All experiments with animals were approved by the Institutional Animal Care and Use Committee (IACUC) of the Nanjing Medical University and the corresponding ethical approval code is IACUC-2112051. The methods were carried out in “accordance” with the approved guidelines.

The 4 main TFs OCT4, SOX2, KLF4, and CMYC from porcine, bovine and murine genome were used to reprogram Bama miniature PFFs and the derived piPSCs were named as ppiPSCs, bpiPSCs and mpiPSCs. The details were as follow.

To generate ppiPSCs and bpiPSCs, Bama miniature PFFs with a tdTomato cassette inserted into the 3′ UTR of the porcine OCT4 locus (POT PFFs) were transfected, which were kindly provided by Prof. Liangxue Lai (Guangzhou Institutes of Biomedicine and Health, Chinese Academy of Sciences) (Lai et al., 2016). The TFs combinations for ppiPSCs and bpiPSCs generation contained 2.0 μg PB-TRE-pOSKM (porcine OCT4, SOX2, KLF4 and CMYC) and 2.0 μg PB-TRE-bOSKM (bovine OCT4, SOX2, KLF4 and CMYC), respectively, and both also with 1.0 μg PB-TRE-pNhL (porcine NANOG and human LIN28), 1.0 μg PB-TRE-hRL (human RARG and LRH1), 1.0 μg PB-EF1a-transposase and 1.0 μg PB-EF1a-rTTA. The PB-TRE-bOSKM was constructed by Prof. Xihe Li’s group (Inner Mongolia University) (Zhao et al., 2021) and the remaining plasmids were kindly provided by Prof. Pentao Liu (University of Hong Kong) (Gao et al., 2019). After transfection, the cells and picked colonies were placed in culture dishes containing STO feeder layers and M15 medium (comprised knockout DMEM (Invitrogen, Carlsbad, CA, United States), 15% FBS, 1 × glutamine penicillin-streptomycin (Invitrogen), 1 × NEAA (Invitrogen) and 0.1 mM 2-mercaptoethanol (Invitrogen), 50 μg/mL Vitamin C (VC) and 10 ng/mL hLIF (cat# LIF1010, Millipore)) with 1 mg/mL doxycycline (DOX) (cat# 04–0,018 Stemgent). The ppiPSCs and bpiPSCs were maintained on STO feeder layers, enzymatically passaged every 3–5 days by a brief DPBS wash followed by treatment with TrypLE select (Gibco) for 3–5 min and incubated at 38.5°C in a 5% CO2 incubator.

For mpiPSCs generation, 8 μg DNA (4 μg TetO-FUW-mOSKM and 4 μg FUW-M2rtTA) was transferred into PFFs using a Lonza nuclear transfer machine under U-023. 24 h later, cells were passaged onto 10 cm dishes, and the same numbers of non-transgenic cells were used as negative control. On the next day, 300 ng/mL zeocin was applied to the culture medium. When the non-transgenic cells were completely drug-screened, the resistant cell colonies were digested in cloning cups and subcultured on mitotically inactivated mouse embryonic fibroblasts (MEFs) and induced by LBX medium. The mpiPSCs were enzymatically passaged and cultured with LBX medium when grown into a normal colony size. The LBX medium was supplemented with gene knockout serum replacement (KOSR) medium, N2B27 medium and supplemented with 16 ng/mL FGF (cat# 100-18B, PeproTech), 2000U/mL LIF and 2 mg/mL DOX. KOSR medium was knockout DMEM supplemented with 20% KSR (Invitrogen), 1% NEAA, 2 mM L-glutamine, 1% PS, 0.1 mM 2-mercaptoethanol, 3 mM CHIR99021 (cat# SML1046, Sigma), 1 mM PD0325901 (cat# PZ0162, Sigma), 2 mM SB431542 (cat# 1,614, Tocris) and 50 ng/mL VC. N2B27 medium was DMEM/F12 and Neurobasal supplemented with 1% N2 (cat# 17502048, Gibco), 2% B27 (cat# 17504044, Gibco), 0.25 mg/mL bovine serum albumin (BSA, cat# A4612, Sigma), 5 mg/mL insulin (Invitrogen), 1% PS, 3 mM CHIR99021, 1 mM PD0325901, 2 mM SB431542, and 50 ng/mL VC.

To obtain piPSCs that were independent of exogenous gene expression, we cultured the ppiPSCs and bpiPSCs in pEPSCs medium with minor modification (Gao et al., 2019), and mpiPSCs in dox-free LBX medium. The pEPSCs medium was N2B27 basal media (Knockout DMEM, 0.5% N2 supplement, 1% B27 supplement, 100 × glutamine penicillin-streptomycin, 100 × NEAA and 0.1 mM 2-mercaptoethanol) supplemented with 0.2 µM CHIR99021, 0.3 µM WH-4-023 (cat# 5413, Tocris), 2.5 µM XAV939 (cat# X3004, Sigma), 65.0 μg/mL VC, 10.0 ng mL hLIF, 20.0 ng/mL Activin A (cat# 120-14 E, PeproTech) and 0.3% FBS (BI, Kibbutz Beit-Haemek, Israel).

To assess the integration of foreign genes, the total genomes were extracted from different piPSCs cell lines harvested at passage 2. Total DNA was isolated from the cells for polymerase chain reaction (PCR) by the Genomic DNA Kit (cat# DP304-03, TIANGEN) according to the manufacturer’s instructions. DNA from PFFs was isolated as negative control and plasmids as positive control. The plasmid specific primers and specific porcine GAPDH primers were designed and used in PCR. The primer oligonucleotide sequences are shown in Supplementary Table S3.

Alkaline phosphatase (AP) staining was performed with NBT/BCIP solution using a method provided by manufacturers (Promega, Madison, WI, United States). The cells were first washed with DPBS, and then fixed with 4% paraformaldehyde (Solarbio, Beijing, China) at room temperature for 10 min. Then, the cells were washed twice with DPBS and incubated with the reagent to avoid light for 15 min. After the staining was terminated with PBS, the cells were observed and photographed.

For immunofluorescence, cells were washed with DPBS and then fixed with 4% paraformaldehyde at room temperature for 10 min. Cells were permeabilized by 1% Triton X-100 (Sigma) for 10 min and blocked with 5% BSA at room temperature for 1 h. Then, the cells were incubated with the primary antibody overnight at 4°C, after three washes with DPBS, the cells were incubated with the secondary antibody for 1 h at room temperature. The nuclei were stained with DAPI (Sigma) and observed by fluorescence microscopy (Ti2, Nikon) and photographed. The primary antibodies were OCT4 (cat# sc5279, Santa), SOX2 (cat# 4900, Cell Signaling Technology), NANOG (cat# sc33759, Santa), βIII-tubulin (cat# ab68193, Abcam), Sma (cat# CBL171, Millipore), Keratin (cat# MAB1677, Millipore), LHX1 (cat# GTX129215, GeneTex), MIXL1 (cat# 22772-1-AP, proteintech), goat anti-rabbit IgG Alexa Fluor 546 (cat# AS007, ABclonal), goat anti-rabbit IgG Alexa Fluor 488 (cat# A11008, Thermo Fisher Scientific), goat anti-mouse IgG Alexa Fluor 546 (cat# A11003, Thermo Fisher Scientific), goat anti-mouse IgG Alexa Fluor 488 (cat# A32723, Thermo Fisher Scientific).

The cells were cultured with stem cell culture medium containing 0.02 μg/mL colchicine (KaryoMAXTM Colcemid^™ Solution in PBS) for 1–1.5 h at 38.5°C in an incubator of 5% CO₂. After digestion and treatment with preheated hypotonic KCl (0.56%, Sigma) for 30min at 37°C, the cells were fixed in a 3:1 mixture of methanol and glacial acetic acid. Suspension was dropped onto clean, precooled slides. After air drying, the chromosomes were stained by Giemsa (1:10 dilution) for 20 min and observed and recorded at 1,000 magnification (80i, Nikon).

To detect the expression of endogenous and exogenous pluripotency genes in the cells, piPSCs were harvested at passage 5. Total RNA was extracted using the Total RNA Kit (cat# LS1040, Promega) according to the manufacturer’s instructions and was reverse transcribed into cDNA using HiScriptII Q RT Super Mix (Vazyme, Nanjing, China). The cDNA obtained from these cell lines was used for Quantitative Real-Time PCR (qPCR) reactions. qPCR was performed using ChamQ SYBR qPCR Master Mix (Vazyme) on a 96-well optical reaction plate in 7,500 Real-time System (Applied Biosystems, Waltham, MA, United States). GAPDH was used as an internal control to standardize 2^−ΔΔCT values of target genes. Overall gene expression was assessed by normalizing the gene expression of the samples to that of parental embryonic porcine fibroblasts. Data were analyzed by Ct method and significance was determined at p ≤ 0.05. The primers used in this study are shown in Supplementary Table S3.

For embryoid bodies (EBs) formation, piPSCs were cultured in hanging drops without bFGF, hLIF, and other small molecule inhibitors for 3 days. Then EBs were placed on a 24-well plate covered with gelatin and cultured in the same conditions for 7–10 days. The differentiation of the three germ layer cells was identified by RT-qPCR and immunofluorescence.

To induce piPSCs differentiating into incipient mesodermal-like cells (iMeLCs), they were first cultured in 12-well plates (1.5–3.0 × 10⁵/well) coated with fibronectin for 2 days. The mpiPSCs, bpiPSCs and ppiPSCs were cultured in the medium containing 50 ng/mL Activin A, 3 μM CHIR99021 and 10 μM Y27632 (cat# 1,254, Tocris). Then, to differentiate cells to pPGCLCs, iMeLCs were digested into single cells and transferred into U-shaped 96 wells at 5 × 10³. They were induced for 6 days in basal medium supplemented with 200 ng/mL BMP4 (cat# 120–05, PeproTech), 10 ng/mL hLIF, 100 ng/mL SCF (cat# 300–07, PeproTech) and 50 ng/mL epidermal growth factor (EGF, cat# PHG0311, Invitrogen). Specific marker genes were detected by qPCR and immunofluorescence.

The CAG-GFP plasmid was transfected into ppiPSCs using Nucleofector™ Kit 1 (cat# VPH-5012, Lonza). For chimeric contribution to mice, 10–15 GFP labeled ppiPSC-9 were injected into the middle post-posterior of E7.5 mouse embryos (ICR) and then cultured in vitro in human serum with KSOM and pEPSCs with DOX mixture medium (1:3) at 37°C in a 5% CO₂ atmosphere for 12 h and 24 h for the evaluation of E8.5 chimerism.

A total amount of 3 mg RNA per sample from mpiPSC line 1 at passage 25, mpiPSC line 2 at passage 21, bpiPSC line 1 at passage 17, bpiPSC line 2 at passage 14, ppiPSC line 4 at passage 17 and ppiPSC line 7 at passage 19 were used as input materials for the RNA sample preparations. The libraries were constructed using TruSeq Stranded mRNA LT Sample Prep Kit (Illumina, San Diego, CA, United States) according to the manufacturer’s instructions. The transcriptome sequencing and analysis were conducted by OE Biotech Co., Ltd. (Shanghai, China). The libraries were sequenced on an Illumina HiSeq X Ten platform and 150 bp paired end reads were generated. p-value < 0.05 and foldchange >2 or foldchange < 0.5 was set as the threshold for significantly differential expression.

All data in this experiment were expressed as mean ± standard deviation (X ± S), and GraphPad software was used for significance analysis. t-test was used to compare the two independent samples and one-way ANOVA was used to compare multiple samples, and p ≤ 0.05 was considered statistically significant. In all graphs, p ≤ 0.05 was marked as *, p ≤ 0.01 was marked as **, p ≤ 0.001 was marked as ***, p ≤ 0.0001 was marked as ****, no statistical significance was marked as ns and n ≥ 3.

The POT PFFs were transfected with three DOX inducible piggyBac plasmids including pOSKM (porcine OCT4, SOX2, KLF4 and CMYC), pNhL (porcine NANOG and human LIN28) and hRL (human RARG and LRH1) to generate ppiPSCs (Figure 1A). The primary iPSC-like colonies were firstly observed on day 6 after DOX induction, which exhibited compact morphology and a high nuclei-to-cytoplasm ratio. These colonies were collected on day 12 and subcultured in single-cell suspension on STO feeder layers in serum-containing medium (M15 medium) plus DOX (Figure 1Ba). Total 15 ppiPSCs lines were obtained and all of them have 8 exogenous TFs integrated and stably expressed (Supplementary Figure S1A). All ppiPSCs cell lines were OCT4-tdTomato⁺ (Figure 1 Bb; Supplementary Figure S1B). The ppiPSCs could be passaged for more than 30 generations in single cells with no change in morphology and proliferation rate, and all were AP positive (Figure 1Bc). The core pluripotency markers such as OCT4, SOX2 and NANOG were positive by immunofluorescence staining (Figure 1C). Analysis of Quantitative Real-Time PCR (qPCR) showed that the 7 endogenous TFs OCT4, SOX2, KLF4, NANOG, LIN28, LRH1 and RARG were activated and expressed except for CMYC in these cell lines (Figure 1D; Supplementary Figure S1C).

The POT PFFs were transfected with three DOX inducible piggyBac plasmids including bOSKM, pNhL and hRL to generate bpiPSCs (Supplementary Figure S2A). The primary iPSC-like colonies were firstly observed on day 6 after DOX induction and colonies exhibiting compact and three-dimensional, well-defined colony morphology were picked on days 10–13 (Supplementary Figure S2B a). The bpiPSCs could be passaged in single-cell suspension and maintained in M15 medium plus DOX. Total 8 cell lines were obtained, which have 8 exogenous TFs integrated and stably expressed (Supplementary Figure S3A). The No. One bpiPSCs cell line (bpiPSCs-1) from POT PFFs were OCT4-tdTomato⁺ (Supplementary Figure S3B b), but the remaining 7 bpiPSCs cell lines were OCT4-tdTomato^- (Supplementary Figure S3B). The bpiPSCs showed AP positive (Supplementary Figure S2B c) as well as intranuclear expression of pluripotency markers including OCT4, SOX2 and NANOG (Supplementary Figure S2C). The qPCR confirmed the expression of eight exogenous TFs and the activation of porcine endogenous counterparts in bpiPSCs. The expression of the endogenous TFs OCT4, SOX2, NANOG, LIN28, LRH1 and RARG were activated in the 8 cell lines although their expression patterns were different among individual cell lines (Supplementary Figure S2D; Supplementary Figure S3C). However, none bpiPSCs cell line initiated the endogenous expression of KLF4 and CMYC. The eight bpiPSCs lines could be stably cultured in vitro for more than 35 passages and maintained normal karyotype.

We transfected the expression vector TetO-FUW-mOSKM and the activation vector FUW-M2rtTA into PFFs, and mpiPSCs were generated through drug screening and DOX induction (Supplementary Figure S4A). The colonies could be observed on day 3 after induction and were picked on day 7. The mpiPSCs exhibited compact and dome-like morphology (Supplementary Figure S4B a) and were able to achieve single-cell colonization after enzymatic treatment. The mpiPSCs had normal karyotype after long-term culture (Supplementary Figure S4B b) and were AP positive (Supplementary Figure S4B c). The mpiPSCs expressed pluripotency markers, including OCT4, SOX2 and NANOG (Supplementary Figure S4C) with upregulated expression of four exogenous TFs. However, except for SOX2, the endogenous TFs OCT4, KLF4 and CMYC remained barely expressed in the mpiPSCs (Supplementary Figure S4D). Ultimately, we obtained 2 cell lines that were able to continuously reproduce for more than 30 passages and maintain a robust self-renewal capacity.

We compared the effect of different species of exogenous TFs on the activation of endogenous pluripotency markers of three kinds piPSCs. The results showed that the average expression levels of the key endogenous pluripotency markers containing OCT4, KLF4, CMYC, NANOG, REX1, TBX3and DPPA5 in ppiPSCs lines were the highest, and they were higher in bpiPSCs line and the lowest in mpiPSCs lines. The LIN28 had the highest expression level in bpiPSCs and then mpiPSCs and ppiPSCs. In contrast, the SOX2 had the highest expression level in mpiPSCs and then ppiPSCs and bpiPSCs. The LRH1 and RARG had the highest expression level in bpiPSCs and then ppiPSCs and mpiPSCs (Figure 2A).

The correlation analysis of the expression level between 8 exogenous TFs and their endogenous counterparts showed the somewhat negative correlation in 62.5% (5/8) of the cell lines of bpiPSCs (Figure 2Ba-c; Supplementary Figure S5A). In contrast, 93.3% (14/15) of ppiPSCs showed a positive correlation, meaning that the activation of endogenous TFs enhanced when the expression level of exogenous counterparts increased. (Figure 2 B d-f; Supplementary Figure S5B).

All three kinds of piPSCs could form EBs-like spheroids when grown in suspension for 7 days (Figure 3Aa-c). Day-4 EB were plated on gelatin-coated dishes under the same conditions, and the cells of EBs showed a markedly differentiated morphology on day 10(Figure 3 A d-f). The EBs from ppiPSCs (pEBs) were able to differentiate into three germ layers (ectoderm ß-III tubulin, mesoderm Sma and endoderm Keratin). EBs from bpiPSCs (bEBs) and mpiPSCs (mEBs) expressed ß-III tubulin and Sma but did not express Keratin (Figure 3A). The expressions of differentiation-related genes, such as ectoderm Nefl, mesoderm Desmin and endoderm Albumin, in EBs-like spheroids derived from all three kinds of piPSCs lines were significantly upregulated compared with undifferentiated counterparts (Figure 3B). In short, all piPSCs had the ability to differentiate into multiple germ layers. However, ppiPSCs exhibited the full-scale differentiation potential in vitro toward the three germ layers.

It has been reported that human PSCs were able to transform into PGCLCs via iMeLCs (Yamashiro et al., 2020). After optimizing the induction system, we developed a procedure to differentiate into PGCLCs from piPSCs (pPGCLCs) (Figure 3C). The ppiPSCs could rapidly and efficiently differentiate into PGCLCs (ppPGCLCs). However, only one of two bpiPSCs cell lines could differentiate toward PGCLCs (bpPGCLCs) and express few reproduction-related markers, but mpiPSCs only differentiated to the iMeLCs could not reach to PGCLCs. Compared to mpiPSCs differentiated cell mass (putative mpPGCLCs), ppPGCLCs and bpPGCLCs proliferated vigorously and both of ppPGCLCs and bpPGCLCs formed clusters with bird’s nest-like colonies, which are typical morphological features of PGCs in vitro. After 2 days of induction, the iMeLCs from three kind of piPSCs expressed mesodermal markers LHX1 and MIXL1 (Figure 3D). qPCR results indicated that the expression of mesodermal markers MIXL1 and GATA3 were significantly upregulated in all pPGCLCs but the expression in ppPGCLCs was at the highest level compare with bpPGCLCs and putative mpPGCLCs (Supplementary Figure S6A). We also analyzed the gene expression dynamics during 8 days of pPGCLCs induction (Figure 3E). During the formation of ppPGCLCs, the expression of the key genes for PGC specification and development were significantly upregulated, including PRDM1, DAZL and VASA. In bpPGCLCs, the expression of PRDM1 and VASA gradually increased, whereas DAZL was significantly increased on day 4 and 6 but downregulated on day 8. For putative mpPGCLCs, PRDM1 and DAZL reached peaks at D4 and D6 respectively and declined significantly thereafter. The expression of SOX17 in all pPGCLCs showed exhibited a trend of high-level expression in the early stage and downregulation in the later stage, which was consistent with the results of PGC formation in mice (Hayashi et al., 2011).

To further investigate the developmental capacity of piPSCs, chimera experiments were performed. First, the ppiPSCs were transfected with the CAG-GFP plasmid (Supplementary Figure S6B). We then injected 10–15 GFP labeled ppiPSCs into the distal region of a late-streak stage embryo and allowed them to develop further for 12 and 24 h in vitro. GFP⁺ cells were detected in the host embryos (Figure 3F) and contributed to the neural ectoderm (Supplementary Figure S6C). Collectively, ppiPSCs could differentiate and contribute to chimeric formation in mouse late embryos.

To obtain piPSCs that can proliferate independently of exogenous TFs expression, we cultured the piPSCs in DOX-free stem cell medium and observed the colony formation on the day 3. The mpiPSCs and bpiPSCs gradually died in 2 passages, whereas ppiPSCs showed no signs of altered morphology or differentiation and could be passaged normally (Figure 4A). The ppiPSCs were positive for AP staining and strongly expressed pluripotency markers including OCT4, SOX2, and NANOG at 5 passages without DOX (Figure 4B).

The RNA sequencing results indicated that ppiPSCs, bpiPSCs and mpiPSCs exhibited the obvious differences in gene expression pattern (Figure 5A). Global gene expression profiling revealed that gene expression in ppiPSCs was distinct from that in mpiPSCs and bpiPSCs (Figure 5B). Pearson correlation analysis confirmed that mpiPSCs and bpiPSCs show a stronger correlation in three kinds of piPSCs (Figure 5C). Compared with bpiPSCs, ppiPSCs had 608 upregulated genes (such as TBX3, NODAL, SOX9) and 434 downregulated genes (such as HAND2, HOXA3, APELA) among the 15,473 identified genes (Figures 5D,E). KEGG enrichment analysis showed that the upregulated genes were mostly enriched in Hippo signaling pathway, PI3K-Akt signaling pathway and the other signaling pathways regulating pluripotency of stem cells (Figure 5G). Compared with mpiPSCs, ppiPSCs had 2,582 upregulated genes (such as POU5F1, NANOG, KLF4, DPPA5, TBX3) and 1,475 downregulated genes (such as BMP6, CCK, POU3F1) among the 15,432 identified genes (Figures 5D,F). KEGG enrichment analysis showed that the upregulated genes were mostly enriched in PI3K-Akt signaling pathway, Hippo signaling pathway and MAPK signaling pathway (Figure 5H). We selected genes related to pluripotency from RNA-seq data for analysis and confirmed that ppiPSCs expressed most endogenous pluripotency markers, and bpiPSCs expressed a few endogenous pluripotency markers, but mpiPSCs lack the expression of most endogenous pluripotency markers (Figure 5I). We also confirmed that ppiPSCs robustly express POU5F1, KLF4, DPPA5, TBX3, ZFP42 and other genes related to pluripotency. This expression pattern was consistent with that of porcine ICM (RNA-seq data obtained previously by our lab) (Zhang et al., 2019), which also highly express these genes (Figure 5J). In addition, we chose the PGC early related genes for analysis, and found that both ppiPSCs cell lines strongly expressed most of these genes, a cell line of bpiPSCs highly expressed these genes, and mpiPSCs barely expressed these genes (Figure 5K).

The similarities and differences of amino acid sequence, protein secondary structure and tertiary structure of the four key TFs OCT4, SOX2, KLF4 and CMYC in mouse, cattle and pig were compared via biological information analysis. Firstly, DNAMAN was used to compare the amino acid sequences, and the results showed that the similarity of OCT4 was 83% between pig and mouse and 97% between pig and cattle; the similarity of SOX2 was 97% among mouse, cattle and pig; of KLF4 was 92% between pig and mouse and 98% for pig and cattle; the similarity of CMYC was 92% for pig and mouse and 96% between pig and cattle (Supplementary Table S1). Secondly, Protean was used to compare protein secondary structure including a-helix, ß-fold and ß-turn amount. The total number of a-helix, ß-fold and ß-turn in OCT4 is 90.9, 91.9 and 88 for pig, cattle and mouse, respectively. They are 78.63, 78.07 and 82.09 in SOX2; 83, 83.93 and 83.67 in KLF4; 87.83, 86.78 and 92.1 in CMYC (Supplementary Table S2). These results revealed OCT4, SOX2, KLF4 and CMYC had more similar amino acid sequence and protein secondary structure between pig and cattle than pig and mouse. Finally, we used the online website SWISS MODEL to perform tertiary modeling and compared the tertiary structures using pyMOL. The RMSD values of mouse and pig at OSKM were 0.004, 0.003, 0.000, and 0.051, respectively and the values were 0.001, 0.003, 0.000, and 0.058 for cattle and pig, respectively (Supplementary Figure S6D). In conclusion, the protein characters of OCT4, SOX2 and KLF4 in pig were closer to cattle than to mouse. CMYC had more similar tertiary structure between pig and mouse.