reflect lineage restriction or premature stem cell/progenitor loss, as was shown to occur during the epithelial-mesenchymal transition in mouse embryos expressing dominant stable b-catenin [35]. Further examination of this issue will require a longer incubation of EBs and the use of differentiation protocols aimed at specific lineages, as well as Tcl1's overexpression in mouse embryos.
It is not clear whether these changes in gene expression were related to the reduced activity of the Wnt pathway. It is possible that Tcl1 affects the gene expression in ES cells via an unidentified pathway. Tcl1 expression is reported to be fairly high in embryos during the early cleavage stage and to gradually decline toward the blastula stage [2]. Its expression level during the peri-implantation period is yet to be investigated, but major roles of Tcl1 may be repression of the trophectoderm fate and promotion of the transition from ICM to primitive ectoderm. The fact that the genes involved in placental function and the genes whose expression diminishes during the transition from ICM to primitive ectoderm were downregulated in Tcl1-expressing ES cells may be consistent with this notion. Interestingly, the latter genes included Dppa3, which is considered a defining marker of the mouse ES cell state and is not expressed in epiblast stem cells [38]. Its expression was reported to be heterogeneous in mouse ES cell cultures. It is thought that the Dppa3-negative mouse ES cells are more 'epiblastlike' but have not been stably committed to this transition since they readily revert to Dppa3-positive [39]. Thus, Tcl1 may be involved in the the metastability and plasticity of ES cells.

Ethics Statement
Experiments involving animals were carried out in accordance with institutional guidelines under protocols (No. 21-089-4) approved by the Animal Care and Use Committee of the Osaka University Graduate School of Medicine.
To assemble the targeting vector, the XhoI-NotI fragment containing the MC1 promoter-driven diphtheria toxin A gene was excised from pMC1-DTA-pA [41] and subcloned into pBluescript II in which the SacII site had been modified to a SwaI site. The amplified left arm was digested with SwaI and EcoRV, ligated to a DNA fragment containing a phosphoglycerate kinase (Pgk) promoterdriven puromycin acetyl transferase gene cassette, and subsequently introduced into the SwaI-NotI site of the above vector, which harbored a diphtheria toxin A cassette. Finally, the NotI-digested right arm was cloned into the NotI site of the targeting vector. The targeting vector was linearized by SwaI digestion and introduced into EB3 ES cells by electroporation. Two days later, positive selection was started with 1.5 mg/ml puromycin (Cat#A11138-03; Life Technologies). Clones resistant to puromycin were screened for homologous recombination by long and accurate PCR using primers P1, 59-TCAGCC-CATCTTGGCACATCTGGCAGATT39 and P2, 59-TACTTC-CATTTGTCACGTCCTGCACGACG-39. Two of the resulting Tcl1 +/2 clones were subjected to a high concentration of puromycin to obtain Tcl1 2/2 colonies. The PCR primers used to detect the loss of heterozygosity were H1, 59-AGGAGCCTGATGATGGTGC-39 and H2, 59-GGTCTGGGTTATTCATCGTT-39. Twenty-eight of 32 clones resistant to the high concentration of puromycin (20,30 mg/ml) were found to be Tcl1 2/2 .

Generation of Tcl1 2/2 (CAG-Tcl1) ES Cells
The Tcl1 ORF was amplified by PCR using Pfx polymerase (Cat#11708; Life Technologies) with the following primers: 59-CGGAATTCCATGGCTACCCAGCGGGCACACAG-39 (EcoRI tailed) and 59-CGGAATTCCGGTCTGGGTTATT-CATCGTTGGAC-39 (EcoRI tailed). The product was digested with EcoRI and cloned into the multiple cloning site of pCAG-IZ [42]. The entire expression cassette was excised with SalI and BamHI and inserted between two tandem repeats of loxP in a pBS246 derivative [43] lacking the EcoRI site. Then, Tcl1 2/2 ES cells were transfected with the linearized vector by electroporation. After the transfection, ES cells were selected in the presence of 20 mg/ml Zeocin (Cat#R250-01; Life Technologies) for 7 days.

Teratoma Formation
For the teratoma-formation assay, 1.0610 6 cells in 75 ml PBS were injected subcutaneously into histocompatible F1 adult mice (C57BL/6J6129/Ola). Four weeks later, the mice were sacrificed, and the tumors were weighed. The tumors were fixed in 20% formaldehyde and processed for paraffin embedding. Sections of paraffin-embedded tumors (5-mm thick) were deparaffinized, stained with hematoxylin-eosin, dehydrated, and examined with a microscope.

Primers used for Semi-quantitative RT-PCR
The primers and RT-PCR conditions for Fgf5, T (Brachyury), and Gata4 were described previously [44]. The primers for Gapdh were purchased (Cat#RPP-401; Toyobo, Osaka, Japan; annealing temperature 60uC, 23 cycles). The primers designed for the current study are listed in Table S2.

Real Time PCR Analysis
Real time PCR was performed on an ABI Prism 7300 Sequence Detection System, using the SYBR Green PCR Core Reagents (Cat#4304886; Applied Biosystems, Foster City, CA) or FastStart Universal SYBR Green Master (Cat#4913850; Roche, Mannheim, Germany) (primer sequences are shown in Table S3). PCR was performed with an initial step of 10 sec or 10 min at 95uC followed by 40 cycles of 5 sec at 95uC and 31 sec at 60uC. The expression levels of targeted genes were normalized to that of bactin. Statistical analysis was performed by Student9s t-test.
MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium Bromide] Assays ES cells were seeded in 96-well plates and incubated for 24 or 48 hours, after which MTT (Cat#M5655; Sigma, St. Louis, MO) reagent was added to the wells. After 4 hours, when purple precipitates were visible under a microscope, detergent reagent was added to the wells, and the cells were incubated overnight. A microplate reader (Bio-Rad, Hercules, CA) was used to determine the optical density of each well at 550 nm.

Statistical Analysis
Statistical manipulations were performed using the open-source statistical environment R (http://www.r-project.org). The P values for the teratoma weights were calculated with exact Wilcoxon rank sum tests in the exacRankTests package, then adjusted with Holm's method. The P values for the TOPflash assay were calculated likewise. Quantitative PCR results are presented as the mean 6 SEM. Statistical analyses were carried out by Student's ttest. A value of P,0.05 was considered statistically significant.