Pluripotent stem cells can differentiate into cells of all lineages and self-renew (Evans and Kaufman, 1981; Martin, 1981). They can be produced through the reprogramming of somatic cells, and induced pluripotent stem cells (iPSCs) are expected to be used as cell therapeutics without immune rejection (Takahashi and Yamanaka, 2006).

However, reprogramming using a viral vector has the potential to cause mutations due to the integration of the transgene into the host genome (Yu et al., 2007). To solve this issue, the use of episome, a non-viral vector (Yu et al., 2009), and some other methods (Jia et al., 2010; Kim et al., 2009; Okita et al., 2011; Woltjen et al., 2009) have been devised, but these methods have a limitation in that residues remain in cells. In addition, all of the above methods use the reprogramming transcription factors Oct4, Sox2, Klf4, and c-Myc, among which c-Myc is a tumorigenic gene and has the potential to cause cancer (Okita et al., 2007).

Therefore, a new reprogramming method is needed to use iPSC as a cell therapy. Chemical reprogramming method using small molecules overcomes the above limitations (Guan et al., 2022; Hou et al., 2013; Long et al., 2015; Ye et al., 2016; Zhao et al., 2015). In this method, reprogramming is induced by adding small molecules that can regulate the target cell signaling pathway and epigenetic modification. Chemical reprogramming is simpler than other reprogramming methods and the use of compounds in various combinations and concentrations allows precise control.

In the process of chemical reprogramming, extra-embryonic endoderm (XEN) cells appear. XEN cells are one of the cell types that appear during development (Zhao et al., 2015) and are multipotent, contributing to the formation of visceral endoderm, parietal endoderm, and yolk sac (Hogan et al., 1980). Chemical reprogramming activates a mechanism that expression of Oct4, which is the core transcription factor for the induction of iPSCs, depends on the increasing expression of XEN marker genes (Gata4, Gata6, Sox17) during XEN cell induction. This indicates that the induction of XEN cells is essential for the induction of iPSCs by chemical reprogramming.

Although this method is simple, most of the reported chemical reprogramming protocols have been performed and repeated in the same group (Guan et al., 2022; Hou et al., 2013; Takeda et al., 2018; Zhao et al., 2015; 2018). In this study, we tried to establish a more detailed protocol assuming that XEN cells are the critical point of chemical reprogramming. Under the assumption that the homogeneity of XEN cells is important in chemical reprogramming, we examined whether the efficiency of inducing pluripotency can be improved by increasing the homogeneity of XEN cells.

Mouse embryonic fibroblasts (MEFs) culture

MEFs were obtained from Max-Planck Institute for Molecular Biomedicine and were cultured in MEF medium consisting of Dulbecco’s modified Eagle’s medium (DMEM; Welgene, Korea) supplemented with 10% fetal bovine serum (FBS; Corning, USA), 0.1 mM non-essential amino acids (NEAA; Gibco, USA), and 1% penicillin-streptomycin (P/S; Welgene). MEFs were plated onto dishes coated with 0.1% gelatin (Sigma-Aldrich, USA)-coated dishes.

CiXEN cells induction

Oct4-GFP reporter MEFs (OG2MEFs) at passage 2 or 3 were seeded in MEF medium into a 12-well plate pre-coated with 0.01% gelatin at a density of 3 × 10⁴ cells per well. After 24 h, the medium was changed to CiXEN induction medium, which was replaced every 4 days. The composition of CiXEN induction medium is described in Supplementary Table S1.

Induction of chemically induced pluripotent stem cells (CiPSCs)

CiXEN cells were seeded into a gelatin pre-coated 12-well plate at a density of 7 × 10⁵ cells per well and were cultured in CiPSC induction step 1 medium for 4 days, then in step 2 medium for 12 days, and finally in step 3 medium for 12 days. Each medium was changed every 4 days. The composition of steps 1, 2, and 3 media is described in Supplementary Table S1. CiPSCs cultured in mouse embryonic stem cell (mESC) medium consisting of Knockout DMEM (Gibco) supplemented with 10% FBS, 10% KnockOut Serum Replacement, 0.1 mM NEAA, 0.55 mM β-mercaptoethanol, 1% P/S, 1% GlutaMax (Gibco) and small molecules (CHIR99021 and PD0325901 or LIF). An overview of the generation of CiXENs and CiPSCs is shown in Supplementary Fig. S1.

Quantitative PCR

Total RNA was isolated using the RNeasy Kit (Qiagen, Germany) according to the manufacturer’s instructions. Total RNA (1 µg) was reverse-transcribed into cDNA using the High Capacity cDNA Reverse Transcription kit (Applied Biosystems, USA) according to the manufacturer’s instructions. Quantitative PCR (qPCR) was performed using SYBR Green (Enzynomics, Korea) on a real-time PCR system (Applied Biosystems). All reactions were run in triplicate and values were normalized to endogenous β-actin control. Primer sequences are listed in Supplementary Table S2.

Alkaline phosphatase staining and immunocytochemistry

Alkaline phosphatase staining was performed using the Leukocyte Alkaline Phosphatase kit (Stemgent, USA) according to the manufacturer’s instructions. For immunocytochemistry, cells were washed with DPBS and fixed in 4% paraformaldehyde for 15 min at room temperature. The fixed cells were permeabilized with 0.25% Triton X-100 (Sigma-Aldrich) in DPBS for 20 min at room temperature and then blocked with 2% bovine serum albumin (Sigma-Aldrich) in DPBS. The cells were then incubated in primary antibody solution overnight at 4°C. After washing with DPBS, the cells were incubated in a secondary antibody for 1 h at room temperature. After washing with DPBS, nuclei were stained with DAPI (Sigma-Aldrich).

Genomic DNA isolation and bisulfite sequencing

To determine the DNA methylation status of CiPSCs, genomic DNA was isolated using a G-spin Total DNA Extraction Kit (iNtRON, Korea). Genomic DNA (1 µg) was modified using an EpiTect Bisulfite Kit (Qiagen) according to the manufacturer’s instructions. The promoter region of the mouse Oct4 gene was amplified by PCR. The amplified products were purified using a Wizard Genomic DNA Purification Kit (Promega, USA) and subcloned using the PCR Cloning Kit (Qiagen) according to the manufacturer’s instructions. Individual clones were sequenced using T7 promoter primers. The data were visualized and aligned using QUMA (Quantification Tool for Methylation Analysis; http://quma.cdb.riken.jp/).

In vivo teratoma formation

CiPSCs were collected into an Eppendorf tube, and the cell pelleted by centrifugation were resuspended in a 1:1 mixture of mESC culture medium and Matrigel (Corning). CiPSCs were injected subcutaneously into an immunodeficient ICR mouse. After 3 weeks, the tetratoma was surgically dissected and fixed in 4% paraformaldehyde (Sigma-Aldrich), embedded in paraffin, and stained using an H&E Staining Kit (Abcam, UK) according to the manufacturer’s instructions.

In vitro differentiation

For embryoid body (EB) formation, CiPSCs were harvested using 0.25% trypsin–EDTA, and 2.0 × 10⁶ cells were suspended in mESC medium. After 5 days, EBs were transferred to a 4-well plate and were cultured in MEF medium for 14 days.

Generation of chimeric mice

Donor C57BL/6N mice were treated with pregnant mare serum gonadotropin and, 48 h after, with human chorionic gonadotropin in order to stimulate superovulation and induce ovulation, respectively. Mice were sacrificed the next day, and blastocysts were collected by flushing the uterus. CiPSCs were treated with trypsin to obtain a single-cell suspension and 10-15 CiPSCs were microinjected into donor blastocysts and incubated at 37°C in 5% CO2 for 1-2 h. Pseudopregnant recipient mice (ICR, E2.5) were used to implant microinjected blastocysts into the uterus. Pups were obtained approximately 17 days afterward and examined for chimerism.

Chemically induced XEN cells generated from MEFs

In this experiment, we used OG2MEFs, which are programmed to express green fluorescent protein (GFP) when Oct4 is expressed. OG2MEFs (3 × 10⁴ cells) were seeded in a gelatin-coated 12-well plate, and the next day changes the CiXEN induction medium. In CiXEN induction medium, cell morphology changed over time (Fig. 1A). However, this change was not uniform (Supplementary Fig. S2), and in particular, cells of mESCs-like colonies did not express both the XEN cell marker gene and pluripotency marker gene (Supplementary Fig. S3). In cells of well that appear typical XEN colonies, the expression of the XEN marker genes (Gata6, Gata4, Sall4, Sox17) gradually increased during the induction (Supplementary Fig. S4). After 12 days in CiXEN induction medium, the cells that had completed the first induction process expressed XEN cell–specific marker genes and proteins (Gata4, Gata6) (Figs. 1B and 1C).

Establishment of a homogeneous CiXEN cell line

CiXEN cells after 12 days of induction had heterogeneous morphology; they were capable of stable subculture and their morphology gradually became homogeneous during continuous subculture. To verify this, we confirmed that each XEN marker gene (Gata4, Gata6, Sall4, Sox17) was expressed at a higher level at passages 4 and above (Fig. 2A) and they have homogeneous morphology (Fig. 2B). Consistently, the expression of Gata4 and Gata6 proteins was also confirmed (Fig. 2C). The marker proteins (Gata4, Gata6) was expressed only in some XEN cells at passage 1, but in most cells at passage 8 (Fig. 2C). Thus, we confirmed that several passages are necessary to obtain a homogenous CiXEN cell population. These data show that the CiXEN cell induction process using chemicals is not uniform and that a single passage is insufficient to produce homogeneous CiXEN cells.

Generation of CiPSCs from homogeneous CiXEN cells

We next obtained CiPSCs from CiXEN cells in 3 steps. When we used unstable heterogeneous CiXEN cells at passage 1, almost all of them did not proliferate well or died during steps 2 and 3 (Supplementary Fig. S5); although the induction experiment was continued with a small number of surviving cells, no Oct4-GFP positive cells appeared. However, the homogeneous CiXEN cells at passage 12 proliferated stably and were able to smoothly proceed to the step 2 and 3 (Fig. 3A). Some of the GFP-positive cells that intermittently appeared during steps 2 and 3 proliferated to form GFP-positive colonies and most of GFP-positive single cells did not proliferate or lost GFP. GFP-positive colonies had a typical morphology of mESC colonies and were able to proliferate and be subcultured when picked up and plated onto MEF feeder cells (Fig. 3B). We induced CiPSCs from CiXEN cells at passages 1 to 20, and none of the passages below passage 4 showed any GFP-positive cells. On the other hand, when passage 4 or higher was used, GFP-positive cells were detected in 13 out of 42 experiments (Fig. 3C), and among them, we established three CiPSCs cell lines. These results indicate that the homogeneity of CiXEN cells is important for inducing CiPSCs, and that successful reprogramming can be achieved only by inducing CiPSCs from CiXENs at passage 4 or higher.

Characterization of CiPSCs

We investigated the properties of GFP-positive CiPSCs derived from a GFP-positive colony. CiPSCs were positive for alkaline phosphatase (Fig. 4A). This cell line has a normal karyotype (40,XY) (Supplementary Fig. S6C). The expression levels of the Oct4, Sox2, Nanog, Cripto, FGF4, Rex1, Utf1, and Esg1 genes in CiPSCs were similar to those in mESC (Fig. 4B), and the pluripotency marker proteins Oct4, Nanog, Sox2, and SSEA1 were also expressed (Fig. 4C). These data confirmed that CiPSCs exhibited genetic characteristics similar to those of mESCs. Next, we performed RNA sequencing to analyze global gene expression. Scatter plots showed that the global gene expression of CiPSCs was more similar to that of mESCs than to that of MEFs (Fig. 4D). Heat map analysis (Fig. 4E) confirmed that CiXEN cells expressed significantly higher levels of XEN markers than MEFs, CiPSCs, and mESCs, whereas pluripotency markers were expressed at similar levels in CiPSCs and mESCs. Hierarchical clustering showed that the global gene expression profile of CiPSCs was similar to that of mESCs and distinct from those of MEFs or CiXEN cells (Fig. 4E), which are cells in the initiation and intermediate stages of chemical reprogramming. To measure the methylation status of the Oct4 promoter, we conducted bisulfite sequencing analysis in MEFs, from which CiPSCs originated, mESCs, and CiPSCs. Oct4 promoter was unmethylated in mESC and CiPSCs but methylated in MEFs (Fig. 4F).

Finally, to confirm the pluripotency of CiPSCs, we analyzed their ability to differentiate into multiple lineages in vitro and in vivo. In an in vitro assay, we examined whether CiPSCs differentiated into three germ layers through EB formation. We observed that cells differentiated from CiPSCs expressed marker proteins of other lineages, MAP2 (ectoderm), α-actin (mesoderm), and AFP (endoderm) (Fig. 5A). In an in vivo assay, we transplanted these cells into immune-deficient mice to generate teratomas. Three weeks after injection, we confirmed the formation of teratoma containing derivatives of all three germ layers, indicating that CiPSCs are pluripotent and capable of differentiation (Fig. 5B). To further evaluate the pluripotency of CiPSCs, we analyzed chimera formation after blastocyst injection with CiPSCs. This analysis revealed that Oct4-GFP-positive cells contributed to germ cells in gonads of chimeric embryos (Fig. 5C). In addition, genotyping of chimeric embryos indicated that CiPSCs contributed to the three germ layers, namely to skin (ectoderm), kidney (mesoderm), and liver (endoderm) cells (Supplementary Fig. S7). We used germline-derived pluripotent stem cells with the Oct4-GFP reporter as a positive control (Lee et al., 2018). Our data show that CiPSCs are pluripotent and can differentiate into three germ layers and germline cells.

We established CiPSCs derived from MEFs, with CiXEN cells as an intermediate stage. In this process, we used chemicals that are used as epigenetic modulators and chemicals that can affect signaling pathways. Although we attempted to generate CiPSCs using the method that induced pluripotency in passage1 CiXEN cells same as Zhao et al. (2015), there were difficulties such as cell death or morphology not changing during the induction, and eventually, CiPSCs induction failed.

The establishment of CiXEN cells was necessary for CiPSC induction, and we found that various cell types appeared in the course of several experiments on CiXEN cell induction (Supplementary Fig. S2). Three types of colony morphology appeared during CiXEN cell induction: mESCs-like colonies, typical XEN colonies, and spread colonies. CiPSCs were successfully induced from typical XEN colonies but not from the other two types. In particular, the mESCs-like colonies not only had different gene expression patterns in comparison with that of typical XEN colonies (Supplementary Fig. S3), but also its proliferative capacity was better than that of the other types, and cells from mESCs-like colonies overwhelmed other cell types when continuously subcultured. Therefore, CiPSCs could not be obtained from mESCs-like colonies, and if this colony type appeared during the experiment, these colonies were physically removed and the experiment was continued. From the results of these experiments, we hypothesized that it would be difficult to induce pluripotency in mixtures of different types of cells. To verify this hypothesis, we induced pluripotency after obtaining homogeneous CiXEN cells by continuously culturing the typical XEN colony type. In homogeneous CiXEN cells, cell death, which had previously been a problem, did not occur, so it was possible to successfully induce pluripotency and establish a CiPSCs line. The reason why the pluripotency induction efficiency was better in homogeneous CiXEN cells seems to be because the expression of the XEN marker gene (Gata4, Gata6, Sox17) capable of inducing Oct4, a major factor in pluripotency induction, was higher than that in heterogeneous CiXEN cells that had undergone fewer than 4 passages. In addition, the expression of Sox17 was particularly high at passages 4, 12, and 18, and the appearance of GFP-positive cells among CiXEN cells of these passages is consistent with the previously demonstrated important role of Sox17 in the chemical reprogramming process (Yang et al., 2020).

In this study, we established three CiPSCs lines (CiPSC Clone 1, CiPSC Clone 2, and CiPSC Clone 3) by physically isolating and culturing mESCs-like colonies. All of them were pluripotent, so it was possible to differentiate them into three germ layers, and they exhibited the same characteristics as mESCs. Despite this, in karyotype analysis, only CiPSC Clone 1 was normal, whereas Clones 2 and 3 had abnormal karyotypes (Supplementary Fig. S6). Cell lines with abnormal karyotypes were both aborted during chimera assay, and the normal-karyotype cell line made germline contribution on embryonic day 14. This is consistent with a study in which the ability for germline contribution was lost when the proportion of euploid cells fell below 50% (Longo et al., 1997). These karyotype abnormalities appeared during CiXEN cell induction. However, we did not determine whether that the karyotype abnormalities appeared in CiXEN cells induction process were inherited by CiPSCs. Therefore, more research is needed to identify the mechanism of karyotype abnormalities caused by chemical treatment.

In addition, studies on the broad-spectrum effects of chemicals are also needed. As shown in Fig. 3C, GFP-positive cells appeared only at passage 4 and above, but they did appear at all passages above 4. In other words, it was not always possible to obtain CiPSCs from homogeneous CiXEN cells. If CiXEN cells were established at passage 4, then GFP-positive cells should have appeared at passages 5 and 6. Also, different types of cells appeared during CiXEN cell induction even when the same chemicals and method were used. These results show that chemical reprogramming does not always produce the same result. The reason for this irreproducibility is that, unlike in the reprogramming method using transcription factors, the chemicals used for chemical reprogramming do not always activate the expression of a specific set of transcription factors required for cell fate transformation because these chemicals have extensive effects on gene expression and epigenetic modifications (Takeda et al., 2018). Therefore, for successful chemical programming, additional research is needed to identify chemicals that can more accurately and specifically control cell fate.

Several chemical cocktails in various combinations have been introduced for reprogramming in mice (Hou et al., 2013; Long et al., 2015; Ye et al., 2016; Zhao et al., 2015) and human cells (Guan et al., 2022). Research by Zhao et al. (2015), shows that XEN cells, which serve as an intermediate stage in this process, are essential for generating CiPSCs. We demonstrated that the success of inducing pluripotency depends on how homogeneous the CiXEN cells are, and presented a method for making them homogeneous. Consequently, in our study, we show that the establishment of a homogeneous XEN cell line is the critical step for the chemical induction of pluripotency.

This research was supported by National Research Foundation of Korea (NRF) grant funded by the Korean government (MEST) (NRF-2021R1F1A1057192) and Korea Environment Industry & Technology Institute (KEITI) through Core Technology Development Project for Environmental Diseases Prevention and Management, funded by Korea Ministry of Environment (MOE) (2021003310002).

D.J., Y.L., and K.K. conceived the study. D.J., Y.L., S.W.L., S.H., N.Y.C., and M.L. conducted the experiments and analyzed the data. G.W. and H.R.S. contributed reagent/material. D.J., Y.L., and K.K. interpreted the results and wrote the manuscript.

The authors have no potential conflicts of interest to disclose.