WNT/β-Catenin signaling pathway regulates non-tumorigenesis of human embryonic stem cells co-cultured with human umbilical cord mesenchymal stem cells

Human pluripotent stem cells harbor hope in regenerative medicine, but have limited application in treating clinical diseases due to teratoma formation. Our previous study has indicated that human umbilical cord mesenchymal stem cells (HUCMSC) can be adopted as non-teratogenenic feeders for human embryonic stem cells (hESC). This work describes the mechanism of non-tumorigenesis of that feeder system. In contrast with the mouse embryonic fibroblast (MEF) feeder, HUCMSC down-regulates the WNT/β-catenin/c-myc signaling in hESC. Thus, adding β-catenin antagonist (FH535 or DKK1) down-regulates β-catenin and c-myc expressions, and suppresses tumorigenesis (3/14 vs. 4/4, p = 0.01) in hESC fed with MEF, while adding the β-catenin enhancer (LiCl or 6-bromoindirubin-3′-oxime) up-regulates the expressions, and has a trend (p = 0.056) to promote tumorigenesis (2/7 vs. 0/21) in hESC fed with HUCMSC. Furthermore, FH535 supplement does not alter the pluripotency of hESC when fed with MEF, as indicated by the differentiation capabilities of the three germ layers. Taken together, this investigation concludes that WNT/β-catenin/c-myc pathway causes the tumorigenesis of hESC on MEF feeder, and β-catenin antagonist may be adopted as a tumor suppressor.

This investigation explores the signaling pathway responsible for the HUCMSC-mediated down-regulation of c-Myc and the non-tumorigenic feature of hESC. We found that β -catenin signaling is the main factor controlling tumorigenesis, and that its inhibition mimics the tumor suppressor activity of HUCMSC.

HUCMSC feeder inhibits tumorigenesis of hESC via the β-catenin/c-myc signaling pathway.
To identify the non-tumorigenetic signaling of HUCMSC-feeder to hESC, the expressions of β -catenin and c-myc in the three types of co-culture feeders were first investigated. Shifting from the MEF feeder to the HUCMSC feeder reduced the expression of mRNA and protein of β -catenin in hESC. The expression of β -catenin rebounded after shifting back to the MEF feeder (Fig. 1b). Similarly, changes to the c-myc expression in hESC depended on the feeder, with down-regulation occurring when using the HUCMSC feeder, and up-regulation upon shifting back to MEF (Fig. 1a,b).
A reporter assay of the key β -catenin target gene TCF/LEF was performed to investigate the downstream target of β -catenin transactivation. The hESC cultured on HUCMSC had a significant lower TCF/LEF transactivating activity than the MEF-feeder. The activity rebounded significantly upon turning back to the MEF feeder (Fig. 1c). A chromatin immunoprecipitation (ChIP) assay further confirmed the binding of β -catenin to the promoter sequences of CMYC in the hESC/MEF culture, and the same reduction in the hESC/HUCMSC culture and rebounding when turning back to MEF feeder (Fig. 1d). The differentiation status of hESC in MEF or HUCMSC feeder in the embryoid body (EB) state was also tested. Experimental results show expressions of genes (c) TCF/LEF activity of hESC cultured on different condition was measured by luciferase assay, and Firefly luciferase activity was normalized to Renilla luciferase activity, which was adopted as internal control. Values are shown as the mean of three replicates ± standard deviations. (d) Real-time PCR analysis of DNA fragments precipitated in a ChIP assay by using a β -catenin antibody. Primers designed for the 5′ promoter of c-myc were adopted to detect specific β -catenin binding. Data are represented as percentage input. Error bars represent SEM. (e) Three germ-layer differentiation gene expressions of embryoid body (EB) derived from hESC cultured on MEF and HUCMSC were compared by qRT-PCR (ectoderm: β-3-tubulin, MAP2, GFAP; endoderm: GATA4; mesoderm: GATA6, Hand1). *p < 0.05, **p < 0.01, ***p < 0.001. All cropped blots were run under the same experimental conditions in (b). of the three germ layers, including beta-3-tubulin, MAP2, GFAP (ectoderm); GATA4 (endoderm); GATA6, Hand1 (mesoderm) were not altered in hESC cultured on either MEf or HUCMSCs feeder (Fig. 1e). Furthermore, the EB of hESC cultured on HUCMSC had even higher expressions of MAP2, GATA4, GATA6 and GFAP than that cultured on MEF.

Lithium Chloride (LiCl) and BIO (6-bromoindirubin-3′-oxime) up-regulated the c-myc in hESC/ HUCMSC in vitro and in vivo.
To determine whether c-myc is up-regulated via the canonical β -catenin signaling pathway in hESC/HUCMSC, LiCl and BIO were applied to increase the cytoplasmic β -catenin. Treatment with 10 mM LiCl treatment did not change the cell morphology of hESC/HUCMSC (Fig. 2a), but significantly increased the protein level of β -catenin, and the mRNA and protein levels of CMYC (Fig. 2b,c). Treatment with BIO also showed the same results (Fig. 2d,e).

FH535 inhibited and LiCl promoted teratoma formation through the β-catenin/c-myc signaling pathway in vivo.
To further confirm the tumorigenic role of β -catenin/c-myc signaling, the hESC/ MEF were treated with the β -catenin signaling inhibitor FH535, and subjected to xenograft in NOD/SCID mice. After 3 months, FH535 significantly inhibited teratoma formation of hESC/MEF. Only 3 of 14 (21.4%) injection sites developed teratoma, compared to 4 of 4 (100%) in the non-treated group (p = 0.011, Table 2). The β -catenin mediated myc effect was also observed in the hESC culture. As demonstrated in Fig. 4, 24 hours treatment with FH535 or DKK1 led to down-regulation of the downstream β -catenin and c-myc protein was evident in hESC/ MEF (Fig. 4), but did not change the morphology of hES cells (Fig. 4a).

Discussion
Our previous study found for the first time that hESC lost their ability of teratoma formation upon co-culturing with HUCMSC, but regained the tumor forming activity after shifting back to the MEF co-culture 3 . The hESC/  HUCMSC coculture indicated down-regulation of c-myc. This study further found that c-myc is down-regulated via the β -catenin signaling pathway. A TCF/LEF reporter assay and ChIP assay were performed to confirm the binding of β -catenin to the promoter of CMYC in hESC/MEF culture. Inhibition β -catenin with FH535 or DKK1 down-regulated both mRNA and protein expression of CMYC, and reduced teratoma formation by 79%. Conversely, activation of β -catenin signaling with LiCl or BIO induced an up-regulation of c-myc in the hESC/ HUCMSC culture, and increased the teratoma formation from 0% to 28.6%. These findings indicate that the β -catenin/c-myc signaling pathway is largely responsible for the tumorigenicity of hESC. Our previous study of hESC co-cultured with HUCMSC found downregulation of both Oct4 and c-myc as compared to hESC with MEF feeder 3 . Oct4 is reportedly involved in the formation of multiple cancers and their stem cells, such as colorectal 18 , liver 19 , cervical 20,21 , oral 22 and ovarian cancer 23 , where its down-regulation is associated with slowed tumor progression and proliferation 24 . This may explain the incomplete recovery of tumorigenesis after adding LiCl to resume c-myc expression.
The GSK3 inhibitor LiCl and BIO stabilize the intracytoplasmic β -catenin protein, preventing its degradation by proteasome 25 . LiCl helps maintain the pluripotency in hESC by enhancing the β -catenin signaling 26 . Consistent with that finding, this study found that treating the HUCMSC-fed hES with LiCl could resume the expression of β -catenin and transcription of CMYC, and eventually resume the teratoma formation activity.
Our result is consistent with a previous report that Wnt/β -catenin signaling pathway plays a fundamental role in modulating the tumorigenicity of ESC-derived retinal progenitors 27 . Their investigation found that WNT signaling-activated TCF7-SOX2-NESTIN cascade was responsible for the tumor formation. Our study further identified c-myc as the major oncogenic effector of Wnt/β -catenin signaling in hESC tumorigenesis.
C-myc is a direct transactivating target of β -catenin 13 , and is one of the earliest-found oncogens. It plays a significant role in regulating cell proliferation, differentiation, stemness, senescence and tumor invasiveness 28 . In embryonic stem cells, c-myc is a universal amplifier of expressed genes 29 . It also plays a significant role in transcriptional regulation and matainance of the pluripotency in hESC 30 . This study found that inhibiting suppresses CMYC is the major mediator tumorigenic signalin of β -catenin in the hESC/MEF culture.
The teratogenicity of pluripotent embryonic stem cells has inhibited their clinical application in regenerative cellular therapy. Inhibition of β -catenin signaling could reduce teratoma formation of hESC, potentially enabling the development of safer cellular therapy than using therapy with hESC with full teratoma formation capability. Although β -catenin pathway also plays a fundamental role in stem cells self-renewal and maintenance of stem cell properties 31 , this study found that inhibition of β -catenin with FH535 does not compromise the pluripotency.
The conventional method of preventing teratoma formation in hESC cellular therapy is to apply this therapy only on the differentiated cells. Undifferentiated cells are eliminated by treating them with chemical inhibitor (YM155) to down-regulate survivin signaling 32 , with antibodies, small molecules, anti-angiogenic agents, or with suicide genes for elimination 33 . This study demonstrated for the first time that FH535 can be utilized to reduce teratogenesis in cultured hESC before induction of differentiation. Adding the beta-catenin inhibitor FH535 reduced teratoma formation by 79%. Meanwhile, researchers have investigated targeting β -catenin signaling as a novel treatment for multiple cancers such as those of the breast 34 , pancreas 35 , esophagus 36 and liver 37 . This study recommends modulating β -catenin to reduce the risk of teratoma formation in hESC transplantation.
The non-tumorigenesis of HUCMSC coculture has many possible factors in the upstream signal of β -catenin. WIF (or sFRP) may be secreted to counteract the Wnt-Frizzled binding and down-regulate beta-catenin signaling 38 . DKK1 and SOST/WISE proteins can also bind to LRP5/6 to prevent Frizzled-LRP6 complex formation 38 . DKK1 secreted from HUCMSC was found to inhibit breast cancer cell growth 39 . This study found that DKK1 could significantly reduce nuclear β -catenin and c-myc expression of hESC/MEF. While the function of molecules in HUCMSC secretome is still largely unknown [40][41][42] , this study found that DKK1 is likely to be the major tumor suppressor secreted in the hESC/HUCMSC coculture.
In summary, this study reveals that the hESC/HUCMSC co-culture can confer a non-tumorigenesis phenotype of hESC. Down-regulation of the β -catenin/c-myc signaling inhibits tumor formation. Inhibition of this signaling by β -catenin inhibitor could markedly reduce the incidence of teratoma formation in the conventional hESC/MEF co-culture system.

Methods
Culture and Passage of hESC. The Research Ethics Committee of Buddhist Tzu Chi General Hospital (IRB 100-166) approved the protocols for collecting and using human umbilical cord. Written informed consent was obtained from the pregnant women before labor. The methods were performed in accordance with the relevant guidelines, including any relevant details. The TW1 hESC was obtained from the Food Industry Research and Development Institute of Taiwan, and maintained on the mitomycin-C treated MEF (hESC/MEF) or HUCMSC (hESC/HUCMSC) following a previously reported protocol 3 . The experiment was performed using knock-out (KO) Dulbecco's modified Eagle's medium (DMEM) and 20% (v/v) KO Serum Replacement containing 2 mM glutamine, 10 nM non-essential amino acids (all from Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA, https://www.thermofisher.com/), 50 μ M B-mercaptoethanol (Sigma-Aldrich, St. Louis, MO, http://www.sigmaaldrich.com) and 4 ng/ml basic fibroblast growth factor (bFGF). The medium was changed daily, and the hESC was passed each week.
To prepare the hESCs with different feeders, hESC was first co-cultured with MEF (hESC/MEF), and the established hESC clusters were transferred to the HUCMSC feeder as the hESC/HUCMSC co-culture. For the reverse co-culture, hESC was further transferred to the MEF feeder as the hESC/MHM. All the feeder transfers were performed after five passages in the previous feeder. The hESC/MHM were also maintained for more than five passages before investigation.
Quantitative RT-PCR and RT-PCR. RNA for all qRT-PCR and RT-PCR analysis was prepared using Trizol (Invitrogen) and quantified. 500ng of RNA was DNAase-treated using DNaseI amplification grade (Invitrogen). The first strand of cDNA was synthesized by a SuperScript III One-Step RT-PCR kit (Invitrogen) following the manufacturer's instructions. All PCR samples were analyzed by electrophoresis on 2% agarose gel containing 0.5 μ g/ml ethidium borome (Sigma). The quantitative RT-PCR (qRT-PCR) analysis adopted FastStart universal SYBR green master (ROX, Roche, Basel, Switzerland, https://lifescience.roche.com) gene expression assays in an ABI Step One Plus system (Applied Biosystems, Thermo Fisher Scientific), with GAPDH as an internal control. Table 3 shows the sequences of primers and product size.  Chromatin immunopreciptation (ChIP) assay. SimpleChIP Enzymatic Chromatin IP kit (Cell Signaling) was adopted for ChIP assay for beta-catenin-CMYC promoter binding. The assay with beta-catenin antibody was performed with 4 × 10 6 hES cells cultured with different feeders, according to the manufacturer's instructions. The bound CMYC sequences were quantified by qPCR after preparing ChIP DNA. The primer sequences of CMYC promoter are listed as below: forward GTG AAT ACA CGT TTG CGG GTT AC; reverse AGA GAC CCT TGT GAA AAA AAC CG.
Extraction of cytoplasmic/nuclear proteins from hESCs. ReadyPrep protein extraction kit (cytoplasmic/nuclear, Bio-Rad) was adopted to isolate protein from the nucleus and cytoplasm. The isolation procedures were conducted according to the manufacturer's instructions. The resulted cytoplasmic/nuclear protein were analysed by Western blot.
TCF/LEF report assay. The TCF/LEF signal reporter assay kit (Qiagen) was adopted to demonstrate the transactivating activity of β -catenin. Briefly, one day before transfection, ES cells were seeded at a density of 30,000 cells per well into a 96-well plate in 100 μ l of medium. Next day, 1 μ l of TCF/LEF luciferase reporter (component A) was transfected from each well into cells. After 24 h of transfection, 55 μ l of Luciferase reagent per well were added and shaken at room temperature for 15 min, and firefly luminescence was measured by a luminometer. 55 μ l of Stop & Glo reagent per well were added and rocking at room temperature for 15 min, and Renilla luminescence was measured. To obtain the normalized luciferase activity for TCF/LEF reporter, the background luminescence was calculated as the ratio of firefly luminenscence from the TCF/LEF reporter to Renilla luminescence, and subtracted from the control Renilla luciferase vector. Graphs were plotted from data obtained as a mean of three independent experiments, with standard deviation shown as error bars.
Xenograft in immune-compromised mice. All animal works were in accordance with protocols approved by the Institutional Animal Care and Use Committee at the Buddhist Tzu Chi General Hospital. All methods were performed in accordance with the relevant guidelines and regulations. For the tumorigenesis assay, hESC were removed from the feeder with mechanical slicing using glass capillaries, then pelleted and resuspended in PBS. For the xenograft, 5 × 10 5 cells mixed with Matrigel (1:1) were injected into the back subcutaneous tissue of 6-8-week-old female non-obese diabetic-severe combined immunodeficiency (NOD-SCID) mice.
Tumor formation was followed up by palpation. The resulting tumors were dissected, fixed, embedded in paraffin and processed for histological examination.