**2. ES cell self-renewal genes that are involved in cancer cell growth**

#### **2.1. Oct3/4**

modifications in DNA methylation. Recent studies have provided evidence that tumor tissue contains a small subset of stem-like cells called cancer stem cells. Cancer stem cells have stem cell-like attributes, namely, self-renewal and differentiation, which enables them to produce tumors by self-renewing and giving rise to differentiated progeny. The concept of cancer stem cells first arose from studies of leukemia stem cells [10] and subsequently of solid tumors [11, 12]. Given that cancer stem cells play a prominent role in cancer cell growth, it may be

Indeed, ES cells are similar to cancer cells in several respects. When injected into immunode‐ ficient mice, ES cells and cancer cells can produce benign and malignant tumors, respectively. Both cell types have a rapid cell cycle, which results in fast proliferation. Telomerase activity is very high in both cell types, which allows them to proliferate indefinitely. Both cell types contain a "side population of cells" with a high drug efflux capacity, which gives rise to their drug-resistance phenotype. In addition, several signal transduction pathways seem to be used in both ES cell self-renewal and cancer cell growth. For example, the STAT3 pathway, which plays a central role in ES cell self-renewal, is activated in several types of cancer cells. On the other hand, the Wnt/β-catenin pathway, whose activation is associated with tumorigenesis in many tissues, is involved in ES cell self-renewal. Moreover, poorly differentiated tumors

The similarities between ES cells and cancer cells raise the possibility that certain molecules that are involved in ES cell self-renewal play important roles in cancer cell growth, while certain oncogenes play critical roles in ES cell self-renewal (Fig. 1). In this chapter, I will provide examples of such molecules and describe their roles in ES cell self-renewal and tumorigenesis.

reasonable to expect similarities between ES cells and cancer cells.

42 Pluripotent Stem Cell Biology - Advances in Mechanisms, Methods and Models

preferentially overexpress genes that are normally enriched in ES cells [13].

**Figure 1.** Similarity between ES cells and cancer cells.

Oct3/4 was originally isolated by three groups [14-16]. Since Oct1 and Oct2 had already been identified, two groups named this protein Oct3, while the other group named it Oct4. Therefore, I will use the term "Oct3/4" to describe this protein. Oct3/4 is a transcription factor belonging to the POU family and is encoded by the *pou5f1* gene. The Oct3/4 protein contains three domains, namely, the N-terminal, POU, and C-terminal domains. The Nand C-terminal domains are transactivation domains with redundant functions, while the POU domain is a bipartite DNA-binding domain consisting of the POU-specific domain and the POU homeodomain. In mouse development, deficiency of this transcription factor results in loss of the ICM [17]. In agreement with this, conditional knockout of this gene in ES cells results in their differentiation into trophectoderm cells [18], indicating that Oct3/4 is a central player in the self-renewal of ES cells. Furthermore, the recent finding that Oct3/4 is one of the four factors required for the production of iPS cells indicates the importance of Oct3/4 for the acquisition of pluripotency [19].

Not only suppression, but also overexpression of Oct3/4 induces ES cell differentiation [18], suggesting that the activity of Oct3/4 needs to be sustained at the correct level to main‐ tain ES cell self-renewal. Oct3/4 expression in ES cells is positively and negatively regulat‐ ed by multiple factors. The upstream region of the *pou5f1* gene contains proximal and distal enhancers, which regulate stem cell-specific expression of Oct3/4 [20]. An orphan nuclear receptor Lrh1 (also known as Nr5a2) binds to the proximal enhancer to maintain Oct3/4 expression [21], whereas Oct3/4 and Sox2 associate with the distal enhancer to stimulate Oct3/4 expression [22]. Negative regulators of Oct3/4 expression include Gcnf, Coup-tfs, and Cdx2, whose expression are induced upon ES cell differentiation. In addition to regulation of Oct3/4 expression, the transcriptional activity of Oct3/4 protein is regulated by Oct3/4-binding proteins. For example, it is well-established that Sox2 is a co-factor of Oct3/4. β-catenin binds to Oct3/4 and functions as a co-activator that enhances the transcrip‐ tional activity of Oct3/4 [23]. The orphan nuclear receptor Dax1 also binds to Oct3/4, but acts as a negative regulator by interfering with the binding of Oct3/4 to DNA [24].

In adult human tissues, expression of Oct3/4 is restricted to germline cells and is very low in other tissues. By contrast, Oct3/4 is expressed in several types of human cancers, including prostate, breast, oral, bladder, and seminoma [25], suggesting the importance of this transcription factor in cancer development. Moreover, Oct3/4 is highly expressed in cancer stem-like cells in breast cancer, lung cancer, and bone sarcoma [26-28]. In lung cancer, Oct3/4 maintains the properties of cancer stem-like cells [28]. In agreement with these findings, high expression of Oct3/4 significantly correlates with poor overall survival of nasopharyng‐ eal carcinoma patients [29].

## **2.2. Sox2**

Sox2 is a transcription factor that belongs to the SRY-related HMG-box protein (Sox) family. The Sox2 protein contains one HMG box, as well as a transactivation domain in its C-terminus. Expression of Sox2 in ES cells is mainly regulated by Oct3/4 and Sox2 itself. In addition, microRNAs (miRNAs) are involved in the control of Sox2 expression. For example, miR-9 binds to the 3'-untranslated region of *Sox2* mRNA and represses Sox2 expression [30]. In addition to regulation at the transcriptional level, the activity of Sox2 protein is regulated by post-translational modifications. Mouse Sox2 protein can be sumoylated at Lys-247, which impairs its binding to the *Fgf4* enhancer [31]. Additionally, acetylation of Lys-75 by p300/CBP promotes the nuclear export of Sox2 [32]. Akt directly interacts with Sox2 to phosphorylate Thr-118, leading to stabilization of the Sox2 protein [33].

Normally, Nanog is expressed at early embryonic stages and in germline stem cells, but not in adult tissues. However, Nanog is expressed at high levels in several types of cancers, including breast, cervical, oral and kidney [48]. Nanog is also highly expressed in germ cell tumors [49], which are characterized by the gain of the short arm of chromosome 12, at which the *Nanog* gene is located. In addition, several reports have suggested that Nanog is involved in the epithelial-mesenchymal transition and metastasis. For example, Nanog stimulates cell migration and invasion in ovarian cancer through downregulation of *E-cadherin* and *Foxj1* [50]. Nanog expression is higher in hepatocellular carcinoma cell lines that are highly metastatic than in those that are lowly metastatic, and the Nodal/Smad3 pathway plays an important role in the Nanog-stimulated epithelial-mesenchymal transition [51]. More importantly, Nanog overexpression is associated with poor prognosis in some types of cancer, such as colorectal,

Embryonic Stem Cells and Oncogenes http://dx.doi.org/10.5772/57614 45

Nanog has 11 highly homologous pseudogenes in human cells. Of these, *NanogP8* encodes a full-length protein with only one amino acid difference from Nanog, and is involved in tumorigenesis [52, 53]. It was recently shown that NanogP8 can function as a reprogramming

STAT3 is a downstream transcription factor of gp130, and is therefore activated by several cytokines, including interleukin (IL)-6, IL-11, and LIF. This transcription factor was initially identified as APRF (acute phase response factor), an inducible DNA-binding protein that binds to the IL-6 response element within the promoters of genes that encode hepatic acute phase proteins [55]. Human and mouse STAT3 proteins both have 770 amino acid residues and are highly homologous to each other (only three amino acid differences). In addition to the DNAbinding domain, STAT3 protein contains a SH2 domain, which facilitates dimer formation upon phosphorylation by upstream kinases including JAK2. STAT3 protein is usually in an inactive form and is localized in the cytoplasm. Upon cytokine stimulation, STAT3 is tyrosinephosphorylated by activated JAKs. Thereafter, the phosphorylated STAT3 protein forms a homodimer or a heterodimer and translocates into the nucleus, where it stimulates the transcription of its target genes. In mouse development, STAT3 activity is detected during early post-implantation development [56] and *Stat3*-deficient mice die prior to gastrulation

STAT3 plays an indispensable role in the self-renewal of mouse ES cells, and deficiency of STAT3 leads to the differentiation of these cells. By contrast, human ES cells do not require STAT3 activation for their self-renewal. This discrepancy in the requirement for STAT3 is most likely due to differences between mouse and human ES cells, as mouse ES cells are derived from the ICM, whereas human ES cells are derived from the epiblast. Extensive studies have identified many self-renewal factors that are downstream of STAT3 signaling, such as

Although the STAT3 protein is normally only activated in response to signals that control cell growth, overactivation of STAT3 protein has been observed in several types of cancer, including breast, prostate, and pancreas, as well as leukemia and lymphoma [58]. Indeed,

[57], suggesting that STAT3 plays an important role in early embryogenesis.

transcription factors, epigenetic regulators, and oncogenes [8].

ovarian, and breast [48]

**2.4. STAT3**

factor, with similar activity to Nanog [54].

*Sox2*-deficient mouse embryos die shortly after implantation [34], indicating that Sox2 is required for early development. In ES cells, Sox2 usually cooperates with Oct3/4 to regulate the expression of several self-renewal factors, including Nanog, and disruption of the *sox2* gene results in the differentiation of ES cells into trophectoderm-like cells [35]. In addition, Sox2 regulates expression of histone acetyltransferases, including Tip60 and Elp3, and forms a positive feedback loop with the polycomb group protein Eed to control the levels of histone acetylation and methylation [36]. Similar to Oct3/4, Sox2 is one of the four factors required for iPS cell production [19].

Besides its pivotal role in maintaining ES cell self-renewal, Sox2 is closely associated with many types of cancer [37]. The *Sox2* gene is located at chromosome 3q26, a region that is frequently amplified in carcinomas. Amplification of the *Sox2* gene has been observed in human squa‐ mous cell carcinomas of the lung and esophagus [38]. Sox2 is involved in the tumorigenesis of several types of tumors, such as lung, breast, skin, prostate, ovarian, and sinonasal. For example, Sox2 is expressed in early-stage breast tumors, and high Sox2 expression is associated with large tumor size [39, 40]. Sox2 expression in breast cancer enhances cancer stem cell-like properties [40]. Sox2 is also reportedly involved in regulation of cancer stem-like cells in ovarian carcinoma [41].

#### **2.3. Nanog**

*Nanog* was originally identified as a gene whose overexpression can bypass the LIF require‐ ment of mouse ES cells for self-renewal [42, 43]. Nanog protein can be ubiquitinated at its PEST domain, resulting in its degradation through the proteasome pathway in ES cells [44]. Phosphorylation of Nanog prevents this ubiquitination, thereby increasing the stability of Nanog [45].

Although *Nanog*-deficient ES cells expand at a slower rate than wild-type cells, they can self-renew and retain expression of Oct3/4 and Sox2 [46]. These observations suggest that Nanog is involved in the growth of ES cells, but is dispensable for ES cell self-renewal. Human Nanog is a reprogramming factor that can produce human iPS cells [5]. In murine cells, mouse Nanog accelerates reprogramming and promotes the transition of pre-iPS cells into mature iPS cells [47].

Normally, Nanog is expressed at early embryonic stages and in germline stem cells, but not in adult tissues. However, Nanog is expressed at high levels in several types of cancers, including breast, cervical, oral and kidney [48]. Nanog is also highly expressed in germ cell tumors [49], which are characterized by the gain of the short arm of chromosome 12, at which the *Nanog* gene is located. In addition, several reports have suggested that Nanog is involved in the epithelial-mesenchymal transition and metastasis. For example, Nanog stimulates cell migration and invasion in ovarian cancer through downregulation of *E-cadherin* and *Foxj1* [50]. Nanog expression is higher in hepatocellular carcinoma cell lines that are highly metastatic than in those that are lowly metastatic, and the Nodal/Smad3 pathway plays an important role in the Nanog-stimulated epithelial-mesenchymal transition [51]. More importantly, Nanog overexpression is associated with poor prognosis in some types of cancer, such as colorectal, ovarian, and breast [48]

Nanog has 11 highly homologous pseudogenes in human cells. Of these, *NanogP8* encodes a full-length protein with only one amino acid difference from Nanog, and is involved in tumorigenesis [52, 53]. It was recently shown that NanogP8 can function as a reprogramming factor, with similar activity to Nanog [54].

#### **2.4. STAT3**

**2.2. Sox2**

iPS cell production [19].

ovarian carcinoma [41].

into mature iPS cells [47].

**2.3. Nanog**

Nanog [45].

Sox2 is a transcription factor that belongs to the SRY-related HMG-box protein (Sox) family. The Sox2 protein contains one HMG box, as well as a transactivation domain in its C-terminus. Expression of Sox2 in ES cells is mainly regulated by Oct3/4 and Sox2 itself. In addition, microRNAs (miRNAs) are involved in the control of Sox2 expression. For example, miR-9 binds to the 3'-untranslated region of *Sox2* mRNA and represses Sox2 expression [30]. In addition to regulation at the transcriptional level, the activity of Sox2 protein is regulated by post-translational modifications. Mouse Sox2 protein can be sumoylated at Lys-247, which impairs its binding to the *Fgf4* enhancer [31]. Additionally, acetylation of Lys-75 by p300/CBP promotes the nuclear export of Sox2 [32]. Akt directly interacts with Sox2 to phosphorylate

*Sox2*-deficient mouse embryos die shortly after implantation [34], indicating that Sox2 is required for early development. In ES cells, Sox2 usually cooperates with Oct3/4 to regulate the expression of several self-renewal factors, including Nanog, and disruption of the *sox2* gene results in the differentiation of ES cells into trophectoderm-like cells [35]. In addition, Sox2 regulates expression of histone acetyltransferases, including Tip60 and Elp3, and forms a positive feedback loop with the polycomb group protein Eed to control the levels of histone acetylation and methylation [36]. Similar to Oct3/4, Sox2 is one of the four factors required for

Besides its pivotal role in maintaining ES cell self-renewal, Sox2 is closely associated with many types of cancer [37]. The *Sox2* gene is located at chromosome 3q26, a region that is frequently amplified in carcinomas. Amplification of the *Sox2* gene has been observed in human squa‐ mous cell carcinomas of the lung and esophagus [38]. Sox2 is involved in the tumorigenesis of several types of tumors, such as lung, breast, skin, prostate, ovarian, and sinonasal. For example, Sox2 is expressed in early-stage breast tumors, and high Sox2 expression is associated with large tumor size [39, 40]. Sox2 expression in breast cancer enhances cancer stem cell-like properties [40]. Sox2 is also reportedly involved in regulation of cancer stem-like cells in

*Nanog* was originally identified as a gene whose overexpression can bypass the LIF require‐ ment of mouse ES cells for self-renewal [42, 43]. Nanog protein can be ubiquitinated at its PEST domain, resulting in its degradation through the proteasome pathway in ES cells [44]. Phosphorylation of Nanog prevents this ubiquitination, thereby increasing the stability of

Although *Nanog*-deficient ES cells expand at a slower rate than wild-type cells, they can self-renew and retain expression of Oct3/4 and Sox2 [46]. These observations suggest that Nanog is involved in the growth of ES cells, but is dispensable for ES cell self-renewal. Human Nanog is a reprogramming factor that can produce human iPS cells [5]. In murine cells, mouse Nanog accelerates reprogramming and promotes the transition of pre-iPS cells

Thr-118, leading to stabilization of the Sox2 protein [33].

44 Pluripotent Stem Cell Biology - Advances in Mechanisms, Methods and Models

STAT3 is a downstream transcription factor of gp130, and is therefore activated by several cytokines, including interleukin (IL)-6, IL-11, and LIF. This transcription factor was initially identified as APRF (acute phase response factor), an inducible DNA-binding protein that binds to the IL-6 response element within the promoters of genes that encode hepatic acute phase proteins [55]. Human and mouse STAT3 proteins both have 770 amino acid residues and are highly homologous to each other (only three amino acid differences). In addition to the DNAbinding domain, STAT3 protein contains a SH2 domain, which facilitates dimer formation upon phosphorylation by upstream kinases including JAK2. STAT3 protein is usually in an inactive form and is localized in the cytoplasm. Upon cytokine stimulation, STAT3 is tyrosinephosphorylated by activated JAKs. Thereafter, the phosphorylated STAT3 protein forms a homodimer or a heterodimer and translocates into the nucleus, where it stimulates the transcription of its target genes. In mouse development, STAT3 activity is detected during early post-implantation development [56] and *Stat3*-deficient mice die prior to gastrulation [57], suggesting that STAT3 plays an important role in early embryogenesis.

STAT3 plays an indispensable role in the self-renewal of mouse ES cells, and deficiency of STAT3 leads to the differentiation of these cells. By contrast, human ES cells do not require STAT3 activation for their self-renewal. This discrepancy in the requirement for STAT3 is most likely due to differences between mouse and human ES cells, as mouse ES cells are derived from the ICM, whereas human ES cells are derived from the epiblast. Extensive studies have identified many self-renewal factors that are downstream of STAT3 signaling, such as transcription factors, epigenetic regulators, and oncogenes [8].

Although the STAT3 protein is normally only activated in response to signals that control cell growth, overactivation of STAT3 protein has been observed in several types of cancer, including breast, prostate, and pancreas, as well as leukemia and lymphoma [58]. Indeed, STAT3 is constitutively activated in nearly 70% of tumors. Because of the high frequency of its overactivation, STAT3 is considered to be a valuable target for anti-cancer therapy.

brain [71, 72]. Loss of the zygotic function of Zfp57 leads to partial lethality, while eliminating both the maternal and zygotic functions of Zfp57 results in complete embryonic lethality [73]. Through its KRAB domain, Zfp57 interacts with KRAB-associated protein 1 (Kap1), a scaffold protein for heterochromatin-inducing factors, and thus participates in genome imprinting by recruiting Kap1 to multiple imprinting control regions [73, 74]. Mutations in the *Zfp57* gene cause transient neonatal diabetes mellitus type 1 [75]. Zfp57 is a downstream molecule of STAT3 and Oct3/4 in ES cells, and is therefore specifically expressed in self-renewing ES cells [76]. Zfp57 deficiency has no effect on the self-renewal or growth in ES cells, suggesting that

Embryonic Stem Cells and Oncogenes http://dx.doi.org/10.5772/57614 47

Based on our prediction that a molecule expressed in self-renewing ES cells may play an important role in cancer cell growth, we recently screened several ES cell-specific transcription factors for their tumor-promoting activity, and found that Zfp57 can promote anchorageindependent growth of human fibrosarcoma HT1080 cells [77]. Zfp57 overexpression enhan‐ ces, while its knockdown suppresses, HT1080 tumor formation in nude mice. Zfp57 regulates the expression of insulin-like growth factor 2, which plays a critical role in Zfp57-induced anchorage-independent growth and tumor formation. Furthermore, overexpression of Zfp57 causes anchorage-independent growth of the mouse immortal fibroblast cell line NIH3T3, and immunohistochemical analysis revealed the overexpression of Zfp57 in several cancers, including pancreatic, gastric, breast, colon, and esophageal. These results suggest that *Zfp57* is an oncogene in some types of cancer. Moreover, we also found that Zfp57 is involved in anchorage-independent growth of ES cells and that *Zfp57*-null ES cells form smaller teratomas than the parental ES cells in immunodeficient mice, suggesting the importance of Zfp57 in

The importance of the Wnt pathway in tumorigenesis was recognized by identification of adenomatous polyposis coli (APC) mutations in familial adenomatous polyposis [78, 79]. In Wnt signaling, the Apc protein functions as a negative regulator and is involved in degradation of β-catenin, the mammalian homologue of *Drosophila* Armadillo. β-catenin contains multiple armadillo repeats in its central region and a transcriptional activator domain in its C-terminal region. Human and mouse β-catenin proteins both have 781 amino acid residues and are almost identical to each other (only one amino acid difference). β-catenin acts as a transcrip‐ tional co-activator and an adaptor protein for intracellular adhesion. In epithelial tissues, βcatenin interacts with cadherins and α-catenin, and regulates epithelial cell growth and intracellular adhesion. By contrast, in Wnt signaling, β-catenin is a major transcriptional modulator and plays a crucial role in embryogenesis. In the absence of Wnt signaling, Apc forms a complex with β-catenin and Axin. This leads to phosphorylation of β-catenin by glycogen synthase kinase (GSK)-3β, which triggers degradation of β-catenin. When Wnt binds to its receptor Frizzled, Disheveled is hyper-phosphorylated, which results in release of

this transcription factor is dispensable for ES cell self-renewal.

**3. Oncogenes that are involved in ES cell self-renewal**

teratoma formation by ES cells.

**3.1. β-catenin**

#### **2.5. Krüppel-like factor (Klf) 4**

Klfs are a family of transcription factors that play important roles in many fundamental biological processes. They were named "Krüppel-like" owing to their strong homology with the *Drosophila* gene product Krüppel, which is involved in segmentation of the develop‐ ing embryo. Klf family proteins contain three C2H2-type zinc fingers that bind DNA. *Klf4* was independently cloned by two groups and named "gut-enriched KLF" and "epithelial zinc finger" owing to its high expression in the intestine and skin epithelium, respective‐ ly [59, 60]. However, it was later discovered that this transcription factor is expressed in several other tissues, such as lung, testis, and thymus. In addition to a C-terminal DNAbinding domain consisting of zinc fingers, Klf4 protein contains an activation domain in its N-terminus and a repressive domain in its central region. Probably owing to this struc‐ ture, Klf4 is a bi-functional transcription factor that can either activate or repress transcrip‐ tion of its target genes. Similar to Oct3/4 and Sox2, Klf4 is one of the four factors that induce reprogramming of murine cells [19].

Klf4 is highly expressed in self-renewing ES cells, but not in differentiated ES cells. Klf4 regulates self-renewal-specific expression of Lefty1, in cooperation with Oct3/4 and Sox2 [61]. Klf4 is also involved in Oct3/4 expression [62]. Overexpression of Klf4 results in the inhibition of ES cell differentiation, possibly through upregulation of *Nanog* [63, 64]. In addition, there is a marked overlap between genes that are regulated by Nanog and those that are regulated by Klf4. These observations suggest the importance of Klf4 in ES cell self-renewal. However, *Klf4*-null mice have no detectable defects during embryogenesis [65]. Furthermore, a recent study reported that the function of Klf4 in ES cell self-renewal is partially redundant because combined knockdown of *Klf4*, *Klf2*, and *Klf5*, but not any one gene individually, results in spontaneous ES cell differentiation [62], suggesting that Klf4 is dispensable for ES cell selfrenewal.

In cancer cells, Klf4 acts as a tumor suppressor or an oncogene, possibly owing to its bifunctionality. Whether Klf4 acts as a tumor suppressor or an oncogene likely depends on the tumor type. For example, Klf4 functions as a tumor suppressor in the intestinal and gastric epithelium, and expression of *Klf4* is downregulated in human colorectal and gastric carcino‐ mas [66, 67]. On the other hand, overexpression of Klf4 in the skin results in squamous epithelial dysplasia, eventually leading to squamous cell carcinoma [68, 69]. A high expression level of Klf4 significantly correlates with a poor prognosis in hepatocellular carcinoma [70].

#### **2.6. Zinc-finger protein (Zfp)-57**

Zfp57 is a transcription factor that was originally identified as an undifferentiated cell-specific gene in F9 embryonal carcinoma cells [71]. Mouse Zfp57 protein contains one Kruppelassociated box (KRAB) domain and five zinc fingers, while human Zfp57 protein has one KRAB domain and seven zinc fingers. In adult mouse, Zfp57 is highly expressed in testis and brain [71, 72]. Loss of the zygotic function of Zfp57 leads to partial lethality, while eliminating both the maternal and zygotic functions of Zfp57 results in complete embryonic lethality [73]. Through its KRAB domain, Zfp57 interacts with KRAB-associated protein 1 (Kap1), a scaffold protein for heterochromatin-inducing factors, and thus participates in genome imprinting by recruiting Kap1 to multiple imprinting control regions [73, 74]. Mutations in the *Zfp57* gene cause transient neonatal diabetes mellitus type 1 [75]. Zfp57 is a downstream molecule of STAT3 and Oct3/4 in ES cells, and is therefore specifically expressed in self-renewing ES cells [76]. Zfp57 deficiency has no effect on the self-renewal or growth in ES cells, suggesting that this transcription factor is dispensable for ES cell self-renewal.

Based on our prediction that a molecule expressed in self-renewing ES cells may play an important role in cancer cell growth, we recently screened several ES cell-specific transcription factors for their tumor-promoting activity, and found that Zfp57 can promote anchorageindependent growth of human fibrosarcoma HT1080 cells [77]. Zfp57 overexpression enhan‐ ces, while its knockdown suppresses, HT1080 tumor formation in nude mice. Zfp57 regulates the expression of insulin-like growth factor 2, which plays a critical role in Zfp57-induced anchorage-independent growth and tumor formation. Furthermore, overexpression of Zfp57 causes anchorage-independent growth of the mouse immortal fibroblast cell line NIH3T3, and immunohistochemical analysis revealed the overexpression of Zfp57 in several cancers, including pancreatic, gastric, breast, colon, and esophageal. These results suggest that *Zfp57* is an oncogene in some types of cancer. Moreover, we also found that Zfp57 is involved in anchorage-independent growth of ES cells and that *Zfp57*-null ES cells form smaller teratomas than the parental ES cells in immunodeficient mice, suggesting the importance of Zfp57 in teratoma formation by ES cells.
